A Plus For Meyer and Moran
I’ve had a revelation. Thanks to a comment by a reader, I’ve discovered that the Meyer and Moran research and AES paper did actually produce something meaningful. It certainly wasn’t what they thought they established (that high-resolution audio and a downconverted CD spec version were audibly indistinguishable). What Meyer and Moran demonstrated in 2007 was that an analog tape transferred to a DSD 64 SACD format disc is indistinguishable from a CD resolution version of the same thing. And it turns out that this is important…and not good news for DSD.
Remember the list of products that were used during the Meyer and Moran project? You can take a look at it here. Every product listed…except one…is an SACD using DSD at 2.8224 MHz. And substantially more than half of them were original analog recordings transferred to DSD digital. The rest were sent through PCM systems to mix, master, and release on SACD.
And remember that SACD using DSD 64 has no more frequency range than a CD. And commercial recordings on analog tape have no more dynamic range than a CD.
This is the basis of my contention that the source materials used in the study were not real high-resolution recordings. They were standard definition sources packaged as DSD. The fidelity of these recordings is dependent on the production techniques used during the original sessions. The Pink Floyd “Dark Side of the Moon” was recorded on an analog 24-track tape deck, mixed to analog tape or PCM digital and then converted to DSD in both stereo and surround. The dynamic range of DSOTM and the others doesn’t eclipse that of a compact disc…nor does the frequency response. In fact, the fidelity of most of the content used in the study is standard definition.
The question is then, what was compared in the Meyer and Moran study? In fact, their results showed that members of their society could not perceive audible differences between a standard resolution SACD (DSD 64) recording of an analog source (or even the new “native DSD tracks) and a CD rendition of the same track. The Meyer and Moran result supports another study I wrote about recently that compared PCM vs. DSD. In that study (also an AES paper), the respondents couldn’t tell the DSD from the PCM.
The belief that DSD is somehow more “analog sounding”, “warmer”, or “more musical” than a PCM version is self delusion and marketing spin used by manufacturers of equipment and download vendors that want you to buy their stuff. It’s as simple as that.
Here’s the bottom line. A standard definition recording made using analog tape (stereo or multichannel) will always be a standard definition recording…it cannot meet or exceed the capabilities of human hearing. Our goal as audio enthusiasts is to get access to the “best” version of our favorite recordings…regardless of format or resolution. If properly transferred to DSD or PCM, the original recording will captured with all of its original fidelity…all of its standard definition fidelity. If you play a high-resolution transfer and compare it to a standard resolution transfer (SACD or CD spec), research has shown that individuals or varying skills and backgrounds cannot tell them apart. Thank you Meyer and Moran for establishing this fact. Pretend as they might, makers of digital equipment using ever-higher sampling rates are not able to improve on the fidelity of these transferred masters.
However, the question of whether music listeners can distinguish between real high-resolution recordings (those originally made using high-resolution capable equipment) and CD spec tracks or even great quality MP3 files is yet to be determined in a rigorous study despite what so-called experts would have you believe. The evidence seems to be on the side of sample rates at 96 or 192 kHz but these are only recent studies and will need to be duplicated.
As I’ve stated previously, I don’t really care whether people can tell the difference between CDs and my HD-Audio files…I know that the fidelity is better at 96 kHz/24-bits than at 44.1 kHz/16-bits. I also believe that someday…hopefully soon…there will be definitive proof.
+++++++++++++++++++
I’m still looking to raise the $3700 needed to fund a booth at the 2015 International CES. I’ve received some very generous contributions but still need to raise additional funds (I’ve received about $3500 so far). Please consider contributing any amount. I write these posts everyday in the hopes that readers will benefit from my network, knowledge and experience. I hope you consider them worth a few dollars. You can get additional information at my post of December 2, 2014. Thanks.
“As I’ve stated previously, I don’t really care whether people can tell the difference between CDs and my HD-Audio files…I know that the fidelity is better at 96 kHz/24-bits than at 44.1 kHz/16-bits. I also believe that someday…hopefully soon…there will be definitive proof. ”
Does that then mean you are ok with advocates of DXD etc believing that some day….hopefully soon… there will be definitive proof that DXD is better?
You seem to spend a lot of effort on this whilst ultimately you admit your views are just “belief” and have no proof . And then don’t actually care whether anyone else hears the difference as long as you can keep your own beliefs.
