MQA: Aliasing, B-Splines, Centers of Gravity

The right thing at the wrong time is the wrong thing.—Joshua Harris

The sampling theory formulated by Claude Shannon in the late 1940s had a key requirement: The signal to be sampled must be band-limited—that is, it must have an absolute upper-frequency limit. With that single constraint, Shannon's work yields a remarkable result: If you sample at twice that rate—two samples per period for the highest frequency the signal contains—you can reproduce that signal perfectly. Perfectly. That result set the foundation for digital audio, right up to the present. Cue the music.

However, in the 69 years since Shannon published "Communication in the Presence of Noise" in the Proceedings of the Institute of Radio Engineers (footnote 1), sampling theory has moved on. The new work began almost immediately, carried out by mathematicians and math-fluent engineers; indeed, some of it had already been done when that paper was first published, in January 1949. But even as Shannon's work was embraced by the digital audio community—in audio, Shannon sampling theory is the foundation for almost everything digital—that post-Shannon work remained hidden. As recently as the 1990s, when post-Shannon sampling was applied to digital signal processing, the focus was almost entirely on visual information—imaging. With just a handful of exceptions, the audio world was oblivious.

Post-Shannon sampling theory relaxes Shannon's requirement that a signal to be sampled—eg, a recording of music—be band-limited to half the sample rate. Relaxing that constraint restores the symmetry between the time and frequency domains that was missing from Shannon's theory. In the newer theory, it's fine to use an antialiasing filter, but it's not required. Post-Shannon sampling accepts aliasing as a matter of course while allowing its impact to be minimized in both the time and frequency domains.

Some years ago, Bob Stuart and Peter Craven—the creators of Master Quality Authenticated, or MQA—began exploring some similar ideas in an audio context. Their first article hinting at the technology that eventually became MQA (footnote 2) referred to some post-Shannon work, and in some later writings—including "MQA: Questions and Answers," published on the Stereophile website in August 2016—the references are fairly explicit. I began to wonder: Is MQA a rigorous application of post-Shannon sampling theory?

Considering the newer theory's relaxation of Shannon's absolute prohibition against aliasing, this seemed a reasonable entry point for a somewhat technical interview with MQA's Bob Stuart on aliasing and its effects in MQA. The interview, carried out mainly by e-mail, is presented here in slightly compressed, lightly edited, occasionally annotated form.

Jim Austin: Is MQA an application of post-Shannon sampling theory to audio coding?

Bob Stuart: Yes, very definitely, it is!

Austin: Is it the first such application?

Stuart: Yes. So far as we know, this is the first.

Austin: Is it a rigorous application of post-Shannon sampling theory?

Stuart: Yes, we believe so. MQA stands on a firm basis which synthesized intuition of desirable characteristics, the mathematics of sampling and reconstruction based on B-splines, losslessly reversible processing (used in flattening)—and was informed by empirical observations, auditory modeling, and hundreds of experiments.

Austin: MQA's critics have often focused on aliasing. In a patent application covering MQA technology, you claimed the invention of "a system . . . wherein . . . the asymmetric component of response of the decimation filter is characterized by an attenuation of at least 32dB at frequencies that would alias to the range 0–7kHz on decimation [footnote 3]. Is that specification—attenuated by at least 32dB at frequencies that would alias to the range 0–7kHz on decimation—realized in MQA's implementation?

Stuart: In all cases the aliasing heard by a listener with an MQA decoder will be well below that implied in the quoted patent claim and will be, we claim, either inaudible or nonexistent.

[To test this claim, beginning with a FLAC file containing white noise at –10dBFS peak, I pasted in 20 seconds from Talking Heads' "Girlfriend Is Better" at the 20s mark, 50dB below the noise, repeating the music and increasing the level by 10dB every 20 seconds. I can detect very faint drums at the one-minute mark, 30dB below the noise level and at least 30dB louder than the aliased-content level allowed by the MQA specification. By 20dB below the noise—40dB above the spec—I could hear the music clearly. My conclusion: The specified level of aliasing is not audible, with a safe safety margin. The use of white noise instead of a 1/f, music-like signal makes this a very conservative test.

Your mileage may vary, so do the test yourself with the embedded audio file. At first you will just hear the noise, but you will then start to hear the music beneath the level of the noise.]



Footnote 1: Proceedings of the IRE, January 1949, Vol.37 No.1, pp.10–21. Reprinted in Proceedings of the IEEE, February 1998, Vol.86 No.2, pp.447–45. See https://web.archive.org/web/20100208112344/http://www.stanford.edu/class/ee104/shannonpaper.pdf.

Footnote 2: J. Robert Stuart and Peter Craven, "A Hierarchical Approach to Archiving and Distribution," AES Paper 9178 (8 October 2014).

Footnote 3: Note that the claim is that aliased content will be attenuated by 32dB, not that it will be 32dB below the regular audio in the specified range. At CD sampling rates, at which aliasing would typically be strongest, aliasing at the top of this range will be reflected down from about 36kHz. In that frequency range, the music in an audio file is already very low in level. I estimate that if this specification is met, the aliased content will be at least 60dB below the musical information at 7kHz.—Jim Austin

COMMENTS
CG's picture

OK... Just how do you know that the "inaudible" aliasing is not the actual perceived benefit of such filters? The alias content is time synchronous with the program content and is hard to pinpoint when played back by itself. That does not mean that it doesn't have an audible effect.

Go find an amplifier or preamp using loop feedback that has marginal phase margin in the amplifier response. Under a variety of conditions into real loads (cables that are not terminated in their characteristic impedance, which they almost never are, as one example), these amplifiers will often produce low level intermodulation distortion products high in the audio band. These are obviously time synchronous with the program content. What's that sound like? An awful lot like the descriptions I've read of how MQA sounds.

As for "Aliasing cannot be a problem if there are no signals to alias.", well, duh! This is why band limiting filters are used in the original A-D conversion. If there are no signals above half the sampling rate, no "anti-alias" filter is required. (Random noise will alias, but that's at the, ahem, noise floor.)

But, it does beg the question of just why you'd need various folding and encoding schemes to allow for higher sample rates if there is no information there.

What's the point of 96 KHz sample rates if there is no audio information above, for example, 20 KHz?

In the mixed signal world outside of audio, higher sample rates are usually employed to make the filtering easier, mostly because passive filters take up space, cost money, and modern engineers don't like to design them!

I'm probably wrong here, but I thought the same reasoning was true in audio. Well performing filters are difficult to implement with analog components, so you move the sampling rate higher to relax the requirements for these filters. There *may* be little or no program material above that magical 20 KHz, but at least the practical filter implementation is more easily attained. If there are useful signals above 20 KHz, they're captured as a kind of bonus.

Again, I'm probably just showing my ignorance, but I fail to see how this is some "post-Shannon" sampling theory. At best, it's clever management of what signals appear where and at what levels. Psychoacoustics, not sampling theory.

MQA may sound exquisite for all I know. The explanations do not.

CG's picture

One other point...

B-spline and similar filter algorithms have been around for a while.

