Reference

Would someone summarize these results in 25 words or less of regular-speak? Thanks!

It is now widely accepted by 0.000000000000000001% of the human population that "hi-rez" audio delivers improved sound quality.
New math perhaps? ;-).
In my universe, SACD and DVD-A were abject failures in the marketplace. Perhaps elsewhere?

Where's the music?

I'm all for anything that will help music sound better, but asking people to get excited about MQA is like asking us to get excited about flying cars. Let me know when they get here.

Can you please stop saying that sound in the air is analog. It is not. Once the signal leaves the loudspeakers it becomes acoustical energy. When we talk we do not make analog sound. Again, it is energy. It is not analog nor digital, these terms relate to the way sound is captured, stored and reproduced.

The math could all be correct, and logical, but the real proof is in listening. I have listened to a couple of tracks from 2L and been unable to tell a difference in my setup, but I have no idea about the provenance. Ideally, an MQA track is recorded in MQA to begin with, or converted knowing the original ADC system. Converting a pre-existing digital track to MQA is not ideal.

Are there some publicly available tracks to listen to MQA recordings in ideal situations, vs. non-MQA that will make me jump for joy?

Best,

Erik

I'm glad this is here, but it really needs curating. The material is better suited for a wiki than an article, but like I said, I'm happy the data is up there for us to examine au plain aire.

Best,

Erik

I've been waiting almost two years for MQA to come to Tidal. When??

Thank you Mr. Stuart et al. for the details around the rationale for MQA as well as some of the technical capabilities.

There is obviously much in the text to parse and understand. However whether the reader agrees that the effects and theory of temporal blurring being discussed is of significance, can practically be accurately "fixed" (especially in complex studio mixes), and the algorithm used by MQA will result in audible benefit is unclear to many at this time. Thank you as well for discussing the noise floor effects; it's clear from the cases demonstrated that MQA is able to encode down to the noise level needed for the music. I know that much of this can be argued back and forth ad nauseum in terms of whether music should be kept as pure PCM and the benefits of encoding as per MQA's psychoacoustic modeling. However I think it is fair to say that whatever digital bit-depth and samplerate is used (certainly beyond the usual 16/44), we are discussing digital representations already close to auditory thresholds (which is my interpretation of the 52.3% accuracy rate in the hi-res vs. standard resolution meta-analysis).

However, in the time domain which MQA primarily focused on as the mechanism by which higher fidelity can be achieved, I must voice my concern around this comment:

"It's useful to know the original sampling rate of the mastering process, because that tells us about the first part of the chain. The DAC is equally important; a chain is as strong as the weakest link. More important, unless the encoder (A/D or mastering) and decoder (plus D/A) processes are complementary, it isn't possible to reach the final result and certainly not at low data rates..."

Why should we focus on the ADC/DAC as significant "weak links" in the overall audio chain (much less the "weakest link")? At least for the consumer end, is it not true that as almost always, the weakest links are the transducers (speakers) and room effects? Compared to typical speaker time-domain variations (step-response variation for example) on the order of multiple milliseconds, it is hard to imagine the significance of comparatively small changes at the level of the DAC including the occasional pre-ringing from poorly low-passed signal through a linear phase filter.

Without a truly "end to end" solution which terminates not at the level of the consumer's DAC, but the actual sound energy that reaches the listener, MQA alone IMO cannot truly achieve the goal of time-domain accuracy for the system. Furthermore, the closed hardware implementation which at this point does not allow access to the decoded digital audio data prevents the more sophisticated audiophile from implementing his/her own optimizations such as DSP room correction to suite the listening environment in an optimally transparent fashion.

Ultimately, in my opinion, if MQA is to suggest that the digital processing algorithm being implemented is an overall improvement, it should be allowing more A/B comparisons in public venues and taking up the offers such as that by Mark Waldrep to aid in "Real-World Comparisons" with actual hi-resolution content. Likely even more beneficial would be releasing high resolution files demonstrating a typical properly dithered 16/44 audio file upsampled to 24/192 (with say a standard linear filter) and the decoded output from an MQA file to 24/192 with the same mastering. This will provide an opportunity for the public to audition the effect for themselves in their own home with existing high-resolution DACs to at least provide an idea of what difference the algorithm can make. (Perhaps more appropriate would be 24/44 vs. MQA as they would have similar bitrates.)

I don't even know where to start. Besides sidestepping the whole DRM issue these articles are a new low not only for Bob Stuart but for audio journalism in itself.

Just one example how dishonestly Stuart answers questions: it "is understood that a digital distribution system (including MQA) can be lossless in distribution and therefore requires lossless delivery."