Much like an argument between different religions about who’s right – a real distraction from getting on with life in a meaningful way for it’s own sake, or in this case a distraction from just making great sounding music HOWEVER you do it! Life is short…
First, DXD is not a format…it’s PCM at excessively high rates. I’ve explained the rationale and company behind it. The research that I’ve done and the people I consult with (Robert Stuart or Meridian, John Siau of Benchmark and others) are knowledgeable about these matters. The best information at the current time is that maybe, just maybe 96 kHz or 192 might be perceptible. There is no case for anything higher than that…other then marketing.
.. but there’s no harm in using it if the converters one records with are capable of using it and that those same converters provide superb quality in the lower rates (so no money wasted there) AND are cost effective because of their other features.
Artists, recording engineers, mastering engineers.. they should all be FREE to choose the format (I call it a format – that you don’t is the semantics of your world versus the semantics of mine, neither of which are correct) that suits them sonically or for efficiency reasons or other reasons.
I simply don’t understand your crusade against people’s freedoms to do what they want. No-one is forced to buy DXD material, or equipment. My future recordings will be available in DXD and people will be free to downsample and listen in any format they wish.
With regard to the rational and company behind it, you seem to have ignored the fact that it was a collaboration with Sony / Philips in order to provide editting facilities for DSD. DSD itself is nothing more than a stream created by the then current 1-bit delta-sigma A/D process in general use without committing to a particular sample rate and there for involving less digital processing/packaging. It was “invented” for in-house archiving of analoge tape, the storing of which is a large financial burden. Why not then offer these newly digitised archives for sale in the exact format they were archived in? And so DSD was born as a consumer format and the need for DXD to edit it.. History and context are absolutely everything !
By the way, did you know that the D/A frequency response of the older DAD AX24 only extends to 40KHz or so by design? I’m sure that’s also true for their newer models, again making the argument of extended frequency range moot.
Lastly, you say “the best information available” with absolutely no irony after having completely dismissed other people’s “best information” e.g. Meyer and Moran’s AES paper. It would seem that “best information” is a euphamism for “things that agree with my belief”. One needs to take ALL information and understand that there is NO best !
Producers use what they know and what their clients want…hit records. They can use cheap recorders, capture an album in a car using a tablet…it’s all good. They use the tools they know and produce recording that please them. I have no axe to grind with that…it certainly not a part of my initiative to expand their knowledge about new formats and techniques. And kudos to you for using high-rate PCM (aka DXD). My friend Morten Lynberg uses it as well and garners Grammy nominations and very positive reviews. But knowing that the format is 90% air and contributes nothing to improved sound is a fact. I write this so that other engineers will be fully informed before spending the money on 352.8 kHz/32-bit PCM audio…96 kHz/24-bits sounds just as good.
Sony and Phillips didn’t develop DXD…they acknowledged it once it was made available by Merging Technologies. The archival system that Sony wanted and needed for their analog tapes is unsuitable as a consumer format (I wrote a few lengthy posts about the history of DSD that trace all of this). Converting to PCM or using analog mixing were the only choices. Offering the co-called native 1-bit streams proved a challenge. 85%^ of the SACD released were either analog transfers or PCM produced…what’s native DSD about that?
You are certainly free to agree with Meyer and Moran…you would not be alone. However, I disagree…audio quality can get better. There may be no absolute “best” but there is accurate information to be considered.
Presumably both the hardware ans software used to convert old analog tapes to digital formats is significantly better than that used in the 80s and even in2007. If one assumes that the best equipment available today can do at least 96/24 (I don’t know if this is a good assumption), there is no reason not to use 96/24 except space and the playback capabilities of the playback device. Would an average (not audiophile) quality DAC reproduce the analog better at 96/24 than 44/16?
BTW, the Pono store still has Neil Young’ s first three solo albums only available at 44/16 – apparently from 2009 remasters. The fourth was mastered for DVD-Audio about a decade ago.
Thanks Vince. The transfers done of older analog masters at the label mastering facilities are 192 kHz/24-bits…which is more than enough to match or exceed the fidelity on the tape. Anyone advocating for 384 or 352.8 kHz/32-bits it playing a numbers game and doesn’t know much about digital audio production.
