Why 192kHz matters?

[LHC]

@@LHC

 

Having both authored and reviewed published work, I can't agree with the 'absolute' view you take. Peer review is not a process to prove or disprove work, or to legitimise it as 'absolutely correct'. If learning and evolving from published experience had to involve additional peer-reviewed work, learning would be an inordinately long process - particularly when what's been published has blatant error, or is of limited relevance.

 

 

(The discussion below is off-topic. Sorry)

 

I think it depends on which 'world' you are from. Kunchur is a scientist and academic, so I am talking specifically about his world. All I did is just outlining the traditional and accepted approach here; I have not presented my own opinion or view on it.

 

Of course published work can contain mistakes and errors. (if you happen to read some of my posts in the Climate Change thread elsewhere in this forum, you would me see rising this point as a reality check to others) When errors are brought to the attention of the authors, and they conceded their mistakes, they would publish an erratum to notify the scientific world of this. Quite commonly others would publish their own work in which they would explain where the errors have occurred (with or without first consulting with the original authors); this is perfectly legitimate and proper way to proceed. If the original authors do not concede that they were in the wrong, they could write a counter paper defending their claims. I have even seem a case where the authors and their detractors debated over their work on the 'Letters to the Editor' page of the journal (over several issues!). Those are not peer reviewed but might be vetted by the Editor.

 

Now if you had said there are advanced knowledge, known to experts, that are not published then I totally agree with you. There are indeed much knowledge not been through the publication process. When someone like Stephen Hawking give a talk or issue a preprint, people in his field would generally 'treat' it like a published piece of work and begin using it. Scientific work is not slowed down by the need to eventually have the work peer reviewed and published. Scientists have been sharing preprints over FTP starting in 1991, they foresaw the need to rapidly share information before publication. This provides opportunities for others to identify any mistakes, and of course the majority of the preprints would eventually be published. The rapid and world wide distribution of unpublished papers did not replace the need for peer reviewed publication, but rather enhanced it (as feedback on flawed work would prevent the authors from even submitting it to a journal). 

 

Kunchur have stated in his FAQ that he went through a similarly contested pre-publication process. He presented his work at conferences where he was met with challenges. It is likely that any questions about blatant errors and limited applicability would have been raised with him before he even submit his papers. Moreover his journal editors and referees would have certainly pull him up for anything that are obviously wrong. I am not saying this process is prefect, far from it, but I am saying the amount of checks-and-balance in the process does provide a certain degree of quality assurance and credibility. 

 

(@@rmpfyf I know you have written heaps in response to my posts. I have not yet find the time to write on them proper, it doesn't mean I am ignoring you or dismissing what you have written.  )

[davewantsmoore]

What is relevant is that both are properly published work and should be accepted as the state of our knowledge to date. 

 

No, that's rubbish.   Not everything published or reviewed is without fault.

 

 

What is interesting though is that Kunchur, in his paper, doesn't say:   Note that the reciprocal of 6 microseconds is 166kHz - indicating that an audio system should be able to process this frequency to satisfy this timing perception"

 

The author of the Yamaha page said that.

 

 

Kunchur does make the error about "timing resolution" (and needing a high sample rate to offer that), although that line of thinking is secondary.   In his statements and conclusions he is much more focussed on the fact that he experimentally demonstrated that fact that the listener could hear the 7us difference in signal timing.    (which is fine, he appears to have done that - which isn't surprising as it's not a controversial result)

 

 

Certainly more rigorous research work in the same vine as Meyer, Moran and Kunchur is desperately needed.

 

I would expect it never gets off the ground, because small trials always show it to be not worth exploring.

[davewantsmoore]

I wonder whether that ever took off. 

 

Unlikely....

[Newman]

On 24/01/2016 at 9:01 PM, davewantsmoore said:

Kunchur does make the error about "timing resolution" (and needing a high sample rate to offer that), although that line of thinking is secondary.   In his statements and conclusions he is much more focussed on the fact that he experimentally demonstrated that fact that the listener could hear the 7us difference in signal timing.    (which is fine, he appears to have done that - which isn't surprising as it's not a controversial result)

 

Totally agree with your perspective Dave, but re the bit in bold, we can hardly conclude that he has shown it while the major issues I listed remain, i.e. his filters changed his 7 kHz fundamentals by a detectable amount (so trials were probably actually proving the detectable change in the fundamental of a 7 kHz tone, and lo and behold the 0.2 dB change his filters made at 7 kHz is very close to detection thresholds), and secondly that amplitude thresholds were ignored despite the fact that Kunchur's test signals have amplitudes that never occur in any natural/musical situation.

