Skip to main content
Topic: Hi Rez vs Redbook in Classical music (Read 15642 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Hi Rez vs Redbook in Classical music

Reply #50
You might find this comparison interesting:
Uploads forum pictures of Sox and CEP resampling

I also don't know what you do there. I suggested a simple resampling setting from higher rate to 44.1. First your pictures are named something with 192kHz and if this impulse from CEP is for 192 down to 44.1 it must have a miserable frequency response or aliasing as hell.
Please do the same graphs that are used on the SRC comparison page for your strange CEP setting.
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

Hi Rez vs Redbook in Classical music

Reply #51
CEP and Audition 2.0 were very similar in terms of performance in most areas including SRC. Audition 2.0 is represented at the Infinite Wave web site, such as it is. AFAIK its so close to CEP 2.1 that there's no need for both results to be published.

The 2 Audition 2.0 impulse responses they have there look nothing like the CEP one you posted.  As Wombat suggests, the corresponding CEP freq. response is unlikely to meet the requirements of Redbook mastering.

Hi Rez vs Redbook in Classical music

Reply #52
My interest in testing is in 44.1/16 **as** 44.1/16, stored in that format and played back as that format, vs. other resolutions (and good MP3 encodings), in **those** formats and played back as those formats. I see the logic of the workaround, but double conversion defeats the purpose as far as I'm concerned.

If the difference is removed in your system by resampling on playback,  then obviously you would use this free real time capability of fb2k. Unless you WANT to hear this unecessary difference.

One difference is removed by substituting another one. At that point I'm not hearing playback of a 44.1/16 playback signal chain.

Hi Rez vs Redbook in Classical music

Reply #53
Please use the new abx plugin that was linked to before. Also since we don't know what your playback chain does to different sample rates the sample i provided back to 24/192 could be fun.

Yikes -- now there's a problem. I now have the newest foobar 1.3.8, and the newest ABX Comparator from here: http://www.foobar2000.org/components/view/foo_abx

And now there's an audible momentary high pitch sound whenever it switches files, just before playback begins, and no audible sound when it doesn't switch, i.e. it's telling me outright when X is the same (or different) as the previous file played. My previous ABX did not do this. So that's a broken test instrument, any workarounds welcome:


The problem is common and the workaround is to upsample both files to the same sample rate after you do the processing of interest.

We even had a variation of this problem with pure hardware ABX comparators back in the late 1970s.

It led to the following comments in Clark's 1982 JAES paper about ABX:

"
REFINEMENTS TO THE A/B TEST
The author's first experience with double-blind audibility testing was as a member of the SMWTMS Audio Club in early 1977. A button was provided which would select at random component A or B. Identifying one of these, the X component was greatly hampered by not having the known A and B available for reference.

This was corrected by using three interlocked pushbuttons, A, B, and X. Once an X was selected, it would remain that particular A or B until it was decided to move on to another random selection.

However, another problem quickly became obvious. There was always an audible relay transition time delay when switching from A to B. When switching from A to X, however, the time delay would be missing if X was really A and present if X was really B. This extraneous cue was removed by inserting a fixed length dropout time when any change was made. The dropout time was selected to be 50 ms which produces a slight consistent click while allowing subjectively instant comparison.
"

There is a hardware ABX comparator, several 100 of which were built and sold by QSC in the 1990s. It had a similar but mechanical tell - I can take mine and read its list of unknowns with no other equipment attached. If you want a DBT put it across the room! ;-)

Excellent, thank you for the information and reference. I can imagine possible workarounds, like an automuting feature combined with a delay. I'm hearing a very short, very high-pitched tone.

Hi Rez vs Redbook in Classical music

Reply #54
Since there seems to be some confusion about the resampling i used please read my explanation.
http://www.hydrogenaud.io/forums/index.php...mp;#entry893857

@UltimateMusicSnop
Out of curiousity. Did you try to abx the 192kHz versions?
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

Hi Rez vs Redbook in Classical music

Reply #55
Indeed, if you are willing to accept that downsampling is a transparent process, it seems strange to believe that upsampling would not be.  Particularly given that your DAC will up sample during playback even when operated at 44.1kHz


?  I think that neither downsampling nor upsampling is transparent.


