This may be of interest to some of you who live near London...
Audio Engineering Society British Section Meetings
Lectures are free and are open to all - members and non-members.
Wednesday 3rd June (extra meeting)
Lecture: Critical listening/evaluation - a path to the future of quality music
George Massenburg, George Massenburg Labs
George Massenburg says of his lecture: "This is pretty audacious but I am hoping to set the current crisis in craft in music recording into a broader, perhaps philosophical context. We'll touch on the tyranny, greed and utter stupidity of big record business in taking down music. Anyone who has worked in a factory will have experienced life as a much-reduced palette - life as formulaic, as a slave to technology. A given technology, once introduced, plays itself out to its logical or natural conclusion. Where technology dominates you end up with machines running the people. Ultimately, the meltdown of the record business might be due in some part to the precipitous decline in quality music and quality music recording.
More down to earth, I will be playing some of the best recordings I can get my hands on. They are not all necessarily conceived as 'high-res' but are definitely deeply scarred by cheap playback methodologies. I've also gotten my hands on some extraordinary multi-track masters, such as Stevie Wonder's Superstition, which we'll play and listen to critically.
My hope for the meeting is to develop ideas between us all as to how we can better demonstrate to students how important quality recording is."
The lecture will be held at The Royal Academy of Engineering, 3 Carlton House Terrace, London, SW1Y 5DG, just off Pall Mall. The meeting starts at 7.00pm with refreshments at 6.30pm.
http://www.aes.org/sections/uk/meetings/index.html (http://www.aes.org/sections/uk/meetings/index.html)
Edit: an mp3 of the lecture is now available:
http://www.hydrogenaudio.org/forums/index....st&p=640868 (http://www.hydrogenaudio.org/forums/index.php?s=&showtopic=72446&view=findpost&p=640868)
Cheers,
David.
Thanks for posting David- good to hear people still do this stuff, I had no idea they did.
If anyone goes, will they post?
- Spike
Hopefully, the MP3 of this presentation will be online in due course, for those of us in the hinterlands! ;-)
IME George Massenburg is one of the "good guys", someone who has spoken out clearly against the hi-rez recording format myth.
I think those Superstition multitracks have been floating around the net. This guy has them:
http://www.youtube.com/watch?v=WryUOXo9sfM (http://www.youtube.com/watch?v=WryUOXo9sfM)
The sad thing is there is no good sounding version of that album. I have an original vinyl, original CD, and had the remastered CD release that all sound poor.
One remarkable thing about this to me is that it looks like he's gonna play the actual Superstition multis. Stevie Wonder has never authorized use of his master tapes for any CDs so far, and his LPs back then were notoriously hard to cut (so I doubt we've ever head this music in best quality). Wish I could hear this.
One remarkable thing about this to me is that it looks like he's gonna play the actual Superstition multis. Stevie Wonder has never authorized use of his master tapes for any CDs so far, and his LPs back then were notoriously hard to cut (so I doubt we've ever head this music in best quality). Wish I could hear this.
I see that there are currently available torrents composed of at least some of the individual tracks.
From practical experience I can tell you that a collection of individually brilliant tracks does not necessarily give you an adequate basis for an excellent mixdown. If some of the tracks are mediocre, then even less so.
IME George Massenburg is one of the "good guys", someone who has spoken out clearly against the hi-rez recording format myth.
That's odd. George and I both appeared on a panel at the 2007 AES Conference in London and were pretty much in agreement on the merits of high sample rates and bit depth greater than 16.
John Atkinson
Editor, Stereophile
With what particular merits were you in George in agreement on?
With what particular merits were you in George in agreement on?
Exactly the right question.
I hope this isn't just the uncontroversial merits of high-bit digital production/processing, or the employment of high SR stages to allow artifact-immune antialiasing and antiimaging filtering. I.e., 'larger numbers' as a means to AVOID introducing artifacts in the audible band, not because the Redbook 'numbers' inherently fail to capture what's audibly important?
Or does Mr. Massenburg think Redbook delivery formats *inherently* 'lack' something important, which has long been the stance among types who champion, say *vinyl*?
With what particular merits were you in George in agreement on?
Most recently, the lack of audible differences due to 96 KHz sampling.
IME George Massenburg is one of the "good guys", someone who has spoken out clearly against the hi-rez recording format myth.
That's odd. George and I both appeared on a panel at the 2007 AES Conference in London and were pretty much in agreement on the merits of high sample rates and bit depth greater than 16.
John Atkinson
Editor, Stereophile
Just another example of John's delusion that just a few nuts, who mostly live in Michigan disagree with him about anything. ;-)
BTW John, you have often been shown to hear whatever you want to hear, and I guess this includes not only listening tests of equipment, but also panel discussions.
If anyone goes, will they post?
Yes.
With what particular merits were you [and] George in agreement on?
That question would appear to be unanswerable on this forum, as any statement of opinion I make would need, by HA's rules, to be supported by the results of blind testing, and I have no intention of being sucked (again) into at that particular maelstrom. If you are really interested in my thoughts and not just trolling, read my essays at http://www.stereophile.com/asweseeit/907awsi/ (http://www.stereophile.com/asweseeit/907awsi/) and http://www.stereophile.com/asweseeit/1007awsi/ (http://www.stereophile.com/asweseeit/1007awsi/) .
John Atkinson
Editor, Stereophile
IME George Massenburg is one of the "good guys", someone who has spoken out clearly against the hi-rez recording format myth.
That's odd. George and I both appeared on a panel at the 2007 AES Conference in London and were pretty much in agreement on the merits of high sample rates and bit depth greater than 16.
Just another example of John's delusion that just a few nuts, who mostly live in Michigan disagree with him about anything. ;-)
I don't believe so, Mr. Krueger. I had thought it apparent from my posting that I was disagreeing with just one nut who lives not "mostly" but 100% in Michigan. :-)
BTW John, you have often been shown to hear whatever you want to hear...
Really? A specific example would be helpful, Mr. Krueger. This is HA, after all, where expression of opion are required to be supported with evidential data.
John Atkinson
Editor, Stereophile
read my essays at http://www.stereophile.com/asweseeit/907awsi/ (http://www.stereophile.com/asweseeit/907awsi/) and http://www.stereophile.com/asweseeit/1007awsi/ (http://www.stereophile.com/asweseeit/1007awsi/) .
Certainly worth reading.
But given all that, you'd think it would be quite easy to set up a blind test and get statistically significant results.
Part of me thinks people just haven't tried hard enough. But part of me knows that certain organisations with a
lot to gain
will have done these tests, and if they had positive results, they would have published them.
Interesting comments about the old 78rpm disc sampled at a higher rate - of course a 78 replayed with the correct de-emphasis curve (e.g. Westrex rather than RIAA) has
blisteringly high levels of high frequency and ultrasonic noise. You couldn't get a worse case for triggering the ringing from anti-alias / imaging filters in ADCs and DACs. It should be inaudible, but any non-linearity in the system
could cause it to appear within the audible range.
Cheers,
David.
With what particular merits were you [and] George in agreement on?
That question would appear to be unanswerable on this forum, as any statement of opinion I make would need, by HA's rules, to be supported by the results of blind testing, and I have no intention of being sucked (again) into at that particular maelstrom. If you are really interested in my thoughts and not just trolling, read my essays at http://www.stereophile.com/asweseeit/907awsi/ (http://www.stereophile.com/asweseeit/907awsi/)
Here's two easy TOS8 violations, which I am free to post because unlike John, I don't have anything to lose by pointing out the obvious - thet are base don sighted evaluations, and are therefore higly questionable:
Meridian's Bob Stuart wrote a decade ago, echoing earlier work by the Acoustic Renaissance for Audio, that without noise-shaping, a minimum of 58kHz sampling and a word length of 18.2 bits were necessary for audio to be PCM-encoded with audible transparency (footnote 1), and there was a consensus at the AES conference that hi-rez audio encoding does offer a sonic improvement compared with the 16-bit/44.1kHz-sampled CD standard.
In his keynote address, for example, preeminent engineer Peter Craven demonstrated how digitally transferring a 78 of an electrical recording of an aria from Puccini's La Bohème, recorded live at Covent Garden on June 8, 1926, improved in sound quality when the sample rate was increased from 44.1kHz to 192kHz. Though it was difficult to hear in the reverberant lecture hall, at the highest sample rate, the inevitable surface noise, crackling, and distortion did indeed seem to float free from the mono image of the singer and orchestra compared to the lower rate, at which the noise and distortion seemed more wedded to the image.
and http://www.stereophile.com/asweseeit/1007awsi/ (http://www.stereophile.com/asweseeit/1007awsi/) .
Let's start with the rediculous straw man and grotesquely flawed simile:
So when you read in the popular press that 128kbps MP3s are indistinguishable from CDs, or that satellite radio, which runs at around 64kbps for two channels, is of "CD quality," think of the implications of Philip's demo. Not the least of these, of course, was that we were aware of the degradation from 24/88.2 to "Red Book" CD data, despite the proclamations from some pundits that the CD medium is audibly transparent.
and two more TOS8 violations which I can of course simply dismiss by saying I don't buy any of the following unsupported claims:
This has been my own experience. I've been recording in high resolution since pianist Hyperion Knight's performance of Gershwin's Rhapsody in Blue, in 1997 (Stereophile STPH010-2), and a constant observation has been that undesirable aspects of the sound that I felt were at or below threshold with the original hi-rez files become more annoying, and less readily resolved, when the data are mastered for the commercial CD release. This was the case with my recording of Robert Silverman performing Beethoven's complete piano sonatas, in which the early reflections from the walls of the small recital hall had more of a deleterious effect with the "Red Book" data than with the original hi-rez data (see Stereophile, January 2001, pp.99–107), thus mandating a remix.
This is currently the case with my most recent recording of the vocal group Cantus: after I'd done all the mixing and equalization at 88.2kHz, the CD versions sounded more muddy and less refined than I'd expected, given the care with which I'd downsampled and noiseshaped the hi-rez data. The mix and EQ choices I'd made at 88.2kHz were not optimal for the 44.1kHz versions. My detectives had been misled by the clues. As a result, the release of the CD is horribly late.
The second example could have a technical leg to stand on, if John screwed things up enough - it is well known that nonlinear distoriton in the digital domain tends to reflect downwards from the nyquist frequency, and thus potentially into a range where it is clearly audible. So, if John exceeded the dynamic range of his software with his processing (conceivably possible with extreme noise shaping), there could well be audible artifacts.
With what particular merits were you [and] George in agreement on?
That question would appear to be unanswerable on this forum, as any statement of opinion I make would need, by HA's rules, to be supported by the results of blind testing, and I have no intention of being sucked (again) into at that particular maelstrom. If you are really interested in my thoughts and not just trolling, read my essays at http://www.stereophile.com/asweseeit/907awsi/ (http://www.stereophile.com/asweseeit/907awsi/)
Here's two easy TOS8 violations...and two more TOS8 violations...
It should go without saying that an opinion expressed in "meatspace" rather than the somewhat rarefied cyberspace of HA is not subject to ToS8.
This has been my own experience. I've been recording in high resolution since pianist Hyperion Knight's performance of Gershwin's Rhapsody in Blue, in 1997 (Stereophile STPH010-2), and a constant observation has been that undesirable aspects of the sound that I felt were at or below threshold with the original hi-rez files become more annoying, and less readily resolved, when the data are mastered for the commercial CD release. This was the case with my recording of Robert Silverman performing Beethoven's complete piano sonatas, in which the early reflections from the walls of the small recital hall had more of a deleterious effect with the "Red Book" data than with the original hi-rez data (see Stereophile, January 2001, pp.99–107), thus mandating a remix.
This is currently the case with my most recent recording of the vocal group Cantus: after I'd done all the mixing and equalization at 88.2kHz, the CD versions sounded more muddy and less refined than I'd expected, given the care with which I'd downsampled and noiseshaped the hi-rez data. The mix and EQ choices I'd made at 88.2kHz were not optimal for the 44.1kHz versions. My detectives had been misled by the clues. As a result, the release of the CD is horribly late.
The second example could have a technical leg to stand on, if John screwed things up enough - it is well known that nonlinear distoriton in the digital domain tends to reflect downwards from the nyquist frequency, and thus potentially into a range where it is clearly audible. So, if John exceeded the dynamic range of his software with his processing (conceivably possible with extreme noise shaping), there could well be audible artifacts.
That was not the case with either of the recordings mentioned, Mr. Krueger.
John Atkinson
Editor, Stereophile
Meridian's Bob Stuart wrote a decade ago, echoing earlier work by the Acoustic Renaissance for Audio, that without noise-shaping, a minimum of 58kHz sampling and a word length of 18.2 bits were necessary for audio to be PCM-encoded with audible transparency (footnote 1), and there was a consensus at the AES conference that hi-rez audio encoding does offer a sonic improvement compared with the 16-bit/44.1kHz-sampled CD standard.
Mr. Atkinson,
These two essays haven't changed since I first saw them: more anecdotal than scientific (Drs Woszczyk and Usher's presentation hasn't been published as peer-reviewed work yet, has it?) . It's surpassingly odd that the 'usual suspects *still*, at this late date, haven't come up with the slam-dunk evidence from controlled listening tests to support this supposed 'consensus'. (And where were Dr. Lipshitz and other skeptics when this 'consensus' was arrived at?)
Besides, since Redbook delivery these days
is typically dithered down with noise shaping from a higher-bitrate production chain, *
what is the problem*?
(2bdecided, Bob Katz did do a more thorough investigation of higher SR -- and found that audible difference, when it exists, is likely due to the filtering, not the SR per se. He talks about it in his book 'Mastering Audio' and on some of the pro sound boards)
With what particular merits were you [and] George in agreement on?
That question would appear to be unanswerable on this forum, as any statement of opinion I make would need, by HA's rules, to be supported by the results of blind testing, and I have no intention of being sucked (again) into at that particular maelstrom. If you are really interested in my thoughts and not just trolling, read my essays at http://www.stereophile.com/asweseeit/907awsi/ (http://www.stereophile.com/asweseeit/907awsi/)
Here's two easy TOS8 violations...and two more TOS8 violations...
It should go without saying that an opinion expressed in "meatspace" rather than the somewhat rarefied cyberspace of HA is not subject to ToS8.
This has been my own experience. I've been recording in high resolution since pianist Hyperion Knight's performance of Gershwin's Rhapsody in Blue, in 1997 (Stereophile STPH010-2), and a constant observation has been that undesirable aspects of the sound that I felt were at or below threshold with the original hi-rez files become more annoying, and less readily resolved, when the data are mastered for the commercial CD release. This was the case with my recording of Robert Silverman performing Beethoven's complete piano sonatas, in which the early reflections from the walls of the small recital hall had more of a deleterious effect with the "Red Book" data than with the original hi-rez data (see Stereophile, January 2001, pp.99–107), thus mandating a remix.
This is currently the case with my most recent recording of the vocal group Cantus: after I'd done all the mixing and equalization at 88.2kHz, the CD versions sounded more muddy and less refined than I'd expected, given the care with which I'd downsampled and noiseshaped the hi-rez data. The mix and EQ choices I'd made at 88.2kHz were not optimal for the 44.1kHz versions. My detectives had been misled by the clues. As a result, the release of the CD is horribly late.
The second example could have a technical leg to stand on, if John screwed things up enough - it is well known that nonlinear distoriton in the digital domain tends to reflect downwards from the nyquist frequency, and thus potentially into a range where it is clearly audible. So, if John exceeded the dynamic range of his software with his processing (conceivably possible with extreme noise shaping), there could well be audible artifacts.
That was not the case with either of the recordings mentioned, Mr. Krueger.
Prove it John. All of it.
It is painfully obvioius to many of us that your lengthy leadership of SP have given you the very unhealthy impression that simply speaking words makes them so. Since you don't sign any paychecks around here, you should be prepared for a little higher standard of support for your various exceptional claims.
I think those Superstition multitracks have been floating around the net. This guy has them:
http://www.youtube.com/watch?v=WryUOXo9sfM (http://www.youtube.com/watch?v=WryUOXo9sfM)
The sad thing is there is no good sounding version of that album. I have an original vinyl, original CD, and had the remastered CD release that all sound poor.
I'm curious about what you find disatisfying about the sound of that album.
The collection of discrete tracks for the title track is not difficult to obtain, and I've done so. Listening to them individually and mixing them at will, I find that they are very interesting. For one thing they are a catalog many of the numerous techical problems common to studios in the days of analog. They do seem to be genuine. There's some funny stuff that is probably masked or edited out from the final mix that most of us are familiar with.
(2bdecided, Bob Katz did do a more thorough investigation of higher SR -- and found that audible difference, when it exists, is likely due to the filtering, not the SR per se. He talks about it in his book 'Mastering Audio' and on some of the pro sound boards)
Yep - read those - all sighted tests - results were as you say (though in the one I read, they listened to just the filtering at one stage).
At this rate, this thread will have to be split.
Cheers,
David.
Listening to them individually and mixing them at will, I find that they are very interesting. For one thing they are a catalog many of the numerous techical problems common to studios in the days of analog.
There was a great programme on BBC 6 Music the other week dissecting some multi-tracks from Roy Wood (The Move, ELO, Wizard) and talking about just how much he had to cram onto one track in the days of 8 or 16 tracks. Really interesting - amazing too how different the final mix sounds from the 8 or 16 tracks that make it.
It was called "The Record Producers - Roy Wood".
Cheers,
David.
"This has been my own experience. I've been recording in high resolution since pianist Hyperion Knight's performance of Gershwin's Rhapsody in Blue, in 1997 (Stereophile STPH010-2), and a constant observation has been that undesirable aspects of the sound that I felt were at or below threshold with the original hi-rez files become more annoying, and less readily resolved, when the data are mastered for the commercial CD release. This was the case with my recording of Robert Silverman performing Beethoven's complete piano sonatas, in which the early reflections from the walls of the small recital hall had more of a deleterious effect with the "Red Book" data than with the original hi-rez data (see Stereophile, January 2001, pp.99–107), thus mandating a remix.
"This is currently the case with my most recent recording of the vocal group Cantus: after I'd done all the mixing and equalization at 88.2kHz, the CD versions sounded more muddy and less refined than I'd expected, given the care with which I'd downsampled and noiseshaped the hi-rez data. The mix and EQ choices I'd made at 88.2kHz were not optimal for the 44.1kHz versions. My detectives had been misled by the clues. As a result, the release of the CD is horribly late." (John Atkinson, quoted from http://www.stereophile.com/asweseeit/1007awsi/) (http://www.stereophile.com/asweseeit/1007awsi/))
The second example could have a technical leg to stand on, if John screwed things up enough - it is well known that nonlinear distoriton in the digital domain tends to reflect downwards from the nyquist frequency, and thus potentially into a range where it is clearly audible. So, if John exceeded the dynamic range of his software with his processing (conceivably possible with extreme noise shaping), there could well be audible artifacts.
That was not the case with either of the recordings mentioned, Mr. Krueger.
Prove it John. All of it.
Why do I have to, Mr. Krueger? I didn't post these statements of mine to HA thus ToS8 doesn't apply. I do know for a fact that I didn't screw up the downsampling and word-length reduction, however. But I will gladly supply samples of the original high-resolution files and the downsampled versions to others who might like to do the comparisons if they email me their street addresses.
John Atkinson
Editor, Stereophile
(2bdecided, Bob Katz did do a more thorough investigation of higher SR -- and found that audible difference, when it exists, is likely due to the filtering, not the SR per se. He talks about it in his book 'Mastering Audio' and on some of the pro sound boards)
Yep - read those - all sighted tests - results were as you say (though in the one I read, they listened to just the filtering at one stage).
At this rate, this thread will have to be split.
Cheers,
David.
You sure they were all sighted? On one list I read, BK recently seemed to make a distinction between SR tests -- which have all showed him 'it's the filtering' -- and tests of bitdepths, which he said he believed should be raised for production steps,
even though he has no DBTs to prove it.
But I will gladly supply samples of the original high-resolution files and the downsampled versions to others who might like to do the comparisons if they email me their street addresses.
My street address is unchanged.
I attended the lecture last night.
Quite a mixture.
George was quite shy / intimidated by the standing-room-only crowd in the hall. He'd set up 2 ATC monitors, Prism converters, Pro Tools, and an Arcam DVD-A player at the front.
I believe the two multi-tracks he played are available via bit torrent. The other recordings he played (and much of the evening was spent listening) were 192kHz 24-bit stereo masters - some unreleased tracks, some recent and not so recent ones. All good recordings.
Too loud, too reverberant, and sat too far back - it's not how I'd choose to listen to music. The reproduction was about as far away from believing the singer was really in the room as you can get. I think the recordings and system were probably great, but who can judge in that environment at listening levels that make your ears ring?
George had some very interesting and sensible things to say about the recording of music. He said he didn't have an agenda, but really wanted to make people think. So he talked about capturing performances "as live" - tell the musicians they're not going to get the chance to punch in - if they do a re-take, it wipes the previous one. Put them all in the same room so that they can see and hear each other - the isolation of putting them all in padded boxes lets an accountant some along and demand you change the bass player next month, but it doesn't let you capture the best performance.
Take time to work out the subtleties of the music, and capture them properly. Good recordings with these subtleties still have more to give on the hundredth listen, decades after they were first recorded.
However, there was plenty said that would make most people here on HA seethe. The start was a look at some codecs, playing 192/24 vs 128kbps mp3 (ProTool High Quality - is that FhG based?) - and playing the difference signal - i.e. subtracting the original from the downsampled > encoded > decoded > upsampled version. He played a lot of examples of the noise added by the codec, saying "how can we do this to music?".
A question from the floor pointed out that an all-pass filter would give an even larger difference signal, but create little or no audible difference (hence the method is flawed in terms of illustrating the audible difference) - George doubted this example.
He also explained how he taught people to listen for artefacts by listening to original, coded, difference - over and over again on a particularly tricky part of the track - until they could hear the artefacts in the coded version.
This is a reasonable approach - but the question was never asked "what if they can't hear a problem with the coded version, despite knowing how much noise has been added?" - the impression was that any added noise was bad - "I can get 1.5Mbps over my DSL - stream that - there's no need to even go down to 256kbps AAC".
George was very unhappy about this paper (which we've discussed at length on HA):
http://www.aes.org/e-lib/browse.cfm?elib=14195 (http://www.aes.org/e-lib/browse.cfm?elib=14195)
The discussion about ABX testing brought out the usual straw men. Interestingly, George and Peter (not sure which Peter!) had written a letter to the AES journal criticising the Meyer and Moran paper, which "they refused to publish". He said they'd asked the wrong question.
George wasn't using a microphone, and asked that no-one was recording because he was playing some pre-release tracks which he didn't want to get onto the internet. I don't know if an mp3 without the musical content will find its way onto the AES website - I'm going to ask.
Cheers,
David.
George was very unhappy about this paper (which we've discussed at length on HA):
http://www.aes.org/e-lib/browse.cfm?elib=14195 (http://www.aes.org/e-lib/browse.cfm?elib=14195)
The discussion about ABX testing brought out the usual straw men. Interestingly, George and Peter (not sure which Peter!) had written a letter to the AES journal criticising the Meyer and Moran paper, which "they refused to publish". He said they'd asked the wrong question.
The letter to the JAES was cosigned by George Massenburg, Peter Craven, Vicki Melchior, and Wieslaw Woszczyk and was indeed rejected for publication, though John Vanderkooy did set up the on-line forum mainly as a result of the internal debate at the AES over the content of the letter.
Thank you for posting the report, David.
John Atkinson
Editor, Stereophile
George was very unhappy about this paper (which we've discussed at length on HA):
http://www.aes.org/e-lib/browse.cfm?elib=14195 (http://www.aes.org/e-lib/browse.cfm?elib=14195)
The discussion about ABX testing brought out the usual straw men. Interestingly, George and Peter (not sure which Peter!) had written a letter to the AES journal criticising the Meyer and Moran paper, which "they refused to publish". He said they'd asked the wrong question.
The letter to the JAES was cosigned by George Massenburg, Peter Craven, Vicki Melchior, and Wieslaw Woszczyk and was indeed rejected for publication, though John Vanderkooy did set up the on-line forum mainly as a result of the internal debate at the AES over the content of the letter.
Why are these guys writing letters when one simple ABX test with positive results would do far more to advance their case? Just one so-called hi-rez wav file of regular music that won't let itself be downsampled without easy detection, that is all it would take. Look at the collection of hi-rez files that Atkinson must have. Not one will ABX from its 44/16 self with a positive result? Not one?
The explanation is simple - even though they claim the difference is obvious, the so-clalled obvious difference magically goes away when people are forced to discern differences by just listening.
We have been here many times before - one example involed talking to plants, and then there was cold fusion.
However, there was plenty said that would make most people here on HA seethe. The start was a look at some codecs, playing 192/24 vs 128kbps mp3 (ProTool High Quality - is that FhG based?) - and playing the difference signal - i.e. subtracting the original from the downsampled > encoded > decoded > upsampled version. He played a lot of examples of the noise added by the codec, saying "how can we do this to music?".
That's a shame. It's the same bogus demo John Atkinson is 'educating' people with in his essay. GM loses some respect from me for that ..talk about 'asking the wrong question'!
And if he has to play a the difference sample 'over and over and over' in order to train people to hear mp3 artifacts...doesn't that tell him something about how easy it is to make mp3s that *don't* sound bad?
The letter to the JAES was cosigned by George Massenburg, Peter Craven, Vicki Melchior, and Wieslaw Woszczyk and was indeed rejected for publication, though John Vanderkooy did set up the on-line forum mainly as a result of the internal debate at the AES over the content of the letter.
which is here (http://www.aes.org/journal/forum/)btw
If you are really interested in my thoughts and not just trolling, read my essays at [Stereophile]
I'll read through them, though I do wish they weren't quite so wordy.
BTW John, you have often been shown to hear whatever you want to hear...
Really? A specific example would be helpful, Mr. Krueger. This is HA, after all, where expression of opinion [is] required to be supported with evidential data.
Only with respect to statements related to matters of subjective sound quality, John
George was very unhappy about this paper (which we've discussed at length on HA):
http://www.aes.org/e-lib/browse.cfm?elib=14195 (http://www.aes.org/e-lib/browse.cfm?elib=14195)
The discussion about ABX testing brought out the usual straw men. Interestingly, George and Peter (not sure which Peter!) had written a letter to the AES journal criticising the Meyer and Moran paper, which "they refused to publish". He said they'd asked the wrong question.
The letter to the JAES was cosigned by George Massenburg, Peter Craven, Vicki Melchior, and Wieslaw Woszczyk and was indeed rejected for publication, though John Vanderkooy did set up the on-line forum mainly as a result of the internal debate at the AES over the content of the letter.
Why are these guys writing letters when one simple ABX test with positive results would do far more to advance their case? Just one so-called hi-rez wav file of regular music that won't let itself be downsampled without easy detection, that is all it would take.
Members of a private mail list that includes the above authors of the letter, as well as many other AES illuminati, have been having an active discussion about designing such a test. Far being it simple as you suggest, Mr. Krueger, it seems very difficult to design a test where the _only_ variable is the sample rate _alone_. Without that being attained, the test produces noisy results, with potential false positives as well as false negatives. Of course, if you are only concerned with obtaining a null result, the design of the test becomes much easier, vide the Meyer-Moran paper. :-)
John Atkinson
Editor, Stereophile
The start was a look at some codecs, playing 192/24 vs 128kbps mp3 (ProTool High Quality - is that FhG based?) - and playing the difference signal - i.e. subtracting the original from the downsampled > encoded > decoded > upsampled version. He played a lot of examples of the noise added by the codec, saying "how can we do this to music?".
I know that this is a trick performed by a lot of people who hate on lossy. What happens, I ask naively, if you take such a difference signal and subtract it from the original signal. Does it end up being something like the lossy encode? Would that be an example of how much information you can throw away without making an audible difference? Or does it not work like that?
The start was a look at some codecs, playing 192/24 vs 128kbps mp3 (ProTool High Quality - is that FhG based?) - and playing the difference signal - i.e. subtracting the original from the downsampled > encoded > decoded > upsampled version. He played a lot of examples of the noise added by the codec, saying "how can we do this to music?".