 

The first issue means test subjects were responding to something other than what he thought, and the second means that a re-test with threshold considerations is needed, to see what is the threshold with naturally-occurring amplitudes.

 

He didn't demonstrate anything.

 

 

Edited December 21, 2017 by Newman
Link needed updating

[davewantsmoore]

When someone like Stephen Hawking give a talk or issue a preprint, people in his field would generally 'treat' it like a published piece of work and begin using it. 

 

Yes, but "using it", doesn't mean accepting it, or relying on it, without question.    By "using it" what actually happens is they confirm whether it works or not.

 

What happens with Kunchur, is he (or Yamaha, whoever) implies that you need a sampling frequency of 166khz to have a <7us timing difference between two channels....   and someone pulls out their DAW, and cues up two tracks with <7us difference using any sample rate they like.    Finished.

 

I am not saying this process is prefect, far from it, but I am saying the amount of checks-and-balance in the process does provide a certain degree of quality assurance and credibility. 

 

He intended to demonstrate whether someone could hear two misaligned sources.   He did that (all fine).    That is where the checks and balances would be.

 

 

He referred to the observations that "people say higher rates sound better"....  and then made an incorrect inference about sampling rates....  suggesting it might lend support to the observation.

 

People "say" a lot of things ..... but all actual results I know of, show that when you reduce the sampling rate of audio (as low as ~40khz), that as long as you avoid the pitfalls, it is not audible.

[davewantsmoore]

Totally agree with your perspective Dave, but re the bit in bold, we can hardly conclude that he has shown it while the major issues I listed remain, i.e. his filters changed his 7 kHz fundamentals by a detectable amount

 

He discusses this in his paper... and he provides references which support the idea that the level changes (0.2dB) in the experiment are not audible.

 

... anyways.    What I'm trying to say, it more fundamental.

 

 

He got results which say X   ......  he should conclude X   (ie.  that people can hear misalignment of 7us)    [yes, he might have missed something like you noted]

 

What he did was, got result X .....    and concluded X....   AND (inferred) that you will need to have a high sampling rate to provide this timing.

[davewantsmoore]

lo and behold the 0.2 dB change his filters made at 7 kHz is very close to detection thresholds

 

JND said to be 0.7dB.     Anyways, not particular relevant to the thread topic.

 

The resolution we actually need to achieve for good reproduction might well be 1000 times better than what his result show .....    but high sampling rates don't help for that.

[Guest rmpfyf]

(The discussion below is off-topic. Sorry)

 

I think it depends on which 'world' you are from. Kunchur is a scientist and academic, so I am talking specifically about his world. All I did is just outlining the traditional and accepted approach here; I have not presented my own opinion or view on it.

 

My opinion comes having been an active part of this 'world'.

 

Of course published work can contain mistakes and errors. (if you happen to read some of my posts in the Climate Change thread elsewhere in this forum, you would me see rising this point as a reality check to others) When errors are brought to the attention of the authors, and they conceded their mistakes, they would publish an erratum to notify the scientific world of this. Quite commonly others would publish their own work in which they would explain where the errors have occurred (with or without first consulting with the original authors); this is perfectly legitimate and proper way to proceed. If the original authors do not concede that they were in the wrong, they could write a counter paper defending their claims. I have even seem a case where the authors and their detractors debated over their work on the 'Letters to the Editor' page of the journal (over several issues!). Those are not peer reviewed but might be vetted by the Editor.

 

Now if you had said there are advanced knowledge, known to experts, that are not published then I totally agree with you. There are indeed much knowledge not been through the publication process. When someone like Stephen Hawking give a talk or issue a preprint, people in his field would generally 'treat' it like a published piece of work and begin using it. Scientific work is not slowed down by the need to eventually have the work peer reviewed and published. Scientists have been sharing preprints over FTP starting in 1991, they foresaw the need to rapidly share information before publication. This provides opportunities for others to identify any mistakes, and of course the majority of the preprints would eventually be published. The rapid and world wide distribution of unpublished papers did not replace the need for peer reviewed publication, but rather enhanced it (as feedback on flawed work would prevent the authors from even submitting it to a journal). 

