This BS study lacked investigatory drive in my view. It was very passive. It seems that the study was conceived and implemented, full stop. I note that its two main conclusions are expressed vaguely, or even cryptically.

I'd have thought that preliminary informal tests should have been done with subjects with outstanding hearing and listening skills. This presumably would have shown up high yield (revealing) sections. Then a variation could have been done by way of substituting a different loudspeaker system, or even headphones to see whether a difference was still in evidence.

The way the two "main" conclusions "offered" are expressed is also passive, and not particularly helpful. It is obvious that audible sounds exist that cannot be transparently encoded at 16/44. Some children can hear a loud sine wave at 23kHz and a standard CD cannot render 23kHz at all. Also if a CD is played back with ear-splitting gain, then dither noise may be audible even to middle-aged adults. It is accordingly quite cryptic and unhelpful to offer as a conclusion: "first, there exist audible signals that cannot be encoded transparently by a standard CD".  Perhaps what was meant was "audible signals contained in conventional music recorded with high definition and listened to at a normal listening level".

In summary,

• the BS experimental study appears to have been executed in a passive way, not taking advantage of preliminary informal testing to pinpoint what was causing an apparent lack of transparency in the filter emulations (particularly the emulation for a sample rate reduction to 44.1kHz), and
• the two main conclusions offered are vaguely expressed and unhelpful, and in the case of the second conclusion (about the need for a hi-fidelity reproduction chain) not actually supported by experiment.

CONCLUSIONS
1. FIR filters that emulate downsampling for sam-
ple rates of 44.1 kHz and 48 kHz can have a dele-
terious effect on the listening experience in a
wideband playback system.
2. 16-bit quantization with and without RPDF
dither can have a deleterious effect on the listen-
ing experience in a wideband playback system.
3. Our findings are consistent with the idea that
filters with long impulse responses blur fine tem-
poral details of signals.
4. Not all pieces of music contain musical features
that demonstrate these losses of transparency.
Possible important features include echoes that
give a sense of the physical space around the
performers.
5. Consideration should be given to the use of psy-
chophysical tests that minimise cognitive load
in studies of this kind.

Further comments
For someone familiar with MATLAB audio filters, it is a walk in the park to alter the parameters. The BS study could have been so much more. In my opinion, it was a narrowly focussed endeavour, with vaguely expressed conclusions.

Whilst it is true (well established in fact) that truncation to 16-bits has been found to be inadequate, this comment in my opinion is passive, and misdirected. Real life CDs are not produced from undithered low-noise truncations to 16 bits.

[Hypothetical sentences:]  We note that in practice standard CDs are not produced with quantization alone (that is to say discarding all bits beyond the 16th by way of truncation) but with dither. Dither allows bits beyond the 16th to be reflected in the audible signal, despite the signal being distributed to the consumer in a 16-bit format. We would stress that the test conditions that involved quantization without accompanying dither were merely a "probe" to serve as a reference point: there is no suggestion that standard CDs are produced in this way.

Alas, no such qualifying sentences appear in this part of the report. I think it is this tenth page of the report that will be fodder for high definition proponents not averse to letting technical accuracy get in the way of condemning the standard CD format for its "insufficient" 16 bits; or simply not understanding the benefits of dither.

Is anybody denying that here? On the contrary, everyone here noticed that years ago, and pointed out the stupidity of the endeavour back then already. Nothing has changed since then.

Are they? It's not unthinkable or impossible for a company to use or create an audio container with DRM and special hardware or software requirements.

Monty has pointed out that in practice hi-res playback might have a lower audio fidelity due to distortions coming from high-frequency content.

People like you always know that the "best quality" is up from whatever is the current standard. That's the oldschool audiophile delusion.

Quote from: amirm on 2014-12-02 15:43:29

5. By a miracle, the music labels have decided to license their studio masters to us.  You don't see that in movies, do you?  This is a gift to be embraced.

I still wonder why you assume hi-res masters to be inherently superior to CD masters. What prohibits engineers from fucking up those, too? They managed to make 16/44 sound like shit, it's just as easily possible with 24/96. Audio technology is not the problem, it's human error.