I know that this is a trick performed by a lot of people who hate on lossy.
Yes, it is like failing the qualification exam for Y2K audio production techniques. If you get why subtrasction is a bad idea, then you have somehow figured out some useful stuff about how modern audio works. If you think subtraction is a relevant test, then you badly need a technical update class.
Ironic that widely-respected dudes like Atkinson and Massenberg make such a big public show of stepping in this doggy-do-do on the sidewalk. Their feet smell of the do-do. Many well-informed teenagers (and older) smell the stink. Guys who were once hip now make themselves look like clueless old turkeys.
Life lesson: Part of growing older is watching some of your heroes start looking like bozos.
Warning about life lesson: its not always what it seems, but in this case it very much appears to be exactly what it seems. Sad.
What happens, I ask naively, if you take such a difference signal and subtract it from the original signal.
Or add it. Doesn't much matter.
Do the math.
O = original, R = recording, and D = O - R. Then R + D = R + O - R = O,
and
O - D = O - O + R = R
Does it end up being something like the lossy encode?
Something like.
Would that be an example of how much information you can throw away without making an audible difference?
Yes, but the idea that you can throw something away is anathema to people who believe that the problem with modern recordings is what is being thrown away.
And at one point they are right. What's wrong with modern recordings *is* what is being thrown away.
The current problem *is* that the throwing away happens *before* the sound is converted into an electrical signal.
After all these many years of there being many faults with the handling of the electrical signal, they can't get the idea that the handling of the electrical signal is good enough that the remaining problems are elsewhere. In fact we can handle the electrical signal with considerably but stategically-designed slop (we call this slop modern lossy compression), and get away with it.
I'm fairly new around here. I don't think I'm as "objectivist" as many here, as I simply have no patience for running ABX tests and listening for insignificant differences when I could be listening to music. I know I don't have the technical knowledge many of you have, as I don't quite understand some of what you're saying. Still, I'm driven here because the nonsense I was reading (and not hearing) on many audiophile forums was maddening. The language is laughable -- musical, euphonic, PRaT -- music is musical, and PRaT? As if your equipment can affect timing of the performance? I sometimes wonder if the inventors of such terms are sad, would-be musicians unsatisfied with their passive role in the art. But the language doesn't need to cross over into such absurdities to be highly suspect. I can't even count the number of times audiophiles and even professional reviewers have referred to the performance of components with simple terms like "muddy" or "harsh," and while these are vague enough, we all pretty much understand that the former means attenuated highs, the latter boosted ones (or distortion associated with them). These are simple characteristics that, if they can be heard, could easily be measured and reported on as distortion and/or frequency response.
But they almost never are.
I think, ultimately, listening is subjective. If you like attenuated highs and rich (distorted) mids, that's fine with me. Buy valves and be happy. If you like exaggerated highs and the illusion of detail they create, I'm good with that too. Buy some Naim gear and speakers with nice metal tweeters; that should do the trick. I won't even require that you ABX test it if it makes you happy. The trouble is, the audiophiles I've run into on most of these boards aren't satisfied with being happy. They have to be right. They have to tell me that their personal choices in tone are the more natural reproduction of the source material.
And I just shake my head in wonder...
When it comes from someone like Mr. Atkinson, I shake my head a bit sadly. He has the resources at his disposal. If he hears muddy, he can verify it and report on it in meaningful terms. He not only has that ability, he has that responsibility. I don't know which came first, the lazy editorial content or the lazy engineering, but somewhere along the line, most of the "high-end" abandoned the pursuit of high-fidelity in favor of the pursuit of tone wrapped in elegant industrial art. And in the meantime, it seems that much of the "mid-fi" has passed them by. If accuracy and transparency is my goal, I would trust an integrated amp from Cambridge Audio or Yamaha (just a couple of many examples, I'm sure), far above the overwhelming majority of high-end audiophile products.
How sad is that?
Sorry for the long post. I needed to get that off my chest.
Tim
The letter to the JAES was cosigned by George Massenburg, Peter Craven, Vicki Melchior, and Wieslaw Woszczyk and was indeed rejected for publication, though John Vanderkooy did set up the on-line forum mainly as a result of the internal debate at the AES over the content of the letter.
which is here (http://www.aes.org/journal/forum/)btw
I checked the link and found 9 comments about "Audibility of a CD-Standard A/DA/A Loop Inserted into High-Resolution Audio Playback". I see none by the authors listed about.
What am I missing?
The language is laughable -- musical, euphonic, PRaT -- music is musical, and PRaT? As if your equipment can affect timing of the performance?
Try experimenting with a traditional analog compressor or expander. While, of course, the timing of the notes isn't affected, the listener's _perception_ of that timing can be by such a device. See also http://www.stereophile.com/reference/23 (http://www.stereophile.com/reference/23) .
And, of course, lossy compressors set to low bit rates can affect the perceived rhythm of the music. The transient information will be smeared in time. In the recent blind comparisons of uncompressed and lossy-compressed music that I performed at public seminars in Colorado - see the "Why We Need Audiophiles" thread for details - the listeners in 10 separate presentations in 3 different cities spontaneously offered words like "lifeless" "uninvolving," and "flat" to describe the lowest-rate version after the test had concluded, all of which suggest a change in the music's perceived dynamics and/or rhythm.
When it comes from someone like Mr. Atkinson, I shake my head a bit sadly. He has the resources at his disposal. If he hears muddy, he can verify it and report on it in meaningful terms.
I do try to correlate the sonic descriptions of the audio components we review with their measured performance, extensively. Perhaps you are you not a regular reader of Stereophile? If not, you can find a glossary of the descriptors used in subjective reviewing in a consistent manner in Stereophile at http://www.stereophile.com/reference/50/ (http://www.stereophile.com/reference/50/) .
But to address your point, the problem is that even a simple descriptor like "muddy" can be due to many different measured errors. It could be due to a loudspeaker cabinet panel resonance, an internal air-space resonance, an underdamped woofer alignment, a response aberration due to a power amplifier with a pathologically high output impedance, an unfortunate coincidence between the tuning frequency of the speaker's woofer/port and a room resonance, a high level of even-order harmonic distortion at low frequencies, and on and on...or even to _all_ of these to a greater or lesser degree.
As this is the case for every descriptor, the task you wish for is not so simple or obvious as you seem to be suggesting.
John Atkinson
Editor, Stereophile
Try experimenting with a traditional analog compressor or expander. While, of course, the timing of the notes isn't affected, the listener's _perception_ of that timing can be by such a device.
Well, given that what audiophiles and reviewers are typically talking about when they use the term PRaT is things like amplifiers and DACs, not the effects of analog studio gear, I'm not sure this is relevant. And even then, I'm almost certain that the use of the terms "pace, rhythm and timing" are inappropriate on the verge of absurd. Thousands of musicians have learned material off of cassette decks and turntables, devices so imprecise that instruments had to be re-tuned slightly to even play along, and it didn't degrade their ability to follow the pace, rhythm and timing of the music in the least.
And, of course, lossy compressors set to low bit rates can affect the perceived rhythm of the music. The transient information will be smeared in time.
There we go...there was already a term for the minor effects you're talking about...transient response. So why do you suppose the audiophile community felt compelled to invent a new, self-important and grossly imprecise term? What is it about precision of measurement and language that does not fill their needs?
I do try to correlate the sonic descriptions of the audio components we review with their measured performance, extensively. Perhaps you are you not a regular reader of Stereophile?
Glad to hear it. No, I'm not a regular reader. I was in the old days, of Julian Hersch (sp?) and Stereo
Review. You guys lost me along the way, coincident, I think, to my discovery that pro studio gear handily out-performed much of the stuff that you were enthusiastically recommending at many times its cost.
But to address your point, the problem is that even a simple descriptor like "muddy" can be due to many different measured errors.
Yes, and what I'm suggesting is that they be described in relation to their causes and
by their effects. In other words, rather than simply saying something sounds muddy, which is pretty meaningless, say that the pathologically high impedance causes a frequency response dip of X db from X khz to X khz, resulting in...and then, as far as I'm concerned, you can characterize it as you please. Muddy. Gooey. Stanky. Once you've given me the cause and effect, have at it with the descriptive prose. But as I said, I'm not a reader. If you are already doing this in your reviews, I apologize for referring to your post in my original diatribe.
I'm sure you're right that what I wish for is not as simple as I'd like it to be, but it seems that this hobby gave it a much greater effort in the 70s than it does today, and that manufacturers responded by pursuing fidelity instead of style. A good thing, in my estimation. Maybe I've been reading too many forums and not enough professional reviews. This much I'm pretty sure of: I've been listening, and I'm left with the impression that music lovers would get as good, if not better performance from a $1000 CA or Yamaha as they would from most grotesquely expensive high-end gear. If your magazine is running the pricey boutique stuff head-to-head against the best of what audiophiles look down their noses at and call "mid-fi," comprehensively measuring them against each other, and framing your findings in the science and engineering of sound before you indulge in the kind of audiophile poetry I've seen so much of, sign me up for a subscription.
Tim
What happens, I ask naively, if you take such a difference signal and subtract it from the original signal.
Or add it. Doesn't much matter.
Do the math.
Well, the point of my question was that I know I don't know enough maths to do the math.
Does it end up being something like the lossy encode?
Something like.
Am I worthy of further explanation?
What happens, I ask naively, if you take such a difference signal and subtract it from the original signal.
Or add it. Doesn't much matter.
Do the math.
Well, the point of my question was that I know I don't know enough maths to do the math.
Does it end up being something like the lossy encode?
Something like.
Am I worthy of further explanation?
I am very, very sorry. I have 4 goals when I post (which I probably fail at terribly).
(1) *Not* talk over people's heads, talk down, or seem pompous.
(2) Be economical with my time and words because I do have a life, appearances notwithstaning.
(3) Not waste time with trolls and arrogant people, other than maybe to have a little fun.
(4) Do help people understand audio, and BTW help myself by learning from other people regardless of where they are coming from.
I did not consider the possibility that some people who post here might not be comfortable with any algebra. My bad.
O = original, R = recording, and D = O - R.
IOW, the difference (D) is defined here to be the origional (O) minus the recording ®. That's a choice I made. If I say that the difference is the Recording minus the Origional, then that difference is inverted from what I said. No biggie.
Then R + D = R + O - R = O,
If I add my Difference to the recording then I get back the Origional.
and
O - D = O - O + R = R
And if I subtract my Difference from the Original then I get the Recording, which I believe was your question.
So the answer to your question is "Yes".
I do try to correlate the sonic descriptions of the audio components we review with their measured performance, extensively.
this is the step where you should think twice.
even a simple descriptor like "muddy" can be due to many different measured errors. It could be due to a loudspeaker cabinet panel resonance, an internal air-space resonance, an underdamped woofer alignment, a response aberration due to a power amplifier with a pathologically high output impedance, an unfortunate coincidence between the tuning frequency of the speaker's woofer/port and a room resonance, a high level of even-order harmonic distortion at low frequencies, and on and on...or even to _all_ of these to a greater or lesser degree.
dear HA readers could think that the words like this may be related to specific phenomena;
http://dictionary.classic.reference.com/browse/egotism (http://dictionary.classic.reference.com/browse/egotism)
smokey pokey,
HA, very important
video-section coordinator
edit: better url
I do try to correlate the sonic descriptions of the audio components we review with their measured performance, extensively.
this is the step where you should think twice.
Thank you. I was merely pointing out to "tfarney" that a) he was suggesting something that the magazine already does, to the extent that it is possible, and b) his criticism was not actually based on firsthand knowledge of the magazine's content.
even a simple descriptor like "muddy" can be due to many different measured errors. It could be due to a loudspeaker cabinet panel resonance, an internal air-space resonance, an underdamped woofer alignment, a response aberration due to a power amplifier with a pathologically high output impedance, an unfortunate coincidence between the tuning frequency of the speaker's woofer/port and a room resonance, a high level of even-order harmonic distortion at low frequencies, and on and on...or even to _all_ of these to a greater or lesser degree.
dear HA readers could think that the words like this may be related to specific phenomena;
http://dictionary.classic.reference.com/browse/egotism (http://dictionary.classic.reference.com/browse/egotism)
My apologies,but I fail to see the connection between my posting and your comment. Assuming that you are not just making an ad hominem comment, are you suggesting that the measurable aberrations I listed do _not_ have any correlation with a listener's description of a sound being "muddy" (the specific descriptor mentioned by "tfarney")? Of course they do, but to pin down _which_ is actually the culprit could take considerably more time than is possible for a monthly review magazine. If you read Stereophile's reviews, you will see that we a) describe the sound quality of a component, using a consistent and transportable terminology as listed in the article I linked to, and b) we comment on the measured performance where that might suggest possible reasons for the sonic description. Why would anyone have a problem with that?
John Atkinson
Editor, Stereophile
If you read Stereophile's reviews, you will see that we a) describe the sound quality of a component, using a consistent and transportable terminology as listed in the article I linked to, and b)...
sorry, i can't read that...., now if this was web only, you would be forgiven, but you are actually killing trees to print that...
The letter to the JAES was cosigned by George Massenburg, Peter Craven, Vicki Melchior, and Wieslaw Woszczyk and was indeed rejected for publication, though John Vanderkooy did set up the on-line forum mainly as a result of the internal debate at the AES over the content of the letter.
which is here (http://www.aes.org/journal/forum/)btw
I checked the link and found 9 comments about "Audibility of a CD-Standard A/DA/A Loop Inserted into High-Resolution Audio Playback". I see none by the authors listed about.
What am I missing?
'here' meant 'the forum is here', not, 'the letter is here'.
And, of course, lossy compressors set to low bit rates can affect the perceived rhythm of the music. The transient information will be smeared in time. In the recent blind comparisons of uncompressed and lossy-compressed music that I performed at public seminars in Colorado - see the "Why We Need Audiophiles" thread for details - the listeners in 10 separate presentations in 3 different cities spontaneously offered words like "lifeless" "uninvolving," and "flat" to describe the lowest-rate version after the test had concluded, all of which suggest a change in the music's perceived dynamics and/or rhythm.
To actually affect the perceived
rhythm of a track the smearing would have to
pretty damn bad. And *dynamics* is something else completely. Your demo was dubious for reasons cited already. What codecs, what rates, what samples were used? How were listener and presenter bias accounted for? Was this before or after the bogus and *highly* biasing 'look a what MP3s leave out!' demonstration?
When it comes from someone like Mr. Atkinson, I shake my head a bit sadly.
You're not the only one, tfarney.
Wait, rhythm?
This is a joke, right???
Members of a private mail list that includes the above authors of the letter, as well as many other AES illuminati, have been having an active discussion about designing such a test. Far being it simple as you suggest, Mr. Krueger, it seems very difficult to design a test where the _only_ variable is the sample rate _alone_.
Only in some people's minds. No doubt the real problem is political, laced with a heavy dose of fear of the dark, along with fear of unfavorable illumination.
Without that being attained, the test produces noisy results,
If the goals for the test have not been attained, how can there be any discusison of results?
The only noise I hear is silence from people who can't back up their claims with a practical test that proves their point.
So-called Hi-rez recordings have been on the market for over 5 years, thousands of titles have been released, and yet nobody can come forward with even one example of a recording that sounds unambigiously different when cleanly downsampled.
... with potential false positives as well as false negatives.
IOW people are using the classic strategy for keeping the test from ever happening - demanding perfection.
Of course, if you are only concerned with obtaining a null result, the design of the test becomes much easier, vide the Meyer-Moran paper. :-)
As usual John, you're ignoring all of the vocal claims from your side of the fence that the audible benefits of higher sample rates are so obvious that only nearly deaf people listening to horrifically made recordings on incredibly bad systems, can't hear them.
The fact of the matter John is that you can't provide a hi-rez example of a typical kind of musical recording with none of the usual tricks that your side likes to pull like evaluating the end of a fade-out starting at -40 dB, or using a recording where the peaks are at -40 dB, etc., that can't survive downsampling without there being an audible difference.
Of course, in any fair evaluation, your side won't be able bring their favorite crutches, like sighted evaluations, unmatched levels, and non-sychronized comparisons.
IOW, ethical and mental cripples need not apply.
If you read Stereophile's reviews, you will see that we a) describe the sound quality of a component, using a consistent and transportable terminology as listed in the article I linked to, and b)...
sorry, i can't read that...., now if this was web only, you would be forgiven, but you are actually killing trees to print that...
I did give the link to the Web reprint of the article earlier in the thread: http://www.stereophile.com/reference/50/ (http://www.stereophile.com/reference/50/) .
And the Web archive has currently approaching 1000 reviews that illustrate the point I was making: http://www.stereophile.com/equipmentreviews/ (http://www.stereophile.com/equipmentreviews/) .
John Atkinson
Editor, Stereophile
And, of course, lossy compressors set to low bit rates can affect the perceived rhythm of the music. The transient information will be smeared in time. In the recent blind comparisons of uncompressed and lossy-compressed music that I performed at public seminars in Colorado - see the "Why We Need Audiophiles" thread for details - the listeners in 10 separate presentations in 3 different cities spontaneously offered words like "lifeless" "uninvolving," and "flat" to describe the lowest-rate version after the test had concluded, all of which suggest a change in the music's perceived dynamics and/or rhythm.
To actually affect the perceived rhythm of a track the smearing would have to pretty damn bad.
It doesn't have be dreadful, just enough to be noticed. Tony Faulkner wrote an article on this subject in the January 2009 issue of HiFi Critic, BTW. Worth reading.
Your demo was dubious for reasons cited already.
Why was it dubious? Levels were matched precisely, the lossy compressed and sample-rate-reduced files were reconverted to 24 bits and 88.2kHz sample rate so that the DAC would be behaving identically with all the examples.
What codecs, what rates, what samples were used? How were listener and presenter bias accounted for?
I have answered all your questions before, Mr. Sullivan, in the "audiophiles" thread. Either I can cut'npaste my reponses or you can do the work of looking them up for yourself. (My preference is for the latter.)
Was this before or after the bogus and *highly* biasing 'look a what MP3s leave out!' demonstration?
You mean like George Massenburg's? Because many more people signed up for my demonstrations than the organizers had predicted, I had to omit that example. The demo of lossy compression consisted of 2 parts: first a "learning" section, where I identified which was the full resolution file and which was the data-reduced file and pointed out aspects of the sound that had changed; second, a single-blind section, using a different piece of music, which followed Phil Hobbs' protocol, where the data density was progressively reduced through the 5 minutes of the piece. The listeners were not aware of what they were listening to and stood to the side, out of the line of sight while the music played. At the conclusion of the presentation, I asked them to describe what they heard. Almost everyone said they felt the sound got worse, as I described, no-one felt it sounded better as the piece progressed..
Feel free to disregard the results. But around 200 audiophiles took part in the 10 sessions, which meant that all the listeners could get a good sound, and the overall reaction was very positive, in that people appreciate learning things heuristically under non-threatening conditions.
John Atkinson
Editor, Stereophile
Members of a private mail list that includes the above authors of the letter, as well as many other AES illuminati, have been having an active discussion about designing such a test. Far being it simple as you suggest, Mr. Krueger, it seems very difficult to design a test where the _only_ variable is the sample rate _alone_.
Only in some people's minds. No doubt the real problem is political, laced with a heavy dose of fear of the dark, along with fear of unfavorable illumination.
<snip.
As usual John, you're ignoring all of the vocal claims from your side of the fence that the audible benefits of higher sample rates are so obvious that only nearly deaf people listening to horrifically made recordings on incredibly bad systems, can't hear them.
The fact of the matter John is that you can't provide a hi-rez example of a typical kind of musical recording with none of the usual tricks that your side likes to pull...IOW, ethical and mental cripples need not apply.
Is it really not possible for you to discuss anything without maintaining a barrage of ad hominem statements, Mr. Krueger? The people involved in this effort are hardly "ethical and mental cripples," but include several AES Fellows, some university professors, some well-known recording and mastering engineers, and even JJ. And no-one involved, least of all me, has made any accusations of deafness, or having "incredibly bad systems," or using "horrifically made recordings" about people like you who claim not to hear any advantage from bit depths greater than 16 or sample rates higher than 44.1kHz. Those are projections on your part.
John Atkinson
Editor, Stereophile
The demo of lossy compression consisted of 2 parts: first a "learning" section, where I identified which was the full resolution file and which was the data-reduced file and pointed out aspects of the sound that had changed; second, a single-blind section, using a different piece of music, which followed Phil Hobbs' protocol, where the data density was progressively reduced through the 5 minutes of the piece. The listeners were not aware of what they were listening to and stood to the side, out of the line of sight while the music played. At the conclusion of the presentation, I asked them to describe what they heard. Almost everyone said they felt the sound got worse, as I described, no-one felt it sounded better as the piece progressed..
Glad you are mentioning this, Mr. Atkinson. I attended Phil Hobbs' "DVD-A - CD - MP3 192 kbps - MP3 96 kbps" demo at the 31st AES International Conference in London two years ago (http://www.aes.org/events/reports/31stConference.pdf) (actually, George Massenburg sat right behind me back then, remarking twice afterwards, "that's a great demo!"). I assume you followed the same format configuration? Did you also prepare a reverse demo, i.e. one with "CD followed by DVD-A"? It would be very interesting to see if listeners could then still hear the difference between the CD quality and the hi-res reproduction. If so, there might actually be an audible difference between the two formats. If not, one could argue that a perceived difference, i.e. sound degradation, between "DVD-A followed by CD quality" was just due to auditory fatigue kicking in.
Side note: In the Hobbs demo I attended, I thought I heard a difference between DVD-A and CD, and I definitely heard a difference between MP3 192 kbps and MP3 96 kbps.
Chris
ok, so now we have egotism, different commercials and HA rule breaking in one place already, can we please do something like, either:
a. ban somebody
b. close this thread
c. move the thread to recycle bin
Members of a private mail list that includes the above authors of the letter, as well as many other AES illuminati, have been having an active discussion about designing such a test. Far being it simple as you suggest, Mr. Krueger, it seems very difficult to design a test where the _only_ variable is the sample rate _alone.
Only in some people's minds. No doubt the real problem is political, laced with a heavy dose of fear of the dark, along with fear of unfavorable illumination.
<snip>
As usual John, you're ignoring all of the vocal claims from your side of the fence that the audible benefits of higher sample rates are so obvious that only nearly deaf people listening to horrifically made recordings on incredibly bad systems, can't hear them.
The fact of the matter John is that you can't provide a hi-rez example of a typical kind of musical recording with none of the usual tricks that your side likes to pull...IOW, ethical and mental cripples need not apply.
Is it really not possible for you to discuss anything without maintaining a barrage of ad hominem statements, Mr. Krueger?
So John, you again obfuscate your inability to deliver that simple little recording by whining about the kinds of name-calling that you yourself frequently descend into?
The people involved in this effort are hardly "ethical and mental cripples,"
I would sincerily hope so - which is what I meant by: "...ethical and mental cripples need not apply..."
...but include several AES Fellows, some university professors, some well-known recording and mastering engineers, and even JJ.
Right John, so there is a reasonble hope that the kind of highly fallible listening evaluation procedures that you've built your ragazine on won't be allowed.
And no-one involved, least of all me, has made any accusations of deafness, or having "incredibly bad systems," or using "horrifically made recordings" about people like you who claim not to hear any advantage from bit depths greater than 16 or sample rates higher than 44.1kHz.
First off John, you're making a negative hypothesis - you're projecting that you know every word that every one of these people have said at every time in their lives. This is just as rediculous as your claims that nobody from Stereopile posts under aliases. A reasonable person would admit that you really have no way to know for sure what these people have said and done for every second of their lives.
Secondly John, I never said that *everybody* who is on that task force or that you yourself have ever said such a thing. I said that some people on your side of the fence have said things like that. So now John, are you expecting us to believe that you know what each and every person who is on your side of this issue has ever said?
John you need to admit that can't provide a recording that unambigiously justifies your religous belief in hi rez recordings. If you could, you obviously would have by now. But you can't bring yourself to admit it, presumably not even to yourself. When it comes to your projections of your omniscience or your projections of the validity of your religous beliefs about audio, you are as fact-challenged as many of the openly religious people that you criticize.
Members of a private mail list that includes the above authors of the letter, as well as many other AES illuminati, have been having an active discussion about designing such a test. Far being it simple as you suggest, Mr. Krueger, it seems very difficult to design a test where the _only_ variable is the sample rate _alone_.
This seems to imply that 'high-resolution' files - which have been commercially available for some not inconsiderable time - have never been robustly tested against 16bit/44.1kHz files. If that is the case, isn't promoting these files as notionally 'better' than 16bit/44.1kHz potentially doing a disservice to listeners?
If and when such a test becomes available, will you publish the findings whatever the results? If the results of such investigation determines that there is no audible advantage over 16bit/44.1kHz files, you will have to make a fundamental
volte face in print. Specifically - we heard something that wasn't really there.
Personally, I wouldn't want to stake my reputation on something so provisional.
The demo of lossy compression consisted of 2 parts: first a "learning" section, where I identified which was the full resolution file and which was the data-reduced file and pointed out aspects of the sound that had changed; second, a single-blind section, using a different piece of music, which followed Phil Hobbs' protocol, where the data density was progressively reduced through the 5 minutes of the piece. The listeners were not aware of what they were listening to and stood to the side, out of the line of sight while the music played. At the conclusion of the presentation, I asked them to describe what they heard. Almost everyone said they felt the sound got worse, as I described, no-one felt it sounded better as the piece progressed..
Glad you are mentioning this, Mr. Atkinson. I attended Phil Hobbs' "DVD-A - CD - MP3 192 kbps - MP3 96 kbps" demo at the 31st AES International Conference in London two years ago (http://www.aes.org/events/reports/31stConference.pdf) (actually, George Massenburg sat right behind me back then, remarking twice afterwards, "that's a great demo!"). I assume you followed the same format configuration? Did you also prepare a reverse demo, i.e. one with "CD followed by DVD-A"?
Time was limited, but in most of the Colorado presentations last month, while I did not play a reverse version in full, I did follow the 16/44.1k/128kbps MP3 with 24/88.2k non-compressed.
I was also at Philip's London dem. You can find my resultant comments at http://www.stereophile.com/asweseeit/1007awsi/ (http://www.stereophile.com/asweseeit/1007awsi/) . (My apologies to the HA moderators if my providing links is regarded as "commercials," but it really does make more sense to provide a URL to my prior published comments rather than resort to lengthy and time-consuming cut'n'pastes.)
john Atkinson
Editor, Stereophile
Members of a private mail list that includes the above authors of the letter, as well as many other AES illuminati, have been having an active discussion about designing such a test. Far being it simple as you suggest, Mr. Krueger, it seems very difficult to design a test where the _only_ variable is the sample rate _alone.
Only in some people's minds. No doubt the real problem is political, laced with a heavy dose of fear of the dark, along with fear of unfavorable illumination.
<snip>
As usual John, you're ignoring all of the vocal claims from your side of the fence that the audible benefits of higher sample rates are so obvious that only nearly deaf people listening to horrifically made recordings on incredibly bad systems, can't hear them.
The fact of the matter John is that you can't provide a hi-rez example of a typical kind of musical recording with none of the usual tricks that your side likes to pull...IOW, ethical and mental cripples need not apply.
Is it really not possible for you to discuss anything without maintaining a barrage of ad hominem statements, Mr. Krueger?
So John, you again obfuscate your inability to deliver that simple little recording by whining about the kinds of name-calling that you yourself frequently descend into?
I try very hard to address the argument, not the arguer, Mr. Krueger. If people examine this or the "Audiophiles..." thread on HA, they will note that I have almost entirely been successful at achieving that goal, despite you and others resorting to insults and name-calling, as in the examples I quoted above.
The people involved in this effort are hardly "ethical and mental cripples," but include several AES Fellows, some university professors, some well-known recording and mastering engineers, and even JJ...And no-one involved, least of all me, has made any accusations of deafness, or having "incredibly bad systems," or using "horrifically made recordings" about people like you who claim not to hear any advantage from bit depths greater than 16 or sample rates higher than 44.1kHz.