Wadia was perhaps the first audiophile oriented user of the concept. They were applying it to DACs back in the late 80's. Of the last century. Note that the Wadia products were developed about 35 years after Shannon, Hartley, and Nyquist were publishing papers about signal theory. We're now almost 35 years after Wadia...

https://www.stereophile.com/digitalprocessors/wadia_digital_1000_decoding_computer

http://www.gammaelectronics.xyz/s_1996-12_Wadia.html

http://www.hifi-advice.com/blog/specials/wadia-digital-company-special/wadia-digital-history/

(In the last of these links, there's an article snippet about the ups and downs of Wadia as a corporation. If nothing else, always remember that a lot of audio companies are businesses with investors who expect a profitable return...)

Pioneer had something they called Legato Link that came soon thereafter. Several others followed. A search on Google on these subjects can quickly consume your morning.

Oh! One last thing.

If you look carefully at the Measurements sections of Stereophile reviews of DAC products, there's a couple plots that show distortion products for both harmonics as well as two tone IMD. Yet, you see all sorts of other products that seem to not be related. But, if you do the arithmetic (not complicated math) you quickly learn that these really are simple intermodulation distortion products caused by non-linearities most likely in the analog amplification portions of the DACs. You see, unless you have an analog low pass filter that cuts off just above the audio band, there's all sorts of imaging products that are the result of the digital to analog conversion process.

Here's some really nice pictorials of the process:

http://www.dspguide.com/ch3/3.htm

The takeaway to consider is that linearity outside the traditional audio band matters with digital to analog conversion. (Unless there's that analog filter right after the converter chip itself). If opamps or other amplification with loads of feedback are used, their linearity often deteriorates above the audio band due to the falling open loop gain. "Open loop" amplifiers also deteriorate as you go higher in frequency, but because the designers can't rely on feedback to correct for non-linearity, the amplifiers usually are linear over a wider frequency range. This doesn't mean that high feedback systems can't be made to work - you just have to plan and design accordingly. Keep in mind that the same argument holds for not only whatever amplifiers are in the DAC itself, but in whatever preamp and amplifiers you might be using in the chain as well.

leedom's picture

MQA: A Review of controversies, concerns, and cautions
https://www.computeraudiophile.com/ca/reviews/mqa-a-review-of-controversies-concerns-and-cautions-r701/

jherbert's picture

...dann verwirre sie. German saying.

DH's picture

Record at very high sample rates 192K and above. Microphones basically don't pick up anything over 100k and certainly recordings can be made without info with extremely high frequencies.
Problem of band limiting, aliasing, etc eliminated. NO MQA needed.

dalethorn's picture

I read this through a few times, looked up the term equivocation in the dictionary, and thought: hmmmm....

B-splines and the like have a long history in noise reduction, so that part is understandable enough.

"If properly managed, upward imaging need have no negative impact on the sound, especially if the images are beyond the frequency range...." - OK, but is the 'If' always properly managed?

"....In the real world we do not have impulses in air." - hmmmm.

"Even this small error is with a highly unrealistic test signal. With actual music, ...." - hmmmm.

I wonder - even though we can't take possession of a proprietary MQA encoder to test on our own computers, it seems fair that we could get some people to sit down in a closed lab with such encoders to run some independent ad-hoc tests, which would likely reveal weaknesses in the MQA encoder(s). All complex software has bugs, and revealing those bugs - even some of them - would enable us to search more intelligently for specific failures in commercial music releases.

The problem here with MQA is that I'm not getting software to run on my computer on my own data, as when I'm using facebook or a Web browser, etc. In those cases, the public gets to know the bugs. With MQA, I'm getting only the canned output - the music - and the bug list is hidden from all of us.

spacehound's picture

Shannon's sampling theorEM (it's not a theorY as Jim falsely states) is PERFECT. No greater authority than Jim Austin himself just said so. Twice. Even if he did try to con us into this "theorY" nonsense.
And 'perfect' can't be improved on.

And time and frequency 'domains' are exactly the same thing, just looked at in two different ways. Just as saying "Cut this twenty inch stick so one of the two parts is twelve inches long" has the same result as saying "Cut this twenty inch stick so one of the two parts is eight inches long". It's that simple.

So just WHY does Stereophile keep pushing Stuart's/MQA's marketing BS?

Reticuli's picture

In the sciences, "theory" is often considered interchangeable with "theorem" and even "law", and this is obviously considered above hypothesis.

The problem with Nyquist & Shannon's sampling and resampling stuff is it's impossible to engineer perfectly, even if the math technically is a proof of possible in-situ data perfection, because the digital audio info does not (necessarily) come from numbers nor can it stay as purely numbers -- it's got to become analog to be useful eventually. It's not possible to engineer perfect digital audio sampling, resampling (in any but the most simple scenarios), or conversion back to analog (what we're talking about here).

It is possible now to get much better compared to the early days of digital audio, but it requires very computationally-intensive non-realtime techniques using feed-forward and feed-back (a reason why it's not realtime anymore). It's still not perfect, though, and expensive and complicated to pull off the state-of-the-art. With the amount of stuff staying in the digital domain, now, the number of stages necessary to actually conform to such complicated, expensive methods is not inconsequential.

As others have mentioned elsewhere, just using higher sampling rates does do a fairly effective, economical, and time-efficient job of pushing much of the problems out of the most audible bands. It's amazing how even going just a little over 44.1khz can make a noticeable difference in easing out higher fidelity in cheaper and/or real-time applications.

dce22's picture

Jim Austin: Is MQA an application of post-Shannon sampling theory to audio coding?

Bob Stuart: Yes, very definitely, it is!

That is it people 2+2 no longer is 4 the king has spoken.

Bob Stuart: The specified level of aliasing is not audible, with a safe safety margin.

GTFO and take your distortion making machines with you.

Anton's picture

"Perfect" has been an ongoing digital code word, like, forever.

spacehound's picture

It's simple mathematics, cleverly figured out by Shannon. And it's as provably correct as Pythagoras' theorem. The only people who call it a theory are those who don't understand it or charlatans trying to sell us something.

Anton's picture

Remember “perfect sound forever?”

It was a lie.

So is MQA.

spacehound's picture

Sure, anyone who does basic tests on MQA quickly finds out its claims are false. Even the 'authentication', which is merely administrative, rarely happens in reality. And there is plenty of proof that the MQA people often 'authenticate' it anyway, even when the artists/producers point blank refuse to have anything to do with it. And a quick look with a 'scope and a spectrum analyser will prove the technical claims to be rubbish.

CDs? Hmmm. The original CDs were extremely good. They do last 'forever'. Well, up till now, at least :-). Philips, then one of Europe's largest corporations, truly believed in this format and placed strict quality controls on the few plants that made them, followed up with regular inspections.

Later, when they became 'mass market', made by all sorts of outfits worldwide, and sold in huge numbers, Philips could no longer control the quality so it gradually deteriorated.
Now, even in the average household room the later ones can corrode quite quickly due to poor application of the protective varnish (which is on the label side) and the now common direct label printing on that varnish, allowing the thin aluminium layer with the 'pits' to become partially unreadable, causing clicks, pops, and skips. And if that occurs on the index it won't play at all.