There are two kinds of transparency. One involves numeric perfection, and one involves subjective perfection.

I can guarantee you that not a lot of meaningful processing is numerically perfect, which is BTW also called "Bit perfect".

I can similarly guarantee you that a lot of meaningful processing is subjectively perfect, which is BTW also called "Sonic Transparency"

Quote
Downconverting entails a lot of parameter choices in Izotope and other tools. This actually raises a much broader aspect. Rather than starting with 192/24 and downconverting, an ideal comparison would be between two pristine captures of the same acoustic signal produced live in a single test performance, one entering the microphone and going into the 192/24, the other entering precisely the same microphone and going into the 44.1/16. Same performance, same pristine microphone/s (stereo capture), perfect signal splitting (not possible, but probably close enough), and then the processing differs at A/D stages which are *not* doing precisely the same operations. THAT would be interesting to test, but I haven't come across such a pair.


Its been done many times on the recording forums.

In sighted evaluations these files always sound different.

In double blind, level matched, time-synched tests, not so  much.

Quote
On the practical side, recording engineers do actually capture at higher resolutions, and DAWs perform their signal manipulations at higher resolutions, and the product is then downconverted for distribution.


OK it can and often is done. BTW I'm a part time professional recording engineer, plying the  academic music festival and religious live performance trade.

The relevant question to me is it sonically significant or is it just shuffling a lot of numbers unnecessarily?

Thanks for the comments!  Do you happen to know where I could download a matched pair of files for a single-microphone, signal-split capture? i.e., both files native to their formats and no down-converting?


Hi Rez vs Redbook in Classical music

Reply #56
UltimateMusicSnob:  Do you realize that no modern ADC actually operates at the sampling rate it outputs?  Everything is resampled down from the rate it is actually recorded at (typically MHz in a modern device).  Splitting a wire into two ADCs is both unnecessary and probably a bad idea given that it will add more confounding variables that you will have to measure and control.

Hi Rez vs Redbook in Classical music

Reply #57
Since there seems to be some confusion about the resampling i used please read my explanation.
http://www.hydrogenaud.io/forums/index.php...mp;#entry893857

@UltimateMusicSnop
Out of curiousity. Did you try to abx the 192kHz versions?

Yes, but I was using the "old" ABX Comparator. In those few rounds I got nowhere in my tries before I caught up to the new ABX. I tried opening seconds, closing seconds, trumpet blast, and percussion hit areas with no success, random results.
Then when I installed the new ABX, of course, I started having the un-ignorable giveaway cues.

Still learning SoX tool, tips welcome. I did read that description before, but thanks for the re-link.

Hi Rez vs Redbook in Classical music

Reply #58
UltimateMusicSnob:  Do you realize that no modern ADC actually operates at the sampling rate it outputs?  Everything is resampled down from the rate it is actually recorded at (typically MHz in a modern device).  Splitting a wire into two ADCs is both unnecessary and probably a bad idea given that it will add more confounding variables that you will have to measure and control.

Yes, confounding variables are everywhere. 24-bits produced output from recording likely results from an oversampled capture in high multiples of the underlying rate would be one area--parameters to optimize. That's one reason I'm just thinking about the whole signal chain as I actually have it instantiated. I don't have the time/money/equipment to isolate all the variables; the only results which speak to rigorous demonstration would happen in a controlled experimental context in any case (not online forums); nothing is generalizable, any results obtain go to *these* input files into *this* equipment configured *this* way, into my ears. But, since that's how I actually do listen (my program material, through my equipment), it matters to me if my setup handles different resolutions with different aural results.