First off John, you're making a negative hypothesis - you're projecting that you know every word that every one of these people have said at every time in their lives.
No, I am saying that neither me nor any of the people in the mailing list, which I keep archived, has said anything like you accuse them of in that list's content, Mr. Kreuger. In addition, not one of them, many of whom have been interviewed in Stereophile, has said anything like you have accused them of in print. And not one of them has said anything like it in the many personal conversations I have had with them over the years. If you have evidence to the contrary, please produce it.
This is just as [ridiculous] as your claims that nobody from Stereopile posts under aliases.
No-one at Stereophile posts to this or any other group other than using their own name, or in the case of Sam Tellig, his long-used non-de-plume. In the, what, 10 years you have been publicly making this accusation, Mr. Krueger, you have not managed to produce one iota of evidence that I am incorrect. Please either produce your evidence or admit that you have no such evidence, Mr. Krueger. Thank you.
John Atkinson
Editor, Stereophile
And if I subtract my Difference from the Original then I get the Recording, which I believe was your question.
So the answer to your question is "Yes".
Thank you, Arnold. I really didn't know whether the maths was as simple as addition and subtraction--that's what I thought, but my mathematical naivety is positively virginal.
Does this, then, mean that you could confound the people who play difference files by playing a file consisting of original - difference, in the confident expectation that there would be no audible distinguishability, assuming you were using a suitable quality of lossy compression for the start of the whole charade.
You could also con people into doing the test single blind by announcing which file was which, but "accidentally" getting it the wrong way round, and correcting yourself later. Ethics? I've heard of them.
b. close this thread
c. move the thread to recycle bin
Please keep the few posts related to the original topic.
There's a common arguing technique from John - discuss an audible problem (e.g. Pace Rhythm and Timing, noise) - apply it to something like an amplifier where it's usually inaudible, but then justify it using an example where it's clearly audible - e.g. a compressor, or low bitrate mp3.
The implication is that because the problem is audible with the example, it may also be audible in the amplifier.
The idea that something really can be "good enough" is an anathema.
This approach is the antithesis of real engineering, where problems are quantified, and reduced down to the level at which they are no longer problems.
But where's the fun in that? Let's play people 96kps or 128kbps mp3 instead.
(I note people rarely use lame --vbr-new -V2 for demos where they want to demonize mp3 - despite the fact that many of us here can ABX 320kbps! Heaven forbid it should get out that we are far more sensitive to these artefacts than self-proclaimed DBT-shunning audio critics)
Cheers,
David.
Members of a private mail list that includes the above authors of the letter, as well as many other AES illuminati, have been having an active discussion about designing such a test. Far being it simple as you suggest, Mr. Krueger, it seems very difficult to design a test where the _only_ variable is the sample rate _alone_.
This seems to imply that 'high-resolution' files - which have been commercially available for some not inconsiderable time - have never been robustly tested against 16bit/44.1kHz files. If that is the case, isn't promoting these files as notionally 'better' than 16bit/44.1kHz potentially doing a disservice to listeners?
Don't be fooled. The testing has been very robust, actually too robust for the promoters of so-called hi-rez audio. That's the problem. They are trying to finagle a deal, but with people like JJ on the scene, their finagling will be to no avail.
Testing where the only variable is the minimum sample rate and word length that the music has been recorded at is the only varable is actually quite simple. Let the high sample rate musical selection be your reference. There's no problem obtaining a computer with an audio interface that will effectively play it.
Now, take the high sample rate music and downsample it. Then upsample it back to its origional sample rate. Verify that timing and signal levels have remained within tolerances (usually automatic). 5 minutes work with any number of different audio editors. CoolEdit Pro 2.x, which has been around since 2001 or so, will do the job.
Now, you have two high sample rate files to compare. If you can hear a difference then there are all sorts of questions. Was it the downsampling? Was it the upsampling? However, if you can hear no difference, then there are no questions. The processing where the sample rate was the only variable, was sonically innocent.
So, all you need is these two files that took 5 minutes to produce, a copy of WinABX or Foobar or other similar software, and a computer with an appropriate audio interface and monitoring system.
How robust does this test need to be?
Dozens if not 100s of people have done tests like these. The results are that if the test has no other glitches, then the results are null.
But see here, the test I described above is all science.
What about life in the consumer marketplace? Both SACD and DVD-A were on the market for years and had the opportunity to impress or not impress millions of consumers. They both failed dismally. At best they have collapsed into minor niche products. They were solutions looking for a problem to solve that turned out to be non-existent. Their brands have no impact and no wider recognition. The function of high-rez audio has been subsumed into the DTS and Dolby hi rez offerings. New media such as Blu Ray are just branded with the DTS or Dolby brands and few if anybody knows or cares what is under the brand name. Why should anybody care?
Meanwhile
Does this, then, mean that you could confound the people who play difference files by playing a file consisting of original - difference, in the confident expectation that there would be no audible distinguishability, assuming you were using a suitable quality of lossy compression for the start of the whole charade.
Yes. As long as you preserve the files accurately at the bit-by-bit level, you can do the arithmetic and have the desired results.
You could also con people into doing the test single blind by announcing which file was which, but "accidentally" getting it the wrong way round, and correcting yourself later. Ethics? I've heard of them.
This has been done. Obviously the ethical thing to do is just be up front. But tricks have been played on people.
One such trick happened at a meeting of the audio club to which I belong - SMWTMS.
At once clubmeeting some years back, our host was a very successful car dealer who had made a major investment in high end audio components. He was eagerly demonstrating his new equipment to anybody who would listen. At one point his teenaged son secretly substituted a cheap Pioneer receiver for the high end electronics. His father continued to eagerly demonstrate his "new equipment" to anybody who would listen, with no clue that the change had taken place. Nobody else heard any difference, either.
I'm pleased to say that I recently received an email from the son, so no harm befell him as a consequence of his little adventure.
to affect rhythm of a track the smearing would have to pretty damn bad.
It doesn't have be dreadful, just enough to be noticed. Tony Faulkner wrote an article on this subject in the January 2009 issue of HiFi Critic, BTW. Worth reading.
Don't see it on the ToC for the Jan/March 09 issue (http://www.hificritic.com/downloads/CONTENTS13.PDF), but I'll take your word for it that it was there. But I can't access it. So, what was his proof of this remarkable claim that the actual *rhythm* of a piece of music is affected by lossy encoding, that would make it worth reading?
Your demo was dubious for reasons cited already.
Why was it dubious? Levels were matched precisely, the lossy compressed and sample-rate-reduced files were reconverted to 24 bits and 88.2kHz sample rate so that the DAC would be behaving identically with all the examples.
I have answered all your questions before, Mr. Sullivan, in the "audiophiles" thread. Either I can cut'npaste my reponses or you can do the work of looking them up for yourself. (My preference is for the latter.)
I don't bookmark your posts, Mr. Atkinson -- most of which appear to be piddling arguments with Arny anyway -- and here on HA, when you are going to claim *anywhere* that A sounded worse than B, it's incumbent upon *you* to give the details either directly or by link. Do you just not
remember at least what codec and bitrates and sample rates and formats you used, or what?
Was this before or after the bogus and *highly* biasing 'look a what MP3s leave out!' demonstration?
You mean like George Massenburg's?
Yes, I do mean like George Massenburg's. His demo,as described by 2bdecided, was misleading too, as I wrote earlier in this thread (http://www.hydrogenaudio.org/forums/index.php?showtopic=72446&view=findpost&p=639263).
Unfortunately, his demo, as described, was another illustration of the fact that being a revered mastering engineer does not automatically make you qualified to conduct good psychoacoustics tests. Worse still because I suspect Mr. Massenburg really does know how mp3s 'work', and he should know 'why' they work too...and therefore why 'difference signal' demos are misleading.
Why don't you try to imagine how a demo of 'what mp3 does ' would have been designed and run by someone like , oh, say , *JJ*?
Because many more people signed up for my demonstrations than the organizers had predicted, I had to omit that example. The demo of lossy compression consisted of 2 parts: first a "learning" section, where I identified which was the full resolution file and which was the data-reduced file and pointed out aspects of the sound that had changed; second, a single-blind section, using a different piece of music, which followed Phil Hobbs' protocol, where the data density was progressively reduced through the 5 minutes of the piece. The listeners were not aware of what they were listening to and stood to the side, out of the line of sight while the music played. At the conclusion of the presentation, I asked them to describe what they heard. Almost everyone said they felt the sound got worse, as I described, no-one felt it sounded better as the piece progressed..
So here, you're willing to go into detail on protocol, but get cranky when asked what codec(s) and bitrates and samples were used?
Feel free to disregard the results. But around 200 audiophiles took part in the 10 sessions, which meant that all the listeners could get a good sound, and the overall reaction was very positive, in that people appreciate learning things heuristically under non-threatening conditions.
Probably no surprise to you, but
I could care less what '200 audiophiles' thought of such a demo, Mr. Atkinson.
I'm unclear how this test was considered blind, yet apparently consisted of tracks that were *always* progressing from lossless to increasingly lossy. *OF COURSE* the sound would get worse, if you progressed far enough down through lossy bitrates. And of course the codec would matter too, as would the samples used.
So, the test makes no sense as described, as a means to demonstrate the idea that lossy always audibly degrades lossless. Did you compare tracks that stayed lossless to the progressively lossy ones, blind (preferably '*double* -- you standing off to the side really isn't double-blind)? Did you mark at which bitrates the report of 'difference' kicked in on the progressively lossy ones in the demo you DID do?
There's a common arguing technique from John - discuss an audible problem (e.g. Pace Rhythm and Timing, noise) - apply it to something like an amplifier where it's usually inaudible, but then justify it using an example where it's clearly audible - e.g. a compressor, or low bitrate mp3.
The implication is that because the problem is audible with the example, it may also be audible in the amplifier.
I agree, but I see a larger problem - the idea that if something could be imagined to be true, then it surely is true.
One variation is the idea that if an artifact is measurable, then it is of course audible.
Then there's the antithesis - artifacts either not audible or perhaps even desirable if a property of vinyl or tubed electronics.
The idea that something really can be "good enough" is an anathema.
The problem here is that once something becomes good enough, it is tough to justify upgrades to it.
A question from the floor pointed out that an all-pass filter would give an even larger difference signal, but create little or no audible difference (hence the method is flawed in terms of illustrating the audible difference) - George doubted this example.
I find it astonishing that George would not understand this.
He also should have used 192 kbps, or even 256 unless he has an agenda. Oh wait, he said he doesn't have an agenda.
--Ethan
Try experimenting with a traditional analog compressor or expander. While, of course, the timing of the notes isn't affected, the listener's _perception_ of that timing can be by such a device.
I agree with Tim this is an irrelevant comparison. Yes, compressors have their own timing as they raise and lower the volume level.
And, of course, lossy compressors set to low bit rates can affect the perceived rhythm of the music. The transient information will be smeared in time.
This too is irrelevant in the context of musical tempo. John, as a musician yourself I can't believe you'd defend PRaT when it's clearly a BS description used by people who have no idea what they're even saying. Indeed, audio already has enough perfectly usable - and better - descriptions.
the problem is that even a simple descriptor like "muddy" can be due to many different measured errors. It could be due to a loudspeaker cabinet panel resonance, an internal air-space resonance, an underdamped woofer alignment
Exactly! So why "muddy" the waters with nonsense descriptions when the sensible solution is to use terms such as frequency response and distortion and ringing? If our goal is to educate, then let's do a proper job which means use the correct terms. Using PRaT in a discussion about audio is as silly as using the word pee-pee in a sex-ed class.
--Ethan
Tony Faulkner wrote an article on this subject in the January 2009 issue of HiFi Critic, BTW. Worth reading.
Don't see it on the ToC for the Jan/March 09 issue (http://www.hificritic.com/downloads/CONTENTS13.PDF), but I'll take your word for it that it was there. But I can't access it.
My apologies, I misremembered. It was the December 2007 issue.
I have answered all your questions before, Mr. Sullivan, in the "audiophiles" thread. Either I can cut'npaste my reponses or you can do the work of looking them up for yourself. (My preference is for the latter.)
I don't bookmark your posts, Mr. Atkinson...
I had assumed, Mr. Sullivan, that as you were actively involved in the thread where I offered this information - "Why We Need Audiophiles..." between April 25 and April 28 - and had even asked me questions to do with this then forthcoming Colorado presentation, as well as with the blind tests of amplifiers in which I had been involved, you had made note of the information I offered.
here on HA, when you are going to claim *anywhere* that A sounded worse than B, it's incumbent upon *you* to give the details either directly or by link.
I had assumed that as I had already posted all the information on HA, that the answer to a follow-up question such as yours did not require _re_posting all that information. Perhaps this assumption was incorrect, but otherwise those more interested in trolling than in genuinely seeking information could tie someone up in knots by insisting on strict Tos#8 adherence in _every_ related post.
Do you just not remember at least what codec and bitrates and sample rates and formats you used, or what?
The sighted "learning" comparison was between LPCM at 24-bits/88.2kHz and AAC at 128kbps, derived from a Red Book version of the data, downsampled using the SRC in Bias Peak 5 at its highest-quality setting, which independent tests had shown to be one of the best-performing (link available on request), and reduced to 16 bits using the POWR-2 redithering algorithm.
The single-blind comparison (using a different piece of music) used first, the original PCM at 24-bits/88.2kHz, Red Book PCM prepared in the same manner just described, AAC at 320kbps prepared from the Red BooK PCM, and an MP3 at 128kbps, again derived from the Red Book PCM version using the Fraunhofer codec in Adobe Audition 1.0. The splices between the 4 different versions were seamless, and there was no indication to listeners other than the possible change in sound that anything had changed. Levels were matched precisely, the lossy compressed and sample-rate-reduced files were reconverted to 24 bits and 88.2kHz sample rate so that the DAC would be behaving identically with all the examples. As I said, this was the exact methodology used by Linn's Philips Hobbs at the 2007 AES Conference.
In the discussion in which you were involved at the end of April, Mr. Sullivan, I told you directly in response to a question from you that I was not intending to present formal DBTs, explaining that "given that these demonstrations are open to the public there may by up to 20 people in the listening room, a formal DBT is out of the question. And please note that, as I keep saying, these are demonstrations, not tests. There will be no scoring of listeners' preference. As I have said, I am only interested in exposing listeners to the the various formats. This is so that they can decide for themselves whether a) hi-rez formats are necessary, b) whether CD is good enough for serious listening, and c) whether the lossy versions are sonically compromised or not. Who could argue that that would be a bad or, in your emotionally loaded term, a 'shameful' thing." (Quoted from my response to you dated April 26, 2009, 10:40am.)
As I explained in later messages in the same thread, I was a) allowing listeners to hear for themselves under optimal conditions if the widespead use of the term "CD quality" to describe lossy-compressed audio was appropriate and b) allowing them to audition hi-rez PCM data under optimal circumstances.
The listeners were not aware of what they were listening to and stood to the side, out of the line of sight while the music played. At the conclusion of the presentation, I asked them to describe what they heard. Almost everyone said they felt the sound got worse, as I described, no-one felt it sounded better as the piece progressed.
I'm unclear how this test was considered blind, yet apparently consisted of tracks that were *always* progressing from lossless to increasingly lossy. *OF COURSE* the sound would get worse, if you progressed far enough down through lossy bitrates. And of course the codec would matter too, as would the samples used. So, the test makes no sense as described, as a means to demonstrate the idea that lossy always audibly degrades lossless. Did you compare tracks that stayed lossless to the progressively lossy ones, blind (preferably '*double* -- you standing off to the side really isn't double-blind)? Did you mark at which bitrates the report of 'difference' kicked in on the progressively lossy ones in the demo you DID do?
I made no claim that the second comparison was double-blind. As I wrote, it was single-blind because while I knew what was being played in the second demonstration, the listeners had no clue to what they were listening, just as in Philip Hobbs' test. The only identification was in the first "learning" example I described. And thank you for agreeing - "of course" - that the degradation in quality should have been audible.
Feel free to disregard the results. But around 200 audiophiles took part in the 10 sessions, which meant that all the listeners could get a good sound, and the overall reaction was very positive, in that people appreciate learning things heuristically under non-threatening conditions.
Probably no surprise to you, but I could care less what '200 audiophiles' thought of such a demo, Mr. Atkinson.
Oh well. It must remain a mystery, therefore, why you are so concerned with the details of the test, Mr. Sullivan.
John Atkinson
Editor, Stereophile
...the problem is that even a simple descriptor like "muddy" can be due to many different measured errors. It could be due to a loudspeaker cabinet panel resonance, an internal air-space resonance, an underdamped woofer alignment
Exactly! So why "muddy" the waters with nonsense descriptions when the sensible solution is to use terms such as frequency response and distortion and ringing?
Because you are confusing cause and effect, Ethan. The "cause" may indeed be "frequency response and distortion and ringing" but people do not have any means of directly detecting and identifying those technical abberations. By contrast, the "effect" is what listeners perceive and describe using a familiar vocabulary. This dichotomy was examined, if I remember correctly, in an early 1980s AES paper authored by Subir Pramanik and Soren Bech, in the context of phono cartridge behavior. (I was one of the subjects in the series of blind tests used to generate the raw data for the paper.)
I'll try to find the specific reference.
John Atkinson
Editor, Stereophile
Because you are confusing cause and effect, Ethan. The "cause" may indeed be "frequency response and distortion and ringing" but people do not have any means of directly detecting and identifying those technical abberations. By contrast, the "effect" is what listeners perceive and describe using a familiar vocabulary.
I don't buy it that PRaT is "familiar vocabulary" John. Those are just made-up nonsense words that should have been rejected by the mainstream hi-fi press rather than repeated. Do we even know where those terms originated? Regardless, they are no more descriptive than "aura" and "energy meridians" and "chakra" etc as often used by AM practitioners.
Perhaps you or I should write an article to explain what the real terms are and what they mean, and include MP3 examples. It's easy to convey rolled-off and boosted highs, mids, and lows, and my Resonance (http://www.realtraps.com/video_resonance.htm) demo video explain ringing, though I admit I probably should do that again using better video gear.
Various types of distortion could be explained and illustrated using audio clips too. I'm serious. This really needs to be done unless you know of such an article or video that exists already. Would you be interested in collaborating on a video like this? Maybe if you and I were
both involved it would be more universally accepted.
--Ethan
I was also at Philip's London dem. You can find my resultant comments at http://www.stereophile.com/asweseeit/1007awsi/ (http://www.stereophile.com/asweseeit/1007awsi/) .
Thanks for the link! Now, with all due respect:
He had prepared four versions of the chorus—the original 24-bit/88.2kHz data transcoded straight from the DSD master; a version sample-rate–converted and decimated to 16/44.1 CD data; an MP3 version at 320kbps; and, finally, an MP3 version at 192kbps—and spliced them together in that order.
What a coincidence that I remember the demo just as you do (even your thoughts during the demo, as you describe them, largely coincide with mine),
except for this critical piece of infomation. As I wrote earlier, the MP3s were at 192 and 96 kbps,
not 320 and 192. And yes, I'm 100% sure I remember that correctly (and no, I don't have a proof). And yes, there is no reason for me to believe that we heard different configurations at that conference, or that Mr. Hobbs told us different things during the demo. And yes, "as we see it", there is a
significant quality difference between 192 and 96 kbps. Coincidence? Accident? Typo?
You are beginning to open my eyes regarding the content of your magazine. So thank you again for the link! No further questions.
To all: sorry for being off-topic again, but I just had to post this.
Chris
I was also at Philip's London dem. You can find my resultant comments at http://www.stereophile.com/asweseeit/1007awsi/ (http://www.stereophile.com/asweseeit/1007awsi/) .
Thanks for the link! Now, with all due respect:
He had prepared four versions of the chorus—the original 24-bit/88.2kHz data transcoded straight from the DSD master; a version sample-rate–converted and decimated to 16/44.1 CD data; an MP3 version at 320kbps; and, finally, an MP3 version at 192kbps—and spliced them together in that order.
What a coincidence that I remember the demo just as you do (even your thoughts during the demo, as you describe them, largely coincide with mine), except for this critical piece of infomation. As I wrote earlier, the MP3s were at 192 and 96 kbps, not 320 and 192.
I did check both the identity of the music example and the lossy file bitrates with Philip via email once I had returned to the US before I wrote the linked essay. He did confirm that I had them correct. So I don't know why we came away from the conference with different impressions.
John Atkinson
Editor, Stereophile
Because you are confusing cause and effect, Ethan. The "cause" may indeed be "frequency response and distortion and ringing" but people do not have any means of directly detecting and identifying those technical [aberrations]. By contrast, the "effect" is what listeners perceive and describe using a familiar vocabulary.
I don't buy it that PRaT is "familiar vocabulary" John. Those are just made-up nonsense words that should have been rejected by the mainstream hi-fi press rather than repeated. Do we even know where those terms originated?
I did post a link to an article on this subject earlier in this thread, Ethan - http://www.stereophile.com/reference/23 (http://www.stereophile.com/reference/23) - which examines this subject in considerable detail.
John Atkinson
Editor, Stereophile
Because you are confusing cause and effect, Ethan. The "cause" may indeed be "frequency response and distortion and ringing" but people do not have any means of directly detecting and identifying those technical [aberrations]. By contrast, the "effect" is what listeners perceive and describe using a familiar vocabulary.
I don't buy it that PRaT is "familiar vocabulary" John. Those are just made-up nonsense words that should have been rejected by the mainstream hi-fi press rather than repeated. Do we even know where those terms originated?
I did post a link to an article on this subject earlier in this thread, Ethan - http://www.stereophile.com/reference/23 (http://www.stereophile.com/reference/23) - which examines this subject in considerable detail.
John, exactly what is this religious chanting (taken from the 1992 reference), and how does it relate to audio in 2009?
"For all its quantifiable technical faults, easily identified in the laboratory when compared with the measured near-perfection of CD, the vinyl LP disc possesses a powerful and effortlessly musical content, with an easy, fundamental rhythmic stability and solidity. Interestingly, this innate character seems to be quite robust, more so than digital. Subjectively rewarding results may be obtained from analog sources without much trouble. Many well-established but not necessarily high-priced components may be assembled to produce musically satisfying results. With analog, one can listen through the blemishes and be aware of a strong musical message, one in which the music's flow, pace, and tempo are well conveyed, and into which the listener is drawn."
Based on what is now known about human perception of music, the author's condition is easy to diagnose. He is acclimatized to hearing music with all of the noise of distortion that is inherent in the LP format already added in. The complex path that music takes through his brain apparently only reaches his pleasure centers when they match up with his memories of years and years of listening to vinyl. His memories include the same audible noise and distortion that modern humans find undesirable. Modern human beings generally lack memories and paths related to the audible aritfacts involved with listening to vinyl. Therefore, even in 1992 the author was writing for only a narrow minority of music lovers who are, by modern standards not interested in high fidelity in the true sense of those words.
The start was a look at some codecs, playing 192/24 vs 128kbps mp3 (ProTool High Quality - is that FhG based?) - and playing the difference signal - i.e. subtracting the original from the downsampled > encoded > decoded > upsampled version. He played a lot of examples of the noise added by the codec, saying "how can we do this to music?".
It's a pity that he didn't continue with another test: listen to the difference signal from a 24/96 or 24/192 source and its 16/44.1 SRC'd version. Very illustrative IMHO.
Based on what is now known about human perception of music, the author's condition is easy to diagnose. He is acclimatized to hearing music with all of the noise of distortion that is inherent in the LP format already added in. The complex path that music takes through his brain apparently only reaches his pleasure centers when they match up with his memories of years and years of listening to vinyl. His memories include the same audible noise and distortion that modern humans find undesirable. Modern human beings generally lack memories and paths related to the audible aritfacts involved with listening to vinyl. Therefore, even in 1992 the author was writing for only a narrow minority of music lovers who are, by modern standards not interested in high fidelity in the true sense of those words.
Good point, and one which would also explain why those who listen to tube amps seem to fall mainly into one of three categories:
a) people of pensionable age (and who therefore spent their formative musical years using tube amps)
b) people who were given dad's tube amp when they started listening to music, or
c) people who are (or were) guitarists, used to playing their guitars through tube amps
"With analog, one can listen through the blemishes and be aware of a strong musical message, one in which the music's flow, pace, and tempo are well conveyed, and into which the listener is drawn."
[Emphasis Added]
Or one can listen to a CD so they don't have to concentrate on listening "through the blemishes"!
It's a pity that he didn't continue with another test: listen to the difference signal from a 24/96 or 24/192 source and its 16/44.1 SRC'd version. Very illustrative IMHO.
Excellent suggestion, but of course they would never do this, because it would be an example of intellectual and methodological rigour.
Based on what is now known about human perception of music, the author's condition is easy to diagnose. He is acclimatized to hearing music with all of the noise of distortion that is inherent in the LP format already added in. The complex path that music takes through his brain apparently only reaches his pleasure centers when they match up with his memories of years and years of listening to vinyl. His memories include the same audible noise and distortion that modern humans find undesirable. Modern human beings generally lack memories and paths related to the audible aritfacts involved with listening to vinyl. Therefore, even in 1992 the author was writing for only a narrow minority of music lovers who are, by modern standards not interested in high fidelity in the true sense of those words.
Good point, and one which would also explain why those who listen to tube amps seem to fall mainly into one of three categories:
a) people of pensionable age (and who therefore spent their formative musical years using tube amps)
b) people who were given dad's tube amp when they started listening to music, or
c) people who are (or were) guitarists, used to playing their guitars through tube amps
Of course there is at least one other explanation - and that is people who are mislead by all of the hype surrounding tubed amps.
I've got a little patch of psoriasis, and I think that I read someplace that listening to music through tubed amps would cure it! ;-)
Of course there is at least one other explanation - and that is people who are mislead by all of the hype surrounding tubed amps.
I've got a little patch of psoriasis, and I think that I read someplace that listening to music through tubed amps would cure it! ;-)
Only if you use the tube amp in homeopathic dosages.
The start was a look at some codecs, playing 192/24 vs 128kbps mp3 (ProTool High Quality - is that FhG based?) - and playing the difference signal - i.e. subtracting the original from the downsampled > encoded > decoded > upsampled version. He played a lot of examples of the noise added by the codec, saying "how can we do this to music?".
It's a pity that he didn't continue with another test: listen to the difference signal from a 24/96 or 24/192 source and its 16/44.1 SRC'd version. Very illustrative IMHO.
I did do this demonstration at the 2008 Rocky Mountain Audio Fest. I showed that the difference between 88.2kHz/24-bit data and a Red Book derived from those data was only audible at the listening position if you boosted its level by 24dB. So the question devotes to: Does adding something that is inaudible on its own to Red-Book audio result in something that is perceived as being og higher quality? The answer may not neccessarily be "no."
John Atkinson
Editor, Stereophile
I did do this demonstration at the 2008 Rocky Mountain Audio Fest. I showed that the difference between 88.2kHz/24-bit data and a Red Book derived from those data was only audible at the listening position if you boosted its level by 24dB. So the question devotes to: Does adding something that is inaudible on its own to Red-Book audio result in something that is perceived as being og higher quality? The answer may not neccessarily be "no."
No, the question devotes to: When 96db (Redbook's SNR) + 24db (Atkinson's revolutionary finding) = 120 db SNR and there is no audible difference below that, does Atkinson promote listening at levels that may seriously injure your ears at prolonged exposure?
Seriously, even the difference between 32 bit audio (199db SNR) and 33 bit (205db SNR) audio would be audible if you just turn the volume up enough. Lets promote 33 bit audio, Stereophile!
The real question devotes to: Where do we draw a reasonable line with regard to the limits of human perception, maximum listening levels, and maximum dynamic range found in content. Where are commercial DVD-A and SACD releases with actually more dynamic range than 90db?
I did do this demonstration at the 2008 Rocky Mountain Audio Fest. I showed that the difference between 88.2kHz/24-bit data and a Red Book derived from those data was only audible at the listening position if you boosted its level by 24dB. So the question devotes to: Does adding something that is inaudible on its own to Red-Book audio result in something that is perceived as being og higher quality? The answer may not neccessarily be "no."
