AES Convention Day 2
I finally got to the AES Exhibit floor yesterday in between paper sessions and a meeting of the High-Resolution Audio working group. The entire show was contained within one hall of the LA Convention Center, which allowed me to survey pretty much all of the vendors in a few hours.
I found it interesting to see large groups of individuals sitting in two of the corners of the open space attending seminars on “Project Studios” and “Live Sound”. How could they hope to put on seminars amidst the noise and movement of the main hall? They provided wireless headphones to all of the attendees. Mick Guzauski, an 11-time Grammy winner, was there at the “Project Studio” session talking about the use of limiters in his mixing without a PA system…just the local broadcast to those wearing the special headphones.
Today, I thought I would give a quick overview of the day. I plan to explore the individual topics in full posts over the next week or so.
I arrived just prior to the start of the high-resolution audio technical working group chaired by Vicki Melchior. There were about 15 people in attendance. Unfortunately, there didn’t seem to be much of an agenda to direct the meeting. There were ad hoc discussions of the benefits of XRCD, vinyl LPs, and direct to disc recording, which I thought wasn’t the best use of the short time we had. My contribution was a discussion or the JAS logo and whether the AES should step up and endorse (or reject) the specifications and logo being considered by the CEA.
Steve Green, the Business Development Manager for the AES organization commented about the inclusion of the DEG High-Resolution Audio sessions that were held on Friday. And he suggested that the entire convention in New York next fall would be focused on “high-resolution audio”. That was all it took for me to exit the room with him, exchange business cards, and have a very brief chat about getting the programming committee to stand apart from the DEG and get some serious discussions going about the topic. Sure it’s great to have celebrity recording, mixing and mastering engineers as members of a few panels, but it’s equally important to get accurate and meaningful information to the membership. This was a very positive sign.
Following the HRA working group meeting, I happened upon the Benchmark Media booth. There was John Siau himself with display table showing off the AHB2 Amplifiers (which he promised would be available very soon…the assembly of the units has begun), their new bookshelf speakers, the ADC, and DAC2 converters. John and I sat down and talked for about 30 minutes about a variety of topics, which is always a pleasure. One topic that he and I had briefly discussed previously was about intersample overages in PCM sampling systems…especially as it relates to Sample Rate Conversion. He explained why it is necessary to reduce the amplitude of your output digital signals by 3.5 dB to ensure that the converted signal doesn’t eclipse the bit maximum of the new file. I’ll talk about this in more detail shortly…but I found it very interesting.
I finished the afternoon by attending a few paper sessions. The first was titled, “The Audibility of Typical Digital Audio Filters in a High-Fidelity Playback System”. Although it may not be obvious from the title of the paper, this is the first AES publication that refutes the Meyer/Moran research that has been so often quoted as “proof” that CD specification PCM audio is enough for music reproduction (Meyer and Moran’s research has been widely discredited including by myself because of the lack of real high-resolution content used during the study).
Robert Stuart and his colleagues conclude: “first there exist audible signals that cannot be encoded transparently by a standard CD; and second, an audio chain used for such experiments must be capable of high-fidelity reproduction.” In other words, CDs aren’t good enough. This paper was given the top award by the AES organization. This is a very important finding.
His second paper, “A Hierarchical Approach to Archiving and Distribution” written with his long time collaborator Peter G. Craven focused on the possibility of using varying sample rates and bit depths to capture high-fidelity audio more efficiently than currently methods. Again, I’ll get back in to this topic as well.
It was a great day…meeting friends, seeing some of my students, and learning new things. I’m headed back for day three.
More details on the papers Dr. Waldrep referenced: http://www.aes.org/events/137/papers/?ID=4058
Thanks Dennis…I should have put the link up. It’s been a very busy weekend.
I’m rather disappointed you feel the M&M paper has been “discredited” in any credible manner, or that you feel the Stuart paper here “refutes” the M&M paper. The abstract for the Stuart paper has be available for a while, and it’s clearly a different study – nowhere in the abstract do they claim to have tried to reproduce the M&M results. And nowhere in the M&M paper do those authors claim to have proof or have proven the “CD specification PCM audio is enough for music reproduction”.
I have been a big supporter of your efforts in the past, but if you abandon or refuse to use the scientific method in your work, you’ll lose me and many others. I agree that M&M did not test what you wanted them to test. But no one – not you, nor anyone else – has refuted the actual conclusions that they drew, nor repeated their test as they scoped it and come up with different results. No one. Many have concocted strawmans from the M&M study and rejected those strawmans, but that is neither impressive nor valid.
Many of your opponents use strawmans and other distortions to make their points and you regularly out them for such, and I applaud that. I can’t encourage you enough to not use strawmans and similar tactics yourself.
The Meyer and Moran paper has no credibility for reasons that I and others (including Robert Stuart) have identified. Every one of the tracks they evaluated contained no more frequencies or dynamic range that an equivalent CD (or the downconverted version). You can brush this aside and still hope the Meyer and Moran holds water but it simply doesn’t. They conclusion in M&M was that listeners did no better than chance in being able to detect a so-called “high-resolution recording” vs. a downconverted CD spec version. Perhaps I could have worded my statement better…but the meaning is the same.
Robert Stuart believes as I do that ultrasonics matter…and his paper establishes that there is a perceptible difference between a source file at 192 kHz/24-bits and a filtered version (with a single filter chain) at 44.1 kHz. His conclusion that “there exist audible signals that cannot be encoded transparently by a standard CD” runs counter to the M&M research.
I feel that my statements and the reasons are not strawmans as you say. The conclusions of Meyer and Moran are meaningless because the methodology was fatally flawed. Explain to me how they could have gotten any other result when they used the sources and methods that they did.