Hi Rez vs Redbook in Classical music

Reply #59
Two files were downloaded from here: http://www.naimlabel.com/musicstore-test-files.aspx

The hi-rez was the Super Hi Definition FLAC    24bit 192kHz, the Redbook was the CD Quality WAV    16bit 44.1kHz

Foobar ABX was used for testing. Results as follows:

foo_abx 1.3.4 report
foobar2000 v1.2.8
2015/03/27 20:47:22

File A: C:\Users\DAW\Documents\Golden_Ears\naim-test-2-flac-24-192000.flac
File B: C:\Users\DAW\Documents\Golden_Ears\naim-test-2-flac-16-44100.wav

20:47:22 : Test started.
20:57:05 : 00/01  100.0%
20:57:10 : 01/02  75.0%
20:57:16 : 02/03  50.0%
20:57:28 : 03/04  31.3%
20:57:46 : 03/05  50.0%
20:57:52 : 04/06  34.4%
20:57:57 : 05/07  22.7%
20:58:12 : 05/08  36.3%
20:58:28 : 06/09  25.4%
20:58:34 : 07/10  17.2%
20:58:40 : 08/11  11.3%
20:58:56 : 08/12  19.4%
20:59:06 : 09/13  13.3%
20:59:12 : 10/14  9.0%
20:59:25 : 10/15  15.1%
20:59:33 : 11/16  10.5%
20:59:53 : 12/17  7.2%
21:00:06 : 13/18  4.8%
21:00:24 : 14/19  3.2%
21:00:50 : 15/20  2.1%
21:00:58 : Test finished.

----------
Total: 15/20 (2.1%)

Playback was over Neumann KH120 nearfield monitors in small studio.

Stuck.
Thanks for the help, everyone. My ABX Comparator is unworkable--delete and re-install did not help, so I can't test anything now, new or old. No point in addressing further files and setup until the testing problem is fixed. Appreciate the help, especially the external references.


Hi Rez vs Redbook in Classical music

Reply #61
CEP and Audition 2.0 were very similar in terms of performance in most areas including SRC. Audition 2.0 is represented at the Infinite Wave web site, such as it is. AFAIK its so close to CEP 2.1 that there's no need for both results to be published.

The 2 Audition 2.0 impulse responses they have there look nothing like the CEP one you posted.


As documented in the picture I probably used a vastly different setting (the minimum limit) for quality. 

Quote
As Wombat suggests, the corresponding CEP freq. response is unlikely to meet the requirements of Redbook mastering.


Suggestions aren't proof.  TOS 8, anybody? ;-)

My investigations suggest that most audio interfaces have relatively wide transition bands compared to Sox with recommended parameters.

I see no evidence that hyper-narrow transition bands are required for sonically transparent resampling.

Let's say that for 44.1 KHz downsampling the use of really narrow transition bands create audible artifacts.

How do we find out that is true or false?

If it is true, why are we shooting ourselves in the foot this way?

If it is true, the alternatives that are being stuffed down our throats such as 192 KHz upsampling and Apodizing seem to be far more for complex than using a wider transition band and just doing what sounds best.

Hi Rez vs Redbook in Classical music

Reply #62
Thanks for the help, everyone. My ABX Comparator is unworkable--delete and re-install did not help, so I can't test anything now, new or old. No point in addressing further files and setup until the testing problem is fixed. Appreciate the help, especially the external references.


I don't know how far you have gone in terms of investigating many easy low cost alternatives.

You seem to have already rejected a very reasonable alternative:  upsampling.

There are other audio interfaces including the on-board audio interface that came with the PC that may work out well with or without upsampling.


Hi Rez vs Redbook in Classical music

Reply #64
Quote
As Wombat suggests, the corresponding CEP freq. response is unlikely to meet the requirements of Redbook mastering.


Suggestions aren't proof.  TOS 8, anybody? ;-)

Good grief, you are the one who suggested that filters would or wouldn't produce ringing based on screen shots of the impulse responses!

The main resampling requirements for mastering to Redbook are generally considered to be a flat response to 20kHz and artefact level lower than that of shaped dither at 16-bit, not "what's the shortest filter I could probably get away with". 

Quote
My investigations suggest that most audio interfaces have relatively wide transition bands compared to Sox with recommended parameters.

Again, you seem to be conflating the cost/performance considerations real-time/hardware vs. off-line/software resampling.


Hi Rez vs Redbook in Classical music

Reply #65
Quote
As Wombat suggests, the corresponding CEP freq. response is unlikely to meet the requirements of Redbook mastering.