John Atkinson
Editor, Stereophile
This explanation sounds like an argument from ignorance. You are proposing that since you couldn't find evidence at regular listening volume of any difference in the files, therefore it is possible that there is a difference in the files! A lack of evidence can't be used as proof that there is evidence.
Do you always listen to CDs 24 dB louder than your DVD-Audio discs or SACDs?
The real question devotes to: Where do we draw a reasonable line with regard to the limits of human perception, maximum listening levels, and maximum dynamic range found in content. Where are commercial DVD-A and SACD with actually more dynamic range than 90db?
The problem with CD isn't related at all to the technical capabilities of the format. The problem is how the format has been used for the last decade, i.e. with horrendous mastering standards that don't exploit the format's dynamic range.
It is sad that audiophile magazines often write glowing reviews of horrendously mastered CDs, while providing incoherent lectures on how technically flawed the format is!
I did do this demonstration at the 2008 Rocky Mountain Audio Fest. I showed that the difference between 88.2kHz/24-bit data and a Red Book derived from those data was only audible at the listening position if you boosted its level by 24dB. So the question devotes to: Does adding something that is inaudible on its own to Red-Book audio result in something that is perceived as being [of] higher quality? The answer may not [necessarily] be "no."
This explanation sounds like an argument from ignorance.
Ah, what would HA be without the usual ad hominem stuff. All I am doing is trying to get people to put aside their preconceived notions and think a little more deeply about this matter.
Do you always listen to CDs 24 dB louder than your DVD-Audio discs or SACDs?
No, and I didn't suggest that I did. As I wrote, I first played people the difference signal between 88.2kHz/24-bit data and a Red Book derived from those data at the same playback level as the originals. (They had previously auditioned the originals.) No-one could hear it. I then showed that I had to add 24dB gain to make that difference audible at the same playback level, ie, that while there _was_ content in the difference file, that content was indeed inaudible without that additional gain. I fail to see what is controversional or ignorant about this demonstration.
You are proposing that since you couldn't find evidence at regular listening volume of any difference in the files, therefore it is possible that there is a difference in the files! A lack of evidence can't be used as proof that there is evidence.
I didn't propose any such thing. First, I rephrased the question concerning the supposed superiority of hi-rez media in terms of the difference signal that had been introduced by another poster. I then offered the opinion that what might be thought the superficially obvious answer to that question might not automatically be correct. I think that remains to be proved, as demonstrated by the posting I made a few days back pointing out the surprising complexity of designing a blind test where the difference in sample rate is the _only_ variable being examined.
John Atkinson
Editor, Stereophile
The sighted "learning" comparison was between LPCM at 24-bits/88.2kHz and AAC at 128kbps, derived from a Red Book version of the data, downsampled using the SRC in Bias Peak 5 at its highest-quality setting, which independent tests had shown to be one of the best-performing (link available on request), and reduced to 16 bits using the POWR-2 redithering algorithm.
The single-blind comparison (using a different piece of music) used first, the original PCM at 24-bits/88.2kHz, Red Book PCM prepared in the same manner just described, AAC at 320kbps prepared from the Red BooK PCM, and an MP3 at 128kbps, again derived from the Red Book PCM version using the Fraunhofer codec in Adobe Audition 1.0. The splices between the 4 different versions were seamless, and there was no indication to listeners other than the possible change in sound that anything had changed. Levels were matched precisely, the lossy compressed and sample-rate-reduced files were reconverted to 24 bits and 88.2kHz sample rate so that the DAC would be behaving identically with all the examples. As I said, this was the exact methodology used by Linn's Philips Hobbs at the 2007 AES Conference.
So, help me understand correctly: you split a file into four segments, and each segment encoded differently : the 24/88.2 original, the same downconverted to 16/44.1, the downcoverted file encoded at 320, and to 128kbps (CBR or VBR?) with the FgH codec of Audition 1.0 (vintage 2003). Then the latter three were upconverted to 24/88.2 (involving an mp3-->wav conversion for the lossy files), and concatenated to the high-rez segment. Thus listeners heard a single file that progressively decreased in objective quality in three steps.
Is that right? If so, some notes:
First, the Audition 1.0 Fraunhofer codec is hardly SOTA in 2009 (http://en.wikipedia.org/wiki/Codec_listening_test) (arguably it wasn't in 2003, either) -- and IIRC you were told this already on that other thread.
Second, nowadays you can convert directly from 88.2/24 to mp3, with LAME (and other codecs, for all I know; I have personal experience with LAME)
Third, you say the listeners 'were not aware of what they were listening to' but what
did the listeners know about the track they were listening to? How and when did they signal their opinions of sound quality? How could htey compare, say, the hi-rez to the Redbook segment particularly? Or were they just asked what they felt when the track was over? I wasn't at Phillip Hobbs' AES demo in 2007 so you'll have to bear with me.
Fourth, the playback wasn't over a PA type system, was it?
Fifth, AIUI there are
potential (and avoidable) clipping issues when converting an mp3 file to a wav file. Perhaps someone on HA could verify or correct me. If so such potential degradation would have to be accounted for.
And thank you for agreeing - "of course" - that the degradation in quality should have been audible.
You're misquoting me -- I wrote (http://www.hydrogenaudio.org/forums/index.php?showtopic=72446&view=findpost&p=640056)that "*OF COURSE* the sound would get worse,
if you progressed far enough down through lossy bitrates." (emphasis added) The degradation in quality in a conversion to 320kbps should NOT be routinely audible, under normal conditions. Even 128 kbps (with a good codec) has a good chance of sounding 'transparent' to many listeners under normal conditions -- 200 'audiophiles' included.
Oh well. It must remain a mystery, therefore, why you are so concerned with the details of the test, Mr. Sullivan.
John Atkinson
Editor, Stereophile
Just trying to keep the 'authorities' in the high end
honest, Mr. Atkinson.
I fail to see what is controversional or ignorant about this demonstration.
It's not the ignorance of the demonstration, it's ignorance over the meaning of the demonstration.
I think that remains to be proved, as demonstrated by the posting I made a few days back pointing out the surprising complexity of designing a blind test where the difference in sample rate is the _only_ variable being examined.
...especially by someone who seems to be unwilling (if not utterly clueless about how) to conduct a proper double blind test for even something that doesn't necessarily have to have such complexity.
It's actually quite easy to put together a test comparing different sample rates. Are you suggesting that you'll somehow get a statistically significant result that would otherwise be hidden by the steps required to downsample and then upsample one of your signals? What are you afraid of, Stereoeditor?
...So the question devotes to: Does adding something that is inaudible on its own to Red-Book audio result in something that is perceived as being [of] higher quality? The answer may not [necessarily] be "no."
This explanation sounds like an argument from ignorance.
Ah, what would HA be without the usual ad hominem stuff. All I am doing is trying to get people to put aside their preconceived notions and think a little more deeply about this matter.
Pointing out that your explanation is
argumentum ad ignorantium does not make an
argumentum ad hominem in the process. A textbook example of an appeal to ignorance would be to suggest a premise is true simply because it has not been proven false, and your above statement is remarkably close to that example.
Shouldn't trying to get people to "think a little more deeply about this matter" also include rejecting formal fallacies like appeals to ignorance?
The sighted "learning" comparison was between LPCM at 24-bits/88.2kHz and AAC at 128kbps, derived from a Red Book version of the data, downsampled using the SRC in Bias Peak 5 at its highest-quality setting, which independent tests had shown to be one of the best-performing (link available on request), and reduced to 16 bits using the POWR-2 redithering algorithm.
The single-blind comparison (using a different piece of music) used first, the original PCM at 24-bits/88.2kHz, Red Book PCM prepared in the same manner just described, AAC at 320kbps prepared from the Red BooK PCM, and an MP3 at 128kbps, again derived from the Red Book PCM version using the Fraunhofer codec in Adobe Audition 1.0. The splices between the 4 different versions were seamless, and there was no indication to listeners other than the possible change in sound that anything had changed. Levels were matched precisely, the lossy compressed and sample-rate-reduced files were reconverted to 24 bits and 88.2kHz sample rate so that the DAC would be behaving identically with all the examples. As I said, this was the exact methodology used by Linn's Philips Hobbs at the 2007 AES Conference.
So, help me understand correctly: you split a file into four segments, and each segment encoded differently : the 24/88.2 original, the same downconverted to 16/44.1, the downcoverted file encoded at 320, and to 128kbps (CBR or VBR?) with the FgH codec of Audition 1.0 (vintage 2003). Then the latter three were upconverted to 24/88.2 (involving an mp3-->wav conversion for the lossy files), and concatenated to the high-rez segment. Thus listeners heard a single file that progressively decreased in objective quality in three steps.
Is that right?
Yes.
First, the Audition 1.0 Fraunhofer codec is hardly SOTA in 2009 (http://en.wikipedia.org/wiki/Codec_listening_test) (arguably it wasn't in 2003, either) -- and IIRC you were told this already on that other thread.
And as I said, I wasn't concerned with presenting MP3 at its absolutely best possible, only at a typical level of quality, ie, 128kps CBR.
Second, nowadays you can convert directly from 88.2/24 to mp3, with LAME (and other codecs, for all I know; I have personal experience with LAME)
Indeed. But that is not currently SOP for the commercial production of MP3s. I was interested in _typical_ performance.
Third, you say the listeners 'were not aware of what they were listening to' but what did the listeners know about the track they were listening to?
Nothing, other than it was a recording of Handel's Messiah.
How and when did they signal their opinions of sound quality?
I asked them what they had perceived after the demonstration had concluded. There was no talking or other communication while the music was playing.
Fourth, the playback wasn't over a PA type system, was it?
I used the word "optimal" and the system was accordingly very high quality: B&W 802D loudspeakers driven by Ayre or Classe amplification; source was a MacBook running iTunes 8 (with AudioMidi utility set for the correct sample rate and bit depth) feeding an Ayre QB-9 D/A processor via asynchronous USB.
Fifth, AIUI there are potential (and avoidable) clipping issues when converting an mp3 file to a wav file. Perhaps someone on HA could verify or correct me. If so such potential degradation would have to be accounted for.
There was no clipping of any of the versions of the original. Messiah has a fairly low average level and the peaks rarely approach -3dBFS, let alone 0dBFS. In any case, I verified the transcodes by comparing the waveform displays when I was splicing the segments together.
Just trying to keep the 'authorities' in the high end honest, Mr. Atkinson.
Whatever.
John Atkinson
Editor, Stereophile
As I wrote, I first played people the difference signal between 88.2kHz/24-bit data and a Red Book derived from those data at the same playback level as the originals. (They had previously auditioned the originals.) No-one could hear it. I then showed that I had to add 24dB gain to make that difference audible at the same playback level, ie, that while there _was_ content in the difference file, that content was indeed inaudible without that additional gain. I fail to see what is controversional or ignorant about this demonstration.
It's not the ignorance of the demonstration, it's ignorance over the meaning of the demonstration.
I have made no statement about the meaning. I rephrased the question over the putative benefit of high-resolution audio and offered the opinion that the obvious answer to that question may not be correct. That is the issue that still needs to be tested.
I think that remains to be proved, as demonstrated by the posting I made a few days back pointing out the surprising complexity of designing a blind test where the difference in sample rate is the _only_ variable being examined.
...especially by someone who seems to be unwilling (if not utterly clueless about how) to conduct a proper double blind test for even something that doesn't necessarily have to have such complexity.
As a moderator, greynol, perhaps it behoves you to refrain from insults. It is not me who is saying that designing such a test is complex; instead I have been reporting what those involved in doing so - "AES Fellows, some university professors, some well-known recording and mastering engineers, and even JJ" - are finding during the planning of such a test. If you are offering the opinion that those people are "clueless," you need to rethink that position. And if you are saying that _I_ am "clueless" about how to conduct a "proper double blind test," I fail to see anything I have written in this or any other thread that supports your opinion.
It's actually quite easy to put together a test comparing different sample rates.
Not according to those I mentioned. The simple procedure appears to be prone to both false positives and negatives.
John Atkinson
Editor, Stereophile
As an act of good faith, feel free to conduct some double blind tests, Stereoeditor, even if they're personal tests.
And as I said, I wasn't concerned with presenting MP3 at its absolutely best possible, only at a typical level of quality, ie, 128kps CBR.
But,as you've been told , all codecs are not created equal. Therefore all 128kbps mp3s of a track are not created equal. As the wiki page I linked to will demonstrate if you mine down into the links therein.
128kbps FgH mp3 circa 2003 is not typical in 2009. And arguably 128 kbps CBR itself isn't even 'typical'. The iTunes Store went to higher bitrates awhile ago, for example.
Second, nowadays you can convert directly from 88.2/24 to mp3, with LAME (and other codecs, for all I know; I have personal experience with LAME)
Indeed. But that is not currently SOP for the commercial production of MP3s. I was interested in _typical_ performance.
As of 2003, apparently.
Third, you say the listeners 'were not aware of what they were listening to' but what did the listeners know about the track they were listening to?
Nothing, other than it was a recording of Handel's Messiah.
How and when did they signal their opinions of sound quality?
I asked them what they had perceived after the demonstration had concluded. There was no talking or other communication while the music was playing.
Very unclear to me here -- these 200 audiophiles were attending ...what? What did they think they were in for? Didn't they know this was a 'demonstration' of the effects of mp3, or a comparison of some sort?
And how could one know from this whether there was *progressive* subjective difference? How would you know , from this demo, that listeners thought , e.g., Redbook was inferior to hi rez? Or 320kbps inferior to Redbook? Versus , say, 320 vs 128 (the comparison most likely to be within reasonable bounds of audio memory in your 'demonstration')?
Fourth, the playback wasn't over a PA type system, was it?
I used the word "optimal" and the system was accordingly very high quality: B&W 802D loudspeakers driven by Ayre or Classe amplification; source was a MacBook running iTunes 8 (with AudioMidi utility set for the correct sample rate and bit depth) feeding an Ayre QB-9 D/A processor via asynchronous USB.
Good enough. MP3s aren't necessarily 'tuned' to sound good over public address systems.
Fifth, AIUI there are potential (and avoidable) clipping issues when converting an mp3 file to a wav file. Perhaps someone on HA could verify or correct me. If so such potential degradation would have to be accounted for.
There was no clipping of any of the versions of the original. Messiah has a fairly low average level and the peaks rarely approach -3dBFS, let alone 0dBFS. In any case, I verified the transcodes by comparing the waveform displays when I was splicing the segments together.
AIUI, it's not a result of clipping on the original, it's peaks near 0dBFS that can become clipped. But -3dBFS peaks would probably guard against that.
So this leaves basically just a few issues (besides the whopper of a premise that this was a 'demonstration' of anything serious about mp3s): 1) codecs used for lossy, and 2) what sort of expectations the listeners had, and how results were related to the presentation. I might also add ,how the results were tallied.
As a moderator, greynol, perhaps it behoves you to refrain from insults. It is not me who is saying that designing such a test is complex; instead I have been reporting what those involved in doing so - "AES Fellows, some university professors, some well-known recording and mastering engineers, and even JJ" - are finding during the planning of such a test. If you are offering the opinion that those people are "clueless," you need to rethink that position. And if you are saying that _I_ am "clueless" about how to conduct a "proper double blind test," I fail to see anything I have written in this or any other thread that supports your opinion.
I am would wager that JJ would not dream of 'demonstrating' performance of different formats the way you or Mr. Massenburg did.
As an act of good faith, feel free to conduct some double blind tests, Stereoeditor, even if they're personal tests.
Why do I have to? I haven't violated Tos#8 by posting personal opinions concerning sound quality to this forum. Yes, my published opinions on the benefits of high-resolution _have_ been posted to this forum, but that was done by Arny Krueger, not me. I don't see that his doing so obliges me to suppoort those opinions on HA. Otherwise, I would be vulnerable to every troll on the Internet.
John Atkinson
Editor, Stereophile
... who can judge in that environment at listening levels that make your ears ring?
[..]
He played a lot of examples of the noise added by the codec, saying "how can we do this to music?".
David.
Hmm, when he played back the "error noise" on the same level as the original, that would explain the reason for loud levels. The noise would be not very impressive unless turned up.
(I wish Mr. Krueger and others would not make such a personal thing here about a disagreement, as long as everybody sticks to the ToS of this forum)
It is not me who is saying that designing such a test is complex; instead I have been reporting what those involved in doing so - "AES Fellows, some university professors, some well-known recording and mastering engineers, and even JJ" - are finding during the planning of such a test. If you are offering the opinion that those people are "clueless," you need to rethink that position.
I don't think that you are completely honest here. You know the basic procedure: Compare a plain 88.2 kHz record to a time synchronized 88.2 kHz -> 44.1 kHz -> 88 kHz record. If that shows, that no difference can be heard under double blind conditions, you already have very solid results and it doesn't get any more complicated than that.
It would
only get complicated
if a difference could be heard! But not before that, and that's what you are not honest about. And tell me one robust experiment, that would have shown that. Your own non-DBT show-biz experiment mixed up with 16bit/24bit differences at high volume levels excluded.
How and when did they signal their opinions of sound quality?
I asked them what they had perceived after the demonstration had concluded. There was no talking or other communication while the music was playing.
Very unclear to me here -- these 200 audiophiles were attending ...what? What did they think they were in for?
The 200 audiophiles took part in 3 music evenings presented at 3 Colorado audio stores, under the umbrella title "Music Matters." From the events' promotional material: "Billed as 'three special evenings devoted to a celebration of recorded music,' each Music Matters event will be highlighted by presentations from Stereophile editor John Atkinson and AudioQuest's Joe Harley. Other Music Matters participants will include: Steve Silberman (Ayre Acoustics), Dave Baker (B&W Loudspeakers), Dave Nauber (Classé Audio), Carl Porter (McIntosh), David Solomon (Peachtree Audio), Dave Ellington (Sonus Faber, Pro-Ject), and Martin Cooper (Wadia)." There was nothing published prior to the event about my own presentation, other than, IIRC, the discussion on HA.
The two demonstrations of the possible artefacts of data reduction (first sighted, then single-blind, as explained earlier in this thread) took up about a third of the 45 minutes time allocated for me. The rest of my presentation involved playing examples of my many commercially released recordings from the original 24-bit/88.2kHz masters, which no-one would have heard before, and demonstrating how mike techniques and mixing strategies affect soundstage perception. As I said before, I gave 10 presentations over the 3 days to an average of 20 people each time.
Didn't they know this was a 'demonstration' of the effects of mp3, or a comparison of some sort?
Not before they came in. They did know in the first "learning" dem, though they did not know what was to follow. With the subsequent blind presentation, they had no idea to what they were listening to until after it had concluded, as I said before.
And how could one know from this whether there was *progressive* subjective difference?
Because that is what most of the listeners said they had perceived in the subsequent discussion.
How would you know , from this demo, that listeners thought , e.g., Redbook was inferior to hi rez? Or 320kbps inferior to Redbook? Versus , say, 320 vs 128 (the comparison most likely to be within reasonable bounds of audio memory in your 'demonstration')?
There was no way for anyone to identify when exactly the degradation set in, but that was not the point. All I was interested in was whether or not listeners detected the progressive change in quality.
So this leaves basically just a few issues (besides the whopper of a premise that this was a 'demonstration' of anything serious about mp3s): 1) codecs used for lossy, and 2) what sort of expectations the listeners had, and how results were related to the presentation. I might also add ,how the results were tallied.
You have this annoying habit, Mr. Sullivan, of asking questions that have already been answered. I know you stated yesterday that you don't bookmark my postings, but do you really _not_ remember the answers I have offered you in the past? Again, there was _no_ tallying of results. All I did was ask the listeners what, if anything, they had perceived during the prior 5 minutes or so of music. Following that, I explained what they had been listening to and answered any questions people may have had about file formats and about computer audio in general. I then moved on to the next part of my presentation, as explained above.
John Atkinson
Editor, Stereophile
It is not me who is saying that designing such a test is complex; instead I have been reporting what those involved in doing so - "AES Fellows, some university professors, some well-known recording and mastering engineers, and even JJ" - are finding during the planning of such a test.
Or so you say, John. You have a track record for making highly questionable claims, so in the absence of *any* confirming evidence we should discount this as yet another highly questionable claim from a person who has being making questionable claims in print for several decades.
It is not me who is saying that designing such a test is complex; instead I have been reporting what those involved in doing so - "AES Fellows, some university professors, some well-known recording and mastering engineers, and even JJ" - are finding during the planning of such a test. If you are offering the opinion that those people are "clueless," you need to rethink that position.
I don't think that you are completely honest here. You know the basic procedure: Compare a plain 88.2 kHz record to a time synchronized 88.2 kHz -> 44.1 kHz -> 88 kHz record. If that shows, that no difference can be heard under double blind conditions, you already have very solid results and it doesn't get any more complicated than that.
But you haven't eliminated all variables and the result may still be a false negative, thus not transportable. All you can conclude is with _that_ recording_ with _that_ hardware and _that_ testing protocol, no difference could be identified to a given degree of statistical certainty. But you haven't proved anything concerning sample rate _alone_.
John Atkinson
Editor, Stereophile
As an act of good faith, feel free to conduct some double blind tests, Stereoeditor, even if they're personal tests.
Why do I have to? I haven't violated Tos#8 by posting personal opinions concerning sound quality to this forum. Yes, my published opinions on the benefits of high-resolution _have_ been posted to this forum, but that was done by Arny Krueger, not me. I don't see that his doing so obliges me to suppoort those opinions on HA. Otherwise, I would be vulnerable to every troll on the Internet.
John, isn't it true that you're only vulnerable because of all the questionable things that you've said?
If you don't support your opinions when you are asked questions about them, isn't that the same as distancing yourself from them?
It is not me who is saying that designing such a test is complex; instead I have been reporting what those involved in doing so - "AES Fellows, some university professors, some well-known recording and mastering engineers, and even JJ" - are finding during the planning of such a test.
Or so you say, John. You have a track record for making highly questionable claims, so in the absence of *any* confirming evidence we should discount this as yet another highly questionable claim from a person who has being making questionable claims in print for several decades.
Circular reasoning at its finest, Mr. Krueger. You say I can't be trusted because you don't trust me and you don't trust me because I can't be trusted. :-)
Yet I am the one about to celebrate 27 years of a career as the editor in chief of two very successful audio magazines...perhaps my readers and my employers know something you don't, Mr. Krueger?
John Atkinson
Editor, Stereophile
I haven't violated Tos#8 by posting personal opinions concerning sound quality to this forum. Yes, my published opinions on the benefits of high-resolution _have_ been posted to this forum, but that was done by Arny Krueger, not me. I don't see that his doing so obliges me to suppoort those opinions on HA. Otherwise, I would be vulnerable to every troll on the Internet.
John, isn't it true that you're only vulnerable because of all the questionable things that you've said?
No.
If you don't support your opinions when you are asked questions about them, isn't that the same as distancing yourself from them?
Not at all. On this forum, I am complying with its ToS#8. Outside this forum, in what is generally called the "real world," ToS#8 doesn't apply.
John Atkinson
Editor, Stereophile
Second, nowadays you can convert directly from 88.2/24 to mp3, with LAME (and other codecs, for all I know; I have personal experience with LAME)
Maybe a bit offtopic, but: LAME resamples such input to 48kHz. And I think that LAME internal resampler was designed to be fast (and simple), and there are better (in terms of absolute quality) samplerate converters.
But you haven't eliminated all variables and the result may still be a false negative, thus not transportable.
Don't make me laugh!
1. You would have already eliminated
many more variables than any test concerning the matter, that you have presented so far.
All you can conclude is with _that_ recording_ with _that_ hardware and _that_ testing protocol, no difference could be identified to a given degree of statistical certainty.
2. In a very strict sense this can be said about any scientific induction.
Mr. Atkinson, you are equipped with considerable ressources, regarding the availability of highest end equipment, knowledgeable people as possible advisors, and financially. A solid demonstration, that high rez really
was superior, like your magazine has always claimed, would surely pay-off in every respect.
Just show us one example where a difference between 88.1 kHz & 44.1 kHz is audible. And I will shut up regarding this forever.
The fact, that people with your ressources and supposed motivation have failed to produce this until this day, is itself already a pretty sold result.
But you haven't eliminated all variables and the result may still be a false negative, thus not transportable. All you can conclude is with _that_ recording_ with _that_ hardware and _that_ testing protocol, no difference could be identified to a given degree of statistical certainty. But you haven't proved anything concerning sample rate _alone_.
Yes, additional variables that will only lead to further degradation are going to result in false negatives.
This was more or less my point earlier, but it would appear that you would rather dodge the subject.
BTW, in our "unreal" world of requiring objective measurement to demonstrate that differences in sound quality aren't a figment of your imagination, there is no such thing as a false negative.
I don't think that you are completely honest here. You know the basic procedure: Compare a plain 88.2 kHz record to a time synchronized 88.2 kHz -> 44.1 kHz -> 88 kHz record. If that shows, that no difference can be heard under double blind conditions, you already have very solid results and it doesn't get any more complicated than that.
But you haven't eliminated all variables
John, you are again not being fair or honest with us. You say that something is true with absolutely no supporting detail or independent support. You expect us to bow to the authority of just your personal word over us, which of course is a very arrogant and insulting thing for you to do since you have no authority over us.
and the result may still be a false negative, thus not transportable.
Again John, just you saying so does not make it so. It appears to me that you have been dealing with poorly-informed people who have suspended disbelief so long that you are incapable of holding a decent conversation with people who have not pledged allegance to your religious leadership.
All you can conclude is with _that_ recording_ with _that_ hardware and _that_ testing protocol, no difference could be identified to a given degree of statistical certainty.
If we followed your logic John, we'd have to destory every home in the world with an atomic bomb to prove that atomic bombs can destroy homes. As long as there is one home that hasn't been destroyed, we really don't know for sure that it can be destroyed by an atomic bomb, right? Why that house may be the perfect house and undestroyable!
So it is your claim John that the benefits of so-called hi-rez recordings completely and totally lack generality, such that you cannot specify a recording or a test system where they are effective without us first testing thousands or millions of recordings and systems?
But you haven't proved anything concerning sample rate _alone_.
One positive test outcome John, is all it takes. Isn't your continual backpedalling away from your past published claims, and your present obfuscating about test producedure sufficient evidence to cause us to believe that not even you John Atkinson have any faith at all in your past claims about their effectiveness?
But you haven't eliminated all variables and the result may still be a false negative, thus not transportable. All you can conclude is with _that_ recording_ with _that_ hardware and _that_ testing protocol, no difference could be identified to a given degree of statistical certainty. But you haven't proved anything concerning sample rate _alone_.
Yes, additional variables that will only lead to further degradation are going to result in false negatives.
This was more or less my point earlier, but it would appear that you would rather dodge the subject.
BTW, in our "unreal" world of requiring objective measurement to demonstrate that differences in sound quality aren't a figment of your imagination, there is no such thing as a false negative.
It's not just our world where there are no false negatives. In the world of science, every experimental outcome obtained under relevant and documented circumstances is true as far as its inhrent truth goes.
I find it tremendously curious that John can't simply provide us with one of his hi-rez recordings and specify a test environment where his claims about the effectiveness of hi rez recordings would be clearly and unambigiously heard. Not one of the great many at his disposal. Apparently such recordings and equipment has never darkened the hallowed doors of Stereophile.
Apparently, the rigorous standards of generally-accepted standards body publications such as BS 1116 are not sufficient.
John is implicitly telling us that if you have an audio system and trained listeners that conform to the relatively lofty standards of BS 1116, he has no cofidence that one of his recordings would support his published claims.
One has to ask what sort of shakey technology demands such an exceptional listening environment and listeners that not even John Atkinson can specify what or who it is, for hi-rez recordings to be effective!
But you haven't eliminated all variables and the result may still be a false negative, thus not transportable. All you can conclude is with _that_ recording_ with _that_ hardware and _that_ testing protocol, no difference could be identified to a given degree of statistical certainty. But you haven't proved anything concerning sample rate _alone_.
John Atkinson
Editor, Stereophile
Then you simply don't trust to science. There is always one more variable to make up.