Lossy delivery can be acceptable with proper mastering. The CD releases you get today are (depending on genre) hardly more than 12bit/32khz anyway.

See above. The conjecture that audio engineers will magically improve their craft just due to the existence of a hi-res audio format is absurd. Again, it is already possible to produce proper releases on CD today (which has been demonstrated in the classical and soundtrack genres), it's just not done.

Quote from: amirm on 2014-12-02 15:43:29

[audiophile gibberish]

Get lost already. Nobody gains any insight from reading your FUD nonsense and propaganda.

Don't go responding to this post because per above, you don't matter.  AJ doesn't matter.  Krab doesn't matter.  None of you do.  Get a sense of reality, conduct yourself more professionally and join the audiophiles elsewhere in increases the choice of formats we have.  Heaven knows I am getting bored interacting with you all.

You are wrong and that can be trivially shown. There are many high-resolution stereo masters that are distributed prior to loudness compression of CD/MP3. There is nothing delusional about their improved fidelity. You speak from point of view of forum facts, not real life facts.

That is signal to noise ratio of cassette tape.

See the separate mastering for HD Tracks?  What we want is happening.

As if your reply to me had an ounce of insight or data?

Quote

Monty has pointed out that in practice hi-res playback might have a lower audio fidelity due to distortions coming from high-frequency content.

I have known Monty more than many of you.  He is a super smart guy but has little expertise in this area.  I have debated his write-up with him on WBF Forum and have shown the many flaws in his blog on AVS Forum.  He has improved his write-up some but at the end of the day, it is yet another person writing about a field that is not his specialty.  He doesn't understand the business side of this field, nor does he have electrical engineering experience to understand the hardware side.  You want to the right version of that, read Stuart's JAES paper, coding for high-resolution audio.  Anything beyond that which Monty says is yet another online blogger making arguments with no personal data or professional experience.

Quote

I still wonder why you assume hi-res masters to be inherently superior to CD masters. What prohibits engineers from fucking up those, too? They managed to make 16/44 sound like shit, it's just as easily possible with 24/96. Audio technology is not the problem, it's human error.

I assume it because I personally know the people who create them.  And the founder of the biggest distributor of such (HD tracks).  You don't have that kind of real life data and relationship, right?  So what you hypothesize is just FUD and random arguments thrown out there.  What if I got a steak house and they feed me hamburger they bought from Burger King.  Yes, it could happen but come back when you know it has.

I will be leaving shortly as I have never seen so many unprofessional posters in any audio forum.

Quote from: amirm on 2014-12-02 16:57:58

See the separate mastering for HD Tracks?  What we want is happening.

There is no need for separate mastering. All you need to do to cerate a CD master from a hi-res master is to decimate to 16bits using proper noise shaped dithering and resample to 44.1kHz. Anybody who's doing it differently doesn't have a clue what he's doing.

Quote from: amirm on 2014-12-02 15:43:29

Don't go responding to this post because per above, you don't matter.  AJ doesn't matter.  Krab doesn't matter.  None of you do.  Get a sense of reality, conduct yourself more professionally and join the audiophiles elsewhere in increases the choice of formats we have.  Heaven knows I am getting bored interacting with you all.

And when you are king.... oh wait, in your mind, you already *are*. 

I have known Monty more than many of you.  He is a super smart guy but has little expertise in this area.  I have debated his write-up with him on WBF Forum and have shown the many flaws in his blog on AVS Forum.  He has improved his write-up some but at the end of the day, it is yet another person writing about a field that is not his specialty.  He doesn't understand the business side of this field, nor does he have electrical engineering experience to understand the hardware side.  You want to the right version of that, read Stuart's JAES paper, coding for high-resolution audio.  Anything beyond that which Monty says is yet another online blogger making arguments with no personal data or professional experience.

I have known Monty more than many of you.  He is a super smart guy but has little expertise in this area.  I have debated his write-up with him on WBF Forum and have shown the many flaws in his blog on AVS Forum.  He has improved his write-up some but at the end of the day, it is yet another person writing about a field that is not his specialty.