But 'Shannon' is totally solid. As I said earlier, it's a theorem, not a theory. If used as instructed (sample rate a minimum of twice the highest frequency you want to capture) it actually DOES result in perfect digital capture of all the possible analog waveforms. It's non-intuitive but it's true.
So if people will accept that adult humans simply CAN'T hear above about 20KHz then the 44.1K of CD is fine. Though if I had been 'in charge' I would have gone for 48K simply to eliminate all argument.

dalethorn's picture

I would add a very old saying: "A chain is only as strong as its weakest link". When you consider the number of processes in the recording and playback chain, and the number of hands (or 'bots) the music passes through before it gets to your ears, you'd want to have constant attention on all of that to maintain maximum purity of the data. My biggest (but not only) ongoing concern about MQA is that it appears to me they're trying to make it a truly robotic process so that every album can be 'processed' by automation, with a veneer of human interaction that's just vapor.

Maybe everyone is so used to automation that they aren't concerned about this aspect of it. I'm concerned about *every* aspect of automation that has a potential impact on the music.

In the past few weeks, I've come to the tentative conclusion that no matter how smart they make a filter (MQA), applying it robotically to every album is a bad idea. And I wonder how many people have thought this through, putting aside all of the other arguments for the moment.

Bogolu Haranath's picture

Hello everybody ........Why am I writing this comment on this forum about MQA? Read on ........ I have recently seen an "infomercial" about a new kind of microphone with a "triangular" shaped transducer diaphragm on Pro audio review journal. The company is called Ehrlund Microphones. If you are interested, check out their web site. They have a video posted on the web site how the microphones work. Essentially they say a triangular shaped diaphragm is superior compared to a round diaphragm which is commonly used because of superior "impulse response". They claim that the suboptimal impulse response causes distortion of the audio signal recorded. Again, check out their website for more detailed explanation. Now about MQA ....... Is "blurring", time domain errors caused by microphones? Is some how MQA correcting them as a "side benefit"? Probably JA is reading these comments along with Jim Austin. May be they could comment on this topic. JA is a recording engineer/producer. He may be interested in these triangular diaphragm microphones. May be in the near future somebody could work on triangular shaped loud speaker and head phone drivers as well.

Bogolu Haranath's picture

JVS ........ Great music recommendations. Bought most of them. Keep them coming. Thanks.

Jason Victor Serinus's picture

Not being a recording engineer, I have no comments on these microphones other than to say that words are words, measurements are measurements, and sonic results are more or less related to the same, but in ways that are not entirely predictable due to the limitations of measurements and the fact that, no matter how hard they pretend, human beings do not know everything.

spacehound's picture

Knowing they don't know everything is the very basis of all science. Science is a DISbelief system. Many laymen seem to have difficulty in understanding that.

Another problem is that perfect 'impulses' don't exist in reality so they can neither be measured nor heard.

And there's not much point in measuring imperfect impulses :-)

spacehound's picture

Nobody is listening.

Except for a tiny minority people are either against it or totally indifferent. And that's everywhere, it's not just Stereophile readers.

Is Stereophile too blind to see that?

ok's picture

..aliasing can't be audible since ultrasonic – but ultrasonic content should be audible despite being ultrasonic :-)

CG's picture

Very curious, isn't it? :8^)

spacehound's picture

It is the result of the recording process recording ALL the output of the instruments, audible or not, being recorded. and it's not a 'fault', it's an 'effect'.

Some instruments produce inaudible high frequencies which gets 'reflected' back into the audible range dues to the Shannon sample rate/frequency limitation, and is correctly termed an 'image' or 'imaging'.

Unfortunately MQA relies on weak' filters that don't prevent this imaging for the process to work at all. So on an MQA file this imaging can be heard quite clearly (again depending on the instruments) if you know what to listen for.

It tends to give a fake 'liveliness' to the music that would not have been on the original. Some people, including the Stereophile reviewers seem to like that, or pretend to.

dalethorn's picture

I'd like to build a list of tracks that have audible imaging like that.

Gumbo2000's picture

Stereophile might want to lay low on MQA for a while. Every new article makes MQA look more ridiculous.

Jason Victor Serinus's picture

Is the alternative to calling something "fake news" to now call it "too real news"? After all, what Jim did was to ask Bob Stuart a series of informed, highly technical questions, and then transcribe the answers.

Would the press make you happier if it did not ask questions, or only asked questions about issues that you are comfortable with? I pose these questions not to you in particular, but to everyone reading this.

The fact is, more companies are adding MQA decoding to their devices. That includes some mass market gear. Another batch of MQA CDs has been released in Asia. Major labels are onboard. The entire Warner catalog is going on Tidal in MQA. MQA is making its presence known.

To my mind, asking questions of the co-creator of MQA is essential. Note that there is no effort on Stereophile's part to control the responses to those answers, unless they cross the line to personal attack or downright irresponsible accusations. But everyone reading these comments knows that every time we discuss the acronym MQA, the naysayers will come out in force, posting again and again in hopes that if they say what they have to say enough times, people will ignore everything else and simply believe them. What we are doing is presenting the facts and sharing our experiences, and allowing people to read, listen, investigate, and then decide for themselves.

Gumbo2000's picture

more articles like the current one only continue to muddy the waters with BS (not referring to Bob). To paraphrase "With friends like Stereophile, who need enemies!"

spacehound's picture

The problem is believing the answers you get are true.

Stuart is quite clever. He is careful to not actually lie. He 'dissembles' instead. Some of his answers above are careful dissembling and some are mere gibberish.

Also.
What companies are doing is irrelevant. The buyer, and only the buyer, will decide. See '8 Track', 'DAT', etc.

Buyers are not going for Tidal, which remains 'microscopic' (though I use it myself, not because of MQA, but because of the '44.1' regular stuff).

And to date less than 0.02 percent of Tidal's catalogue is MQA. And it's decreasing, as far more new issue recordings are non-MQA than are MQA.

Warners and others just add MQA on afterwards. So the claimed 'authentication' can't happen. So what's MQA actually for?

rt66indierock's picture

Are the major labels really on board? You wouldn’t get much of breakfast at the Blue Moose Café with what they each paid for their MQA Ltd stock.

And why would Warner give any more MQA content to Tidal? I got a hold of a Universal royalty statement from June 2017. According to that statement Tidal only has 170k HiFi subscribers. Norway has the third highest number of subscribers behind the United States and Poland. In Norway Tono (a royalty collection organization) has gone to police and the Norwegian arm of the INFI has started an audit as has have the Danes. And of course you did notice Tidal has not denied it is behind on royalty payments. Wouldn’t the prudent thing to do be to let Tidal twist on the vine and collect whatever they can? More content MQA won’t bail Tidal out.

Bogolu Haranath's picture

Thanks for responding to my comments .......... No, it is not about measurements ....... It is about "blurring " and about time domain errors caused by round diaphragm microphone "ringing" ....... This phenomenon may be similar to what MQA is supposed to be correcting ....... Since most microphones are made with round diaphragm, which are being used for at least 100 years or more, the so called "blurring" may be a distortion caused by microphones themselves and not entirely digital artifact ........I am not an engineer either ......... As I mentioned, if you visit the Ehrlund microphone web site and watch the video, you will know what I am talking about ......... The video is appox. 30 minutes long ........ Thanks again ..........