Of course, all it takes to prove your side of the argument is _one_ recording, _one_ hardware setup and _one_ testing protocol. This should be particularly easy if the difference is apparent (which to you it seemingly is). If you don't do this, all your claims about this subject are as valid as someone saying there are aquatic pink elephants that all turn into herrings the moment someone looks.
Didn't they know this was a 'demonstration' of the effects of mp3, or a comparison of some sort?
Not before they came in. They did know in the first "learning" dem, though they did not know what was to follow. With the subsequent blind presentation, they had no idea to what they were listening to until after it had concluded, as I said before.
And how could one know from this whether there was *progressive* subjective difference?
Because that is what most of the listeners said they had perceived in the subsequent discussion.
So, let me see again if I have it straight...they were given a 'dem' about possible artefacts of data reduction (involving showing/playing difference tracks..possibly at high dBs? Or did only G. Massenburg do that?) ...and then you played them the Frankenstein track that started with lossless 'hi rez' and ended with a 2003 vintage 128kbps encode. No spoken intro, just something like, 'Now listen to this'. And then afterwards you asked them something like 'what did that sound like'? And during the 'discussion' some number of people said something like 'well, it started out great, but sounded worse and worse".
Close?
How would you know , from this demo, that listeners thought , e.g., Redbook was inferior to hi rez? Or 320kbps inferior to Redbook? Versus , say, 320 vs 128 (the comparison most likely to be within reasonable bounds of audio memory in your 'demonstration')?
There was no way for anyone to identify when exactly the degradation set in, but that was not the point. All I was interested in was whether or not listeners detected the progressive change in quality.
Did any of the listeners note a three-stage reduction in quality? Did any just note that it sounded worse at the end than at the start?
So this leaves basically just a few issues (besides the whopper of a premise that this was a 'demonstration' of anything serious about mp3s): 1) codecs used for lossy, and 2) what sort of expectations the listeners had, and how results were related to the presentation. I might also add ,how the results were tallied.
You have this annoying habit, Mr. Sullivan, of asking questions that have already been answered. I know you stated yesterday that you don't bookmark my postings, but do you really _not_ remember the answers I have offered you in the past? Again, there was _no_ tallying of results. All I did was ask the listeners what, if anything, they had perceived during the prior 5 minutes or so of music.
So very sorry to have annoyed you, Mr. Atkinson. But I am trying to keep all the pertinent facts at hand; absent links, the alternative is to restate them here. It can't hurt, and can certainly help, from a reader' POV.
Now, let me see if I have it finally straight:
-you don't think the number of listeners who replied ...and how they replied...matters as much as the fact that 10 groups of 20 'audiophiles' were involved
-you seem to think your 'teaching demo' (which may have grossly exaggerated the audible effects) did not load any expectations into the 'audiophile' audience during the 'single blind listening demo'.
So, any recordings of this event? I'd
love to hear the actual back and forth between yourself and the audience. I'll even count the numbers of respondents and classify their responses for you, if you like.
in our "unreal" world of requiring objective measurement to demonstrate that differences in sound quality aren't a figment of your imagination, there is no such thing as a false negative.
Really? No such thing as a Type II error? That the results of a bias-free test may still have failed to identify a real difference? Color me ( and the professor I learned statistics from)l amazed.
Thank you for the clarification.
John Atkinson
Editor, Stereophile
in our "unreal" world of requiring objective measurement to demonstrate that differences in sound quality aren't a figment of your imagination, there is no such thing as a false negative.
Really? No such thing as a Type II error? That the results of a bias-free test may still have failed to identify a real difference? Color me ( and the professor I learned statistics from)l amazed.
Thank you for the clarification.
John Atkinson
Editor, Stereophile
You have not carefully read TOS8 or you do not understand it. It is there only to remove false positives. If a person claims to hear a difference between A and B under a certain condition, it is implausible if not impossible that just the effect of making him unaware of which is which in any other way than hearing the recordings causes a false negative.
He's just being evasive again. I'm sure removal of my "BTW" was not unintentional, not to mention his failure to directly address the common sense points that have been raised by other people.
I did post a link to an article on this subject earlier in this thread, Ethan - http://www.stereophile.com/reference/23 (http://www.stereophile.com/reference/23) - which examines this subject in considerable detail.
But that article is full of the same misuse of those words:
What many listeners fail to note is that weaknesses in the audio chain can give rise to errors which, in combination, suggest that the musical pace has become slower
I've never noticed audio equipment having any effect on musical tempo.
The ear has an extraordinary ability to recognize playing which is not on or aligned to the beat. This includes deliberately time-shifted or syncopated playing, as unmusical errors here quickly destroy meaning.
Again, this is not a function of the gear.
And it goes downhill from there.
So I'll take that as a No, that you're not interested in collaborating on a video where we could explain this stuff properly? It's a shame because this sort of clarification is really needed, and coming from different angles (yours and mine) we could provide a real service to the audio community.
--Ethan
Really? No such thing as a Type II error? That the results of a bias-free test may still have failed to identify a real difference? Color me ( and the professor I learned statistics from)l amazed.
But what is "real" difference? There are so many differences that are inaudible.
Convert your favorite piece of music from 24 bit to 16 bit (with dither) twice. Dither adds random noise so you'll get 2 different WAV files. If someone cannot distinguish one file from another, is this false negative?
in our "unreal" world of requiring objective measurement to demonstrate that differences in sound quality aren't a figment of your imagination, there is no such thing as a false negative.
Really? No such thing as a Type II error? That the results of a bias-free test may still have failed to identify a real difference? Color me ( and the professor I learned statistics from)l amazed.
Thank you for the clarification.
You have not carefully read TOS8 or you do not understand it. It is there only to remove false positives. If a person claims to hear a difference between A and B under a certain condition, it is implausible if not impossible that just the effect of making him unaware of which is which in any other way than hearing the recordings causes a false negative.
I read it but no, I did not fully grasp the implication that it gives a free pass to tests that may throw up false negatives. Just because a test may be double-blind does mean _by itself_ that it is free from either Type 1 or Type II error, particularly when the effect being tested for is small, inconsistently audible (ie, there is a dependency between its audibility and another variable, such as the program etc) or both. The test needs to be free from both kinds of errors, if it is desired that it be scientifically rigorous. I can't believe that you are arguing this point. If HA subscribes to scientific rigor, then it needs to be judged by the appropriate scientific tenets. You can't pick and choose just those aspects of rigor that align with your beliefs.
John Atkinson
Editor, Stereophile
So, let me see again if I have it straight...they were given a 'dem' about possible artefacts of data reduction...
Yes. With unfamiliar recordings played on an unfamiliar system in an unfamiliar listening room, a learning session is pretty much essential.
involving showing/playing difference tracks..possibly at high dBs? Or did only G. Massenburg do that?
No, that was George Massenburg. Any reference I have made in this thread to difference tracks involved some dems I did last year.
...and then you played them the Frankenstein track that started with lossless 'hi rez' and ended with a 2003 vintage 128kbps encode. No spoken intro, just something like, 'Now listen to this'. And then afterwards you asked them something like 'what did that sound like'? And during the 'discussion' some number of people said something like 'well, it started out great, but sounded worse and worse".
Close?
But no cigar. I described what happened is precise terms. I refer others to that description.
Did any of the listeners note a three-stage reduction in quality? Did any just note that it sounded worse at the end than at the start?
As I wrote, most listeners felt that there had been a degradation. That was all I felt necessary. given the premises underlying the presentations that I have described in earlier postings.
Now, let me see if I have it finally straight:
-you don't think the number of listeners who replied ...and how they replied...matters as much as the fact that 10 groups of 20 'audiophiles' were involved
I have not said anything like that. I refer you to my previous posts.
-you seem to think your 'teaching demo' (which may have grossly exaggerated the audible effects) did not load any expectations into the 'audiophile' audience during the 'single blind listening demo'.
Again I didn't say that.
So, any recordings of this event?
Sorry, no.
John Atkinson
Editor, Stereophile
I'll take that as a No, that you're not interested in collaborating on a video where we could explain this stuff properly?
Actually, I have been working for some time on a series of "rich" Web articles that does just what you propose, Ethan, based on a series of articles called "The Sonic Bridge" that appeared in the print magazine in the 1990s. I don't see the need for a collaborator, however.
John Atkinson
Editor, Stereophile
I can't believe that you are arguing this point. If HA subscribes to scientific rigor, then it needs to be judged by the appropriate scientific tenets. You can't pick and choose just those aspects of rigor that align with your beliefs.
John Atkinson
Editor, Stereophile
Just do it like Atkinson, he knows! Because a proper* sample-rate DBT can have a theoretical 0.000005 % probability of false negatives, he does the testing his way against all accepted scientific standards with a double digit probability of false positives. Just to come here afterwards and educate people about how to do it right.
*88.2 kHz vs. 88.2 kHz -> 44.1 kHz -> 88.2. kHz (with a SRC of at least SSRC's quality), time synched, same DAC, double blind. In this case the high quality sample stays untouched and only the low res sample is degraded twice. Any sane person would think, that degrading the low res source twice would
increase the probability to identify a difference against the untouched hi res source. Not Mr. Atkinson! What he is trying to sell, is his theory that this would instead increase the probability of false negatives!
Understand how flawed that logic is: When comparing two identical samples (in absolutely every aspect), according to Atkinson it could increase the probability that people report them to be actually identical if you degrade one of them twice through sample rate conversion.
Just one statistically sound and positive DBT result can show with a high probability that a difference can be audible. One negative result, of course, cannot show the same. But, as said, failure to produce just one positive result for over a decade is a clear indication, also with very high probability, that there probably are no audible differences. Please, Mr. Atkinson, just one!
Really? No such thing as a Type II error? That the results of a bias-free test may still have failed to identify a real difference? Color me ( and the professor I learned statistics from)l amazed.
But what is "real" difference? There are so many differences that are inaudible.
Convert your favorite piece of music from 24 bit to 16 bit (with dither) twice. Dither adds random noise so you'll get 2 different WAV files. If someone cannot distinguish one file from another, is this false negative?
I am sorry, I have no idea. I am not trying to be evasive, as greynol seems to think, but your question is so open-ended as to be unanswerable.
John Atkinson
Editor, Stereophile
I can't believe that you are arguing this point. If HA subscribes to scientific rigor, then it needs to be judged by the appropriate scientific tenets. You can't pick and choose just those aspects of rigor that align with your beliefs.
Just do it like Atkinson, he knows! Because a proper* sample-rate DBT can have a theoretical 0.000005 % probability of false negatives, he does the testing his way against all accepted scientific standards with a double digit probability of false positives. Just to come here afterwards and educate people about how to do it right.
Stereophile makes no claim concerning the scientific rigor behind its published opinions, hence I am not concerned if those opinions are felt to be nothing more than false positives by HA denizens. (Doesn't mean I agree with that opinion, however.) But if those same HA denizens are claiming the scientific high ground, as they are, then I certainly feel it appropriate to judge their writings by the appropriate scientific standard. Otherwise the argument devolves to a matter of opposed beliefs and I don't see the need to debate beliefs.
John Atkinson
Editor, Stereophile
But if those same HA denizens are claiming the scientific high ground, as they are, then I certainly feel it appropriate to judge their writings by the appropriate scientific standard.
You mean in the same way as you are trying to claim scientific high ground at Hydrogenaudio by word while failing to present anything but show-biz A/B setups by deed?
in our "unreal" world of requiring objective measurement to demonstrate that differences in sound quality aren't a figment of your imagination, there is no such thing as a false negative.
Really? No such thing as a Type II error?
False negative does not always mean identically the same as Type II error.
Furthermore, the alleged Type II errors that we hear about from golden ears are not proven facts, but just hypothetical wishes and dreams.
That the results of a bias-free test may still have failed to identify a real difference?
A test that fails to detect a real difference is not bias-free.
Now getting back to the chase John, why can't you provide us with unambigious evidence that high rez recording techniques provide even a different sound, let alone a better sound.
in our "unreal" world of requiring objective measurement to demonstrate that differences in sound quality aren't a figment of your imagination, there is no such thing as a false negative.
Really? No such thing as a Type II error?
False negative does not always mean identically the same as Type II error.
Others disagree: http://en.wikipedia.org/wiki/Type_I_and_type_II_errors (http://en.wikipedia.org/wiki/Type_I_and_type_II_errors)
Now getting back to the chase John, why can't you provide us with unambigious evidence that high rez recording techniques provide even a different sound, let alone a better sound.
As I said in the message to which you originally posted, Mr. Krueger, a group of AES illuminati are working on exactly that project. Why am I obliged to duplicate their efforts? In the meantime, I shall continue to make my recordings - recordings with sound quality that others claim to appreciate - with high sample rates and bit depths greater than 16. The only people to whom I need to prove anything to concerning those recordings are myself and those who employ my services.
John Atkinson
Editor, Stereophile
As an act of good faith, feel free to conduct some double blind tests, Stereoeditor, even if they're personal tests.
Why do I have to? I haven't violated Tos#8 by posting personal opinions concerning sound quality to this forum. Yes, my published opinions on the benefits of high-resolution _have_ been posted to this forum, but that was done by Arny Krueger, not me. I don't see that his doing so obliges me to suppoort those opinions on HA. Otherwise, I would be vulnerable to every troll on the Internet.
John Atkinson
Editor, Stereophile
Whoa, I always try to be civil, but you're saying you can't be arsed to take the (relatively little) time to do ABX testing in spite of having spent hours at the keyboard on HA. Perhaps no TOS#8 violation as such, but bad faith aplenty. And if the assertions of HA members are little better than those of internet trolls, then -- without meaning to sound nasty -- why are you bothering to post here in the first place?
I did do this demonstration at the 2008 Rocky Mountain Audio Fest. I showed that the difference between 88.2kHz/24-bit data and a Red Book derived from those data was only audible at the listening position if you boosted its level by 24dB. So the question devotes to: Does adding something that is inaudible on its own to Red-Book audio result in something that is perceived as being [of] higher quality? The answer may not [necessarily] be "no."
This explanation sounds like an argument from ignorance.
Ah, what would HA be without the usual ad hominem stuff. All I am doing is trying to get people to put aside their preconceived notions and think a little more deeply about this matter.
LOL! Well done mate. You can't demonstrate why your conclusions are sound, so you pretend that my reply was an ad hominem argument. That in itself is an ad hominem argument against me, because it completely misrepresents what I asserted, which is simply that you are drawing conclusions without there being any evidence for you to draw such conclusions.
He's just being evasive again. I'm sure removal of my "BTW" was not unintentional, not to mention his failure to directly address the common sense points that have been raised by other people.
I'm a bit distressed that people with something of real value to offer are spending so much time on this thread. I guess they're so involved in the topic that they're not really seeing how useless the discussion is. Or it's just the old story of red rags and bulls.
Anyhow, people, you are, in all seriousness, wasting your time. Atkinson is a weaselling troll. He's an intelligent, well educated, and articulate troll, but that just makes him a bigger time-waster. If you think about the kind of responses he's giving, and the failure to to engage with the serious content of the debate(s) (which people have remarked on, several times, in this thread), you'll see it. Unless it's giving you some kind of perverse pleasure, the world will be well served if he's just ignored.
No, this is not
ad hominem; it's a serious characterisation of the mode of argument employed.
I read it but no, I did not fully grasp the implication that it gives a free pass to tests that may throw up false negatives. Just because a test may be double-blind does mean _by itself_ that it is free from either Type 1 or Type II error, particularly when the effect being tested for is small, inconsistently audible (ie, there is a dependency between its audibility and another variable, such as the program etc) or both. The test needs to be free from both kinds of errors, if it is desired that it be scientifically rigorous. I can't believe that you are arguing this point. If HA subscribes to scientific rigor, then it needs to be judged by the appropriate scientific tenets. You can't pick and choose just those aspects of rigor that align with your beliefs.
John Atkinson
Editor, Stereophile
So you still don't get it. It does not give a free pass, because the person already claims to consistently hear a difference under a certain condition. Nothing prevents that person to recreate that condition under a DBT. If you do not understand why it is almost impossible to get a false negative this way, I'm afraid you do not know one thing about scientific rigor or tenets.
I'm a bit distressed that people with something of real value to offer are spending so much time on this thread. I guess they're so involved in the topic that they're not really seeing how useless the discussion is. Or it's just the old story of red rags and bulls.
Anyhow, people, you are, in all seriousness, wasting your time. Atkinson is a weaselling troll. He's an intelligent, well educated, and articulate troll, but that just makes him a bigger time-waster. If you think about the kind of responses he's giving, and the failure to to engage with the serious content of the debate(s) (which people have remarked on, several times, in this thread), you'll see it. Unless it's giving you some kind of perverse pleasure, the world will be well served if he's just ignored.
No, this is not ad hominem; it's a serious characterisation of the mode of argument employed.
Calling Mr Atkinson a troll is probably taking things a bit far. This is something of a hostile environment for the man and he's hardly likely to convert HA members to
Stereophile readers. That in and of itself deserves some kudos, at least compared to those who hide behind their editorial pulpits.
But I also agree with you. Is the editor of an audio magazine's duty of care really to educate their readers, or to reinforce their belief systems? My guess is it's the latter. This is understandable - unless it's an academic journal, people choose to read something that supports, not challenges, their own feelings toward a subject. No one reads
Hot Rod magazine to learn about hybrid technology or electric cars for example, and if an editor of that magazine suddenly turned 'green', the readers would probably see red.
Magazines frequently run features that border on challenging the beliefs of the readership, usually when it becomes almost inevitable not to do otherwise. Photo magazines were singularly resistant to the ingress of digital imaging (because the readership was openly hostile to the concept) but then quickly changed direction to embrace digital, when it became clear film photography was dying off in the consumer sector. Those self-same magazines that suggested for decades that you spent the winter in the darkroom suddenly discovered the horrors of an army of readers flushing away gallons of toxic chemicals down their respective sinks. But no one challenges the fundamentals - you'll never see a feature like 'why a cellphone is the box brownie of the 21st century' or 'why camcorders are better than DSLRs for most people' in a photo magazine - because that would disenfranchise the regular readers.
Which is why those who read
Stereophile see it as an authority on sound, while we see something more concerned with presenting the perception of authority.
Curiously, I do think this is reader-led, rather than advertising-led, because of magazines like
Hi-Fi Critic in the UK. This relies on subscription and takes no advertising, yet holds to precisely the same audiophile mind-set seen in all the other magazines. Were it to challenge this mind-set head-on, I suspect it would have no subscription base.
Stereophile makes no claim concerning the scientific rigor behind its published opinions...
IME Stereophile is reliatvely unique among modern secular publications in that it routinely publishes opinions that grotesquely and shamelessly violate our society's norms for scientific reasonbleness and correctness, while making a big show out of its scientific correctness.
Normally, enthusiast magazines publish opinons that are factually true or at least reasonably close to being true.
For example, if a camping magazine publishes an opinon that a water purifier purifies water, there is an excellent possibility that the device in question will in fact take bad-tasting swamp water and make it safe and pallatable.
However, if Stereophile publishes an opinion that an interconnect or speaker cable will make your audio system sound better, there is an excellent possibility that no such thing will ever happen. Furthermore, there is an excellent possibility that there no known scientific means by which such a thing could ever happen.
Perhaps Stereophile should take its lead from Analog Science Fiction magazine and bill itself as a journal of Science Fiction and Science Fact!
False negative does not always mean identically the same as Type II error.
Others disagree: http://en.wikipedia.org/wiki/Type_I_and_type_II_errors (http://en.wikipedia.org/wiki/Type_I_and_type_II_errors)
As usual John, you misrepresent the references you cite. Just because a type two error is on occasion called a "false negative" does not mean that the only meaning of "false negative" is a type 2 error. In the recent past you've even made a big show out of observing in a response to me that just because all house cats are felines does not mean that all felines are house cats. I guess that not only is your understanding of logic and rehetoric fading, you can't even remember things youve written in the recent past.
Now getting back to the chase John, why can't you provide us with unambigious evidence that high rez recording techniques provide even a different sound, let alone a better sound.
As I said in the message to which you originally posted, Mr. Krueger, a group of AES illuminati are working on exactly that project.
So what? I understand that those people are also eating and sleeping, but I'm not depending on them to do my eating and sleeping for me.
Furthermore, just because those people are working on a project does not mean that they will ever complete it with an unambigious result. As I understand it, this is a voluteer project - none of them are being paid by the AES for their time and talents. Therfore, they are under no obligation to make this project a high priority, or give it any priority at all.
Why am I obliged to duplicate their efforts?
Because John your past claims and comments have painted the picture that this should be a relatively easy project. You and/or your magazine has never said that it would take a team of AES illuminati to set up an audio system or make a recording that would unambigiously demonstrate the benefits of high sample rate recordings. You've made representations that would seem to say the exact opposite.
John, are you now recanting on your past statements and now saying that only AES illuminati can set up audio systems and make recordings that unambigiously demonstrate the benefits of high sample rate recordings?
John, are you now recanting on your past statements and now saying that only with great difficulty and extreme expense can audiophiles and music lovers set up audio systems and make recordings that unambigiously demonstrate the benefits of high sample rate recordings?
No, that was George Massenburg. Any reference I have made in this thread to difference tracks involved some dems I did last year.
OK, so the 'dem' we're talking about now featured you saying things like 'notice the change in rhythm here...the loss of low-level detail here..' as you presented A and B...things like that? Feel free to be specific and correct me if I'm wrong.
...and then you played them the Frankenstein track that started with lossless 'hi rez' and ended with a 2003 vintage 128kbps encode. No spoken intro, just something like, 'Now listen to this'. And then afterwards you asked them something like 'what did that sound like'? And during the 'discussion' some number of people said something like 'well, it started out great, but sounded worse and worse".
Close?
But no cigar. I described what happened is precise terms. I refer others to that description.
I'd be happy to refer readers of this thread to precise terms too... how about a link? Or just restate them.
Did any of the listeners note a three-stage reduction in quality? Did any just note that it sounded worse at the end than at the start?
As I wrote, most listeners felt that there had been a degradation. That was all I felt necessary. given the premises underlying the presentations that I have described in earlier postings.
So, you said 'What did that sound like to you?" and they said 'That sounded worse at the end than it did at the beginning?' Or what? Feel free to be specific
(post split in two because of quote restriction)
(reply continued from above)
Now, let me see if I have it finally straight:
-you don't think the number of listeners who replied ...and how they replied...matters as much as the fact that 10 groups of 20 'audiophiles' were involved
I have not said anything like that. I refer you to my previous posts.
Where, IIRC you indicate it was groups of 20, and a total of 200, but no clear breakdown of who said what. Let's use this thread to get all the specifics in one place, shall we? Because it's you who seems to think this 'dem' -- and particularly the audience response -- proved something about the sound of the samples you presented. Let's interrogate that claim.
-you seem to think your 'teaching demo' (which may have grossly exaggerated the audible effects) did not load any expectations into the 'audiophile' audience during the 'single blind listening demo'.
Again I didn't say that.
Hence the word 'seem'. But does this mean you think you may have loaded the audience of audiophiles with at least some expectations of difference?
So, any recordings of this event?
Sorry, no.
Pity. Memory can be so unreliable.
Oh,
great. Another massive slugfest between JA and Arny. Just what I always wanted to read.
So far, the biggest take home points I'm seeing are
- JA does not understand what an ad hominem attack is
- Arny does a great Statler/Waldorf impression
- Everybody seems to be insulting everybody else in this thread
- even widely respected audio engineers like Massenberg can get the theory wrong (which meshes nicely with the debate over Dr. Kunchur's papers at the Stereophile forums, suggesting that PhDs in physics can also get the theory wrong)
- PRaT actually does have a justifiable meaning in the case of compressors and other nonlinear devices that operate primarily in the time domain over a period of many milliseconds. But virtually all linear filters (to say nothing of other nonlinear effects like quantization error) don't fit that description
- The AES refused to print a letter of criticism about Meyer/Moran from acknowledged pros? Wow, that kinda supports the whole 'cabal' conspiracy theory some people keep ranting on about. Bad form.
Beyond that, I'm not seeing any new points in this thread that haven't been discussed before with JA here; AFAIK I am not aware that these objections have ever been satisfactorily resolved:
- The usage of difference signals as a methodology for identifying sound quality (and more specifically the lack of quality in lossy codecs) is fundamentally flawed, due to the explicit rejection of masking effects (which are the entire point of lossy encoding in the first place)
- The MP3 encoders available in Audition and Pro Tools are very strongly believed to not be representative of the sound quality of commonly available lossy encodings today, and are strongly believed to be substantially worse - thus using them is something of a strawman
- While type II error does exist, IMHO it is strongly exaggerated in most listening situations, and might even be under 5% for many ABX tests conducted today
Calling Mr Atkinson a troll is probably taking things a bit far. This is something of a hostile environment for the man and he's hardly likely to convert HA members to Stereophile readers. That in and of itself deserves some kudos, at least compared to those who hide behind their editorial pulpits.
Hi, HHGTTG man: I'm not talking about Atkinson's life-role, just his participation here.
The issue is very simple: however much perceived differences might require metaphorical and imprecise language to describe in terms of subjective perception, DBT is the only way to be absolutely certain that the differences in subjective experience are the result of audio differences, and not other factors. Anybody who denies this just ain't interested in the truth, and in the worlds of wine tasting and the grading of student work in English Literature (both fields not normally noted for scientism), something like DBT is accepted as a way of checking on results.
It is, however, always open to someone to claim that a particular method of DBT is in some way flawed, and doesn't permit expert listeners to perform at their best. In such a case, what someone interested in approximating to the truth would do would be to work on some mode of DBT which isolates the audio component of experience, whilst not interfering with proper critical listening. In a somewhat co-operative world, people on both sides of the debate would work out a number of possible methods of DBT, in an attempt to establish the limits of human hearing, and the abilities of those few people we know to exist who have statistical outlier hearing.
Atkinson is just prolonging this thread by quibbling about minutiae, verbal details of posts, and the other things beloved of fifteen-year-old competitive debaters.
I guess Atkinson may be in part motivated by a desire for revenge for participation by HA members in woo forums, where, doubtless, some participants see the cold voice of reason as a form of trolling. I've got a certain sympathy for that view, since such forums seem to contain people who are bolstering a flaky self image by claims for super hearing powers, and the willingness to spend a lot of money. They are sincere, in that they are not driven by ulterior motives, though not, in the Existentialist sense, authentic: at some level they've got to have doubts about what they're saying. And, of course, some HA people can get pretty abrasive in their pursuit of truth, and sometimes mistake subtleties, nuances, and genuine questions for endorsement of woo.
I actually think Atkinson probably knows woo when he sees it. At the most favourable, you could see his position as being like that of a steely-minded Vatican bureaucrat forced to grit his teeth and go along with the canonization of Padre Pio, because saying what he really thinks would disturb the faithful. But that is probably too charitable a view of where he really stands.
The issue is very simple: however much perceived differences might require metaphorical and imprecise language to describe in terms of subjective perception, DBT is the only way to be absolutely certain that the differences in subjective experience are the result of audio differences, and not other factors. Anybody who denies this just ain't interested in the truth, and in the worlds of wine tasting and the grading of student work in English Literature (both fields not normally noted for scientism), something like DBT is accepted as a way of checking on results.
Never use the words 'absolutely certain' if you are going to rely on audio DBTs -- or scientific experiments generally. The results are going to be probabilities, not
absolute certainties. That's plenty good enough for science...and BETTER than sighted comparison!
The issue is very simple: however much perceived differences might require metaphorical and imprecise language to describe in terms of subjective perception, DBT is the only way to be absolutely certain that the differences in subjective experience are the result of audio differences, and not other factors. Anybody who denies this just ain't interested in the truth, and in the worlds of wine tasting and the grading of student work in English Literature (both fields not normally noted for scientism), something like DBT is accepted as a way of checking on results.
A DBT is the most robust method we currently have of determining whether there is or is not a difference in audio performance. Does that mean it will remain the only method? Who knows? Someone may come up with an equally valid method of testing today, tomorrow or in 20 years time that demonstrates superiority over DBTs in key areas.
