SACD vs Standard CD

[Bryn]

1 bit 2.8224 MHz (note MHz not kHz).

Though it should be stressed that this is using DSD, not LPCM, and that the recordings on many SACDs originate from 96/24, 192/24 or higher resolution LPCM, rather than being original DSD recordings which are cnverted to DSD as part of the authoring.

[Mal]

...section 4 - "... the SACD ... recordings sounded better than most CDs—sometimes much better."

If you take a recent SACD of Mozart symphonies and compare it to DG's 1983 Bernstein CD then the SACD is likely to sound better. But this isn't because it's a SACD, it's because the engineers made more effort, or used more modern techniques, to produce an audiophile product. They did this because audiophiles will make a big stink if a SACD sounds worse than a CD!

Actually the SACD layer might sound better than the CD layer on the same physical product... the engineers might botch the conversion to make the company more £££. If the CD layer sounds the same as the SACD layer to audiophiles performing double blind testing then they will say, why are we paying extra £££ for SACD?

"Had we not “degraded” the sound to CD quality and blind-tested for audible differences, we would have been tempted to ascribe this sonic superiority to the recording processes used to make them." Why didn't the experiment compare the original, unmodified SACD recordings with CD?

Are you asking why they didn't compare a SACD release to a CD release of the same performance? The answer is, they (rightly) didn't trust the engineers to do a good job. They trusted *themselves* to do a good job, because they are top research scientists, not average engineers suffering from commercial & time pressures.

Look at fig. 1. This shows they compared "original unmodified" SACD (track A...) with that SACD dowgraded to CD (track B...). This seems to me exactly what the experiment calls for.

[Bryn]

This link is only directly to an abstract, but this later paper comes to different conclusions regarding the discrimination of higher sampling rates:

http://www.aes.org/e-lib/browse.cfm?elib=15398

[Mal]

Here's Moran, a decade after that paper taking on the editor of stereophile and fellow academics:

"You might think instead that someone would publish their own test showing the dramatic difference hi-rez makes. But in almost a decade now, much less longer, nooooo.

So come on, Atkinson, the commercially interested Stuart, Reiss, all others, do the test, train the listeners, force the choice, don't use crap filters as in the Stuart preprint, and just do it. It ain't that hard. We did it, perfectly or imperfectly. Have at it. Conclusively blind-show the superiority of hi-rez.

Ah, I thought so. Bwok bwok bwok."

Read more at https://www.stereophile.com/content/...3TpXThK1efj.99

[Mal]

This link is only directly to an abstract, but this later paper comes to different conclusions regarding the discrimination of higher sampling rates:

http://www.aes.org/e-lib/browse.cfm?elib=15398

Here's the full paper:

https://www.researchgate.net/publication/257068631_Sampling_Rate_Discrimination_441_kHz_vs_882_kHz

Blimey, I thought I'd retired from scientific research :)

OK, I might accept that with this partiicular, perfect recording of Bruckner 6, differences might be heard by *some* "super experts" in the prime of life. Federers of listening...

"Sixteen expert listeners, fifteen males and one female1, with a mean age of 30 (SD = 7.1), took part in the study and received CDN$20 per hour for their participation. All participants reported having studio experience in sound engineering for an average of 8 years (SD = 5.6). Six reported working as professional sound engineers in Montreal and ten were Sound Recording students at McGill University. All participants except one had musical training (15 years on average, SD = 5.5)."

But some failed:

"While we observed audible differences between sample rates of 88.2 and 44.1 kHz, they remain very subtle and difficult to detect. It is difficult to interpret why three out of 16 participants significantly picked the wrong answer... "

Does this show the *dramatic* difference that Moran calls for in post #94?

P.S David Moran pops up in another audiophile thread(!) criticising this experiment, which makes me inclined to retract my attempt at acceptance:

https://www.sa-cd.net/showthread/58757/60081

[jayne lee wilson]

Here's Moran, a decade after that paper taking on the editor of stereophile and fellow academics:

"You might think instead that someone would publish their own test showing the dramatic difference hi-rez makes. But in almost a decade now, much less longer, nooooo.

So come on, Atkinson, the commercially interested Stuart, Reiss, all others, do the test, train the listeners, force the choice, don't use crap filters as in the Stuart preprint, and just do it. It ain't that hard. We did it, perfectly or imperfectly. Have at it. Conclusively blind-show the superiority of hi-rez.