 

Kunchur have stated in his FAQ that he went through a similarly contested pre-publication process. He presented his work at conferences where he was met with challenges. It is likely that any questions about blatant errors and limited applicability would have been raised with him before he even submit his papers. Moreover his journal editors and referees would have certainly pull him up for anything that are obviously wrong. I am not saying this process is prefect, far from it, but I am saying the amount of checks-and-balance in the process does provide a certain degree of quality assurance and credibility. 

 

I'll go another way.

 

There are many avenues to publish work. Criticism and critique are about as absolute as suppositions and inferences in submission for review... in other words, they're quite far from absolute. It takes a lot for reviewing authors to suggest a paper be not approved for publication. Strong suggestions might be made, some additional work and proofs may be sought, but to withhold work from publishing is rare, and would bring the submitting university dept into disrepute. I've seen some extremely borderline work published, even in moderately good journals and even despite a significant and deserved hazing in open forums. My Prof's most succinct quotation to this end goes a little like 'the body of published work doubles every two years, and 90% of it is crap'. 

 

To suggest that the supposed rigour of the process for publication is a proof of legitimacy of work is neither true nor the point of that process.

 

I'm suggesting that the above debate isn't even the point - as others have written, a lot of what Kunchur has observed and stated as a result of his research in this field makes a lot of sense and isn't at all contentious. We can hear very small sample delay: he evaluated this much and there's a positive evidential proof to this end. It makes a lot of sense, it's consistent with relevant neuroscience. There is a separate debate as pointed out by @@Newman as to the legitimacy of the test (some of which I'd support), but the nature of the results is neither news nor surprising (as @@Newman @@davewantsmoore rightly point out). 

 

I'm going to be honest (academic hat on) and suggest both the tests and the methods are fairly rudimentary, and the conclusions made are largely speculative - particularly in his 2007 paper. To suggest that the above results infer a higher sampling rate is necessary is at best an extrapolation of the results Kunchur obtained:

 

• Despite the suggestion, he didn't actually evaluate this.
• He neither proved nor disproved that a higher sample rate makes the difference audible.
• What he's suggesting is inconsistent with sampling theory for >=2 channel audio. Probably not picked up in review as it's not the point of the paper. 
• What he's suggesting has been proven for decades to be not possible psychoacoustically for single-channel audio. 
• His speculative conclusions concern largely the impact for musical tones with spectral content exceeding psychoacoustic limits... which he references... but incredibly (see first point) doesn't evaluate. At all. Despite having five subjects, speakers, reference microphones, amplifiers etc onsite. Whilst not a trivial experiment, it wouldn't have been too much of a stretch either, and would have done much for his speculation.

 

@@LHC, that's pretty rubbish for a scientific paper - conclusions typically reflect exactly what's studied, not what's speculated from it.

 

That proponents of hires anything in turn speculate from these speculations... isn't surprising.

 

As stressed previously - he does mention attack, or 'steep onset', where a cymbal can crack 120dB in around 10us, and is a waveform with real spectral content exceeding Redbook. I'm not sure why we're overlooking this - it's the only obvious part of his speculations as to the potential merit of hires audio that makes any sense, and deserves unpicking. If he'd followed this up with a paper testing this much, we'd have something with teeth. I'd look forwards to it.

[JSmith]

I'll have a bit if intermodulation distortion with my music please...

 

"192kHz digital music files offer no benefits. They're not quite neutral either; practical fidelity is slightly worse. The ultrasonics are a liability during playback.

 

Neither audio transducers nor power amplifiers are free of distortion, and distortion tends to increase rapidly at the lowest and highest frequencies. If the same transducer reproduces ultrasonics along with audible content, any nonlinearity will shift some of the ultrasonic content down into the audible range as an uncontrolled spray of intermodulation distortion products covering the entire audible spectrum. Nonlinearity in a power amplifier will produce the same effect. The effect is very slight, but listening tests have confirmed that both effects can be audible."

 

https://xiph.org/~xiphmont/demo/neil-young.html

 

"Indeed, there’s little evidence that High-Resolution Audio is perceptibly better than CD audio. Quite the contrary, the evidence appears to suggest otherwise.

 

A 2007 paper by Brad E. Mayer and David R. Moran described a double-blind trial where audio engineers, “dedicated audiophilesâ€, and audio-recording university students were played CD-quality audio and self-described High-Resolution Audio.

 

The result? No perceptible difference between CD-quality audio and High-Resolution Audio."