Well the question was if MQA is a lossless FORMAT. But Mr. Stuart sidesteps this question by speaking not about formats but about "lossless delivery."

He then continues to essentially argue that MQA is more lossless than lossless by being lossy.

There are plenty more gems like this in the marketing material spread out here. I certainly hope that Stereophile gets some serious experts and engineers to fact-check Stuarts claims. That is what good journalism is about and what the audiophile community deserves.

Until this happens I'd venture to call Bob Stuart the Donald Trump of the Audio world.

Absent from Mr. Stuart's rather abstruse and lengthy essay are answers to the issues raised in Archimago's first comment ("Thank you for the series") on this thread. I don't think it's the first time they've been raised. They appear to me to be germane, and whether enthusiastic, skeptical or merely curious about MQA, I believe readers would benefit from a clear and direct response from Mr. Stuart.

All of these downloading/streaming articles make me wonder what the automotive equivalent might be.

Wouldn't be fascinating that I would have to wade through such opaque dissertations to know where to fill my car with gas, where to change the oil, where, perchance, to have my factory transmission replaced for some other contrarian engineering standard which might have exceeded that which came with my car?

All of these "dueling" standards over which standard to stream and download leave me glad for my CD purchases. CDs may be inferior, but it's a reliable and tradable format. If I don't like a CD, I sell it; there's a ready market for such misfires in my purchases. I don't know where I'd sell my Studebaker-like MQA, MP3, WAV, FLAC, AIFF, ALAC, IRAQ... files on my Seagate hard-drive, and soon-to-be-obsolete tablet menu software.

Let these interested parties in the downloading/streaming world debate what appear to be varying versions of the Confederate Dollar (no offense to audiophiles south of the Mason-Dixon line). This technological hair-splitting (all to hear songs that only an even smaller audience will like) puts the audiophile pursuit--already a marginal endeavor--even further to the outer reaches of the Solar System, Kuiper Belt, or Oort Cloud, of ever more constrained disposable income, and dwindling patience of we sympathizers of the hobby.

Keep us posted when the doctoral dissertation critiques, and standards of the week, settle on something that many of us would bother to purchase. Keep us posted when the "latest and greatest" will not be the"has-been standard" of the same forum of audio enthusiasts two years hence. VHS -vs- Betamax, Toshiba HD -vs- BluRay was nothing compared to this audio Ecumenical Council.

I look forward to the latest CD releases and recommendations from Stereophile, Amazon, Arkivmusc, Allmusic, etc. while the academic debate on the competing "internal combustion engine designs" rages on elsewhere.

When you have decided if the "new car" should have 4 or 6 or 8 or 3 wheels, let us know. When you know if the steering wheel will be on the left, right, or center, let us know, When you know, whether the "new car" runs on diesel, gasoline, kerosene, coal, wood chips, cow chips, sugar cane chaff, Uranium, moonshine, corncobs, let the rest of us know.

Until the next article, we all wait.

I skipped over the math, but it being publicly asserted, I accept that it will hold up and the "high res container" concept will prove itself. Crypto people were talking about steganography back in the 90's before the War on Terror, and now we know that it works well enough that our intel agencies depend more on the old-fashioned ear to the ground and social engineering than they do being able to crack crypto codes. I don't see any reason why MQA licensing should be more intrusive than Dolby or any of the hundreds of license holders you see in the iPhone 'About/Legal' screens.

It is most hard for your folks or at Meridian to argue with folks who live in a country that ranks 28th? in the world in Math Ability. A sad state of affairs for sure. The Math presented here IS most difficult and as we are learning there is much going on in digital audio that is not fully fully understood even yet, and more will be uncovered in the future. Think back to when the issue of jitter was first present and that is easily understood now and what an important issue it was and is, but is only a small part of the digital process, and differences audibly discernable.

When you sit in middle school math classes as a teacher like I do every day with students who can't add, subtract, or multiply with any proficiency I have problems of explaining things mathematically. I taught high school math for 6 years and came back to MS to try and fix was has gone horribly wrong in our math world. I have a wonderful lady down the hall with a masters from Georgia Tech teaching the brighter math students to take them to new heights. I have more job security. lol She often asks me how I can do what I do.

Even forgetting the very high level math going on, it always comes down to the bottom line of whether you can HEAR the improvements or not and all too many trained listeners can and have, and reported it here. No one can say that they doubt the finding just because they don't understand the math, which is incredibly foolish and only exposes yourself as one who disavows what they don't understand. It is alright to not understand all the math going on and how it relates to what we hear. The folks at Meridian are doing that for us.