Suggestions aren't proof.  TOS 8, anybody? ;-)

Good grief, you are the one who suggested that filters would or wouldn't produce ringing based on screen shots of the impulse responses!


What is wrong with testing using impulse responses for technical tests or even listening tests?  Isn't an impulse like one heck of a sharp castanet click? ;-)  I mean if there are strong practical reasons to be less demanding, then OK but where is the evidence of that?

So far what I've heard is that the impulse response of what some believe to be an ideal filter is Sinc(x), given that their idea of ideal has a rectangular frequency response curve. No surprise here - I learned about Sinc(x)  in 1972 and I haven't forgotten it, I don't think.

But who says that an ideal rectangle with sharp edges and perpendicular sides is the ideal FR curve and why? Study of what structures in the human ear says that? What DBTs say that?  I'm merely ignorant here, but I don't know of anybody who actually seems to know better.

Quote
The main resampling requirements for mastering to Redbook are generally considered to be a flat response to 20kHz and artifact level lower than that of shaped dither at 16-bit, not "what's the shortest filter I could probably get away with".


Straw man argument. My suggestion is that rather than make arbitrary choices based on limited information about what's audible, we instead listen to the results of what we do critically, and reliably.

Quote
Quote
My investigations suggest that most audio interfaces have relatively wide transition bands compared to Sox with recommended parameters.

Again, you seem to be conflating the cost/performance considerations real-time/hardware vs. off-line/software resampling.


What's the difference? They are both in the signal chain we end up listening through both of them. The right answer for either seems like the right answer for both. I think that circuits on silicon is getting cheap enough that we can afford to do it right. 

Thing is, I think we've had the hardware right for a long time.  I see the possibility that the use of long-ringing filters by some is being used to justify weird science like Apodizing. Meridian and Dolby seem to be trying to make a business of it.  I've got a radical idea - just do it right the first time and don't bother with yet another layer of upsampling and tacked on filters.

In case there is any confusion I'm not against upsampling when it makes engineering sense. I'm talking about this: Dolby TrueHD with 96 KHz upsampling

Hi Rez vs Redbook in Classical music

Reply #66
Arny's had an epiphany. The thread that started it got closed, and I don't think we've discussed it properly since. This thread probably isn't the place. I think what's happened is that, having digested the results of this paper, and/or maybe some other source, Arny now believes that ultra-sonic filters can be audible under some circumstances, and that it might be wise to reduce their length (in as much as you can without compromising anything important) to avoid them being ABXable.

I think the whole thing needs more research and more discussion.

Hi Rez vs Redbook in Classical music

Reply #67
The right answer for either seems like the right answer for both.


Read Shannon's proof for the sampling theorem. It says what is right for replay.

Quote
I've got a radical idea - just do it right the first time and don't bother with upsampling and tacked on filters.


Upoversampling is a requirement to attain the right playback filter.

The one interesting question is: what is the best recording filter?


Hi Rez vs Redbook in Classical music

Reply #68
My interest in testing is in 44.1/16 **as** 44.1/16, stored in that format and played back as that format, vs. other resolutions (and good MP3 encodings), in **those** formats and played back as those formats. I see the logic of the workaround, but double conversion defeats the purpose as far as I'm concerned.

If the difference is removed in your system by resampling on playback,  then obviously you would use this free real time capability of fb2k. Unless you WANT to hear this unecessary difference.

One difference is removed by substituting another one. At that point I'm not hearing playback of a 44.1/16 playback signal chain.

As others have explained, there are almost no 44.1kHz/16 "playback chains". The 44.1kHz almost never stays as 44.1kHz, the 16-bits almost never stay as 16-bits. Certainly the playback chain you are using is not native 44.1kHz 16-bit.

The only question is whether you leave the upconversion to chance, or choose the first and most important stage of it yourself. If you leave the upconversion to chance, this doesn't make your experiment somehow more "pure" - it reduces it's relevance, such that it only applies to exactly your playback chain (which in the worst circumstances means the specific combination of OS, drivers, soundcard/output and DAC). That may be useful for demonstrating some unwanted audible property of that combination, but doesn't say anything about the 44.1kHz/16-bit source.