 

http://www.makeuseof.com/tag/high-resolution-audio-future-music-scam/

 

... anyone for some tomato in their source? 

 

JSmith 

[gcgreg]

The Xiph article is very good on a number of fronts, especially when read with the footnotes and linked supporting material. I've just played with the intermodulation files with my computer direct to DAC. It showed two things in fact: my system doesn't exhibit any audible intermodulation distortion; I definitely cannot hear 30khz, 33khz or even an entire audio clip of ultrasonic content.

Edit: spelling

Edited January 26, 2016 by gcgreg

[LHC]

The Xiph article is very good on a number of fronts, especially when read with the footnotes and linked supporting material. I've just played with the intermodulation files with my computer direct to DAC. It showed two things in fact: my system doesn't exhibit any audible intermodulation distortion; I definitely cannot hear 30khz, 33khz or even an entire audio clip of ultrasonic content.

Edit: spelling

 

So you are saying intermodulation distortion is not an universal problem, it certainly didn't apply to your system. 

[gcgreg]

So you are saying intermodulation distortion is not an universal problem, it certainly didn't apply to your system.

Yes, and that is what the author points out in one of the footnotes making a distinction between "can" and "will".

My further point was that there was nothing at all audible in any of the tracks. Some people may need to chew on the implication of that...

[LHC]

Yes, and that is what the author points out in one of the footnotes making a distinction between "can" and "will".

My further point was that there was nothing at all audible in any of the tracks. Some people may need to chew on the implication of that...

 

People can do and say whatever they like, its a free world. 

 

Not everyone can hear the benefit of hi-res, and not everyone will be troubled by intermodulation distortions. It is really a case of 'your mileage will vary'. 

[firedog]

"Indeed, there’s little evidence that High-Resolution Audio is perceptibly better than CD audio. Quite the contrary, the evidence appears to suggest otherwise.

 

A 2007 paper by Brad E. Mayer and David R. Moran described a double-blind trial where audio engineers, “dedicated audiophilesâ€, and audio-recording university students were played CD-quality audio and self-described High-Resolution Audio.

 

The result? No perceptible difference between CD-quality audio and High-Resolution Audio."

 

http://www.makeuseof.com/tag/high-resolution-audio-future-music-scam/

 

... anyone for some tomato in their source? 

 

JSmith 

 

 

Really? The Meyer - Moran paper has been shown to have multiple serious flaws. So much so that it doesn't prove anything. 

 

Just one example: They didn't check the source of the SACDs they used as sources for "high -res". Some of them had been produced from 16/44.1 recordings that were converted to DSD/SACD. Result: you can't show using these sources that listeners can't tell a difference between "hi-res" and Redbook, if your source isn't hi-res in the first place. 

 

That's a fatal flaw in their experiment. There are lots of other flaws which make the whole paper nothing more than a curiosity, IMO. 

[LHC]

Really? The Meyer - Moran paper has been shown to have multiple serious flaws. So much so that it doesn't prove anything. 

 

Just one example: They didn't check the source of the SACDs they used as sources for "high -res". Some of them had been produced from 16/44.1 recordings that were converted to DSD/SACD. Result: you can't show using these sources that listeners can't tell a difference between "hi-res" and Redbook, if your source isn't hi-res in the first place. 

 

That's a fatal flaw in their experiment. There are lots of other flaws which make the whole paper nothing more than a curiosity, IMO. 

 

Consistent with what I wrote earlier, I can't agree the Meyer-Moran paper is merely a curiosity. It is a properly peer reviewed published paper and therefore it is a legitimate part of our knowledge. Yes there may be questionable aspects of their work, flaws if you will, but that means one needs to exercise extreme care to interpret the results and drawn conclusions. Their results may be of little use, I don't know. I know that Meyer have issued a statement defending his work against online criticisms. 

 

I am not saying you can't express your own opinion, of course you can. But I insists that the only proper way to deal with this is for someone else to rebut their work through the same peer reviewed publication process. 

[davewantsmoore]

So much so that it doesn't prove anything. 

 

The results of their test show that a lot of people couldn't hear an AD/DA process added.  <shrug>  Prove is such an absolute word.

 

These are just some results, and similar but different tests can be (and are) done by people all the time, which may address any number of the tests flaws.