Originally Posted by amirm

If what I present doesn't meet some high bar, so be it....
The comparison I performed was using a Mark Levinson No360S against the on-board DACs in five to six DVD-A and SACD players, all playing the same time sync'ed CD. In other words, I would listen to the analog output of the player while its digital output would feed the ML DAC. All front panel lights were turned off in addition to video circuits (yes, all of that made a difference in fidelity).
The two sources were fed to the dual inputs of a Stax "earspeaker" electrostatic headphone amp. If you are not familiar with Stax, you can read learn more about them here: http://www.stax.co.jp/Export/ExportProducts.html. I have three of their units and results are consistent across the board although the highest end unit does make the job a bit easier. Using headphones allowed me to completely eliminate the room and take advantage of the amazing transparency of these headphones to listen for the slightest differences. To latter point, I would often listen to material at levels well above what I would use for listening to music, allowing me to hear detail that would otherwise be lost.
I then picked material that made it easier to detect differences between DACs. I am not going to disclose what constitutes such content. Without such material, the job can range from difficult to impossible. One has to know what could be damaged by a DAC and then use music that has such content. To give you an example, when you compress music, it is the transients that suffer. So something like guitar music is much more revealing than say, violin as the latter is much more harmonic than the sharp impulses of a guitar. Voices play the same role. None of these are useful for testing DACs though so don’t use that as a hint to the question posed . You can’t test the cornering of a car if you just drive it straight….
The comparison was then conducted without knowing which input is which, sitting in front of the headphone amp and toggling back and forth. When necessary, I would go back and re-listen. Once I found which one sounded worse, I would then repeat the exercise by randomizing the inputs and seeing if I could still identify which one was worse. My success rate was 100% in the second test (i.e. could always verify that the first result was not by chance). This testing was repeated a number of times comparing the different sources against each other and the ML.
I did not level match anything. However, once I found one source was worse than the other, I would then turn up the volume to counter any effect there. Indeed, doing so would close the gap some but it never changed the outcome. Note that the elevated level clearly made that source sound louder than the other. So the advantage was put on the losing side.
The results above were later objectively shown to be backed by some science in Stereophile magazine.

He doesn't understand the business side of this field

But How Does it Sound - By Amir Majidimehr
In comparison testing I have done, switching amplifiers using the classic class D configuration always sport incredible low frequency control and power. They beat out linear class AB amplifiers almost regardless of price. What they give up though is high frequency fidelity which I find somewhat harsh. The distortion is highly non-linear and challenging to spot but it is there. The Mark Levinson No 53 is the first switching amplifier I have heard which does not have this compromise. Its bass is amazingly authoritative: tight and powerful. Yet the rest of the response is absolutely neutral and pleasant.
If you have not heard these unique $50k amplifiers, I highly encourage you to come into our showroom for a listen. We have a pair on hand driving our Revel speakers. I am confident that they will improve the sound of your current speakers given the ease with which they can drive any load regardless of how difficult they might be (and many high-end speakers are difficult to drive). We are happy to let you evaluate them with your own system to see the benefits of this technology.  Hearing this amplifier was an eye-opener for me.  I think it will be for you too.

Indeed that first conclusion--and btw, the two conclusions in the Abstract are *not* the same as the ones in the 'Conclusion' section --  is really rather curious. It was certainly already known that there are audible signals that CD cannot encode transparently.  So why trumpet that as a 'conclusion' here?  This is what some of us here wondered about as soon as we read the abstract, even before the paper was made public.

...

But note again, those 'conclusions' are not offered in the actual conclusion section of the paper -- there, there are 5 conclusions that are more specific. If I reviewed this I would have told the authors to bring their Abstract more in line with the Conclusions, or vice-versa.

CONCLUSIONS
1. FIR filters that emulate downsampling for sam-
ple rates of 44.1 kHz and 48 kHz can have a dele-
terious effect on the listening experience in a
wideband playback system. [This is a bit of a leap. It is one thing to have established that a difference was correctly identified some 60% of the time. It is another to accept subjective remarks made by one or more of the 8 test subjects that the effect was of a degree and nature to be fairly described as having a "deleterious effect on the listening experience". The report provides no detail on how many of the eight test subjects made negative comments, and whether they were test subjects with high scores.]