T-NYC's picture

... may equal Post - Stereophile. Enough PR BS. Stuart may need the money although his retirement should be secure via Meridian; notably Peter Craven is never part of these infomercials. Please, get back to reviewing magic cables & racks -- at least these do not debase the source material. And for an informed, impartial look at MQA try Archimago's Musings -- he's not selling anything -- not MQA licenses nor advertising.

dalethorn's picture

Archimaga is not selling anything? I dispute that. After all, the head of that site paid for his "article".

T-NYC's picture

Your comment is disingenuous. What does it have to do with his multiple objective reports containing lots of hard experimental data, both measurements and controlled listening tests, by himself and others? I mentioned his site purely because his articles are very well done, and link to many others doing similar experiments and thus are a much faster way to find many sources of hard data than Google. I suggest you read them and then ask him how much he earns from the site. My guess is not enough to pay for test gear; that this is truly an informed enthusiast's passion, unlike Stereophile which stopped being a thought-leader in the 1990s and has primarily served an industry PR function since. I'm okay with that except when it becomes ridiculously blatant as it has with MQA.

dalethorn's picture

Disingenuous you say? No - I challenged your assertion that "he isn't selling anything". Whether he's merely making a living and his employer is the actual seller makes no difference to the issue.

Archimaga had his chance to publish a clean article, but instead packed it with lots of personal angst. Now he can go back and clean all of that up and make it suitable for AES or IEEE or whatever, but the intent was and is clear from the beginning.

And besides, nobody I know is going to take the time to validate any of those "proofs". What we want to know is what effect is has on the sound, and is it likely to become a DRM tool. And as I posted elsewhere on this forum, it would be great if Bob and friends would organize a test event so many of us could test his encoder for bugs. We would NOT need access to any code or algorithms or other secrets - just the computers under Bob's watchful eye so that we can take a non-MQA file to the MQA coded state and see what the possibilities are.

T-NYC's picture

I am going to take your concerns at face value because you are not selling anything ;-0 You need to read his MQA thread and follow the links to others; not one article but roughly twenty. A pain because blogs show them in reverse order. You will find ample evidence that sound is affected. "Bob and his friends" are trying to sell this. If you read my AES post above you'll see -- no controlled listening tests THREE YEARS after the announcement. This is a scam, and if it were just more magic cable type nonsense it wouldn't matter because we can make choices about what we buy. MQA affects the source material. The record companies are enthused because MQA (like DSD) has a theme-song: "Buy It Again, Sam". However, this time around Lo-Fi is built in through a compressive algorithm of very limited application value, originally intended only to improve streaming sound and now being applied to CDs. Why would you compress a CD, adding artifacts?. Money, that's why!

dalethorn's picture

"You need to read his MQA thread..."

Erm, no, because it comes from a source that's not only unprofessional, but has unethical behaviors.

T-NYC's picture

he's a degreed EE and what is unethical about him?

dalethorn's picture

The source, boss, the source...

The CA guy, his "biggest" anti-MQA threads in the Whole World, and the unethical behavior that's been described at length here in previous months' topics.

I'm sure you can easily find what you need.

Archimago's picture

Re: Dale's comment.

1. I have no ties to the industry. My employer services the public and my day job is not audio related though my background training does allow me to say a few things related to this topic matter.

2. The "proofs" were reviewed by 3 others who have backgrounds in audio engineering and computer science.

3. Please... I don't think my writing was all that "packed with personal angst". I expressed what I thought needed to be said based on facts, reasonable opinions and I hope logical conclusions. I think the audiophile public can appreciate my perspective in the service of consumer freedoms amongst the sea of uncritical support from the audio magazines as if this were some kind of true "revolution" rather than an attempt at control, using techniques which actually decreased potential fidelity of playback. If the press actually addressed these issues directly and fostered debate, I would not need to say anything, would I?

dalethorn's picture

The problem here is not your facts. I've seen the charts and graphs and don't dispute them. How they are interpreted could be a problem, and if I remember correctly, your famous article added a lot of interpretation.

Archimago's picture

What's the point of charts and graphs if not for providing insights into the meanings, interpretations and explaining implications?

Whether one agrees with the interpretations I leave to the readers. But if one disagrees then by all means discuss the concerns rather than vague insinuations and insults.

dalethorn's picture

Your interpretations disagree wildly with others who've published articles here, which leads me to think that you'd get more respect for your articles if you could build some info-bridges to the customers here who might read those articles. I don't find them useful, albeit I am a tech person. [Flame deleted by John Atkinson]

Archimago's picture

Dale, the "info-bridge" has always been there. While there are elements to the purely subjective philosophy that I do not agree with, I can appreciate the opinions of those who hold on to that stance even if I disagree. No need to get angry.

The issue with MQA has to do mainly with the product itself and the way it's promoted. Just because it exists doesn't mean it's any good.

Flames are unnecessary Dale.

dalethorn's picture

Info-bridge means getting around the fact that your opinions disagree wildly with what I see published here. Anger and emotions have nothing to do with your disagreements, it's about the facts only.

spacehound's picture

….You try to discredit Archimago by saying he is 'selling' things. Or if he is not doing that directly, his employer is 'selling' his audio-related work.
We who have 'followed' him for a long time are fully aware that his 'day job' and his audio articles have no financial connection whatsoever. Some of us are also aware of what he does for a living, though as he does not disclose that in his 'audio' articles (and why should he?) we are not going to disclose it either. Though it is not a formal 'secret'.
It's no different from me, employed in the computer industry, writing some articles on fly fishing for trout and being paid for them. Which I have done and I was paid for them too, as all magazine writers, even 'casual' ones such as myself, are.

So FAIL on your part.

And having totally failed on your first attempt to discredit his audio work, you later turn to comparing Archimago with the editor and regular writers here, all of whom are ultimately paid out of advertising revenue. As of course was Archimago for his CA article, but writing such articles is not his 'employment', not even part of it. Unlike, for example, John Atkinson or even JVS, who's income seems to be derived from 'music' and related stuff, which HiFi is.

And the inevitable result of their method of payment is that the industry must not be criticised in anything greater than a 'trivial' manner. So their 'tests' and 'reviews' are not as searching as they might be. Hence your Archimago "disagrees wildly with others"

Similarly, your "I don't know anyone who will repeat these tests" Lots of people have. It's quite simple to do, Archimago having laid the path, and given the relevant equipment. You just don't know them.

Furthermore, Archimago's CA article was peer reviewed by several known figures. Nothing here is. Even the editor says the writers are totally free to write whatever they want.

So FAIL for a second time on your part.

"Credibility on MQA's part".
No. the fact remains that Shannon's theorEM (it's not a theorY as some deliberately mis-call it, it's a fact, as all theorems are, by definition) must be obeyed if analog dats is to be digitally recorded 'perfectly' (other than microphone and other analog imperfections on the front end).