This is not so far-fetched as it first appears. What if a method of longitudinal testing was found to return more consistent, reliable and repeatable results that DBTs, for example? Would you jump to dismiss such findings as not truth-seeking?
Note, this does not let simple sighted-test subjectivism in by the back door. That's already dismissed as flawed. Note also that even the most stringent test does not guarantee certainty, because any scientific test is subject to refutation and falsifiability.
It is, however, always open to someone to claim that a particular method of DBT is in some way flawed, and doesn't permit expert listeners to perform at their best. In such a case, what someone interested in approximating to the truth would do would be to work on some mode of DBT which isolates the audio component of experience, whilst not interfering with proper critical listening. In a somewhat co-operative world, people on both sides of the debate would work out a number of possible methods of DBT, in an attempt to establish the limits of human hearing, and the abilities of those few people we know to exist who have statistical outlier hearing.
Here's how the claims that a particular method of DBT is in some way flawed come to be. The claimant *knows* (based on sighted evaluations and other equally falliable means) that the audible difference exists. The DBT does not support his belief. Therefore the DBT is flawed. Thats how it was when we first started presenting DBTs to high end audiophiles back in the late 1970s, and that is how it is today.
Here is how to estimate the sensitivity of a particular method of doing listening tests. Contrive a series of tests where the audible difference can be reduced in increments, and see how the listener sensitivity for the audible difference stacks up in comparisons with other means of doing tests. I've done this many times, and if anything ABX equals or betters other testing methods.
I've done this many times, and if anything ABX equals or betters other testing methods.
What are the (DBT) alternatives to ABX ?
Are there significant differences in the results between various protocols ?
A DBT is the most robust method we currently have of determining whether there is or is not a difference in audio performance. Does that mean it will remain the only method? Who knows? Someone may come up with an equally valid method of testing today, tomorrow or in 20 years time that demonstrates superiority over DBTs in key areas.
This is not so far-fetched as it first appears. What if a method of longitudinal testing was found to return more consistent, reliable and repeatable results that DBTs, for example? Would you jump to dismiss such findings as not truth-seeking?
Perhaps I overstated the case (I'm absolutely certain that "absolutely certain" was overstated, though "practically certain" seems pretty right). However, given that what we are concerned with is the audibility of phenomena to a listening subject; and given that we know that all sorts of factors can affect the listening experience, including expectations about the item(s) under test: then some way of testing using a human subject who does not know the identity of the items really would seem inescapable.
There might very well be ways of assuring this ignorance more subtle and more reliable than current ABX testing procedures (I think of the famous power cord test as an example), but this requirement for judgement devoid of expectations seems to be required by the nature of the situation, not by limitations of technology and human ingenuity. I mean, you could, conceivably, do brain scans of subjects listening to 24/96, but then, supposing you found measurable differences in brain function, you'd have to validate any conclusion that these differences related to subjectively important differences in audio experience. And to do that, surely you'd need to have a DBT in the loop somewhere?
And why might all this matter? I think, ultimately, it's a matter of ethics. Resources are limited. Although extravagance can be justified, waste can't. Wanting the best possible audio experience is extravagance, but we live with that. But applying resources to, say, high-priced interconnects, rather than to a nice chair or good Scotch, both of which would have a more soundly based effect on the audio experience, is Waste, and Waste is Bad.
I've done this many times, and if anything ABX equals or betters other testing methods.
What are the (DBT) alternatives to ABX ?
The leading alternative to ABX is ABC/hr. ABC/hr is not exactly an alternative to ABX because it is a test for degree of impairment, not a pure test for audible differences. However, if two alternatives have statisticaly different degrees of impairment, that is very strong evidence that they do sound different.
The ABX test that you see used in articles about things like articulation of speech differs from our ABX in that there is only one sequence of AXB per trial, instead as many sequences of AX and BX as the listener desires. Again, it is testing for something a little different than what we are testing for.
The concept of ABX can be extended to more different alternatives being compared in one trial, giving you something like a ABCX or ABCDX test. The people who do taste tests on food have a number of other variations. Krab has written about them here in the past, I think.
There's a testing style that you may have encountered in a audiologist's office where you push a button whenever you start hearing a tone that is increasing in intensity, and when you push the button the tone starts getting softer. This can be extended to any audible effect that you can smoothly increase and decrease at will. This kind of test was used in the Dolby labs jitter tests that were written up in the JAES some years back. This test is not completely foolproof, but its pretty good if your subjects are honest and don't try to spoof it.
Are there significant differences in the results between various protocols ?
There are DBT protocols that are possible, but do put the listener at a big disadvantage. We devised several of them before we hit on ABX. For example, we had a protocol that was based on just listening to a sequence of Xs, and requiring the listener to state whether each was A or B. That one put much to much burden on the listener's memory of what A and B actually sounded like. We tried it once and the next step was ABX.
Any protocol that involves manual switching is at best suspect because the switching delays are usually so long. Since this is how most audiophiles do their listening tests, besides all of the false positives due to sighted listening, no level-matching, and no time-synch, the remaining responses are probably false negatives due to long switching delays.
If they are executed properly, and can possibly produce comparable results, then DBT test results strongly tend to converge. The strongest determiners in the sensitivity of good listening test results are listener training and choice of music. The results of good listening tests also converge with results that can be developed by other means, such as analysis of the structure of the human ear.
Hmm, when he played back the "error noise" on the same level as the original, that would explain the reason for loud levels. The noise would be not very impressive unless turned up.
I don't think it was "turned up" to be louder than it would have been. He literally had the original, and the mp3, on two adjacent tracks in Pro Tools. He inverted the original, and played it at the same time as the mp3. There was no level fiddling (except at one part, where I think he inadvertently knocked one of the faders, and you could hear the music itself didn't quite cancel - he immediately heard and corrected this).
Cheers,
David.
Thanks Andy.
Wouldn't it be a good idea for important tests to use a "placebo control group" for which stimulus B=A ? If the test is well setup, the result should be close to 50%, right ?
A few days ago there was an interesting post by Bob Katz on a mailing list about audibility of jitter (internal/external clocking of DACs). He used a somewhat different approach, but still DBT. To me it looks like a correct test, although the conclusions are perhaps a bit optimistic according to HA criteria. Any comments ? NB: the topic was about audible differences between different pressings of data-identical cd's.
>Always willing to have the 'ears' prove science wrong.
>We synced two Master Disk CD players into a common DtoA so we could A/B the playing CDs.
>No one was able to consistently pick the SHM-CD while blind swapping disc between players.
That's one of the toughest tests in the world to pass blind. And I feel you've put an obstacle in by the very nature of your test. The prime problem is the listening method you used for the blind test. What you chose was the "serial listening" method, which totally screws up because it assumes that music is identical at the switch point! You're not listening to pink noise :-). Music is fluid, it's constantly changing. The ear can get completely confused, even by small changes in dynamics at the switch point. For example, if you cue into the drumbeat, we "expect" that the drum beat will sound the same, but if it's a human drummer, each drum beat is minutely different, and so, you get fooled. If you cue into the ambience, and the drummer hits the drum just slightly harder at the switch point, the ambience will go up! It's nearly impossible to make anything but random conclusions serially!*
The most effective method I've found for conducting blind testing of extremely subtle differences is to:
a---find a perfect 30 second passage. This passage should be "reasonably consistent" and not have any momentary dynamic surprises in it to confuse the listener
b) familiarize the listeners with the loudspeakers, room, and especially the test material
c) TRAIN the listeners on the test material
d) Now, decide amongst yourselves which of A or B you THINK you prefer. That's what you are now trying to prove: A is better than B, or vice versa. Now, let's assume for the moment that you decided A is better, sighted. Prior to the blind test, you're comfortable and as convinced as you can be that you can tell A from B and that you prefer A.
e) For each trial, play 30 seconds of A. Then play 30 seconds of A again. Now play 30 seconds of X (the unknown). Then 30 seconds of X again. Then finish with 30 seconds of A. That's 5 repetitions, 30 seconds each. For each trial. Note how the order of the repetitions helps the listener to feel comfortable about his decision. The fifth repetition is a further assurance. All this is geared to helping the listener to make and feel comfortable with his decision. This works, try it!
How do I know it works? Well, I recently conducted a jitter listening test for 8 listeners (5 in one session, 3 in another), doing exactly the above testing method. 10 trials times 8 subjects equals 80 trials. The score was 60% (48 trials out of 80), which is above chance, but an amazing result for such an extremely difficult and (you would admit) subtle test. I personally scored 7 out of 10 and feel I would have gotten better if I could ust maintain my concentration! The test requires continuous concentration and it is no fun!!!!
What was the test: Listening to the Pro Tools HD 192 DAC with a stereo recording, and comparing Pro Tools on internal clock versus the Grimm CC 1 clock. The Grimm clock won and all listeners felt it improved the depth and stereo image. Plus, with a p of 2.5, the assurance that we got a valid result is pretty high. p of 2.5 means there's only a 2.5% chance that the listeners got these reults "out of luck" or by chance.
That said, I've never been able to pass a blind test on data-identical CDs which I feel (sighted) sound different. Jitter tests on "data-identical CDs" are the most difficult blind test to conduct (see the difficulty of trying to synchronize two players and play just a 30 second pasage!!!!!) and so the controversy lingers on.
* serial is an "ok" method for us mastering engineers who are working on eqs and such, though I try to average it out over time and if possible, rewind and play the passage again with the different EQ, because we can also easily get fooled by short term changes in the music getting confused with changes in EQ!
Bob Katz
[/size]
It is not me who is saying that designing such a test is complex; instead I have been reporting what those involved in doing so - "AES Fellows, some university professors, some well-known recording and mastering engineers, and even JJ" - are finding during the planning of such a test. If you are offering the opinion that those people are "clueless," you need to rethink that position.
I don't think that you are completely honest here. You know the basic procedure: Compare a plain 88.2 kHz record to a time synchronized 88.2 kHz -> 44.1 kHz -> 88 kHz record. If that shows, that no difference can be heard under double blind conditions, you already have very solid results and it doesn't get any more complicated than that.
But you haven't eliminated all variables and the result may still be a false negative, thus not transportable. All you can conclude is with _that_ recording_ with _that_ hardware and _that_ testing protocol, no difference could be identified to a given degree of statistical certainty.
Presumably you'd use a combination of recording and hardware that your expert listeners say demonstrates the huge advantage hi-res has over CD. So, these are irrelevant criticisms (given that you've previously said you trust your expert listeners).
So what you're saying is that "all" you can conclude is that you may prove that "with _that_ testing protocol, no difference could be identified to a given degree of statistical certainty."
I can't believe the protagonists in this debate have the energy to run the obvious subsequent arguments again.
Cheers,
David.
Wouldn't it be a good idea for important tests to use a "placebo control group" for which stimulus B=A ? If the test is well setup, the result should be close to 50%, right ?
Yes. I have definately done tests where I could get very high scores even when A was nominally the same as B. In fact I can do it with the QSC ABX box any time I want to, with nothing attached to it at all. The relay noise is not random. To be effective it has to be acoustically separated from the listeners.
The most effective method I've found for conducting blind testing of extremely subtle differences is to:
a---find a perfect 30 second passage. This passage should be "reasonably consistent" and not have any momentary dynamic surprises in it to confuse the listener
b) familiarize the listeners with the loudspeakers, room, and especially the test material
c) TRAIN the listeners on the test material
This begs the question of how to teach people to hear a difference that they may not be able to hear. My preferred methodolgy is to come up with a means for varying the strength of the artifact being tested, and start people out with super strong and work down to the actual test of interest.
d) Now, decide amongst yourselves which of A or B you THINK you prefer. That's what you are now trying to prove: A is better than B, or vice versa. Now, let's assume for the moment that you decided A is better, sighted. Prior to the blind test, you're comfortable and as convinced as you can be that you can tell A from B and that you prefer A.
This violates a basic rule of testing - communiation among the listeners violates the basic assumptions of most statistical analysis techniques. One of the symptoms of this problem is scores that are worse than random guessing.
e) For each trial, play 30 seconds of A. Then play 30 seconds of A again. Now play 30 seconds of X (the unknown). Then 30 seconds of X again. Then finish with 30 seconds of A. That's 5 repetitions, 30 seconds each. For each trial. Note how the order of the repetitions helps the listener to feel comfortable about his decision. The fifth repetition is a further assurance. All this is geared to helping the listener to make and feel comfortable with his decision. This works, try it!
I can show that in general samples this long are far less sensitive for small differences than samples on the order of just a few seconds.
I suspect that the switching was not quick enough and not under the control of the listeners.
How do I know it works? Well, I recently conducted a jitter listening test for 8 listeners (5 in one session, 3 in another), doing exactly the above testing method. 10 trials times 8 subjects equals 80 trials. The score was 60% (48 trials out of 80), which is above chance, but an amazing result for such an extremely difficult and (you would admit) subtle test. I personally scored 7 out of 10 and feel I would have gotten better if I could ust maintain my concentration! The test requires continuous concentration and it is no fun!!!!
What was the test: Listening to the Pro Tools HD 192 DAC with a stereo recording, and comparing Pro Tools on internal clock versus the Grimm CC 1 clock. The Grimm clock won and all listeners felt it improved the depth and stereo image. Plus, with a p of 2.5, the assurance that we got a valid result is pretty high. p of 2.5 means there's only a 2.5% chance that the listeners got these reults "out of luck" or by chance.
I've got my doubts on the grounds of questionable double blindness.
A single blind test is just a defective DBT.
General comment - these guys seem to be making the classic mistake of going for a "world class" restult without working their way up with some easier tests. This is just like trying to do a 24/96 test without working your way up from 12/32. IOW, they are trying to do the audio equivalent of breaking the 4 minute mile in their first middle school track meet.
An mp3 of the lecture (without pre-release musical extracts) is now available on-line:
www . mediafire . com/?8exbl1v60gcw2tz
(remove the spaces from the URL to make it work)
Poor (dictaphone!) quality - though it's still easier to hear what George is saying on this recording than it was where I was sat.
The criticism of mp3 starts about 12 minutes in (though the music examples are missing - replaced by short beeps)
The post-lecture discussion of the Myer and Moran paper starts at 1:16:40
There were many other interesting parts, but I don't have a spare 2 hours to listen to the whole thing and index them.
Cheers,
David.
Presumably you'd use a combination of recording and hardware that your expert listeners say demonstrates the huge advantage hi-res has over CD. So, these are irrelevant criticisms (given that you've previously said you trust your expert listeners).
I think that Atkinson is implying that using equipment that is all Stereophile Class "A" would not be good enough.
I see his concern there because his Class A list includes some real question marks like all that vacuum tube stuff, especially the SETs.
He's also implying that his staff would not be good enough.
I'm seeing his concern there because he's recently had to overrule one of his lead columnists - Fremer. If I've got their ages pegged right I'd have more than a few doubts of my own.
So what you're saying is that "all" you can conclude is that you may prove that "with _that_ testing protocol, no difference could be identified to a given degree of statistical certainty."
Of course we don't know exactly what *that* protocol is or would be. Based on past performance, Atkinson's choice of test protocol would be non-standard and have a few question marks of its own.
I can't believe the protagonists in this debate have the energy to run the obvious subsequent arguments again.
Different nuances? ;-)
I described what happened is precise terms. I refer others to that description.
I'd be happy to refer readers of this thread to precise terms too... how about a link? Or just restate them.
Sorry for the tardy response, Mr. Sullivan. Sometimes increasing the influence and success of what Arny Krueger has referred to as the "Evil High-End Audio Establishment" and what more level-headed fellows call "Stereophile magazine" has to take precedence over posting to Internet forums. :-)
I have to say that I don't comprehend what you are trying to achieve in this thread by posting and reposting the same questions, often offering your own incorrect paraphrase of what I said or your unsupported conjectures concerning my motivations and conclusions. I have described the presentations I gave in Colorado in considerable detail in this thread - I refer you to postings numbers #37, #49, #55, #65, #69, #85, #93, and #112, which I know you read because you responded to many of them.
Where, IIRC you indicate it was groups of 20, and a total of 200, but no clear breakdown of who said what. Let's use this thread to get all the specifics in one place, shall we?
Reread the numbered posts of mine, in this very thread, that I listed above, Mr. Sullivan, and you will find a full description of what the events were and what happened at each one.
Because it's you who seems to think this 'dem' -- and particularly the audience response -- proved something about the sound of the samples you presented.
Really? I don't see any claim I made concerning "proof" of anything, Mr. Sullivan. All I did was to describe the reactions of listeners to the demonstration. AS to what I "seemed to think," I wrote back in April on HA (on April 26, 2009, 10:40am, in the "Why We Need Audiophiles..." thread) in response to the questions you were putting to me at that time what my motives were: I wanted the listeners to audition hi-rez PCM data under optimal circumstances and I was also interested in exposing listeners to various data-reduced formats so that they could decide for _themselves_ whether hi-rez formats are necessary, whether CD is good enough for serious listening, and whether the lossy versions are sonically compromised or not. There was no scoring, no tallying of results, and no other detail other than what I have already offered you.
If you feel that to be insufficient, my apologies. All I can suggest is that if and when I repeat the event somewhere closer to where you live, you attend and ask questions in person at that time (the same invitation I offered to all HA subscribers back in April).
John Atkinson
Editor, Stereophile
An mp3 of the lecture (without pre-release musical extracts) is now available on-line:
http://www.mediafire.com/?0omijnwtzty (http://www.mediafire.com/?0omijnwtzty)
Thank you for posting the link, David. I shall listen to George's comments with interest.
John Atkinson
Editor, Stereophile
d) Now, decide amongst yourselves which of A or B you THINK you prefer. That's what you are now trying to prove: A is better than B, or vice versa. Now, let's assume for the moment that you decided A is better, sighted. Prior to the blind test, you're comfortable and as convinced as you can be that you can tell A from B and that you prefer A.
This violates a basic rule of testing - communication among the listeners violates the basic assumptions of most statistical analysis techniques. One of the symptoms of this problem is scores that are worse than random guessing.
I don't think Katz is saying anyone should communicate with anyone. He's not a novice at conducting blind tests. Katz is just saying that a subject should believe that A and B really do sound different, before proceeding to try to identify them 'blind'.
d) Now, decide amongst yourselves which of A or B you THINK you prefer. That's what you are now trying to prove: A is better than B, or vice versa. Now, let's assume for the moment that you decided A is better, sighted. Prior to the blind test, you're comfortable and as convinced as you can be that you can tell A from B and that you prefer A.
This violates a basic rule of testing - communication among the listeners violates the basic assumptions of most statistical analysis techniques. One of the symptoms of this problem is scores that are worse than random guessing.
I don't think Katz is saying anyone should communicate with anyone. He's not a novice at conducting blind tests. Katz is just saying that a subject should believe that A and B really do sound different, before proceeding to try to identify them 'blind'.
Which, btw, is really good advice.
Mr. Atkinson, your claims about this 'dem' have been spread over two threads. You are the one who touted it as a demonstration of
something. It has been a chore getting a full description of your 'dem' from you, and even now aspects of it remain unclear. So please don't get cranky when I suggest responses be kept pertinent and comprehensively informative to the questions asked. All you have done in the most recent reply to me is refer me to previous posts in this thread, all of which raised methodological questions of their own, which were not fully answered. The purpose of the 'progressive' degradation demo , and whether it was actually perceived as such, remains unclear; the reason given for using an obsolete codec were frankly,
lame; the role of expectation bias in the 'results' was essentially unaddressed, and the means by which subject response was gauged, seems spotty as described.
As for not trying to 'prove' anything, the first mention of your 'dem' on this thread ((post #37 (http://www.hydrogenaudio.org/forums/index.php?showtopic=72446&view=findpost&p=639711))) was as evidence that low bitrate lossy affects perceived 'rhythm' and 'dynamics' of a track. If not to support a point, why would you mention it at all?
AS to what I "seemed to think," I wrote back in April on HA (on April 26, 2009, 10:40am, in the "Why We Need Audiophiles..." thread) in response to the questions you were putting to me at that time what my motives were: I wanted the listeners to audition hi-rez PCM data under optimal circumstances
and I was also interested in exposing listeners to various data-reduced formats so that they could decide for _themselves_ whether hi-rez formats are necessary, whether CD is good enough for serious listening, and whether the lossy versions are sonically compromised or not. There was no scoring, no tallying of results, and no other detail other than what I have already offered you.
If you feel that to be insufficient, my apologies. All I can suggest is that if and when I repeat the event somewhere closer to where you live, you attend and ask questions in person at that time (the same invitation I offered to all HA subscribers back in April).
John Atkinson
Editor, Stereophile
It cannot possibly have been 'optimal' listening conditions, as described, and yes, we already know what you claimed to be demonstrating.
Let me try to phrase this in a way that you cannot possibly wiggle away from on specious semantic grounds (a mighty challenge, btw):
You
still seem to think that your dem offered a valid means to compare whether hi-rez formats were better than CD, and whether lossy versions are sonically compromised. However, you do
not claim anything was
proved by it.
Correct?
d) Now, decide amongst yourselves which of A or B you THINK you prefer. That's what you are now trying to prove: A is better than B, or vice versa. Now, let's assume for the moment that you decided A is better, sighted. Prior to the blind test, you're comfortable and as convinced as you can be that you can tell A from B and that you prefer A.
This violates a basic rule of testing - communication among the listeners violates the basic assumptions of most statistical analysis techniques. One of the symptoms of this problem is scores that are worse than random guessing.
I don't think Katz is saying anyone should communicate with anyone. He's not a novice at conducting blind tests. Katz is just saying that a subject should believe that A and B really do sound different, before proceeding to try to identify them 'blind'.
Which, btw, is really good advice.
I'd say it required, otherwise why bother with the blind test?
d) Now, decide amongst yourselves which of A or B you THINK you prefer. That's what you are now trying to prove: A is better than B, or vice versa. Now, let's assume for the moment that you decided A is better, sighted. Prior to the blind test, you're comfortable and as convinced as you can be that you can tell A from B and that you prefer A.
This violates a basic rule of testing - communication among the listeners violates the basic assumptions of most statistical analysis techniques. One of the symptoms of this problem is scores that are worse than random guessing.
I don't think Katz is saying anyone should communicate with anyone. He's not a novice at conducting blind tests. Katz is just saying that a subject should believe that A and B really do sound different, before proceeding to try to identify them 'blind'.
I see your point. There's a possibility that he means this happens before the test and not during the test. If it happens before the test, then no harm, no foul. We did something like this before all of the amplifier and CD player tests we did for Stereo Review.
Katz is just saying that a subject should believe that A and B really do sound different, before proceeding to try to identify them 'blind'.
Which, btw, is really good advice.
I'd say it required, otherwise why bother with the blind test?
So you don't think that people who are agnostic about the outcome of the test should could or would do the test?
I've definately seen cases where people who did not in the slightest believe that the difference existed produced positive results. In some cases they didn't believe that they reliably heard anything even after the test!
IME at the threshold of audibility, it is not clear that a diffference exists. However, the so-called guesses are more reliable than chance.
There can be a professional approach to listening where people simply do their best and let the results be what they will be.
On balance, believing that the difference exist might make people more enthusiastic, and therefore find the test easier.
But again, belief that the difference exists can and definately has lead to overconfidence and poor results. Sometimes a test with a questionable outcome leads to a sense of determination that gives better results the next time around.
I don't see the need for a collaborator, however.
Well, I don't (http://www.ethanwiner.com/articles.html) need (http://www.realtraps.com/articles.htm) a collaborator (http://www.realtraps.com/videos.htm) either. I just thought it would be fun, and maybe help to reach more people.
--Ethan
Mr. Atkinson, your claims about this 'dem' have been spread over two threads.
I haven't made any claims as such, Mr. Sullivan. I have offered descriptions of the dems in response to questions from you and others.
You are the one who touted it as a demonstration of something. It has been a chore getting a full description of your 'dem' from you, and even now aspects of it remain unclear. So please don't get cranky when I suggest responses be kept pertinent and comprehensively informative to the questions asked.
I have now three times described the purpose of of these presentations, Mr. Sullivan. For a fourth time: I wanted the listeners to audition hi-rez PCM data under optimal circumstances, mainly using the 24-bit/88.2kHz masters of my own commercially released CDs, and I was also interested in exposing listeners to various data-reduced formats so that they could decide for themselves whether hi-rez formats are necessary, whether CD is good enough for serious listening, and whether the lossy versions are sonically compromised or not. I also demonstrated how miking and mixing techniques affected the listener's perception of the recorded soundstage. I have described all the relevant details; I fail to grasp what "remains unclear" to you and, frankly, I am not concerned.
All you have done in the most recent reply to me is refer me to previous posts in this thread, all of which raised methodological questions of their own, which were not fully answered. The purpose of the 'progressive' degradation demo , and whether it was actually perceived as such, remains unclear; the reason given for using an obsolete codec were frankly, lame; the role of expectation bias in the 'results' was essentially unaddressed, and the means by which subject response was gauged, seems spotty as described.
All in your opinion, Mr. Sullivan, and as I have said before, I see no point in arguing with other's opinions. That I disagree with your characterizations is all that needs to be said, I would have thought.
As for not trying to 'prove' anything, the first mention of your 'dem' on this thread ((post #37 (http://www.hydrogenaudio.org/forums/index.php?showtopic=72446&view=findpost&p=639711))) was as evidence that low bitrate lossy affects perceived 'rhythm' and 'dynamics' of a track. If not to support a point, why would you mention it at all?
I offered it as anecdotal support for the point being made in response to a question from another poster. It certainly wasn't intended as "proof," otherwise I would have identified as such. The word "proof" was your own interpolation, Mr. Sullivan, again it seems a pointless exercise for me to argue in support of the words you put in my mouth.
As to what I "seemed to think," I wrote back in April on HA (on April 26, 2009, 10:40am, in the "Why We Need Audiophiles..." thread) in response to the questions you were putting to me at that time what my motives were: I wanted the listeners to audition hi-rez PCM data under optimal circumstances...
It cannot possibly have been 'optimal' listening conditions, as described...
Why not? The system was top-class, the rooms were acoustically well-designed and _very_ quiet and as I have already mentioned, I eliminated as many interfering variables in the comparisons as was possible. Yes, having up to 20 listeners at a time in the room meant that only a fraction could be optimally seated to perceive the stereo imaging, but as I sadi, the public turnout was much larger than anyone had predicted.
Let me try to phrase this in a way that you cannot possibly wiggle away from on specious semantic grounds (a mighty challenge, btw)
Ah, the obligatory cheap shot...Seriously, as a professional and successful writer I choose the words I use with care and I do try to convey the maximum meaning in the most direct manner. That you regard my comments as "wiggling" on "specious semantic grounds" does suggest that I have failed to achieve that goal. But as I said above, what I _do_ refuse to address are the many instances where you, or others like Arny Krueger, have put words in my mouth by saying that if I said "A," then I must have meant "B." Please note that if I said, "A," then "A" was what I meant. If the meaning of "A" escapes you, then you either didn't read what I wrote carefully enough or you didn't comprehend it.
You still seem to think that your dem offered a valid means to compare whether hi-rez formats were better than CD, and whether lossy versions are sonically compromised.
Yes. Having had considerable experience of demonstrations organized by others over the many years I have been involved in this subject, I think that the circumstances of these demonstrations were of appropriately high quality for differences, if they existed, to be perceived.
However, you do not claim anything was proved by it.
Correct?
That is correct. I have now 4 times offered my motives for doing these demonstrations and the related goals. Those goals were met. But as "proof" of anything? I have no idea whether the answers to the questions I listed were overall positive or negative, nor does the outcome concern me. As I have repeatedly told you, these werent' formal double-blind tests nor have I claimed they were - that was a projection by other posters - nor was there any scoring nor did I tally any results. All I was doing was allowing listeners to hear under optimal conditions the examples I had prepared, which they wouldn't have had access to otherwise, and to decide for themselves whether there were differences that were important to them or not.
John Atkinson
Editor, Stereophile
I'm bemused by the hostility to JA on this particular point.
I mean, he played 128kbps CBR FastEnc after some other stuff, and some people described the quality as going down hill.