Ah, I thought so. Bwok bwok bwok."

Read more at https://www.stereophile.com/content/...3TpXThK1efj.99

Said all I have to say really, I don't think blind tests, especially short-excerpt ABX types like the one you link to, work at all well or offer the best opportunity to human perception/discrimination. (However carefully set up).
But I would just remind everyone that HiFiNews analyses a number of hi-res releases (24/48/96 etc) in the magazine every month and publishes the graphic results; which do indeed usually show content above the lossless limit. They also show where a given file has been up sampled, converted from DSD to PCM or vice versa etc. etc.

They've been dong this very beneficially for some years. Some of us measure such files in Audacity etc at home...then - listen and learn.....

The point I made which Bryn restated, that many SACDs have been derived from PCM masters is a very vital one; it makes the whole process of comparison much more difficult, especially off-disc. And the oft-overlooked point about converters within players, that some will convert back to PCM or effect a compromise, is important too. The DSD/PCM processing paths have to be discrete for any listening tests to make any sense. More DACs are available now which allow this easily.

Some earlier, very highly regarded Esoterics for example, did convert the DSD back to 24/96 PCM. Which seemed to do the sound no harm at all, quite the reverse.
I only wish I'd got my hands on one...

Still one would always prefer the native 24/96 file if you can get it. In principle and, where I made the comparison, on audition. Hence my choice back in the day to stick with hires file/stream, and not offer shelfspace to an SACD Player, excellent music-players though many of them are.

[johnb]

I hesitate to enter into a discussion about CD Red Book vs HiRes vs SACD vs DSD because so many peoples opinions (including mine) are based on their perceptions (rather than hard concrete evidence) and we all know that perceptions can be misleading. (I am not saying that members of any camp are right or wrong.)

However, I think the main advantage of HiRes recordings are that the Nyquist frequency (half the sampling rate) is moved well away from the audible spectrum. This means that any "compromising" effect of the complex filters used in DACs or CD players is also moved well beyond human hearing. I suspect this might well be main reason that people find that HiRes audio is superior.

I do agree with Jayne that short excerpt ABX comparisons are problematic, but that doesn't mean that all blind comparisons can be safely ignored. Personally, I do tend to prefer HiRes recordings but I am also very aware of my own expectation bias.

[Dave2002]

Though it should be stressed that this is using DSD, not LPCM, and that the recordings on many SACDs originate from 96/24, 192/24 or higher resolution LPCM, rather than being original DSD recordings which are converted to DSD as part of the authoring.

It’d be interesting to have an “accurate” history of SACD. When it first came to my attention it was supposedly a 1 bit sampling method at a very high clock frequency (DSD). I assumed that DSD recorders were available. In the early days of CD we know that LPCM encoding wasn’t always very good, so maybe DSD was developed to overcome some of the deficiencies. Later, SACD/DSD may have been promoted as a way of archiving digital recordings. Now we have higher bit rate LPCM recordings with 24 bit (or more) resolution, and we now “learn” that many SACDs are in fact generated from LPCM recordings. So what is/was the “truth”? Perhaps it was different at different times in CD and SACD evolution.

Earlier explanations of DSD referred to noise shaping as a justification for its supposed superiority over CD recording at the time.

I heard “rumours” that some recordings were made using some hybrid process, with (say) a 4 bit sampler/encoder running at very high frequency.

[Ian_of_glos]

If you take a recent SACD of Mozart symphonies and compare it to DG's 1983 Bernstein CD then the SACD is likely to sound better. But this isn't because it's a SACD, it's because the engineers made more effort, or used more modern techniques, to produce an audiophile product. They did this because audiophiles will make a big stink if a SACD sounds worse than a CD!

Actually the SACD layer might sound better than the CD layer on the same physical product... the engineers might botch the conversion to make the company more £££. If the CD layer sounds the same as the SACD layer to audiophiles performing double blind testing then they will say, why are we paying extra £££ for SACD?



Are you asking why they didn't compare a SACD release to a CD release of the same performance? The answer is, they (rightly) didn't trust the engineers to do a good job. They trusted *themselves* to do a good job, because they are top research scientists, not average engineers suffering from commercial & time pressures.

Look at fig. 1. This shows they compared "original unmodified" SACD (track A...) with that SACD dowgraded to CD (track B...). This seems to me exactly what the experiment calls for.