The other issue, and the main one to me, is our own personal hearing flaws and differences that has become more apparent to me in my quest to find some great headphones that work for ME and MY hearing deficiencies, which are great at 69. Accurate to whom? My hearing is not the same as JA's, Tylls', or anyone else, or Bob Stuart's. What changes in frequency response I need will not work for anyone else. I can tell you that what is accurate or pleasing to them does not work for me and my HF loss. So if you, or anyone else is trying to discern the audibility of 16/44.1, 24/96, 24/192, or DSD...on what are you listening, where it is set up and does it match up to YOU and your hearing accuracy (loss)? Is your room a sonic mess? What is bright to some would not be bright to me, I can assure you, and so my level of discernment of "BETTER" will not be the same for anyone at Stereophile or anyone with near perfect hearing, whatever that is. I have learned my lesson with my Focal Spirit Pro headphones and their reduced HF energy that make then sound so dull to me. Buying them off a great review was a mistake as the reviewer and I do not HEAR the same, but they can easily work for someone with no hearing loss.

If all, or most of the folks who have attended the listening sessions for MQA can hear an improvement of what MQA can do ( and I do believe them to be honest folks, by the way) why should I doubt it even if I don't understand all the math (I don't by the way) if the result IS an "audible improvement"? If even 4 out of 5 agree, 80% does win in my math world. If the 1 out of 5 does not have near perfect hearing why would I trust his opinion anyway? I might not even know that they are not a music lover or even care about the music selected for the audition which WOULD matter as it is one of the parameters (variables) of the trial that can't be discounted as a factor. Everything matters even a + .1db change in loudness often has one preferring the "louder" as the better which is often not the case.

It is sad to read some of the comments here. I do not believe that one can present all that is going on with MQA that can be presented in a simple way, as it is not simple, but for those like me, all that matters is can I hear the improvement that all their hard work has achieved? If I can't, my only decision is to become a customer or not. My choice. It does not discount their work or their achievement. I must accept the humbling fact that I am the problem if I can't hear the improvement, too often a hard thing for people to do. Hearing small improvements is hard work, but after reading the comments of those hearing the demonstrations of the trials, these are NOT small improvements.

For what it's worth, I think it's not an unreasonable expectation that editors and columnists in a professional publication model respectful and professional civil debate on its comment boards. Perhaps Mr. Atkinson, whom I think does, might have a gentle word with some of his colleagues?

Whew! It's nice to see all the scientific background about MQA, and interesting to read the objections posed. But it seems to me there are two, and only two, questions that need to be resolved about MQA:

1. Does MQA sound as good or better than the best other processes for making music recordings? In other words, does it sound as good or better than DSD or high-resolution PCM recordings?

and

2. Are files produced by the MQA recording process significantly smaller than other formats?

If the answer to both of those questions is Yes, then MQA is good and should be widely adopted. If the answer to only the first question is Yes, MQA is good but whether it should be adopted depends on how much better it sounds than other recording processes. If the answer to the first question is No, then why bother with a process that sounds worse than what we already have?

Vade Forrester
Reviewer, The Absolute Sound and SoundStage! Network

“If you can’t hear it it’s all academic.” Paul Klipsch

Taken one way, I have nine albums that I use to test. Until Sony and Universal are MQA licenses and release six albums they own the rights to I can’t test MQA. Warner owns the rights to three albums so I wonder when and where I can download them. So does Warner in their latest SEC filing one of their business risks is “our dependence on a limited number of digital music services for the online sale of our music recordings and their ability to significantly influence the pricing structure for online music stores.” I can’t buy the music I want as MQA files so for me it is academic until I can.

Taken a second way I listened to MQA files in Meridian’s room at T.H.E. Show this year. Michael Lavorgna posted his thoughts about the Doors “Riders on the Storm” on MQA May 26, 2016. He found the rain more realistic than his hi-res version. I was allowed an A-B test with “Riders on the Storm.” Michael and I agree the rain is more “realistic” on the MQA file but the sound of rain produced by a Fender Rhodes is more like how Michael described it “akin to glass beads hitting plexiglass” on the hi-res file so the MQA rain played on the Rhodes is not as accurate. This is a small difference but it will require more testing. And I was listening in a hotel room to unfamiliar equipment so I can’t make any conclusions until hear MQA in familiar surroundings and with familiar equipment. So again academic until I compare file formats to see if I can hear any differences.

One comment, don’t forget there are a lot of hi-res files that sound worse than their CD counterparts.

One question what is the difference between “white glove” MQA files and those converted by automated processes?

Thank you, John and Michael, for posting this wonderfully clear and insightful explanation of so many interesting topics that are not easily understood except by experts in this field. It's answered lots of perplexing questions for me on the first reading, and will probably answer more as I go back over it. John, you and your team are setting the standard in the industry for the work you're doing, month by month. Your readers have a debt of gratitude to each of you for that.