It surprises me that people heard a difference a) because it's hard to set up a system in a room and let lots of people hear decent audio from it, and b) because most people are deaf and/or don't care and/or don't know what to listen for.
Both both of those factors can be overcome - if this happened, is anyone really surprised that people heard the quality go down hill?
(Not to mention, I'm not following properly, but I'm guessing the encoder was FhG FastEnc - I think the non-buggy versions are good for CBR, but at least two different buggy versions are available, and either of these would be painfully obvious).
You can't extrapolate at all that the CD quality stuff sounded worse than 24/96 - but as described, the 128kbps mp3 was played last - if that was the only thing that sounded different, the quality would still have been heard to "go down hill".
Cheers,
David.
I'm bemused by the hostility to JA on this particular point.
I mean, he played 128kbps CBR FastEnc after some other stuff, and some people described the quality as going down hill.
But while you and I might understand what different codecs can mean and how they can 'skew' a demosntration...Mr. Atkinson apparently did not explain that to his audience. And the 128kps presentation was not the only 'dem' Mr. Atkinson claims to be performing -- it purported to compare four different levels of sound quality, though it is unclear what he told the audience in this regard. Nor does his method rise to HA standards for a 'dem' of even 128kps quality. So why the bemusement? You should be kinda 'hostile' too, to such flimflammery. I can't see it helping *clarify* the relationship of mp3 sound to lossless, in the minds of 'audiophiles', it can only prejudice them more.
It surprises me that people heard a difference a) because it's hard to set up a system in a room and let lots of people hear decent audio from it, and b) because most people are deaf and/or don't care and/or don't know what to listen for.
Both both of those factors can be overcome - if this happened, is anyone really surprised that people heard the quality go down hill?
There is a third factor in this 'dem' -- it's hard to determine if people really heard the difference, or whether they were primed to expect some audible degradation between start and end. If an audience is primed to hear a difference , is anyone really surprised that they do?
And if the dem was presented as described, would anyone be surprised if the listeners 'decided' that mp3 'typically' sounds that way in 2009?
Are you comfortable with someone presented as an authority on audio, 'educating' listeners in such fashion? If so, count *me* as the one bemused.
(Not to mention, I'm not following properly, but I'm guessing the encoder was FhG FastEnc - I think the non-buggy versions are good for CBR, but at least two different buggy versions are available, and either of these would be painfully obvious).
According to Mr. Atkinson, the codec was the one supplied with Audition 1.0. That makes it a FgH codec from no later than 2003.
You can't extrapolate at all that the CD quality stuff sounded worse than 24/96 - but as described, the 128kbps mp3 was played last - if that was the only thing that sounded different, the quality would still have been heard to "go down hill".
Indeed, a point I made previously (though Mr. Atkinson attempted to twist (http://www.hydrogenaudio.org/forums/index.php?showtopic=72446&view=findpost&p=640351)it): no one here disagrees that a 128kbps encode COULD sound worse than its lossy source...and all here would agree there are good and bad ways to demonstrate this, as well as ways to help 'ensure' that a difference is audible to the average listener (like, bringing the bitrate down
really low..or using a 'killer' sample...or using a bad codec at a middling bitrate...
none of these would be a fair 'dem' of how mp3 'sounds'.)
We have no details on how long each segment was, but we know that the 128kbps segment followed a 320kbps segment (same codec), and that both had been reconverted back to wav, then upconverted to 88.2/24 bits. We know that subject responses didn't come until after the whoel track was played; we know that responses weren't tallied. We know that some number felt the music sounded 'lifeless' afterwards.
We know that Mr. Atkinson presented this a demonstration of some sort (still unspecified as to what JA actually told them before he played to tracks) for the listeners to 'decide for themselves'.
Are you comfortable with someone presented as an authority on audio, 'educating' listeners in such fashion? If so, count *me* as the one bemused.
I'm not "comfortable" with JA's last two decades' work.
I just don't see the point of picking this demo apart in detail. It's a demo which some people will claim or infer proved a lot, but in fact proved nothing. I think we got that point a while back
Now, here's an interesting thing: the demo in the lecture that this thread is nominally about was even more flawed. You can listen to the discussion in the link I provided.
I was in that audience. There were a few hundred people there - my guess is only five of us realised a) it was flawed, and b) the audio would have had real deficiencies that could have been ABXed over headphones, but any differences that people thought they heard in
that public listening environment were almost certainly imagined.
99% of the people there accepted that mp3 sounds awful (full stop!), that the difference signal accurately demonstrated how awful it sounded, and that they could really hear a clear difference between the original and the mp3 over speakers in a hall with several hundred people. Oh, and that ABX testing is fundamentally flawed.
This wasn't the audiophools subjectivists society - this was the audio engineering society.
You see, there's more serious work to be done than arguing about a JA demo.
Stereophile
might just be a symptom, rather than a cause.
Cheers,
David.
EDIT: The reason I
mind the debate about JA's lecture is because JA is a master debater, and solicits spiteful sounding responses which make HA look weak and unfriendly. The responses are born out of frustration and incredulity, not to mention truth and a trust in science, but a casual observer might think the opposite.
Master debater JA does have a charmingly parsimonious attitude towards doling out details, and a real talent for the semantic two-step, I grant him that. I've tried to be polite about it, in a low pH sort of way. In any case, we all seem to agree now that whatever his demonstration was supposed to be about, it can't honestly be said to have demonstrated much about hi rez vs CD vs 320 vs 128.
I already weighed in twice on the tragic nature of Massenburg's demo, and I doubt listening to the lecture (which I plan to do) is going to make me more sanguine about it. But he's not here defending it, while Atkinson came here, talking up his. If GM posted here, I don't doubt there would be some questions for him, and if he answered in Atkinsonian fashion, I don't doubt the questions would grow pointed. However, he doesn't have Stereophile's illustrious history as a propaganda arm of the 'high end' attached to him, the way JA does, so perhaps it would be a less 'unfriendly' reception. ( I also wonder what audience you are concerned about here...those casual HA observers...don't they *always* have a learning curve to attend to?)
Meanwhile, if you're an AES member, a letter to JAES expressing exactly the concerns you voiced above, is one possible recourse. I would imagine JJ and any AES members that post regularly here, could have a dog in that hunt too.
(I'm still trying to reconcile how the AES can be run by a 'cabal' that censors critiques of Meyer & Moran's JAES article, yet also hosts demonstrations like Massenburg's. Sounds kinda schizo to me )
(I'm still trying to reconcile how the AES can be run by a 'cabal' that censors critiques of Meyer & Moran's JAES article, yet also hosts demonstrations like Massenburg's. Sounds kinda schizo to me )
I don't think there's a cabal. I think there were many individual members at that meeting who work in the recording industry and unquestioningly accept more bits+samples = better. I think the Meyer & Moran paper solicited such a large number of responses that the AES decided to make an open forum on their website, rather than fill their publication with the discussion. I don't think there's an unseen or seen "controlling mind" - it's the sum total of its members.
I could be wrong.
Cheers,
David.
There is a third factor in this 'dem' -- it's hard to determine if people really heard the difference, or whether they were primed to expect some audible degradation between start and end. If an audience is primed to hear a difference , is anyone really surprised that they do?
Your problem with grasping what I did, Mr. Sullivan, appears to stem from your not reading what I wrote. You asked me about this aspect of the dem 3 days ago, in this thread, in the following exchange in message #85.:
you say the listeners 'were not aware of what they were listening to' but what did the listeners know about the track they were listening to?
Nothing, other than it was a recording of Handel's Messiah.
There was nothing unclear about my response, nothing equivocal, nothing that would be hard to comprehend.
We have no details on how long each segment was...
Again, all this information was included in my responses to your questions: the musical excerpt was from Handel's Messiah, identified as "For unto us a Child is born" in the article I linked to earlier in the thread. I told you that the excerpt lasted around 5 minutes and that there were four different versions spliced together, meaning that each segment was around 75 seconds. (The article I linked to explained why this musical piece was used, in that it had the same music repeated three time, meaning that each of the four segments was the same length and was comprised of the same content.)
Again, I fail to grasp why you feel that my description was incomplete. I am sorry to belabor the point, but you seem curiously incapable of reading what I actually took the time to write in response to your ongoing inquisition.
And to address a point you made in another posting:
I'm still trying to reconcile how the AES can be run by a 'cabal' that censors critiques of Meyer & Moran's JAES article, yet also hosts demonstrations like Massenburg's. Sounds kinda schizo to me...
The AES is not a monolithic entity. It is comprised of people with differing points of views and experiences. I have been a member of the AES since 1981, but that doesn't mean that I necessarily agree with, for example, the conclusions of the Meyer-Moran paper.
John Atkinson
Editor, Stereophile
What sense does it make to argue anymore about Atkinson's show-biz demonstration?
It didn't even adhere to the most basic standards. It used a stone age mp3 encoder for the lossy demonstration, presented a difference between 16 bit and 24 bit at unemployable listening levels (that would show any thinkable amount of obviously wasted bits to be audible), mixed it up with non correlating sample rate differences, and it wasn't double blind.
It sed a stone age mp3 encoder for the lossy demonstration,
This is a common occurrence with people who think all MP3 encoders are bad, they don't differentiate between old and new encoders, because they assume that the format itself is inherently flawed to the extent that it has never been improved.
In fact, many don't even realise that different encoder versions effect quality.
What sense does it make to argue anymore about Atkinson's show-biz demonstration?
It didn't even adhere to the most basic standards. It used a stone age mp3 encoder for the lossy demonstration, presented a difference between 16 bit and 24 bit at unemployable listening levels (that would show any thinkable amount of obviously wasted bits to be audible), mixed it up with non correlating sample rate differences, and it wasn't double blind.
I'm prone to defend *any* demonstration that is presented as a demonstration on the grounds that it is just a demonstration. I used this argument with some folks from Stereophile w/r/t an ABX-related demonstration that several of my friends put on at an AES meeting about 20 years ago. Even though the Stereophile folks still don't have the grace to agree with my viewpoint with respect to demos , I'm going to defend what George did on the same grounds as I defended what my friends did way back when. The asme argiment applies to what John did outside the AES.
It was a demo!
Obviously, there are demonstrations that are done well and there are crappy ones. Anybody who looks for good science in demonstrations is IMO an optimist.
That all said, it is a very good thing when demonstrations are well-run and advance our understanding of how the universe really works. IOW, the demo is as scientific as is possible, given the obvious limitations of demonstrations.
These demos do show as much if not more about the demonstrators than they do about their purported topics, which is actually very sad.
(I'm still trying to reconcile how the AES can be run by a 'cabal' that censors critiques of Meyer & Moran's JAES article, yet also hosts demonstrations like Massenburg's. Sounds kinda schizo to me )
I don't think there's a cabal. I think there were many individual members at that meeting who work in the recording industry and unquestioningly accept more bits+samples = better. I think the Meyer & Moran paper solicited such a large number of responses that the AES decided to make an open forum on their website, rather than fill their publication with the discussion. I don't think there's an unseen or seen "controlling mind" - it's the sum total of its members.
The various AES sectons run pretty independently. It would be hypocritical of me to criticize the AES for the latitude that the UK section is exercising, when the Detroit section had considerable latitude to do what we did with ABX way back when.
IMO the UK section is running science backward, while back then the Detroit section ran science forward. But not everybody agrees with that.
At the bottom of all this is people's personal perceptual models.
While there are big differences in the specific areas of human behavior involved, ABX (really the whole area of design of subjective experiments) and perceptual coding share a common and very strong interest and have developed practical insights into the management of how people perceive things.
The ABX insight was that memory and belief precondition how we perceive sound. The foundation of perceptual coding is the insight is that more external parts of the human hearing apparatus precondition how we perceive sound. There has been very strong synergism in the meeting of these two insights.
The people who rely on single blind and sighted evaluations are simply ignoring very much of what is known about the current model of human perception, especially regarding bias in the brain.
The people who rely on outdated coders and signal subtraction are again simply ignoring very much of what is known about the current perceptual model of human perception, especially regarding bias in the ears.
Therefore, that it would turn out to be the same people that ignore both of these important areas of our current perceptual model of human beings should be no surprise. They are trailing-edge traditionalists while we are far closer to the leading edge.
I suggest that we should have a moment of silence for our weaker brothers, some poorly-informed and some simply boneheaded.
Then we should get on with the rest of our lives which needs to include formulating effective stratgies for the poorly-informed to become informed. The boneheads will always be among us. ;-)
What sense does it make to argue anymore about Atkinson's show-biz demonstration?
None at all. It was what it was. I have no idea why Steven Sullivan has been obsessing about it or why he has such trouble comprehending what was a straightforward description of the conditions.
It ...presented a difference between 16 bit and 24 bit at unemployable listening levels (that would show any thinkable amount of obviously wasted bits to be audible)
Please please read what I actually write. As I repeatedly mentioned, there were _no_ presentations of difference signals at last month's Colorado dems. That was George Massenburg in his London presentation (the nominal topic of this thread). Yes, I _did_ refer to a 2007 demonstration I had performed of the difference signal between hi-rez and Red Book versions of the data, but that was to show that at normal listening levels, the difference signal was _in_audible.
John Atkinson
Editor, Stereophile
The foundation of perceptual coding is the insight is that more external parts of the human hearing apparatus precondition how we perceive sound.
Do you know if advances in perceptual encoders can be directly attributed to the development of a more sophisticated perceptual model?
For example, was the transition from MP3 to AAC based on new evidence of how humans perceive sound, or was AAC simply an attempt to fix known technical flaws in MP3?
Did we have a better model of human hearing when MP3 was developed, but had to dumb the specification down in order to make encoders that could work on early 1990s computer hardware? Or are all the bad aspects of MP3 related to the fact we just didn't know as much back in the 1980s about hearing that we know now?
I completely agree with you that the rejection of perceptual encoding is in effect a rejection of an aspect of human nature, specifically that human perception is in many ways flawed, and that consciousness itself is essentially an extremely elaborate series of perceptual illusions that - thanks to our large brains - interact in an extremely complex way.
Although they will never realise it, people who assume that lossy encoders are inherently flawed are effectively rejecting a fundamental aspect of human nature, that is the fact our senses aren't perfect, they just evolved over millions of years to provide us with
roughly accurate information as quickly as possible. There are lots of other animals that have better smell, touch, hearing or sight than humans, but humans have flawed senses wrapped in a very sophisticated sense of self awareness. The sad thing is that some people refuse to apply that sophisticated sense of self-awareness, to accept that their senses are flawed. They seem to assume that human hearing is in every way perfect, which therefore leads them to assume that there is no way possible that lossy encoders could work.
Yes, I _did_ refer to a 2007 demonstration I had performed of the difference signal between hi-rez and Red Book versions of the data, but that was to show that at normal listening levels, the difference signal was _in_audible.
This seems believable. The difference between a 16 bit version of a 24 bit file and the 24 bit file is always going to be smaller than the 16 bit version's LSB. That in turn will always be 96 dB or more below peak level. That will be masked in all but the quietest imagainable listening room. Certainly, in a conference presentation context, there is very, very little chance of it being heard.
Hence my previous comments about demos.
There is a third factor in this 'dem' -- it's hard to determine if people really heard the difference, or whether they were primed to expect some audible degradation between start and end. If an audience is primed to hear a difference , is anyone really surprised that they do?
Your problem with grasping what I did, Mr. Sullivan, appears to stem from your not reading what I wrote. You asked me about this aspect of the dem 3 days ago, in this thread, in the following exchange in message #85.:
you say the listeners 'were not aware of what they were listening to' but what did the listeners know about the track they were listening to?
Nothing, other than it was a recording of Handel's Messiah.
There was nothing unclear about my response, nothing equivocal, nothing that would be hard to comprehend.
It's not a problem of not reading, thanks. Possibly a problem of comprehending, since I am having trouble ,still, picturing exactly what 'went down', from you piecemeal descriptions so far. So, showed up, did your 'listener training' dem (but not the difference demo, since you had no time), played that recording , and then asked 'so, what did that sound like'? And some unspecified number of people replied something like 'it sounded worse at the end than the beginning'. Or was is 'It kept sounding worse and worse'? So far you have only, AFAICT, described what some unspecified number spontaneously reported about the *last* presented segment -- the one with the most dubious codec use -- , calling it 'lifeless' and 'uninvolving'.
What, btw, was the title of your session, again?
We have no details on how long each segment was...
Again, all this information was included in my responses to your questions: the musical excerpt was from Handel's Messiah, identified as "For unto us a Child is born" in the article I linked to earlier in the thread. I told you that the excerpt lasted around 5 minutes and that there were four different versions spliced together, meaning that each segment was around 75 seconds. (The article I linked to explained why this musical piece was used, in that it had the same music repeated three time, meaning that each of the four segments was the same length and was comprised of the same content.)
Ah, then the 'seamless' joins you mentioned earlier take on a different meaning too. Software ABX tools have an option to either continue the track at the switch point, or go back to start. I took your referfence to 'seamlessness' as the former sort of protocol. So now, at least , it is clear how they could have known when one segment ended and the other began. We're making progress!
Really, sir, the question boils down to, as you must know it does: when presented multiple versions of the same audio sample , was there anything to lead the audience to
expect audible difference, other than the sound itself? To me , it would seem curious on its face that John Atkinson would come and play me the same segment four times, if there *wasn't* some attempt to demonstrate audible difference in the offing. And of course there's that 'listener training' preamble.
Again, I fail to grasp why you feel that my description was incomplete. I am sorry to belabor the point, but you seem curiously incapable of reading what I actually took the time to write in response to your ongoing inquisition.
Or perhaps I am inclined to get you to put down the information here, in one place, and in as full detail as possible. Is that so wrong?
And to address a point you made in another posting:
I'm still trying to reconcile how the AES can be run by a 'cabal' that censors critiques of Meyer & Moran's JAES article, yet also hosts demonstrations like Massenburg's. Sounds kinda schizo to me...
The AES is not a monolithic entity. It is comprised of people with differing points of views and experiences. I have been a member of the AES since 1981, but that doesn't mean that I necessarily agree with, for example, the conclusions of the Meyer-Moran paper.
John Atkinson
Editor, Stereophile
Tongue in cheek, sir, tongue in cheek.
The foundation of perceptual coding is the insight is that more external parts of the human hearing apparatus precondition how we perceive sound.
Do you know if advances in perceptual encoders can be directly attributed to the development of a more sophisticated perceptual model?
Some, but not all.
For example, was the transition from MP3 to AAC based on new evidence of how humans perceive sound, or was AAC simply an attempt to fix known technical flaws in MP3?
AFAIK, both kinds of changes were involved.
You can find details about the genesis of various modern methods for perceptual coding here:
http://citeseerx.ist.psu.edu/viewdoc/summa...=10.1.1.26.5956 (http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.26.5956) - on this page press the cached button in the upper right corner.
(BTW this is a JAES paper that can luckily be back-doored for free)
Did we have a better model of human hearing when MP3 was developed, but had to dumb the specification down in order to make encoders that could work on early 1990s computer hardware? Or are all the bad aspects of MP3 related to the fact we just didn't know as much back in the 1980s about hearing that we know now?
I don't know.
I completely agree with you that the rejection of perceptual encoding is in effect a rejection of an aspect of human nature, specifically that human perception is in many ways flawed, and that consciousness itself is essentially an extremely elaborate series of perceptual illusions that - thanks to our large brains - interact in an extremely complex way.
Rejection of modern subjective testing techniques is also a rejection of a what is now a well-known aspect of human nature. My studies of DBTs shows that DBTs were well-understood and being applied to real world experiments no later than the early 1950s. DBTs were being applied to tests relating to hearing long and showing up in JASA articles long before we started doing ABX tests of audio gear. The JASA is significant because the study of acoustics is very close to audio, and both the JASA amd JAES are read by many of the same people.
Although they will never realise it, people who assume that lossy encoders are inherently flawed are effectively rejecting a fundamental aspect of human nature, that is the fact our senses aren't perfect, they just evolved over millions of years to provide us with roughly accurate information as quickly as possible. There are lots of other animals that have better smell, touch, hearing or sight than humans, but humans have flawed senses wrapped in a very sophisticated sense of self awareness.
Totally agreed. This is all so well known that it is surprising that audio's high end can still get away with what they try to get away with. Their model of hearing is that no only can we hear anything that is measurable, but there are things that we can clearly hear that are as yet immeasurable.
The sad thing is that some people refuse to apply that sophisticated sense of self-awareness, to accept that their senses are flawed. They seem to assume that human hearing is in every way perfect, which therefore leads them to assume that there is no way possible that lossy encoders could work.
I've listened to part of the Massenburg UK AES talk that relates to this. Now, this recording is so messsed up that for me, tying to get something out of it is more demanging than some ABX tests! However, what I seem to hear him saying is that he justifies signal subtraction on the grounds that it has worked well for him in other areas.
IOW, tradition rules his value system.
In a way, that is also what we seem to be hearing from Atkinson. After all, if sighted listening was good enough for JGH in 1958 or 1982, it is still good enough for SP in 2009.
What sense does it make to argue anymore about Atkinson's show-biz demonstration?
It didn't even adhere to the most basic standards. It used a stone age mp3 encoder for the lossy demonstration, presented a difference between 16 bit and 24 bit at unemployable listening levels (that would show any thinkable amount of obviously wasted bits to be audible)
Actually, he reports (http://www.hydrogenaudio.org/forums/index.php?showtopic=72446&view=findpost&p=639919)that he didn't have time to that. So let's not accuse him of uncommitted sins. It makes us look weak.
The people who rely on single blind and sighted evaluations are simply ignoring very much of what is known about the current model of human perception, especially regarding bias in the brain.
The people who rely on outdated coders and signal subtraction are again simply ignoring very much of what is known about the current perceptual model of human perception, especially regarding bias in the ears.
Therefore, that it would turn out to be the same people that ignore both of these important areas of our current perceptual model of human beings should be no surprise. They are trailing-edge traditionalists while we are far closer to the leading edge.
Which wouldn't matter and wouldn't be worth 'obsessing' about, as Mr. Atkinson now characterizes it, if they weren't seen as authorities on such matters...when they have proven (to us at least) that they are not.
None at all. It was what it was. I have no idea why Steven Sullivan has been obsessing about it or why he has such trouble comprehending what was a straightforward description of the conditions.
Now *that* is a bit of an overstatement. And btw, when people use that awful 'it is what it is' construction, I reach for my gun.
Please please read what I actually write. As I repeatedly mentioned, there were _no_ presentations of difference signals at last month's Colorado dems. That was George Massenburg in his London presentation (the nominal topic of this thread). Yes, I _did_ refer to a 2007 demonstration I had performed of the difference signal between hi-rez and Red Book versions of the data, but that was to show that at normal listening levels, the difference signal was _in_audible.
Your statements have been many, distributed over more than one thread,a nd have involved anecdotes about different events. Some confusion is perhaps understandable, and I seem not to be alone in being not quite able to keep it all tied up in a nice package in my head. I *have* been trying to get a complete picture in one place, but it's been a bit of a *dental* undertaking.
The people who rely on single blind and sighted evaluations are simply ignoring very much of what is known about the current model of human perception, especially regarding bias in the brain.
The people who rely on outdated coders and signal subtraction are again simply ignoring very much of what is known about the current perceptual model of human perception, especially regarding bias in the ears.
Therefore, that it would turn out to be the same people that ignore both of these important areas of our current perceptual model of human beings should be no surprise. They are trailing-edge traditionalists while we are far closer to the leading edge.
Which wouldn't matter and wouldn't be worth 'obsessing' about, as Mr. Atkinson now characterizes it, if they weren't seen as authorities on such matters...when they have proven (to us at least) that they are not.
The irony of Atkinson criticizing people for obsessing should not be lost on anyone! ;-) We both know this is just more of his obfuscatory behavior.
The larger view is that this is just another example of one of the things that happens as we age - we find out that are heroes are clowns, or at least that they have feet of clay.
There is another trend I've noticed. It seems harder to motivate people who have had good sucess to move beyond the technology or other behavior patterns that they obtained that success with.
According to Wikipedia, Massenberg was born within about a year of me. I admit it - I've never done for/with audio what he has. Not even 0.1%. Perhaps my lack of success motivates me to keep trying to learn. ;-)
Your problem with grasping what I did, Mr. Sullivan, appears to stem from your not reading what I wrote....There was nothing unclear about my response, nothing equivocal, nothing that would be hard to comprehend.
It's not a problem of not reading, thanks. Possibly a problem of comprehending, since I am having trouble ,still, picturing exactly what 'went down', from you piecemeal descriptions so far. So, with no explanation or preamble, you just showed up, played that recording...
No, that is not what I described, Mr. Sullivan. Are you really this obtuse? The presentation involved my playing a number of recordings, including, immediately prior to the Handel, comparing one of my hi-rez masters with an lossy-encoded version. You know this because I have explained it to you several times.
I then, as explained previously, then informed the audience that I was going to play another excerpt, an extract from Handel's Messiah that had been recorded by Philip Hobbs of Linn. Period.
... and then asked 'so, what did that sound like'? And some unspecified number of people replied something like 'it sounded worse at the end than the beginning'. Or was is 'It kept sounding worse and worse'? So far you have only, AFAICT, described what some unspecified number spontaneously reported about the *last* presented segment -- the one with the most dubioukls codec use -- , calling it 'lifeless' and 'uninvolving'.
I described the reaction of the listeners several days back, Mr. Sullivan. I refer you to that response.
What, btw, was the title of your session, again?
You have already asked me this, Mr. Sullivan, and I offered you an answer. Do you really not remember?
Again, all this information was included in my responses to your questions: the musical excerpt was from Handel's Messiah, identified as "For unto us a Child is born" in the article I linked to earlier in the thread. I told you that the excerpt lasted around 5 minutes and that there were four different versions spliced together, meaning that each segment was around 75 seconds. (The article I linked to explained why this musical piece was used, in that it had the same music repeated three time, meaning that each of the four segments was the same length and was comprised of the same content.)
Ah, then the 'seamless' joins you mentioned earlier take on a different meaning too.
Not at all. The splices between the four sections of the piece were seamless so that the listeners were presented with one continuous piece of music lasting around 5 minutes with no indication that anything was changing, other than the possible change in sound quality itself.
Software ABX tools have an option to either continue the track at the switch point, or go back to start. I took your referfence to 'seamlessness' as the former sort of protocol. So now, at least , it is clear how they could have known when one segment ended and the other began.
As I have repeatedly said but you don't appear to comprehend, Mr. Sullivan, there was _no_ overt indication that anything was changing. All the listeners knew was that they were listening to a recording of "For Unto Us a Child is born" from Handel's Messiah. That is the elegance of Philip Hobbs's protocol: there is no indication that anything is changing as the music progresses nor are the listeners aware prior to the audition that there is anything unusual about what they are about to hear. Thus there are no expectation biases.
Really, sir, the question boils down to, as you must know it does: when presented multiple versions of the same audio sample , was there anything to lead the audience to expect audible difference, other than the sound itself?
None. Nothing. On the face of things, the presentation of the Messiah excerpt was just another in a series of musical pieces I was playing as part of the presentation.
To me , it would seem curious on its face that John Atkinson would come and play me the same segment four times, if there *wasn't* some attempt to demonstrate audible difference in the offing.
It wasn't the same segment played four times. As explained in the article to which I linked and reiterated in the earlier response, Philip had chosen this piece of music because the same music - not the same _segment_ - repeats three times. ie, there are 4 choruses. I used the same musical example as Philip because of this aspect of the music, ie, the audience would hear the chorus 4 times without being alerted by that fact that something unusual was happening, thus bypassing the expectation effect you mention.
Or perhaps I am inclined to get you to put down the information here, in one place, and in as full detail as possible. Is that so wrong?
It seems redundant given that you have random access to every posting I have made in this thread, offering the same answers to the same questions you keep coming up with, Mr. Sullivan. You have stated that you don't bookmark my responses, but if you are _really_ that interested in this subject and not just trolling, then I don't grasp why you are so resistant to doing the legwork and rereading those responses.