Thank you for your explanation, but it is all rather technical for my simple brain to comprehend. However I am pretty sure that all my SACDs are recent recordings, rather than old recordings presented in the SACD format just to make the little blue light come on when they are played.
Are you saying that the CD layer is not as good as it could be so that people like me can hear a difference and thus gravitate towards SACD recordings in future?

[Stunsworth]

Are you saying that the CD layer is not as good as it could be so that people like me can hear a difference and thus gravitate towards SACD recordings in future?

Certainly in the pop/rock/jazz world the two layers are often not from the same master. The CD layer will be the standard master used for CD production while the SACD layer can be a new master.

[Ian_of_glos]

Certainly in the pop/rock/jazz world the two layers are often not from the same master. The CD layer will be the standard master used for CD production while the SACD layer can be a new master.

I think that might explain what I am hearing (even though I don't listen to much pop or jazz).
Many of the articles I have read argue that there is little point in taking an old recording and presenting it as an SACD, because it is not possible to add something that was not there on the original. However, I think the record companies are just looking for an opportunity to make some more money by reselling old recordings on SACD (eg Pink Floyd Dark Side of the Moon) in the same way as they did when CDs were first released.

[Mal]

Are you saying that the CD layer is not as good as it could be so that people like me can hear a difference and thus gravitate towards SACD recordings in future?

I'm only saying it's possible. It requires proper double blind testing to check for this.

[LHC]

I think that might explain what I am hearing (even though I don't listen to much pop or jazz).
Many of the articles I have read argue that there is little point in taking an old recording and presenting it as an SACD, because it is not possible to add something that was not there on the original. However, I think the record companies are just looking for an opportunity to make some more money by reselling old recordings on SACD (eg Pink Floyd Dark Side of the Moon) in the same way as they did when CDs were first released.

Its interesting that you mention the Dark Side of the Moon SACD. the CD layer on this disc (and indeed all other CD versions of DSOTM) was produced from the original two-track master tape from 1973. However, the SACD multi-channel mix was an entirely new mix made from the original multitrack elements, which were kept in the analog domain until converted to DSD for the SACD. So in this case, purchasers of the SACD were getting something different, and not just a DSD conversion of a recording that was already available in multiple formats.

The use of the original multitrack elements rather than the 1973 two channel downmix could also be one of the reasons why the SACD layer sounds better than the CD layer, or any previous CD version of DSOTM.

[richardfinegold]

Said all I have to say really, I don't think blind tests, especially short-excerpt ABX types like the one you link to, work at all well or offer the best opportunity to human perception/discrimination. (However carefully set up).
But I would just remind everyone that HiFiNews analyses a number of hi-res releases (24/48/96 etc) in the magazine every month and publishes the graphic results; which do indeed usually show content above the lossless limit. They also show where a given file has been up sampled, converted from DSD to PCM or vice versa etc. etc.

They've been dong this very beneficially for some years. Some of us measure such files in Audacity etc at home...then - listen and learn.....

The point I made which Bryn restated, that many SACDs have been derived from PCM masters is a very vital one; it makes the whole process of comparison much more difficult, especially off-disc. And the oft-overlooked point about converters within players, that some will convert back to PCM or effect a compromise, is important too. The DSD/PCM processing paths have to be discrete for any listening tests to make any sense. More DACs are available now which allow this easily.

Some earlier, very highly regarded Esoterics for example, did convert the DSD back to 24/96 PCM. Which seemed to do the sound no harm at all, quite the reverse.
I only wish I'd got my hands on one...

Still one would always prefer the native 24/96 file if you can get it. In principle and, where I made the comparison, on audition. Hence my choice back in the day to stick with hires file/stream, and not offer shelfspace to an SACD Player, excellent music-players though many of them are.

As a Physician, I wouldn't prescribe anything that was validated as efficacious in randomized, double blind studies.
As an Audiophile, such tests are irrelevant, for all of the reasons that jlw has previously outlined. Audiophilia is an Art, not a Science, and these trials are not just meaningless, but dumb

[HighlandDougie]

As a Physician, I wouldn't prescribe anything that was validated as efficacious in randomized, double blind studies.
As an Audiophile, such tests are irrelevant, for all of the reasons that jlw has previously outlined. Audiophilia is an Art, not a Science, and these trials are not just meaningless, but dumb