If that is the intention then 'post Shannon' simply does not exist.
It was never 'hidden' as some imply. MQA is not 'new', similar methods have been tried before. They were not followed up because they all produce a degraded result, as does MQA. And as they disobey the theorEM, they must.

PS: Examples of audible imaging when MQA is used? Try Petty's 'Free Fallin'' and 'Learning to Fly'. You might like it at first, I did. But it gets 'tedious' as 'obvious if accidental DSP' after several plays.

dalethorn's picture

You have it all backwards of course. I went into this with honest intent, as an ordinary person under my real name, and was abused by some dishonest people at CA, not just on the MQA topic but others. Then the very owner of CA, who made it clear that he was *totally* biased and vindictive, posted my *private* message to him publicly. Adding to all of that, one of his sycophants made threats to me on this forum and another, and continues trolling here.

You'd have to be blind to miss all of that. I do realize that you're against MQA in the extreme, but I can respect that IF you're honest about everything. But the sources for the infamous article are not honest, and your apparently uncritical support of them has me thinking that maybe you can't be trusted either. You know who I am and the work I've done, which is extensive. But all I see of the people you apparently support are 'nyms, and hostile 'nyms at that.

Archimago's picture

Yeah I got a few buck for writing the article. It was the standard amount Chris gives for headline articles.

I don't think I need to divulge the exact amount just like I don't think anyone here should need to divulge how much they get for their writing.

I can however tell you that I make about 4x the amount per day at my day job. It took many hours to research that piece and time to double check with others 'in the know'.

No sir, money is NOT my motivation.

dalethorn's picture

Your motivation was never in question. Your interpretations of the graphic data can be questioned, not so much your motivation. The biggest problem is where the call for that article came from, and their motivation was made crystal clear as anti-MQA, end of sentence.

Personally, I see a lot of credibility on the pro-MQA side, and some - even if not an equal amount - on the anti-MQA side. My impression is that you're fighting a losing battle, and a big factor in that losing effort is alienating good people who also have serious reservations about it.

Archimago's picture

We'll see ultimately which side of history the debate settles on in the not too distant future I think.

I got a PM about writing an article to summarize concerns raised by a very long thread. Many have been critical of MQA for months if not years. I got a small amount of money for the review article. So what?

argyle_mikey's picture

How about a poll ?
Choices :
a). A monthly MQA article
b). A regular monthly review of a “budget” component (in the last year you’ve reviewed just 7).

Mike

jimtavegia's picture

better is better, but I could only hear the music starting at about 1:19 or the 1:20 point. As someone who does much recording and owns over 16 microphones, some with only 5db of self noise, I work very hard at the venue to listen for room noises and artifacts than can ruin my efforts.

It still seems to me that MQA is more about master control, but if it does ultimately sound better then the market will decide. The problem here is that if most of the major labels subscribe then WE don't get to choose other than buying a MQA capable DAC or not. I am not sure if this is the same as a DSD recording session mixed down for rebook sale. All of my SACD releases do have a rebook layer so I don't need an SACD player unless I want to hear the most. The price of entry into SACD player land is not cheap anymore and certainly not as cheap as gaining entry to hear MQA files. There is still much to be sorted here.

Ortofan's picture

... the Pioneer Elite BDP-80FD DVD unit will play SACD discs and costs about $300. Otherwise there are SACD capable audio players available from Marantz, Yamaha and Denon for about $1200.

adamdea's picture

I think Jim deserves some praise for putting to Bob Stuart some points which others have raised for some time.
I don’t know why Bob eschewed the opportunity to respond to Archimagos article but this filled in some of the gaps.
I have criticised Jim for not asking probing questions in the past, and so it’s only fair to give credit for getting Bob Stuart to address some of the issues. One might I suppose go as far as to draw an analogy with the Frost Nixon interview where, in advance Frost was seen by many as being too soft and lightweight.

In many ways I see Bob’s Stuart’s answers as a remarkable mix of evasion and white flag. Is that lack of pressing follow up questions a weakness or a strength? Perhaps Bob would only go on the record if he had confidence that his responses would be received uncritically.

Well, anyway I thought it was useful. Is it safe to assume that we now know what time smear is? Is it uncharitable to point out that this could not possibly justify the hype? Or that it is somewhat difficult to accept that this could be what Bob meant all along? Or to point out that if quantisation is the issue then cutting 16/44 to noise shaped 13 bit seems an odd sort of progress?

T-NYC's picture

https://www.researchgate.net/post/Is_Nyquist_Shannon_sampling_theorem_not_a_necessary_condition_to_recover_a_band_limited_signal_from_its_sampled_version

Note: compressive sampling based on prediction NOT high-fidelity; very useful for mobile phone and similar where intelligible results not complete data reconstruction (AKA fidelity) is the goal.

T-NYC's picture

At AES 2017, Bob Stuart was asked if there had been any scientific listening tests performed on MQA. The question was from Bruno Putzeys, mathematical physicist & accomplished audio designer. The answer three years after the launch of MQA was no. The room got very quiet.

dalethorn's picture

A far better test (as I noted elsewhere) than merely listening to whatever "output" Bob would lay on the testers' desks would be to allow the testers to apply the codec to PCM files themselves (in Bob's lab of course), then listen to the before and after right there.

Bob would retain all secrets, and the testers would not only be able to validate the before and after results, they'd be able to try different things (within whatever are the allowable encoding parameters) to see what bugs might be hiding in the encoding software.

Archimago's picture

Controlled pre and post-MQA processing is what the supposedly upcoming study from McGill University is addressing. That will be interesting (if the results surface).

I don't think it's about bugs. Rather the differences we can already find on analysis of the decoded signal is by design and clearly are the results of a lossy process necessary for data reduction and the upsampled reconstruction.

dalethorn's picture

It's about bugs at some level if the industry becomes less demanding of quality, which is likely. Bad enough to have bad intent (if that's the case) in applying potential DRM, but it's even worse if further corruption creeps in due to bugs.

T-NYC's picture

DT, there is no interest in this (your proposition) at Meridian / MQA. It's been three years (see my post above). This is merde served fresh. BTW, again note Dr Peter Craven long ago disassociated himself from MQA. He has not been publicly involved since AES 2014. Ever wonder why? I do not know of a single physicist nor cognition-researcher who's work is with high-fidelity audio or touches on it who buys into this; quite the opposite.

dalethorn's picture

As Bob Stuart comes from audio and Peter Craven from math, is there something I should look for to tell me why Craven should be involved in MQA beyond the tech aspect of the code? I do recall that there were problems with early MQA releases and noise levels that was fixed, and many albums were remastered to accomodate that fix. I imagine Craven would have been involved in that. So I wonder - has Craven disassociated himself from MQA to the extent that if the codes need more work, he will not do any of that work?

spacehound's picture

…..to software development that just 'coding'.