[firedog]

Consistent with what I wrote earlier, I can't agree the Meyer-Moran paper is merely a curiosity. It is a properly peer reviewed published paper and therefore it is a legitimate part of our knowledge. Yes there may be questionable aspects of their work, flaws if you will, but that means one needs to exercise extreme care to interpret the results and drawn conclusions. Their results may be of little use, I don't know. I know that Meyer have issued a statement defending his work against online criticisms. 

 

I am not saying you can't express your own opinion, of course you can. But I insists that the only proper way to deal with this is for someone else to rebut their work through the same peer reviewed publication process. 

 

Besides the fact that the experiment in the end was sloppy - they made basic errors which showed a real lack of understanding of what they were dealing with - it is endlessly quoted as "proof" that hi-res and Redbook have no audible differences. That's my biggest objection. 

 

The "only proper way to deal with it" is to examine the work and see if it stands up to scrutiny. It doesn't have to be a peer reviewed paper, especially if errors can be shown to have been made that don't need peer review to see them. Sometimes even "peer review" doesn't catch pretty important errors.  Some peer review is pretty weak, and may consist of nothing more than checking math. Clearly the "peers" who reviewed the Meyer-Moran paper didn't know enough about SACDs to know that many (especially from a time before hi-res recording became widespread) were made from Redbook upsampled to DSD. And that was something well known among audiophiles who were classical music fans, for example. 

 

Your point is also a little naive - these tests are difficult and expensive to run - and there aren't many sources for funding them. It's one of the reasons there are few such tests run on any "audiophile" subjects at all. Besides chattering audiophile circles, there isn't anyone interested too much in in these types of studies, and even less money to fund them. So it's doubtful you will ever see  "proper" (as you call it)  studies done at all. 

Edited January 27, 2016 by firedog

[davewantsmoore]

Besides the fact that the experiment in the end was sloppy - they made basic errors which showed a real lack of understanding of what they were dealing with - it is endlessly quoted as "proof" that hi-res and Redbook have no audible differences. That's my biggest objection. 

 

But surely you read their paper, and see that is not exactly what they conclude?!

 

Based on the results that a AD/DA at 16/44 is not audible ..... they propose that any claims which say that resampling (any) audio to 16/44 is audible....  should be backed up with robust evidence.

[davewantsmoore]

these tests are difficult and expensive to run - and there aren't many sources for funding them. It's one of the reasons there are few such tests run on "audiophile" subjects at all. Besides chattering audiophile circles, there isn't anyone interested too much in in these types of studies, and even less money to fund them. So it's doubtful you will ever see  "proper" (as you call it)  studies done at all. 

 

It takes a lot of results/participants to go anywhere near proving the null hypothesis.

 

... and so the only thing which is likely to be published are things which are shown to be audible    (unlike Meyer Moran who couldn't show something was audible).

 

AFAIUI no studies progress past small numbers of participants, as the results are all one way.

[firedog]

But surely you read their paper, and see that is not exactly what they conclude?!

 

Based on the results that a AD/DA at 16/44 is not audible ..... they propose that any claims which say that resampling (any) audio to 16/44 is audible....  should be backed up with robust evidence.

No, they used the output of "hi-res" SACDs - converted to analog by SACD players -  and then reconverted to 16/44 digital to conclude that no difference was audible. Ergo, listeners can' t hear the difference between SACD/DVD-A hi-res and Redbook.

[firedog]

It takes a lot of results/participants to go anywhere near proving the null hypothesis.

 

... and so the only thing which is likely to be published are things which are shown to be audible    (unlike Meyer Moran who couldn't show something was audible).

 

AFAIUI no studies progress past small numbers of participants, as the results are all one way.

 

Try a meta analysis which combined the results of 20 smaller studies (out of 80) found to have suitable data for analysis:

 

Joshua Reiss of Queen Mary University, London, where Dr. Reiss performed meta-analysis on 80 published papers concerning high-resolution audio wherein he found:

 

the overall result was that trained listeners could distinguish between hi-rez recordings and their CD equivalents under blind conditions, and to a high degree of statistical significance.

 

 

 Results were presented at an AES workshop. 

Edited January 27, 2016 by firedog

[davewantsmoore]

No, they used the output of "hi-res" SACDs - converted to analog by SACD players -  and then reconverted to 16/44 digital to conclude that no difference was audible. Ergo, listeners can' t hear the difference between SACD/DVD-A hi-res and Redbook.

 

We need to ber very clear about what they did.... and not overstate their conclusions.

 

What they did was insert a AD/DA (16/44) into a system .... and concluded that the AD/DA was not audible.