I don't have the time/money/equipment to isolate all the variables
Fixing the upconversion can be done for free using your existing software. So that's the money and equipment taken care of.

Whether you have the time to perform another ABX test, only you can decide. Obviously if you haven't got that much spare time, you won't have time to post about audio anywhere ever again. Shame.

Cheers,
David.

Hi Rez vs Redbook in Classical music

Reply #69
The right answer for either seems like the right answer for both.


Read Shannon's proof for the sampling theorem. It says what is right for replay.


Shannon leaves out the human factor.

Quote
Quote
I've got a radical idea - just do it right the first time and don't bother with upsampling and tacked on filters.


Upoversampling is a requirement to attain the right playback filter.


No, its a cheap seemingly effective technique we use a lot right now.  We seem to have done OK in the past without any of that, but it costs more money and doesn't see to work any better.

Quote
The one interesting question is: what is the best recording filter?


The answer to that question needs to include the fact that people listen to music for enjoyment.  That says that Shannon is in some sense necessary but definitely not sufficient.

Hi Rez vs Redbook in Classical music

Reply #70
Arny's had an epiphany. The thread that started it got closed, and I don't think we've discussed it properly since. This thread probably isn't the place. I think what's happened is that, having digested the results of this paper, and/or maybe some other source, Arny now believes that ultra-sonic filters can be audible under some circumstances, and that it might be wise to reduce their length (in as much as you can without compromising anything important) to avoid them being ABXable.


Really close.

I now believe that ultra-sonic filters might be audible under some circumstances, and that it might be wise to reduce their length (in as much as you can without compromising anything more important) to avoid them being ABXable.

I think that if reasonably possible, using filters that aren't ABXable is a really good idea. ;-)

The relevance of the AES paper cited above is that it made some promises in its title that it is easy to see it never delivered on, and followed up on that by making more false claims in its body text. Perhaps the worst of the false claims appears to be claiming that ABX has the same memory load as a 3AFC test, at that some franken-test that makes things harder was somehow more sensitive.

More discussion here: AES forum

Quote
I think the whole thing needs more research and more discussion.


Agreed.

Hi Rez vs Redbook in Classical music

Reply #71
I think Arny is still lacking an understanding or two in this field ...

Hi Rez vs Redbook in Classical music

Reply #72
That thread will remain closed if I have anything to say about it.  Discussion about the decision which, IIRC was made by someone else, must be had privately. Read TOS7.
Is 24-bit/192kHz good enough for your lo-fi vinyl, or do you need 32/384?

Hi Rez vs Redbook in Classical music

Reply #73
That thread will remain closed if I have anything to say about it.  Discussion about the decision which, IIRC was made by someone else, must be had privately. Read TOS7.

Just for my understanding: Does this mean that a discussion of the mentioned AES paper is being frowned upon here in the forum, regardless in which thread it takes place?

Hi Rez vs Redbook in Classical music

Reply #74
Arny's had an epiphany. The thread that started it got closed, and I don't think we've discussed it properly since. This thread probably isn't the place. I think what's happened is that, having digested the results of this paper, and/or maybe some other source, Arny now believes that ultra-sonic filters can be audible under some circumstances, and that it might be wise to reduce their length (in as much as you can without compromising anything important) to avoid them being ABXable.


Really close.

I now believe that ultra-sonic filters might be audible under some circumstances, and that it might be wise to reduce their length (in as much as you can without compromising anything more important) to avoid them being ABXable.

I think that if reasonably possible, using filters that aren't ABXable is a really good idea. ;-)
In the context, that must be a truism.

I don't think we have sufficient evidence to say that such-and-such a change is the one we must make to avoid the possibility of ABXing. If we can't be sure what might be "broken", we can't be sure what might fix it.

Quote
More discussion here: AES forum
There have been two new contributions there since I last looked. One suggests further research is underway.

Usual disclaimer: if there is an audible difference, it's very very very small and nowhere near the "night and day" difference sometimes claimed for Hi-Res.

Cheers,
David.

 
SimplePortal 1.0.0 RC1 © 2008-2019