Typically some guy has a few ideas. Trying to code 'ideas' never works. So it is the software people who do all the 'system analysis'. And all the 'going backwards and forwards to the customer and potential end users'. And all the coding. They have to, all the way through the development, as neither the original 'idea' guy nor the customer/end users have a clear concept of what they actually want.

dalethorn's picture

I'm trying to verify whether Craven has really disassociated himself from MQA, as the guy above noted. If true, it could have profound implications in future updates to MQA.

rt66indierock's picture

I respect that John Atkinson doesn’t want financial information discussed but I don’t think saying that Peter Craven disposed of 121 shares of MQA Ltd stock in 2016 and retains 759 shares as of the most recent publicly available records violates that.

dalethorn's picture

Oh BTW, I did do a few Web searches trying to find something on Craven that wasn't just another discussion by or about Stuart, but I couldn't find anything.

Bogolu Haranath's picture

Since you wrote on this forum that you own many microphones and do many recordings, you may be interested about Ehrlund microphones. The transducer diaphragm in these microphones is "triangular" in shape. Probably the microphones you use have round diaphragm, which is very common. The company claims that their new type of microphone has superior "impulse response" and is better suited for recording. If you are not already aware of it, check out their website. They have a appox. 30 minute video posted on their website which explains the concept in more detail. Again, the name is Ehrlund microphones.

jimtavegia's picture

I am sure that due to high quality engineering that these new mics will sound good, acceptable, but I am concerned that the pistonic action with equal pressure being the driving force behind circular diaphragm designs. Until I had heard and used one I would pass no judgement other than it not being tried or used by other established companies, but that was a choice by their engineers and does in any way alter the performance of a new choice in design. The automobile industry has often had to result in differing speaker driver shapes due to space limitation and most often sound excellent.

I will try and find a dealer for an audition as changes and improvements are always welcome in trying to capture more in a recording event. Thanks.

Bogolu Haranath's picture

Hello ...... Me again ........ There is a new album by artist Morgan Agren, "Through the eyes of a Morgchestra" which I think is recorded using Ehrlund microphones ........ Check it out if you want to ......... I think it came out recently in 2017 .......... I did not explore all the new recordings by all the artists using these microphones (yet) ...................

spacehound's picture

….placing free advertisements for these microphones in this thread?

Stereophile does have a formal process for placing advertisements you know.

jimtavegia's picture

The MP3 format will not allow much of a test. Is there a CD version for sale that you know?

Bogolu Haranath's picture

Thanks for responding to my comments ........ I don't know whether the recording is available in CD or not. You can check the Hi-Res download websites and streaming websites like Tidal ......... I downloaded it from iTunes ........ Well ...... It sounds good even with 256Kbs resolution ..........

dalethorn's picture

The artist says that the album you reference was created with a MIDI, but will be performed later by a symphony orchestra. So how can a person relate the quality of the microphones if there is no real acoustic sound?

Bogolu Haranath's picture

I don't know all the details of the recording ...... From what I understand the drums were recorded using Ehrlund "triangular" diaphragm mics ........ The artist Morgan Agren is a drummer .......... He has a video posted on You Tube and Google search .............

dalethorn's picture

I have a drum recording by 'Tutt-Keltner' that has a variety of high quality drum sounds suitable for evaluating the mics used for the recording. Listening to the album you've referenced, I don't get enough sense of drum "skins" in real space that I could make a judgement with. It just sounds electronic to me, unless you have a specific track from that album in mind that reveals acoustic drum sounds very clearly.

Bogolu Haranath's picture

You probably have lot more experience listening to music and drum sounds and various mics, than I do ...... There are several tracks I can hear drums ........ I can't recommend just one or a few tracks with good drum sound ......... You may wanna check the artist's website and videos and decide for yourself ...... I got this recommendation from a friend ..........

dalethorn's picture

I played quite a few tracks from the various Morgan Agren albums, but didn't hear anything I thought unusual.

On the other hand, Stereophile has (or had) a test CD that has Gordon Holt reading text into a dozen different pro mics, and the differences in their sound has to be heard - fascinating stuff.

Bogolu Haranath's picture

There is another album by the same artist released couple of years earlier ....... It is called "Batterie Deluxe" ....... It has drum sounds .......... I don't know whether it was recorded with the new mics ......... You may wanna check that one too .........

Bogolu Haranath's picture

Yes .... I do have that CD which I got many years ago ........... Sadly JGH is not around ......... So, we don't know how he sounds with these new microphones .......... If you can find any other albums of any other artists using these new mics, let us know .......... I sure will be interested .........

Bogolu Haranath's picture

This type of triangular diaphragm transducer would be great for headphones too .......... The diaphragm is supposed to cover very wide band width ......... This type of diaphragm could also be used for a single driver loudspeakers .......... Imagine that .........

Bogolu Haranath's picture

Till those types of headphones and loudspeakers are available, we may not be able to fully appreciate the full potential of these new types of microphones ............ I am all excited about the future of audio reproduction ............

dalethorn's picture

I've reviewed several hundred headphones, and published EQ curves (i.e. the inverse of a response curve) for most of those. Headphone response (ignoring in-ear types) varies pretty radically between the different models, and that's assuming a perfect fit with no issues.

While the advantages of better drivers (for example, the Focal Utopia) can be heard through the choppy frequency responses, the sonic differences in drivers are a very small factor compared to the ragged responses.

Note that this is not mere perception - if you check Innerfidelity's measured curves, you can even see large response differences between different samples of the same headphone! With speakers, you have much more flexibility in adjusting the speakers, the listening position, and the room treatments (including drapes, curtains, etc.)

Bogolu Haranath's picture

The ragged and choppy frequency response you mentioned is the ringing, I think ........... The round diaphragm may be the reason for this ........ Most of the headphone measurements show this kind of response usually from 1 KHz going up .......... For most of the loudspeakers the crossover is in this region between midrange and tweeter .......... Some of the harshness, smearing etc. audiophiles talk about happens in this region .......... some people attribute this to metallic tweeters ......... Also, some people blame this to digital recording .......... A square diaphragm shows less of this ringing phenomenon than a round diaphragm .......A triangular diaphragm shows even less of this ringing phenomenon than a square diaphragm .......... You can see all this being demonstrated in several videos on Ehrlund website ........ Some of the people listening to MQA, reported that they hear "smoothness" to the sound in this crossover region ........ Probably the MQA DSP is donig this .........Ehrlund website does not talk about MQA of course ...........

Bogolu Haranath's picture

This ringing probably is coming from the center of the round diaphragm, because round diaphragm starts "beaming" like a flash light as the frequency goes up ............ Round diaphragm has a wider area of sound dispersion and coverage at the periphery than the center portion ........ You probably already know all about this sound dispersion phenomenon ..........

dalethorn's picture

Take a look at this Stereophile article on a single-driver speaker. Note especially the reviewer's listening comments on page 2, for example: "The conclusion I eventually reached was that the tension I felt with speakers other than the ClairAudient 2+2 arose from subtle distortions caused by the crossovers and dissimilar drivers."

https://www.stereophile.com/content/audience-clairaudient-22-loudspeaker#y17pShWJDSdRSv1S.97

https://www.stereophile.com/content/audience-clairaudient-22-loudspeaker-page-2

Bogolu Haranath's picture

I read that review a while back ...... The jagged and choppy response you mentioned is probably coming from the ringing ........... As you can see in the headphone measurements this kind of jagged response is higher in frequency say from 1 KHz and up .......... This is usually the crossover region between the midrange and tweeter .......... The ear is very sensitive in this region ........ This jagged frequency response does not happen with a triangular diaphragm, as shown on those videos .........

