|
Critic's Corner Discuss a review. Provide constructive feedback. Talk to the industry. |
|
|
|
Critic's Corner Discuss a review. Provide constructive feedback. Talk to the industry. |
|
In Reply to: RE: It's never been done - ML was the closest posted by Michael Lavorgna on August 24, 2017 at 06:33:30:
Hello Michael,
Beetlemania had asked if there had ever been any level-matched comparisons. I told him that all of the ones that I knew of had flaws, but that yours were the best.
Ideally, one would be able to do the comparisons using music one is familiar with through a system one was familiar with in a listening room one is familiar with. That is by far the easiest way to achieve valid results, and one of the reasons that Stereophile stopped making comments on the sound at show reports decades ago (although it hasn't slowed down TAS one bit).
In my opinion the biggest flaw in your tests is as you note in your reply: "there were not many MQA-ready DACs". You want on to say "I wanted to focus on a group of affordable products", but that is something of a red herring, as those three were the *only* MQA DACs available then, regardless of price. So you had no choice but to listen to 3 relatively low-cost, unfamiliar DACs. Which makes for a less-than-ideal comparison environment.
And this tripped you up in at least one place. On The Doors track "Riders on the Storm", Jim Morrison's lead vocal was double-tracked. This is normally done to make the vocal sound "fat" and "full" and is found (for example) on about 80% of all Beatles tracks - to the point where they had their engineers invent a way to do ADT (Automatic Double Tracking) so they wouldn't have to take the time to do it themselves.
However for "Riders on the Storm", they chose to have the overdub be *whispered*. This is well documented in places like Wikipedia to the first-person accounts from the band members and engineer/producer Bruce Botnick. You can even hear an isolated version of the "whisper" overdub at the link below.
I'm sure if you listen to the original high-res 96/24 version on your "big rig", you will easily hear both vocal tracks and be able to mentally "pull them apart" and hear them as two separate events. However when listening to lower performance DACs apparently you mistook the "whisper" as "edge". The loss in resolution from the MQA process (about 17 bits broadband, but even less resolution above 3 kHz where the "whisper" has the most energy) makes the "whisper" difficult to hear at all, and nearly impossible to make out as a separate event. Your unfamiliarity with those DACs apparently led you to the conclusion that the essentially "whisperless" MQA version was "more dimensional and pleasantly softer".
The problem is that MQA processing is *less accurate*
Above you wrote "I've never heard MQA processing make music sound worse." I have, and so has Rick Rubin, famous producer and (I think still) co-president of Sony Music. Michael Fremer attended the LA Audio Show with Rick and at the MQA demo, Rubin proclaimed the MQA version to sound either "synthetic" or "overly processed" (I can't recall which). In many of the MQA-processed tracks I've heard, there have been inexplicable changes in the tonal balance - usually the 100 to 200Hz octave is elevated, giving the MQA track an extra "warmth" and "richness". This can even be seen in the spectra published by JA:
https://www.stereophile.com/content/inside-mqa
Look at figure 2 where the spectral content of the entire track is shown. Unfortunately the blue trace does not appear on the low-res web graphic, but according to the text, the grey and red are the right channels for the MQA and original files. Look at it very closely and you will see many places where the tonal balance has shifted, the octave from 100 to 200Hz being just one. The point is that we have no idea what MQA is really doing to their files. We know that in some instances that they start with different masters. We know that in some instances the levels are boosted by as much as 2 dB.
We know that in the self-recorded file JA sent to BS that the main ADC had no time smear to correct so that "benefit" is completely erroneous. We know that BS asked for the impulse response of the entire recording chain including the microphones, so clearly he "corrected" for the "time blur" of the microphones (otherwise, why even ask for this information?). We also know that JAs master comprised 6 channels mixed down to 2 for the final release. JA sent all six raw channels to BS, so for all we know, the MQA'd version was a complete remix of JA's recording.
So here we are, nearly 3 years in to MQA with tons of hype, yet very little real understanding and essentially no controlled testing. As noted before you came the closest with your testing, yet still made some errors. Now that software decoding is available, you *might* have a chance to run a true fair test. There is still a caveat however - you would need to change your software player from Roon to the latest Audirvana which has MQA decoding.
Once you have spent enough time with Audirvana and are used to the sound, you can now compare the original high-res files you have with the MQA'd versions, decoded through Audirvana. While this only does the first unfold and doesn't "compensate" for the digital filter in the DAC, in your case there is no digital filter to "correct".
Again, this would be the most fair and valid comparison to date. Again, we have no idea what MQA does to the files inside of its "black box". Ideally you would not only listen to them but submit them to some technically oriented person who could at the very least compare the dynamic range, level, spectral balance, and so forth to make sure that the MQA "improvements" are not just the result of typical tricks used by mastering engineers to make the track sound "better".
The ball's in your court, and I say that with the utmost respect, because as I have noted your listening tests have been the most fair I've seen to date. I'm just asking you to make them even more fair and to also submit the files to someone for independent analysis.
Cheers,
Charles Hansen
This post is made possible by the generous support of people like you and our sponsors:
Follow Ups
- RE: It's never been done - ML was the closest - Charles Hansen 14:33:18 08/24/17 (2)
- RE: It's never been done - ML was the closest - Michael Lavorgna 15:23:56 08/24/17 (1)
- RE: It's never been done - ML was the closest - Charles Hansen 23:14:26 08/24/17 (0)
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2024, Audio Asylum, all rights reserved ]