I spoke to Bob Stuart specifically about their study. He didn’t plan his research to redo their process because there are better methods available today to solve the issue.
I agree with Lazy on this. To state that that the M&M study was “flawed” means you or others have found areas in the study where the conclusions in the study are not supported by the actual test approach and data. I don’t believe anyone has done that. Many have condemned the study on the basis of purported conclusions that the authors never made or dislike for the recordings and equipment used. The former are strawmans and thus not valid, and the latter are calls for a different study that everyone has been free to perform but no one has.
For the record, the authors only asserted inaudibility for the specific recordings and the specific test approach that they used. They even went out of their way to state that lack of audibility differences in their test is NOT proof of inaudibility of hi rez versus CDDA, and stated that such proof would require audibility differences detected via properly-controlled double blind test. Also note that given an accurate read of the M&M conclusions, if such a study were performed and if it reliably detected audible differences, such a study would not refute the M&M study, because M&M never generalized their conclusions as so many people desperately want to believe. Refutation of the M&M study would require duplication of the study’s tests as closely as possible and achieving different results, or finding that double blind protocols wee not followed, or finding computation errors or the like. None of that has happened although people have been free to attempt so.
The concepts of “proof”, “credibility” and “refutation” have been carefully defined within the body of work that defines the scientific method over many years, and I recommend sticking with those. You are welcome to read the summary paper of the actual study that includes the conclusions I mentioned above: http://www.drewdaniels.com/audible.pdf
I’ve read the paper carefully and stand by my assessment of their study. It didn’t establish anything regarding “Aubibility of a CD-Standard A/D/A Loop Inserted into High-Resolution Audio Playback” because none of the sources were in fact, high-resolution. The authors and the members of the BAS may have purchased the latest SACDs and a single DVD-Audio titles but the fact the the labels said they were “high-resolution” doesn’t make it true.
The abstract states, “The authors report on a series of double-blind tests comparing the analog output of high-resolution players playing high-resolution recordings with the same signal passed through a 16-bit/44.1-kHz “bottleneck.” There casual claim that they compared the output of players playing “high-resolution recordings” just isn’t true. They eventually published the list of titles that were reviewed and none of them possessed any sound qualities that a bona fide high-resolution track does.
If they didn’t test material that had any elements that might be audible, then they didn’t test anything.
No one credible assesses a paper based on the abstract or title to the exclusion of what is actually in the paper. M&M defined what they meant by “hi rez” (music offered via SACD and DVD-A media), and while you clearly don’t like that definition (nor do I), you can’t just replace their definition with yours and then assess the paper based on your definition.
M&M even acknowledged that the sound floor was evaluated as better on the SACD and DVD-A music they tested, when the volume was around 99 dB, which was 14 dB higher than the loudness for which they conducted their tests.
I am a big supporter of high resolution music. But disparaging the M&M study on bases that are not consistent with the scientific method don’t help the cause. I actually share your frustration that M&M didn’t select recordings that had frequency content above 22k or use a playback system confirmed to accurately reproduce content above 22k. Given the effort they gave their test, I think it was deliberate on their part, and that really irritates me. So their self-imposed test limits merely show that music recorded with content below 22kHz is inaudibly different when the format is DVD-A, SACD or CDDA. In the hi rez world that’s a big “Duh”. I hope you agree with that frustration and assessment, and I think summarizing this study as a “Big Duh” within the bounds of the scientific method is far more credible than asserting strawmans.
So it is enough to assert that M&M’s results show that format alone does not define high resolution audio for content below 22k (which you’ve stated many times), and that live music spectral content and other factors in defining true high resolution music remain untested from an audibility perspective (I agree and I think M&M would also agree).
M&M didn’t define high-resolution. They assumed that the commercial products that were available at the time as supplied by the members of the BAS. My point is that using the M&M study, as many anti high-resolution people do, to dismiss anything better than CD specification audio is wrongheaded for the reasons that I’ve already provided. It’s not necessary for M&M to accept my definition or any other definition. Their study attempted to compare a bunch of commercially available disc releases to a downconverted version of the same titles. Their conclusion was that the ability to tell them apart was random. If we accept that the commercially available titles didn’t have any fidelity beyond that of the CD downconverted versions then there would be no way to get anything but random results. Case closed…another study is clearly needed.
But those who wish to use M&M to bolster their belief that High-Resolultion Audio is unnecessary should understand the limitations of the study. Thus far…I don’t think that’s happened.
The M&M study did provide evidence for a conclusion about audibility of recordings, it’s just not the conclusion reached by the authors. The study clearly demonstrates that converting a recording with resolution no greater than standard CD quality to a format that supports a higher resolution will not result in a recording of higher resolution than the original recording. The subjects in this study were not able to detect any difference between the CD and the SACD or DVD-Audio formats of non high-resolution recordings. Mark, isn’t this just what you have been saying? Converting a non high-resolution recording to a format that supports high-resolution does not make the recording high-resolution.
Is the paper going to be behind a $20 paywall?
Meridian have a bit of a reputation for white papers that ‘prove’ their expensive technologies (products) are audibly better. I hope this is not another one of those.
Boy, I sure haven’t seen any misrepresentation or heard this previously. I find Robert Stuart and Peter Craven to be among the most thorough individuals I know. I don’t believe they “pump” up their results. I listened to Bob’s presentation and have read both papers…this is the real deal.
Will there be a proceedings for the convention? I’d like to read Meridian’s paper.
Actually, I’d like to read both Meridan’s papers.
Members of the AES can get access to the papers through the site. Others will have to purchase access.