The obsession with DRM arises, I suspect, from an information-technology mentality--specifically from people who believe that "information ought to be free." But there are problems with this point of view--very serious problems.

1. I think some people need to think a bit more about the issue of authentication. I suspect that many of the people who shout "DRM! DRM!" are the same people who complain when they download a high-res file from a prominent high-res music site only to find that it's upsampled CD quality. Beyond iTunes and iPods, I avoided the world of servers and digital audio files for years, for precisely this reason: There were (are?) no reliable sources of high-quality audio files. (I finally built my own server, and ripped all my CDs, this past winter.)

Who wants to pay 24 bucks for a file of which the only difference from CD-resolution is that it takes far longer to download?

We need authentication. Digital files are easy to manipulate--especially when they're in an open format! When we buy something, we need to have a reason to have confidence in what we're buying. And while some of us may enjoy that kind of thing, we shouldn't have to analyze every file we download to make sure it has the spectral content we paid for. We need authentication.

2. I'm guessing that in their other lives (the lives in which, presumably, they are willing to admit to being who they are), pseudonymous critics (talk about authentication issues!) probably like to get paid for their work. They deserve to, and so do musicians. It's no secret that the music industry is far less profitable than it once was--that many artists struggle to get paid for the music they create. The reasons are complex, but they boil down to: The Internet. Free information.

MQA is not DRM. It's just obviously not, in any reasonable or conventional interpretation of that phrase. But--I am speaking for myself only--it clearly IS an attempt to gain a certain amount of leverage over the digital distribution chain: You can do whatever you want with an MQA file, but you can't play it back at optimal quality without an appropriate tool. I enjoy recorded music enough that I'm willing to pay if it means that more artists will be able to make a living. Any new approach that has the potential to help the industry regain a bit of its former health, I'm OK with.

Jim Austin, Contributing Editor
Stereophile

and I am also not worried about free as I believe that artists should be compensated for their work and all too many are not. We have been around this block all too many times. I want artists whose work I admire and want to hear over and over to be bought by me and others so they can make a living doing what they love to do. I would not spin my 3 turntables almost daily if convenient was my thing, but there are now many lps that I don't listen to much anymore as their sound quality is lacking either from the recording or the pressing and I don't agonize over which is the problem as I do care about my time and most of mine is in the bottom of my hour glass.

So, eClassical, Linn, Soundkeeper, BlueCoast, Primephonic, and HD Tracks will continue to get some of my cash as they are about quality. I even have a redbook release from Primephonic that is a superb recording and better than many a redbook CD I own, so quality can be done in redbook somebody just needs to care.

It will be fun to watch how this all plays out and if the masses ever start caring about quality over convenience and free. I'm not going to hold my breath. It will be fun to see what new hardware comes down the pike as well. But remembering that DVD-A and SACD players were not crazy expensive and still did not get the masses on board, sadly.

The answer you gave to the criticism that MQA is lossy is to in fact describe how it's lossy in a good way, hence not lossy. What?

You bent over backwards to imply it's lossless, but that's just word play. MQA is WITHOUT A DOUBT lossy. Can you restore the full 44.1khz 16 bit redbook bit 4 bit? No. Then by definition it's lossy. End of story.

Pulling out what is perceived to be noise while dithering down to 13 bits, then filling the freed up space with extended information to be used elsewhere is a lossy modification that not only looses bit 4 bit accuracy, but is vastly different from the original 16 bit file, hence is not even remotely lossless.

All lossy formats like MP3 throw out what is believed to be inaudible data, which is exactly what you're doing. Saying "MQA's noisefloor is stable throughout a song." or "MQA system can reach..23bits" is just double talk. Within the redbook 44.1khz 16bit stereo container you're irreversibly discarding data believed to be noise (what ALL lossy codecs do to some extent) and filling it up with signal data. It doesn't matter if the noise floor level is the same. It's now different, sounds different and some discernible signals within the noise are now lost. Again, playback of MQA CDs on legacy players isn't even remotely lossless, let alone bit 4 bit recoverable.

Why not just use the same technique to create a new audio CD standard with a 91 vs 74 minute runtime. That is, instead of filling up the freed up bits with unfolded data that has no functional impact on sound quality on legacy player, just discard it and create a new CD standard that is just as good as 16 bit CDs (lossless, or bit 4 bit the same) but with 91 vs 74 minutes of runtime? Of course nobody would make such as standard because after dithering to 13 bits and discarding the noise it's not even remotely the same, let alone losslessly bit 4 bit recoverable.