MQA Ltd.

Commenting during the Munich Show, Bob Stuart said: “We’re delighted with the continued and overwhelming interest in MQA throughout the industry and the reactions at Munich HIGH END have been extremely positive and highly encouraging.” Bob Stuart also observed “MQA is a philosophy and establishing MQA Ltd. as a separate ongoing business allows us to advance and progress all the activities involved in setting up this exciting new approach. Major announcements are planned in the build up to IFA in September at what will be a very significant event for MQA.”
I attended the MQA presentation and also had an opportunity to speak to Bob Stuart afterwards and I still don't believe I have a firm grasp on the full MQA picture. Yes, I can be rather dense.

First off, the MQA demo sounded just great. No doubt about it even though there were no A/B comparisons with non-MQA files. Where my head begins to swim is on the real-world implementation side. But let's talk for a minute about what MQA (Master Quality Authenticated) is as presented by my Cliff Notes version from a bird's eye view seen through a glass darkly.

"MQA is a fundamentally new way of looking at digital audio." In essence, Bob Stuart et al have found that we humans are very sensitive to aberrations in the time domain, much more so than previously believed. Further this sensitivity cannot be adequately addressed within a 16-bit/44.1kHz container. Higher resolutions are needed to capture and reproduce the original timing information contained in recordings. Analog originals, no problemo just capture them at the highest digital resolution available, ideally using an MQA A/D converter. Digital recordings or already A/D'ed analog recordings present a potential problemo since the A/D conversion process introduces anomalies we can hear, according to MQA.

So the ideal MQA story begins in the recording studio where the more information gathered about the specific equipment used, including and most importantly the A/D converter, the more the MQA technology can correct for the sonic anomalies found in these devices. Even ideally-er is for the A/D conversion to happen inside an MQA converter. Another technically elegant aspect of the MQA process is these rather large files are then compressed, for all intents and purposes losslessly according to existing neuroscience-based theory, into much smaller containers roughly the size of a 16/48 file, ready to be sent to your MQA-enabled device(s).

On the playback side, MQA can be implemented in hardware or software but it is essential that MQA known the DAC it is to communicate with. One real-world example would be an MQA app for your iOS or Android device. Another would be a third-party device with a DAC where MQA is embedded in the hardware. The "A" in MQA, remember, stands for "Authenticated" so any third party solution must show when an MQA file is being played. Authenticated.

While I totally get the value proposition for MQA with streaming services like Tidal so that we stream high-res content with practically no additional bandwidth requirements, as well as the overall notion of better capturing and playing back what was originally recorded for a more natural and musically involving experience, it seems to me that MQA has mountains to climb on both ends of the ideal equation. That's not a criticism of the technology which is clearly well over my head and tickles many a knowledgeable engineer's fancy. I simply wonder at the practicality of the need to re-encode every recording for it to meet the MQA standard.

MQA Ltd. have announced they are "in active discussions with over 100 potential business partners" so they are certainly hard at work on the hardware side. I have not read which record labels they have deals with which is obviously the other important side of the MQA coin.

I told you, I'm confused.

I also got to listen to some MQA files and their MP3 equivalents on the above system using the MQA-enabled Meridian Explorer DAC where my confusion turned to musical enjoyment.

COMMENTS
BradleyP's picture

"I also got to listen to some MQA files and their MP3 equivalents on the above system using the MQA-enabled Meridian Explorer DAC where my confusion turned to musical enjoyment."

Why not Redbook vs. MQA, or 96/24 vs. MQA? That would have been a more scientific comparison. mp3 vs. MQA makes me wonder if there is much of a difference. I really, really want the technology to be a big step forward and will be saving my DAC dollars in the event it is.

Venere 2's picture

I really enjoy reading about the supposed research the MQA people have done. I a find it very interesting that they prioritize timing, because based on their research, human beings are extremely sensitive to timing. Timing is probably the most important aspect of music (others have also said this). As MQA says:

"MQA is a fundamentally new way of looking at digital audio." In essence, Bob Stuart et al have found that we humans are very sensitive to aberrations in the time domain, much more so than previously believed. Further this sensitivity cannot be adequately addressed within a 16-bit/44.1kHz container. Higher resolutions are needed to capture and reproduce the original timing information contained in recordings."

Ok, here is where the MQA people lose me and I am compelled to call bullshit. Timing is not affected by bit depth and sample rate. Or, it is not affected to a point that humans can discern. If it was, 128 kbps mp3 files would sound off tempo. Crummy mp3 files have a lot of deficiencies, but timing is not one of them.

Michael Lavorgna's picture
...John Atkinson's "I've Heard the Future of Streaming: Meridian's MQA" for more on this subject.
Venere 2's picture

To add to my previous post. Even if mp3 files had bad timing, there is a big difference between an mp3 file and a CD quality 16 bit 44khz file. Also, a 128kbps 16/44 file will have problems/deficiencies a CD quality 1411kbps 16/44 file will not.

MQA's statement is far too vague. To simply state that 16 bit 44khz files cannot accurately render timing is a blanket statement of rubbish.

alexandrov's picture

I've spent over half an hour with these laptops playing files in mp3 and MQA and... if there was difference, it was barely audible. I couldn't recognize MQA. I'm not impressed at all. But these were 128kbps mp3s! If MQA differs so slightly from 128kbps mp3, it doesn't worth all the buzz.
I attended the MQA presentation too and there were no comparisons between MQA and other formats. All we expected that. And the sound was mediocre, I'm sure all of us in this room had better sound at home than these $$$$$ meridian speakers with MQA files.
My impressions are far below the expectations.

labjr's picture

I believe that DSD has an advantage over even 192Khz PCM because the high sample rate allows more accurate recovery of the timing of transients which is more accurate than the math done to recover the information from PCM data. I think timing is very important and can not be easily captured and recovered accurately the way it was being done. Everyone always argues about the Nyquist frequency and how 16/44 is good enough because we can't hear above 22khz..blah,blah, blah. But all the math in the world to extrapolate the analog signal out of the data stream can't fix the timing problem if it's not done accurately. Maybe MQA finally solves this problem?

tresaino's picture

My disappointment with the MQA demo in Munich was huge, for a number of reasons:
(1) I had high expectations because of John Atkinson's and Jason Victor Serinus' enthusiastic articles in Stereophile last Fall.
(2) no A/B comparisons. Something sounds only good as long as you don't hear something better. A friend once tried to convince us how good his digital high resolution system sounded, and we were indeed impressed .. but only until he put the same songs on his old Garrard 301 turntable. No criticism of high-rez digital audio intended.
(3) the silly arguments provided by the presenter, I paraphrase: "MQA finally allows you to hear these old analog masters again, the way they were intended to sound. This would not be possible otherwise." Come on..
(4) the vagueness of the presenter's response, when I asked why, some 6 to 8 months after we read about MQA, there were still no major deals with the recording industry. Considering how fast things went with TIDAL, the time elapsed after the MQA announcements is enormous.
(5) the sound was simply not convincing. It may have had to do with the setup and the room, but sonics were far from glorious.

In sum, disappointing, am not convinced.

Best,
Roberto

NeutronStar's picture

In spite of the raving article by JA whose technical opinion I respect a lot I still do not digest the MQA idea. Encoding ultrasonic frequencies in the audible band at inaudible level... Does not close mathematically and violates known information theory theorems.
I see another "perfect sound forever"situation.
I remember the early 80' when CD just appeared. Every audio critic at that time endorsed and raved about the sound of CD.
Better stick to as many bits and samples as possible and always 88/24 or better, no lossy compression no information loss.

X