Bogolu Haranath's picture

One of the comments I made somehow disappeared from this forum ( I did not use any profanities) ........ Anyway .......... The comment is about ringing of the round diaphragm versus a square diaphragm versus a triangular diaphragm .......... The triangular diaphragm shows the least amount of ringing ......... All this information is available on Ehrlund microphones website ...........

dalethorn's picture

There's enough confusion here to make some things disappear, and part of that is because we're kinda off-topic with the microphone stuff. I am going to look closer at the Ehrlund microphone site to see how they're progressing.

dalethorn's picture

OK - taking a long look at the Ehrlund mic reviews, some by people I know like the Tapeop and Gearslutz guys, the overall impressions are an excellent mic, comparable to the best, with the most obvious differences from perceived neutrality being the emphasized top end and possibly a slight shyness in the lower treble/presence region. Based on what I know about current mastering and some of the high-end headphone tunings like the Focal series or AudioQuest models use, the Ehrlund looks like a great mic for piano, organ, percussion etc.

One reviewer noted his difficulties with trying to record acoustic guitar with the Ehrlund, and apparently gave up. But another reviewer noted similar difficulties and how he fixed the problem, so perhaps the Ehrlund can work well with acoustic guitars if the right pickup is used. In any case, I was looking for specific gotchas, and didn't find any fatal weaknesses from those reviews.

Bogolu Haranath's picture

Thanks for your response ......... I was trying to connect all the dots ........ We all know that the two weakest links in the audio chain are the microphone at one end and the headphone/loudspeaker at the other end .......... What if ultra low distortion diaphragms are used for both? ......... Won't that be great for audio recording and reproduction .......... Think about it ..........

dalethorn's picture

Apparently there's a lot more to it. For example, although planar headphones supposedly have great advantages over conventional piston drivers, I haven't found any around $1000 or less that compete with $1000 piston driver headphones. Focal's Elear that I purchased has a very unusual driver, and you can watch the amazing excursion it has just by gently blowing on it. But there's more to the Focal design - the headphone has noticeably less reflections in the earcups than any other premium headphone I've used - in a very small way it's like an anechoic chamber.

In the loudspeaker realm we've had electrostatics, transmission-line systems, Heil Air-Motion Transformers, and "time-aligned" drivers like the Dahlquist DQ-10. The main problem I see is when you get to the really impressive systems like PS Audio's giant IRS speakers (reviewed here by JA on 5th March), you have a system that can project a realistic sound of big orchestras, pipe organs, etc. - but with the big drivers and complex crossovers you've moving in a direction away from the low-distortion targets.

https://www.stereophile.com/content/binaural-video-report-sound-ps-audios-irs-vs

Bogolu Haranath's picture

Agree with everything you said .......... Mr.Speaker Aeon flow open and Audeze Lcd-2 classic are both planar magnetic headphones and cost less than $1000 each ......... Both are favorably reviewed on Innerfidelity website along with measurements ......... I have seen and heard the binaural videos on the Stereophile website ........ Most of the people don't have the space or the money for IRS reference speakers which I don't think are in production ........ Used ones may be available ........ Pistonic motion is not limed to only a round diaphragm ....... It could be a triangular or square diaphragm as well ......... Since you said you have reviewed many headphones, if you know any headphone manufacturers you could ask them to try out a triangular diaphragm and build a prototype .......... I don't know whether Ehrlund has any patents on triangular diaphragms .......

dalethorn's picture

I had a $600 MisterSpeakers - that was enough for me. Here's my EQ curve.

http://dalethorn.com/Photos/Audioforge/Mrspeakers_Alpha_Dog.jpg

The response was nearly flat: +/- 3.5 db from below 40 to above 10 khz, yet the sound was 'flat' in that other sense -- like a beer that sits too long on the table and loses its sparkle.

Bogolu Haranath's picture

I have not heard either one of those headphones ........... However I had personal experience with the beer ...... I know what you are talking about ......... About the beer .........

Bogolu Haranath's picture

By the way ....... If you like Focal and if you are willing to spend $1500, check out the new Focal Clear headphones ......... HR reviewed them very favorably in the latest Stereophile ..... It was also very favorably reviewed on Innerfidelity with measurements .......... Tyll called them "could be end-game headphones" ........ I have not listened to them, yet ........ I may soon do that .........

dalethorn's picture

I do believe that the Clear could be one of the best yet. If I decide to go with it, I will need to make sure it has both the cloth and leather/plastic earpads. Then I would have to test both, because earpads make a really big difference to the sound.

Bogolu Haranath's picture

Agreed ...... Earpads do make a difference ......... There is an after-market earpad company recently mentioned on Innerfidelity website ........ Tyll interviewed them during one of his Can-jam show reports ..... I haven't checked them out, yet .........

spacehound's picture

….disguised as 'comment'.

Bogolu Haranath's picture

Agreed ....... I own Focal Utopias which I purchased about a year ago ........ I very much enjoy listening to them ........ I also own Audeze Lcd -4, Lcd-x and recently acquired Lcd-Mx-4 .......... I also own AQ Night owls .......... Lcd-4 is the best in bass reproduction ........... Others are also good in various other ways especially in easy drivability ........ I am a frequent visitor to Innerfidelity website ......... None of the headphones available today have "flat" frequency response from 20Hz to 20 KHz ........ For that matter none of the loudspeakers available today have "flat" frequency response from 20 Hz to 20 KHz ......... These new microphone "triangular" diaphragm transducer is supposed to have "flat" frequency response from below 20 Hz to well above 20 KHz ........... That is why I am excited, if and when some company(s) come up with the "triangular" or some other different shaped transducer(s) for the headphones and loudspeakers we may not experience the full potential of audio recording and reproduction ........ To paraphrase the great JGH, "down with the round", may be not down with "flat" ........

Bogolu Haranath's picture

I forgot to mention ........... I also own a pair of Westone W-60 in-ear universal fit model .......... I am working on a custom fit model ............

Bogolu Haranath's picture

When the ultra low distortion transducers become available, we can always use external EQ and DSP to suit personal listening tastes and preferences ......... It is like putting icing on a great cake ........ The cake has be great to begin with .........

Bogolu Haranath's picture

Think about it ........ "Perfect sound forever" is possible in the very near future ......... If you can believe it ..........

mrkaic's picture

Is this MQA algorithm published anywhere? I'd like to see the equations and analyze it myself.

dalethorn's picture

It's more closely guarded than the schematic for the NASA warp drive.

Indydan's picture

You're confusing NASA with Star Trek...

dalethorn's picture

I'm not confused at all on that subject.

mrkaic's picture

Thanks, did not know it is a secret format. Don’t like that a bit.

dalethorn's picture

My suggestion to MQA, to maybe alleviate suspicions and doubts about their product (and still keep their formulas secret), is they should hold a laboratory day or days and invite audiophiles like us to come in and use their computers (not their internal secrets), to apply the MQA codec to a few masters so we can compare the before and after, just like when they apply the MQA codec themselves to albums.