 

The proposed that because re-sampling an audio signal (to 16/44) was not audible .....  that any claims to the contrary would need to be supported by a similar group of results to theirs  (ie.  reasonably large volume, and reasonably good test procedure). 

 

 

 

Perhaps, they could have used some better high sample rate/depth content.    People are free to repeat the tests with such.

 

They noted specifically that SACDs sounded better than CDs .....   but that an SACD re-sampled with an AD/DA to 16/44 was not audibly different to the original SACD  (ie. it still sounded 'better').

 

 

People routinely misrepresent their results and conclusions....  by using absolutes like "proof" ... and "SACDs don't sound different".   These things are decisive and polarising....   and draw attention away from their interesting (and quite repeatable IME) results.

[davewantsmoore]

Try a meta analysis which combined the results of 20 smaller studies (out of 80) found to have suitable data for analysis:

 

Joshua Reiss of Queen Mary University, London, where Dr. Reiss performed meta-analysis on 80 published papers concerning high-resolution audio wherein he found:

 

 Results were presented at an AES workshop. 

 

 

Cool.    If anyone finds a link to the slides, post it up.

 

Here's a snippet of JAs writting on it:

 

Harper was referring to a 2007 paper by E. Brad Meyer and David R. Moran that "proved" that there was no sonic advantage to high-resolution audio formats (footnote 3). Their conclusion ran counter to the experience of many recording engineers, academics, and audiophiles, but other than doubts over their methodology and the fact that their source material was of unknown provenance, Meyer and Moran's paper seemed to be the final formal word on the matter.

Read more at http://www.stereophile.com/content/simple-everything-appears-simple#Io8aJUudpQLx4PtJ.99

 

 

It doesn't run counter to the experience of anyone I know   (not that this means anything though).

 

 

It's funny that even very smart people can get caught up in mis-representing the results of others, when it pushes people buttons in a polarised discussion.

 

Here's a what they found in the paper:

 

The test results show that the CD-quality A/D/A loop was undetectable at normal-to-loud listening levels, by any of the subjects, on any of the playback systems

 

 

 

... here's what they concluded.

 

Now, it is very difficult to use negative results to prove the inaudibility of any given phenomenon or process. There is always the remote possibility that a different system or more finely attuned pair of ears would reveal a difference. But we have gathered enough data, using sufficiently varied and capable systems and listeners, to state that the burden of proof has now shifted. Further claims that careful 16/44.1 encoding audibly degrades high resolution signals must be supported by properly controlled double-blind tests.

 

 

There might be method problems .... and the tests could be repeated using "better" high rate/depth content......  EVEN using content which is artificially changed to (eg.  adding extra high frequencies, or very quiet sounds) to try and find (induce) the point of audibility.

 

 

.... for now, I'm quite fine re-sampling all my high rate recordings to 44.1.    They sound excellent.

Edited January 28, 2016 by davewantsmoore

[Newman]

We need to ber very clear about what they did.... and not overstate their conclusions.

 

What they did was insert a AD/DA (16/44) into a system .... and concluded that the AD/DA was not audible.<snip>

 

People routinely misrepresent their results and conclusions.... 

 

I spectated at the M&M debates with some bemusement. They definitely opened themselves to debate by using commercial music discs with debatable provenance. If only they had approached a local small studio, explained their project, they could have gotten their hands on some proven high-res, high dynamics master files and gone from there.

 

A couple of points that I think their project makes:

• Even if they were accidentally playing CD-quality music files, their results seem to indicate that ADDA at 1644 is incredibly difficult to detect, on an analog signal with 20-20k bandwidth and 90-odd dB noise floor.
• Given that they seem to have been accidentally playing a mix of CD-resolution and DSD-resolution music pieces, you would expect that if high-res is audibly different, their results would have fallen confusingly into two data sets, one being consistently distinguishable, and the other set consistently indistinguishable. That didn't happen.

They noted specifically that SACDs sounded better than CDs .....   but that an SACD re-sampled with an AD/DA to 16/44 was not audibly different to the original SACD  (ie. it still sounded 'better').

 

 

Noting that this was meant to be a comment that studios are preparing better-sounding masters for SACD release, not that the format has inherently better sound.

[sir sanders zingmore]

Noting that this was meant to be a comment that studios are preparing better-sounding masters for SACD release, not that the format has inherently better sound.

This ^

Edited January 28, 2016 by Sir Sanders Zingmore