John Atkinson
Editor, Stereophile
Your problem with grasping what I did, Mr. Sullivan, appears to stem from your not reading what I wrote....There was nothing unclear about my response, nothing equivocal, nothing that would be hard to comprehend.
It's not a problem of not reading, thanks. Possibly a problem of comprehending, since I am having trouble ,still, picturing exactly what 'went down', from you piecemeal descriptions so far. So, with no explanation or preamble, you just showed up, played that recording...
No, that is not what I described, Mr. Sullivan. Are you really this obtuse? The presentation involved my playing a number of recordings, including, immediately prior to the Handel, comparing one of my hi-rez masters with an lossy-encoded version. You know this because I have explained it to you several times.
Actually, I edited that post within minutes of putting it up, to include note of the 'training' segment. Your reply certainly hadn't appeared yet. It appears you ended up responding to the old version.
I completely agree with you that the rejection of perceptual encoding is in effect a rejection of an aspect of human nature, specifically that human perception is in many ways flawed, and that consciousness itself is essentially an extremely elaborate series of perceptual illusions that - thanks to our large brains - interact in an extremely complex way.
I quite agree, but as this thread is all about rhetoric, it might be well to think about the language here. I'm not actually sure that it's right to say that human perception is "flawed": we don't hear as high frequencies as a bat, or as low volumes as a cat, but that's a selectivity, part of the complex evolved way we perceive the world. An awful lot of making sense of the world, surely, is "knowing" what to ignore. To make this practical, I have two recordings that drive me nuts: Glenn Gould's Bach Two and Three Part Inventions, in which his singalong is too present, and a lute album, which is beautifully recorded (despite the fact that they use a green substrate for the CD because it's an audiophile edition ), but you can hear the mechanical noises of the lute to an extent that I find distracting. In both these cases, hearing less would not be a flaw. The microphone has supplemented a limitation of human hearing, not in a good way.
Equally, to talk about us having perceptual illusions is, I think, misleading, since this implies that we could, in some way, perceive the world in a disillusioned way (and this, probably, leads to the homunculus fallacy). Rather we must accept that the rather elaborately processed events in the brain are the way by which we perceive the world (unless you want to get PoMo on my ass).
This rhetorical point matters, because if one talks of "flaws" and "illusions" it suggests that some people might have perceptions that are less flawed, might have fewer illusions, and so leads to audiopholly and the shameless hucksters who feed it and profit from it.
None at all. It was what it was. I have no idea why Steven Sullivan has been obsessing about it or why he has such trouble comprehending what was a straightforward description of the conditions.
I find this back-and-forth about the conditions of the demo to be weird. There are way too many key variables in a demo situation to catalogue in a "straightforward description". Word choice and verbal intonation aside, even an unwittingly meaningful grin, glance, or reflexive audiophiliac wince or body tic at the transition points between the samples can throw the results off kilter. (A bit OT, but my collie uses eye contact to convey situational meaning in lieu of words, as dog owners may be able to relate.)
It thus strikes me as quite useless to try to extract real proof of anything from such demos. But in the white-room conditions of an ABX test, rhetorical sway and groupthink become useless and the ability to discriminate remains as the only thing that matters.
Do you just not remember at least what codec and bitrates and sample rates and formats you used, or what?
The sighted "learning" comparison was between LPCM at 24-bits/88.2kHz and AAC at 128kbps, derived from a Red Book version of the data, downsampled using the SRC in Bias Peak 5 at its highest-quality setting, which independent tests had shown to be one of the best-performing (link available on request)...
http://src.infinitewave.ca/ (http://src.infinitewave.ca/)
Just dotting the Is and crossing the Ts.
John Atkinson
Editor, Stereophile
Do you just not remember at least what codec and bitrates and sample rates and formats you used, or what?
The sighted "learning" comparison was between LPCM at 24-bits/88.2kHz and AAC at 128kbps, derived from a Red Book version of the data, downsampled using the SRC in Bias Peak 5 at its highest-quality setting, which independent tests had shown to be one of the best-performing (link available on request)...
http://src.infinitewave.ca/ (http://src.infinitewave.ca/)
IOW Bias Peak 5.2.
Looks like Adobe Audion R2 is beter documented and has slightly better performance.
..., downsampled using the SRC in Bias Peak 5 at its highest-quality setting, which independent tests had shown to be one of the best-performing (link available on request)...
http://src.infinitewave.ca/ (http://src.infinitewave.ca/)
Well, it is good enough, but far from being "one of the best performing". Bias Peak 5.2's graph shows that it has a broken low-pass, that aliases into the audible band (up to -20db) instead of suppressing >22kHz frequencies. Of course there are terribly worse performing, but "the best" shouldn't solely be pit against failures as Sony Vegas'.
Compare that to some in the league of the really best perfoming, for example Izotope Steep or the
free Sox VHQ Linear Phase! Those accomplish perfectly black, artifact-free plots without compromising phase accuracy. And for those who don't want linear phase there are also minimum phase versions with comparably impressive sweep results.
..., downsampled using the SRC in Bias Peak 5 at its highest-quality setting, which independent tests had shown to be one of the best-performing (link available on request)...
http://src.infinitewave.ca/ (http://src.infinitewave.ca/)
Well, it is good enough, but far from being "one of the best performing". Bias Peak 5.2's graph shows that it has a broken low-pass, that aliases into the audible band (up to -20db) instead of suppressing >22kHz frequencies.
Are you sure you are looking at the correct graphs? While Peak 5.2 is surpassed by Peak 6.03 (which I didn't buy until after I had prepared the Messiah files), I don't see any aliasing above -140dBFS with Peak 5.2 and it appears very similar on the sweep performance to the Barbabatch SRC, which has long been acclaimed for its performance. You appear to be describing something like the Soundhack SRC.
Compare that to some in the league of the really best perfoming, for example Izotope Steep or the free Sox VHQ Linear Phase! Those accomplish perfectly black, artifact-free plots without compromising phase accuracy.
These are very good SRCs, yes.
John Atkinson
Editor, Stereophile
Well, it is good enough, but far from being "one of the best performing". Bias Peak 5.2's graph shows that it has a broken low-pass, that aliases into the audible band (up to -20db) instead of suppressing >22kHz frequencies.
I doubt it.
http://www.hydrogenaudio.org/forums/index....st&p=537521 (http://www.hydrogenaudio.org/forums/index.php?s=&showtopic=59986&view=findpost&p=537521)
...There's a trade-off between no-aliasing and steep filter artifacts...
...sometimes it may be preferrable, for example, to allow a small amount of aliasing in order to preserve other aspects of signal integrity...
Are you sure you are looking at the correct graphs?
This is what I'm getting displayed for Peak 5.2:
(http://src.infinitewave.ca/images/Sweep/PeakPro52.png)
The reflection at the top is what I'm talking about. Other plots don't have it and it's about -20db.
Well, it is good enough, but far from being "one of the best performing". Bias Peak 5.2's graph shows that it has a broken low-pass, that aliases into the audible band (up to -20db) instead of suppressing >22kHz frequencies.
Are you sure you are looking at the correct graphs?
This is what I'm getting displayed for Peak 5.2:
(http://src.infinitewave.ca/images/Sweep/PeakPro52.png)
The reflection at the top is what I'm talking about. Other plots don't have it and it's about -20db.
Right, but the production of an aliasing product at that level stops when it reaches 20kHz. From then on, at all lower frequencies, the aliasing lies around -140dBFS, and is almost free from higher-order products. Adobe Audition 2 is very similar in this respect. I was misled when you said the aliasing extended into the "audible band." I think it a stretch to describe the 20kHz-22kHz region as "audible," at least at my age :-)
John Atkinson
Editor, Stereophile
...There's a trade-off between no-aliasing and steep filter artifacts...
...sometimes it may be preferrable, for example, to allow a small amount of aliasing in order to preserve other aspects of signal integrity...
I'm well familiar with the relation of steepness/aliasing and also the advantages of minimum phase vs. linear phase filtering. Actually for music I haven't even preferred linear over minimum phase only once and always avoided steep filter slopes as much as posible.
Anyway, SRCs like the mentioned Izotope and Sox employ filters that I wouldn't normally use for 0-20kHz content (steep linear phase), but they still excel at every of the site's chosen quality parameters without exception. If there are qualities worth -20db aliasing, which is a considerable amount of energy, the site should provide measurements able to express these advantages.
I think it a stretch to describe the 20kHz-22kHz region as "audible," at least at my age :-)
Yes, I cannot hear the 20-22kHz range myself (and much lower). Still I would be interested to experiment wether the presence of high energy content between 20-22kHz can cause an audible degradation for the lower half of the spectrum on real world speakers. Probably not as long as it doesn't cause clipping, but maybe worth a try. Foobar is just a click away.
I think it a stretch to describe the 20kHz-22kHz region as "audible," at least at my age :-)
Yes, I cannot hear the 20-22kHz range myself (and much lower). Still I would be interested to experiment whether the presence of high energy content between 20-22kHz can cause an audible degradation for the lower half of the spectrum on real world speakers. Probably not as long as it doesn't cause clipping, but maybe worth a try. Foobar is just a click away.
Everything is a tradeoff. Pretty much all the higher-level >16kHz content on my recordings is percussion, so a touch of aliasing between 20-22kHz is only going to add spectral components that are very similar to those already present.
Thanks for the SoX mention. I checked out their website - http://sox.sourceforge.net/ (http://sox.sourceforge.net/) - and will will try out their program in the next couple of days.
John Atkinson
Editor, Stereophile
Yes, I cannot hear the 20-22kHz range myself (and much lower). Still I would be interested to experiment wether the presence of high energy content between 20-22kHz can cause an audible degradation for the lower half of the spectrum on real world speakers. Probably not as long as it doesn't cause clipping, but maybe worth a try. Foobar is just a click away.
The most powerful reason for not hearing sounds > 16 KHz is masking. There are very very few situations where audio > 16 KHz is not accompanied by audio < 16 KHz that masks *everything* > 16 KHz. Just about the only way to have audio > 16 Khz without audio < 16 Khz is to listen to sine waves > 16 KHz.
The SRC discussion is slightly off-topic, so I have started an additional thread (http://www.hydrogenaudio.org/forums/index.php?showtopic=72756) for it. I have provided samples that stress these findings to the extreme to evaluate wether they can matter at all or not.
I completely agree with you that the rejection of perceptual encoding is in effect a rejection of an aspect of human nature, specifically that human perception is in many ways flawed, and that consciousness itself is essentially an extremely elaborate series of perceptual illusions that - thanks to our large brains - interact in an extremely complex way.
I quite agree, but as this thread is all about rhetoric, it might be well to think about the language here. I'm not actually sure that it's right to say that human perception is "flawed": we don't hear as high frequencies as a bat, or as low volumes as a cat, but that's a selectivity, part of the complex evolved way we perceive the world.
What you refer to is a statement of *limits*. That would not be what I would mean if I write that human perception was flawed .
By flawed, I would mean that perception cannot be assumed to be a particularly accurate (much less perfect) modeler of reality. And by 'perception' I would include the process of integrating the various senses, as well as the cognitive process of linking causes and effects based on those inputs.
"Sighted" perception of audio difference/ quality commonly leads to mistakes in cause-effect relationships (e.g., 'The sound was X because of Y'), because information other than the sound becomes a nuisance factor. *That* is the main 'flaw' in perception as relates to audio, and it is exactly what blind testing helps correct.
It thus strikes me as quite useless to try to extract real proof of anything from such demos. But in the white-room conditions of an ABX test, rhetorical sway and groupthink become useless and the ability to discriminate remains as the only thing that matters.
FWIW, I've trawled through the two threads that discuss JA's 'Music MAtters' demo at the Colorado ListenUp! store, as well as his 2008 Rocky Mountain demo, and gathered what I think is all the descriptive information JA has posted about them. I hope to edit it all into a readable form sometime this week, at which point I'll add it to this thread as a single post, or post it as a new thread if people think that makes a difference.
..., downsampled using the SRC in Bias Peak 5 at its highest-quality setting, which independent tests had shown to be one of the best-performing (link available on request)...
http://src.infinitewave.ca/ (http://src.infinitewave.ca/)
Well, it is good enough, but far from being "one of the best performing". Bias Peak 5.2's graph shows that it has a broken low-pass, that aliases into the audible band (up to -20db) instead of suppressing >22kHz frequencies. Of course there are terribly worse performing, but "the best" shouldn't solely be pit against failures as Sony Vegas'.
One might also ask why JA strove for 'the best' in other aspects of the demo -- 'optimal' listening conditions via a high-end salon rig, high-quality SRC -- but insisted on 'typical' Fraunhofer mp3 encoding circa 2003 (which we know isn't 'typical' either, for 2009) and a demonstration/reporting method that could not hope to reliably interrogate three-step progressive audible degradation of a high-rez source.
By flawed, I would mean that perception cannot be assumed to be a particularly accurate (much less perfect) modeler of reality. And by 'perception' I would include the process of integrating the various senses, as well as the cognitive process of linking causes and effects based on those inputs.
*Yawn* That "flawed" thing you refer to, is capable of much more demanding analization and "linking" of information, than any man-made tool nowadays is able to (and i would say CAN be able to, because of the way how most current man-made machines work). What you are outputting here, is pure ideologic propaganda and in no way short of biases and rethoric.
Just to give you a hint: The difficulty of differentiating between selfgenerated differences, and differences already there when the input arrived, has absolutely nothing to do with humans. It also has nothing to do with human senses. It not even has something to do with consciousness. Exactly the same issue applies to any tool, if it were to have no simultaneus measurements available from a different place, so that it can compare its own data, and the data from the other location. What is so difficult to understand about the simple concept of a processing chain, with something at one end trying to find out which information should be attributed to which part in the processing chain?
Comments like the one quoted above, remind me about why i left ha.org in the first place, and why i shouldn't turn my current "guest-appearence" into a permanent one.
By flawed, I would mean that perception cannot be assumed to be a particularly accurate (much less perfect) modeler of reality. And by 'perception' I would include the process of integrating the various senses, as well as the cognitive process of linking causes and effects based on those inputs.
"Sighted" perception of audio difference/ quality commonly leads to mistakes in cause-effect relationships (e.g., 'The sound was X because of Y'), because information other than the sound becomes a nuisance factor. *That* is the main 'flaw' in perception as relates to audio, and it is exactly what blind testing helps correct.
I agree, of course, but still quibble over the word "flawed" (and it is
only the word I'm worried about). Our complex perception is "flawed" compared with instrumentation if we want to identify single factors, but presumably we've evolved this way because it is in some way advantageous.
All I'm saying is that the word "flawed", as opposed to "limited" or "complexly multifactorial" encourages the belief that somehow some people have hearing that is less flawed, which adds to the neurotic uncertainty which Certain Hucksters feed and feed on.
Blind testing is necessary, of course, but if one is trying to persuade people, it might be easier to say "Your perception of sound is a wondrous and complex thing, but we have to do work to isolate one of the many factors influencing it," rather than "You can't hear for crap, you need to do this sciency stuff." I exaggerate, of course, but maybe that's how some of the convertable might hear it.
I agree, of course, but still quibble over the word "flawed" (and it is only the word I'm worried about). Our complex perception is "flawed" compared with instrumentation if we want to identify single factors, but presumably we've evolved this way because it is in some way advantageous.
(Underlining by Lyx)
Plain and simply: No.
If you feed an instrument - no matter if mental, biological or mechanical - with two inputs which are to be drawn as a single plot, then what you get, is a plot with the combined rating of both inputs. What else did you expect? If you dont want to measure something, you exclude it from the measurement. With a mechanical instrument, you disable certain sensors. With consciousness, you hide/block the undesired input (i.e. product appearance). If you want to measure influence A without influence B, you only measure influence A. If you want to only measure influence B without influence A, then you make sure that only B is measured. And if you want to measure the interaction between both, you measure both. And for this simple logic, it does not matter if the measurement-instrument is mental or mechanical - it applies to both.
There is no difference in this regard, and the distinction between "subjective" and "objective" methods and entities is
in this regard a scientific illusion. (Clarification: What i am saying is not that the choice of methodology is irrelevant. Methods differ in their efficiency depending on the task at hand. What i was saying was that the claimed differences between mental measurement and mechanic measurement do not exist, and that therefore, the on this claim built associations with "subjective" and "objective" as well are bogus.)
If you feed an instrument - no matter if mental, biological or mechanical
The point I am making is that to think of the mind as an instrument is mistaken. The object is to produce a good music experience for humans; along the way there are engineering tasks, for which instrumentation is necessary, and for these purposes the human mind isn't so good (because it is very hard to block all but one channel).
But, if the objective is to produce something pleasing to human perception, it's kind of self-defeating to think of your objective as "flawed." It is what it is.
...along the way there are engineering tasks, for which instrumentation is necessary, and for these purposes the human mind isn't so good (because it is very hard to block all but one channel).
Can you name an example? My problem is: All cases i can think of, are not a matter of filtering, but instead situations, where one needs *different* sensors than humans have (i.e. different spectrum range and sensivity, different interpretation-process, etc.).
By flawed, I would mean that perception cannot be assumed to be a particularly accurate (much less perfect) modeler of reality. And by 'perception' I would include the process of integrating the various senses, as well as the cognitive process of linking causes and effects based on those inputs.
*Yawn* That "flawed" thing you refer to, is capable of much more demanding analization
No doubt, and I guess you're nominating yourself to do it.
and "linking" of information, than any man-made tool nowadays is able to (and i would say CAN be able to, because of the way how most current man-made machines work). What you are outputting here, is pure ideologic propaganda and in no way short of biases and rethoric.
What you are 'outputting' here, appears to be the sound of an axe grinding. Which we all get to do, but really, why so abrasive on *this* of all points?
Just to give you a hint: The difficulty of differentiating between selfgenerated differences, and differences already there when the input arrived, has absolutely nothing to do with humans. It also has nothing to do with human senses. It not even has something to do with consciousness. Exactly the same issue applies to any tool, if it were to have no simultaneus measurements available from a different place, so that it can compare its own data, and the data from the other location. What is so difficult to understand about the simple concept of a processing chain, with something at one end trying to find out which information should be attributed to which part in the processing chain?
I'm actually happy to say I haven't a clue from this why you're so upset over the idea that people are commonly overconfident regarding their perception of causes and effects. That 'ideology' is the basis for controls in scientific experiments.
Comments like the one quoted above, remind me about why i left ha.org in the first place, and why i shouldn't turn my current "guest-appearence" into a permanent one.
And that would be tragic for HA because......?
Blind testing is necessary, of course, but if one is trying to persuade people, it might be easier to say "Your perception of sound is a wondrous and complex thing, but we have to do work to isolate one of the many factors influencing it," rather than "You can't hear for crap, you need to do this sciency stuff." I exaggerate, of course, but maybe that's how some of the convertable might hear it.
I think our only difference is that you would emphasize the wondrousness and
then bring in the need for controls, whereas I'm cutting right to the chase
There are plenty of wonderful examples one can show people to illustrate how easily we can be perceptually 'fooled', not just for sound but for other senses.
I don't think anyone here commonly says 'you can't hear for crap ' (or its milder analogues) -- that tends to be more the sort of thing that 'audiophiles' imply of those who express skepticism over the effects of tweaks. What 'we' here more often say is, you can't implicitly trust your 'hearing' in situations where it's not just your hearing that is in play.
I don't think anyone here commonly says 'you can't hear for crap ' (or its milder analogues) -- that tends to be more the sort of thing that 'audiophiles' imply of those who express skepticism over the effects of tweaks. What 'we' here more often say is, you can't implicitly trust your 'hearing' in situations where it's not just your hearing that is in play.
For sure, indeed. All I'm thinking of is a few cases where people have turned up, have asked questions and have been hazed in various ways, which might give the impression that HA, and the debunking of 'phoolery in general, is about wanting people to hear like instruments. They tend to get p'ed off and go away, even though not in the grip of invincible ignorance.
But then again, I spent a working life teaching Humanities in universities, and I may be excessively prone to suffering fools gladly.
If you feed an instrument - no matter if mental, biological or mechanical
The point I am making is that to think of the mind as an instrument is mistaken. The object is to produce a good music experience for humans; along the way there are engineering tasks, for which instrumentation is necessary, and for these purposes the human mind isn't so good (because it is very hard to block all but one channel).
But, if the objective is to produce something pleasing to human perception, it's kind of self-defeating to think of your objective as "flawed." It is what it is.
I see your point, I think. For example, one might say that human hearing is flawed because of its well-known bandwidth limits and also the way masking keeps certain sounds from being heard.
OTOH, if the human ear had wider bandwidth limits and didn't mask, then the brain would have far more stimulus to deal with, with a corresponding decrease in thinking power availble for other purposes, such as seeing or thinking abstractly at the same time. After all, the physical world is basically zero-sum and the capabilities of the human body are generally zero sum.
This is correct, Arnold. The problem however is: This as well applies to every single mechanical instrument on its own. What i take issue with is the statement "human senses are flawed"... flawed compared to WHAT? Mechanical instruments? No. Those as well each have limits on their own, because - well, one cannot be two different things at the same time :) It would be valid to say "a single instrument on its own, is flawed/suboptimal, compared to having many different instruments available" - sure, but at that point, human senses are just "another measurement tool". To get to the point: The implied normative rating "machine sensors > human sensors" is bogus, UNLESS one considers all possible mechanical sensors an entity of their own, in which case it isn't a fair comparison anymore. (Analogy: It basically would be similiar to saying "Human sensors are flawed compared to the sensors of all animals")
This is correct, Arnold. The problem however is: This as well applies to every single mechanical instrument on its own. What i take issue with is the statement "human senses are flawed"... flawed compared to WHAT? Mechanical instruments?
There need not be a comparison. If visual stimuli can trick the brain into hearing something that is not heard without them then human senses simply are flawed. That such problems also apply to other things, does not mean they are not a problem.
As already said before, this assumption is one of the rationals behind scientific method. If you want to disagree with that, and wish to hold on to some sort of romantic view where the human brain is a wonder of creation, fine. I honestly don't see the point, however.
This is correct, Arnold. The problem however is: This as well applies to every single mechanical instrument on its own. What i take issue with is the statement "human senses are flawed"... flawed compared to WHAT? Mechanical instruments?
If visual stimuli can trick the brain into hearing something that is not heard without them then human senses simply are flawed.
I think you may need to rewrite what you just wrote. There's no way that we can be tricked into hearing something that is not heard.
If perhaps you mean trick us into hearing a sound that does not exist such as intermodulation in the ear. But that is a limitation, but not necessarily a flaw that is inherent in our ears.
I don't think we say that a reasonble limiation is a flaw. For example, an ordinary passenger car is not flawed if it cannot travel down the road at 700 mph.
This is correct, Arnold. The problem however is: This as well applies to every single mechanical instrument on its own. What i take issue with is the statement "human senses are flawed"... flawed compared to WHAT? Mechanical instruments?
If visual stimuli can trick the brain into hearing something that is not heard without them then human senses simply are flawed.
I think you may need to rewrite what you just wrote. There's no way that we can be tricked into hearing something that is not heard.
In the audiophile crowd such occurrences are apparently quite frequent, unless you wish to infer that all such claims are made out of dishonesty.
This is correct, Arnold. The problem however is: This as well applies to every single mechanical instrument on its own. What i take issue with is the statement "human senses are flawed"... flawed compared to WHAT? Mechanical instruments?
If visual stimuli can trick the brain into hearing something that is not heard without them then human senses simply are flawed.
I think you may need to rewrite what you just wrote. There's no way that we can be tricked into hearing something that is not heard.
In the audiophile crowd such occurrences are apparently quite frequent, unless you wish to infer that all such claims are made out of dishonesty.
This is turning into semantic wankery.
By 'hearing something that is not heard' is meant: the 'audible difference' that the listener believes to be real, is not real, in the sense of being due to difference in the sounds A and B. The belief in difference here is due to factors other than the sound.
An extreme example: listeners can be tricked into believing that the same excerpt played twice in succession, is two different 'versions' produced by two different means. This is the so-called 'phantom switch' demonstration.
I think you may need to rewrite what you just wrote. There's no way that we can be tricked into hearing something that is not heard.
Not in terms of hear / not hear - but visual cues certainly change the perceived location of auditory events.
Cheers,
David.
This is correct, Arnold. The problem however is: This as well applies to every single mechanical instrument on its own. What i take issue with is the statement "human senses are flawed"... flawed compared to WHAT? Mechanical instruments?
If visual stimuli can trick the brain into hearing something that is not heard without them then human senses simply are flawed.
I think you may need to rewrite what you just wrote. There's no way that we can be tricked into hearing something that is not heard.
In the audiophile crowd such occurrences are apparently quite frequent, unless you wish to infer that all such claims are made out of dishonesty.
This is turning into semantic wankery.
I think "a bad choice of words" is blunt, but clear and more impersonal.
By 'hearing something that is not heard' is meant: the 'audible difference' that the listener believes to be real, is not real, in the sense of being due to difference in the sounds A and B. The belief in difference here is due to factors other than the sound.
I guess. But as a reader, I don't like to guess.
The *big* problem is that people often use the word "hear" when they really mean perceive.
For example, if one rewrites "hear something that is not heard" as "perceives something that is inaudible", then the meaning seems to me to be far more clear.
An extreme example: listeners can be tricked into believing that the same excerpt played twice in succession, is two different 'versions' produced by two different means. This is the so-called 'phantom switch' demonstration.
Everybody who has ever had to work hard to do well on an ABX test knows about that one! ;-)
The letter to the JAES was cosigned by George Massenburg, Peter Craven, Vicki Melchior, and Wieslaw Woszczyk and was indeed rejected for publication, though John Vanderkooy did set up the on-line forum mainly as a result of the internal debate at the AES over the content of the letter.
I checked the link and found 9 comments about "Audibility of a CD-Standard A/DA/A Loop Inserted into High-Resolution Audio Playback". I see none by the authors listed about.
The letter was quoted in a ProSoundWeb thread today:
http://recforums.prosoundweb.com/index.php...902/#msg_431067 (http://recforums.prosoundweb.com/index.php/m/431067/16902/#msg_431067)
FYI
That's a very civilised thread!
Like how HA used to be
Cheers,
David.
That's a very civilised thread!
As it sits, it has almost zero responses in it.
Before jumping to conclusions, let's see where it is after 50 posts.
Like how HA used to be
No posts to speak of, despite an interesting topic? ;-)
As it sits, it has almost zero responses in it.
I thought the posts from Jon, Krab, and Kees pretty much covered it. You've also made similar point here in the past: the issues they raise apply equally to
any situation where you're asked "which is better" - and yet they only seem to "impair listening" when it's a
blind test. Case closed, as far as I'm concerned.
The case against double blind testing, that is.
Whether there's a real audible ABXable difference when these "pros" are tracking 24/192 vs 16/44.1 is a different case - and it may go far beyond whether there's a difference between 24/192 vs 16/44.1 in optimum systems. Now that I understand how tragically little many of these "pros" understand of the equipment they use, there are hundreds of possible reasons to explain real audible differences - for example, it would only need one plug-in which behaves slightly differently at 192k vs 44.1k to cause an audible difference - and I think the chances of such a plug-in existing are probably near 100%!
Cheers,
David.
Before jumping to conclusions, let's see where it is after 50 posts.
It's a small forum with few off topic posts. Therefore the bandwidth is low and the SNR high. Great or boring ?
Like how HA used to be
I second your feelings, but apparently some like it "molto agitato con fuoco e poco moderato"
That's a very civilised thread!
Like how HA used to be
Cheers,
David.
Be careful what you wish for. ProSoundWeb can get pretty feisty...even about topics like 'hi rez'.
And in context of the bulk of Hydrogenaudio posting every day, recent 'feistiness' here is rather drop-in-the-bucket.
It's great that guys like Lavry and Putzys have their own forums thre though. PSW is one of the highest S/N information resources about audio on the web...*IF* you know whose posts to read (and whose to ignore...)
Bruno P even maintains a thread on his forum, where 'feisty' arguments are sent for timeouts
http://recforums.prosoundweb.com/index.php/t/17678/4549/ (http://recforums.prosoundweb.com/index.php/t/17678/4549/)
It's called 'The Respository'