I should revisit Pono and see how they’re doing. Interesting that you can rediscover the soul of music at CD quality.
Reading the day’s posting and going to the AES online library reading the paper and the Boston Audio Society posting,…. it looks like a more stringent and rigor needs to be done at the next stage of this subject.
Wouldn’t recording a source, simultaneously to the multiple formats in question do the trick? Instead of comparing oranges to grapefruits, albeit some sort of transfer process. This would eliminate generations and maintain the integrity, by being first source for all, rather than a transfer.
It would be a comparison of first source, to first source in a true resolution format comparison. Hope I’m clearly explaining that. If not, I’m looking forward to getting an earful from Mark,… such a good teacher on this subject.
Partev, been there and done that. I was part of a simultaneous recording at Snow Ghost studios a few years ago. We recorded the same mic feeds to analog tape, Sonoma DSD, and 96 kHz/24-bit PCM. The analog tape was the first that we rejected. The DSD and PCM both sounded great…indistinguishable by those present. I opted for the PCM because in looking at the spectra of both the DSD and PCM…the ultrasonics were all about noise in DSD. If they both sounded great and one doesn’t have the noise, PCM is the clear winner.
Doc,
I consider myself a skeptic and have a mind that chooses science over anecdotal information. That being said, I don’t believe everything is explainable-yet. This includes the way humans perceive sound. Please note I said perceive and not hear! I’m postulating that perhaps we process the ultra and sub sonic information in entirely different ways. Further that “inaudible” data either hits our “sweet spots” resulting in pleasure, or for whatever reasons misses the mark causing annoyance. I have transferred the test to my computer and plan to assemble some pairs of ears (trained listeners & not) of different ages to see whether anyone reacts favorably to a certain track. I’ve not previewed them in whole, and only made sure I captured the tracks. What I’d like is your opinion, Mark. I’m thinking of bypassing “why”. I’m going to remove the methodology and let these folks listen to them however they’d like, as many times as they like. I’m trying to come up with a way to BS (temporarily) my guests to remove any emotional bias from it trying to keep things objective. I realize this also isn’t a scientific method with many flies in the ointment. Some will likely not even care about the quality, some prefer a smaller file size for convenience according to some answers I’ve got in the past from these listeners. They would never spend >$200 for a pair of cans unless it’s a fashion statement. These folks are from all walks of life & hope it has some result, even if unusable as a study. Any suggestions are welcome!
Thanks for all you do!
I’ll be interested in hearing about the results. Keep me posted.
1} The resolution of a sound recording depends solely on timing transient response and has nothing to do with frequency range although unconciously one might hear 100 kHz or even higher… Thus, a MP3 track upsampled to 192 kHz will have exactly the same resolution {that is ‘detail’} as a HD track recorded at 192 kHz sampling rate. Hence, SACD at its ~2.8—11.2 MHz should sound obviously livelier than, say, 384 kHz high-definition.
2} Fidelity of digital sound is heavily affected by quantization error, Gibbs phenomenon, filter’s issues & a plenty of nonlinearities {including that of listening at >40dB volumes} which have analog nature & hardly defined by transistor characteristics. So, if a sound is to be recorded using binary numeral system, then highest sampling rates would be of very great importance to reduce quantization noise as much. Meanwhile, recording at a too high sampling rate may be unnecessary just due to upsampling that can make it better, but only if decimal numerals are used.
3} To my ears, the aural difference between a CD track & a MP3 one is substantional. {My speakers are strictly full-range paper, albeit stereo that imparts some phase trouble as when a track has been recorded from a couple of mikes.}
4} Bad audio hardware does not affect the quality of comparison. Moreover, due to timing transients being preserved, a 96 kHz track played through a 48 kHz DAC sounds practically same as when through a 96 kHz DAC.
I can’t say I agree with your points but thanks for the comments.
Jay,
In response to your points
1) Unintelligible pseudoscience mumbo jumbo. Exactly of the kind we have been trying to negate in these pages lately. Scientists know that ‘timing’ translates to frequency. You appear to understand nothing about how sampling and digital signal processing works. Nada.
2) In respect of quantization errors you should have heard of dithering by now (2014). As for the rest, the same comment as above applies
3) Sorry I’m missing the point that you’re making here
4) See 1. above
Sorry if I sound just a bit rude, that was actually my intention since I’m fed up with all this stuff.