2. 16-bit quantization with and without RPDF
dither can have a deleterious effect on the listen-
ing experience in a wideband playback system. [Again, "deleterious effect on the listening experience" may be too strong a phrase. "From time to time perceptible effect, that some listeners may describe as deleterious to their listening experience" might have been more appropriate.]

3. Our findings are consistent with the idea that
filters with long impulse responses blur fine tem-
poral details of signals. [The idea is diminished in persuasiveness given that the filtering was of ultrasonic frequencies using linear phase.]

4. Not all pieces of music contain musical features
that demonstrate these losses of transparency.
Possible important features include echoes that
give a sense of the physical space around the
performers. [It might have assisted to try a less aggressive MATLAB filter, and if results were still audible, look for other possible explanations, e.g. some instability in the tweeters triggered by high frequency content. On the other hand, if this was intended very much as a narrow, pilot study, the limited rather passive approach taken is more understandable; and the paper should not spoken of as a "sea change", but rather as "raising questions for further study".]

5. Consideration should be given to the use of psy-
chophysical tests that minimise cognitive load
in studies of this kind. [Certainly it could help if subjects could control which parts of a musical passage (or “section”) they listened to, just as is done routinely by some members of this forum when using foobar and its ABX plug-in! This begs the question why distinguishable passages have not been identified in the past in this forum when applying a filter for a 44.1kHz, or even 48kHz sample rate, for members with a "wideband playback system" (whatever that is intended to mean exactly, in the context of the BS paper).]

"wideband playback system" (whatever that is intended to mean exactly, in the context of the BS paper).

Indeed that first conclusion--and btw, the two conclusions in the Abstract are *not* the same as the ones in the 'Conclusion' section --  is really rather curious. It was certainly already known that there are audible signals that CD cannot encode transparently.  So why trumpet that as a 'conclusion' here?

the peer reviewers never even saw the abstracts' two "offered conclusions" because it was written (or modified) after the fact!

... I still wonder why you assume hi-res masters to be inherently superior to CD masters. What prohibits engineers from fucking up those, too? They managed to make 16/44 sound like shit, it's just as easily possible with 24/96. Audio technology is not the problem, it's human error. ...

No, it's not human error. They do it deliberately and with full knowledge of what they are doing. sad.gif

The sound quality deficiencies  of contemporary CD recordings are due to human error (intention or ignorance) and have nothing to do with the technological means available.

Quote from: mzil on 2014-12-02 20:35:20

the peer reviewers never even saw the abstracts' two "offered conclusions" because it was written (or modified) after the fact!

That's a pretty bold claim to be making without offering any proof.

Quote from: mzil on 2014-12-02 20:35:20

the peer reviewers never even saw the abstracts' two "offered conclusions" because it was written (or modified) after the fact!

That's a pretty bold claim to be making without offering any proof.

I've made a short audio file with 4 different dither formats to illustrate the differences.
The source is a 1kHz sine fading from -80dBFS to -160dBFS during 8 seconds. The 4 dither versions are:
1) 24-bit TPDF dither
2) 16-bit TPDF dither
3) 16-bit RPDF dither
4) 16-bit truncated

60dB gain was applied to make the result easier to hear at normal playback levels.
Monitor gain has to be reduced by 60 dB to get the original level back.

How did you apply gain? This alone may add own dither or noise. I suggest shifting it by 10 bit.

I've made a short audio file with 4 different dither formats to illustrate the differences.
The source is a 1kHz sine fading from -80dBFS to -160dBFS during 8 seconds. The 4 dither versions are:
1) 24-bit TPDF dither
2) 16-bit TPDF dither
3) 16-bit RPDF dither
4) 16-bit truncated

60dB gain was applied to make the result easier to hear at normal playback levels.
Monitor gain has to be reduced by 60 dB to get the original level back.

Thanks for this. Nice and obvious effects.*

As to his theory of IM distortion, it is just that: some theory.