So we would not see any formulas or software internals - we would just do what their automated processes do to apply the MQA, and then see how it goes. This is no different than any software user of Windows, Mac, manufacturing systems, music systems, or any other software - we would just be "testers" to see if there are any problems or bugs in the process, which MQA would prefer we not know about.

Bogolu Haranath's picture

It is proprietary DSP I think ........ They have patents for it .........

Bogolu Haranath's picture

Do we need a "special counsel" appointed to investigate this? .......... Is there some kind of "collusion" between the audio press and MQA? .............

Bogolu Haranath's picture

The same people who are guarding the secret recipe for KFC fried chicken, are guarding the secret recipe for MQA .......... In the MQA more than 100 herbs and spices were used from around the world .......

spacehound's picture

They get it from a bull.

Bogolu Haranath's picture

I don't know why all this fuss about MQA ......... If you like it, get it ....... If you don't like it, don't get it ....... I don't think anybody is forcing anybody to do anything .......... The last time I checked, this is still a free country ..........

dalethorn's picture

I'll risk confusion here by trying to simplify the issue. My simplification should not be taken as a set of testable facts, but rather as a set of principles, to see if MQA is working toward fulfilling those principles or not.

Music is very complex, and even when headphones and speakers are EQ'd precisely, they often don't sound right due to all of the complexities starting with the mics and mixers right through the chain to your ears. In other words, there are the performers, the engineers, the gear, and then you - lots of places where things can go wrong.

One principle I've discovered in EQ'ing 200 headphones is that once a relatively neutral sound is achieved, the final tweaking is not intended to produce a "flat" test response, but rather a close match to the sound of live music. If that seems counter-intuitive, consider that the advantage of approaching a natural live sound is you're not compensating for individual hearing, since your gear is truly neutral when it sounds like real music. That's a simplification of the actual process, but the principle itself is simple.

My understanding of the MQA sound principle (ignoring authentication for the moment) is that they intend to achieve a more natural sound by correcting problems in the production masters that are predictable by the MQA algorithm (i.e. computer program). I've asked many times about the "hands-on" aspects of remastering in MQA, and I've been assured by the experts here that it's almost entirely an automated process.

So (simplifying again), on one side we have Stuart and MQA assuring us that their remastering algorithm will make the right decisions in fixing those issues in the production masters, because the MQA algorithm really understands what's important in the listening process, and on the other side we have the critics who, with their technical tests et al, are assuring us that MQA is making things worse in some cases, and not helping anything in other cases. And those critics are certain that what their tests show can be heard by audiophiles as bad sonic effects.

I bought a few MQA albums along with the non-MQA versions. Generally, when both versions were released at the same time (Steve Reich's Pulse for example), they sounded nearly identical. In one case (Radka Toneff's Fairytales), the MQA version was much better, but that was a hands-on remaster that cleaned up a lot of things that had nothing to do with MQA. In another case (Buena Vista Social Club), the recent MQA version sounded a little better than the high-res download from a few years prior, but the larger part of "better" was probably something that made the sound a little more "alive" - not necessarily an MQA fix (but I don't know).

So on the MQA side we need to be able to run the algorithm on a few test albums, and then see for ourselves that it works properly, by comparing the before and after results. On the critic side we need specific examples from them of downloadable recordings that we can compare in MQA and non-MQA versions.

The problem in that latter example is getting MQA and non-MQA versions that come from the same production master, otherwise the comparison might not be valid.

Bogolu Haranath's picture

Agreed with everything you said .......... It appears that you are mentioning about MQA applied to the master recording after it was recored in hi-rez, say 24/192 or 24/96 for example ........ My understanding is that, there is another way of recording ........ MQA from end to end, that is AD converter is MQA encoded and later MQA DA conversion ......... Unless, there is simultaneous recording both MQA AD conversion and, say 24/192 AD conversion, it may not be apples to apples comparison ....... In that way you will have 3 different types of recordings ......... See what I am saying? ......... Can you clarify on that? ...........

dalethorn's picture

We know that MQA has already created original masters by recording and/or streaming live events, so in that case it would be difficult to separate what MQA does versus what the rest of that process does.

We know that MQA gets some masters (digital I assume) from the record companies and they apply MQA to those masters, then depending on what rights they agree to for that master, they might distribute it to a streamer like Tidal, or to a download seller like HDTracks. In those cases, if the exact same master that MQA received from the record company is available from a download site, you could possibly buy both and compare them. The problem with that is 1) How would you know that your download is from the exact same master that MQA got from the record co.?, and 2) If MQA is providing their MQA album for streaming only, no comparison is possible.

The best case, in terms of MQA processing anyway, is if the MQA guys get a master that has sonic problems or limitations - issues that nobody else has fixed - and they not only apply MQA to it, but before doing so they re-engineer (restore) the recording first so we get a better product. That's what they did with the Radka Toneff album, but as I understand it, that's going to be a rare thing.

My personal intuition says that if MQA gets the same digital masters that the downloads sites (HDTracks etc.) get, and all they do is apply MQA and jack up the price, that's going to be a bad deal. I've seen those at some download sites, and I won't buy them. OTOH, if MQA albums show up on iTunes at a reasonable price, I might buy those. The Radka Toneff MQA album on iTunes is coded at ~500 kbps AAC, which plays very nicely after conversion to WAV format.

Bogolu Haranath's picture

Holy ****** ......... The whole damn thing is too complicated .......... I do have the Radka Toneff album downloaded from iTunes after it was recommended by Stereophile .......... Mine says 256 Kbps AAC ...... Sounds amazing to me even at that resolution ......... One of the best recommendations by Stereophile .....

dalethorn's picture

The 256 kbps is the old version. The new version labeled MQA is the higher bitrate.

Note that you won't see the higher bitrate at all times, because it's VBR (variable bitrate), and because the music is not all that dynamic. Just watch the meter if you have the MQA** version, and eventually you'll see bitrates in the high 300's or even above 400.

**The iTunes "MQA" version is not actually MQA'd, or perhaps the internal MQA flag bits didn't survive the conversion to lossy AAC files. But it definitely is higher resolution than the old 256 kbps files.

Bogolu Haranath's picture

I just downloaded the MQA version of Radka album from iTunes after your recommendation .......... May be I will compare the new version with the old version tomorrow .......... Thanks for the recommendation ..........

Bogolu Haranath's picture

By the way ....... I did not see the old non-MQA version of Radka album on iTunes website ....... Only the new MQA version is shown and available for download ......... Conspiracy much? ....... It is also possible that the new version being remastered version (regardless of MQA) the old version was removed ........ The sad part is, I may have to re-download all the albums which I have already purchased from iTunes ....... I have to check my bank account ....... The bright side is, I can compare the old ones with the new ones in sound quality ........ Ha ha ha ..........

dalethorn's picture

I just checked the U.S. iTunes site, and indeed the old version has vanished. But the new version looks correct, so I have to wonder if your music player is showing the correct bitrates - the variable bitrate should jump around quite a lot, and at one point you can see 461 kbps.

Bogolu Haranath's picture

I haven't listened to it yet ........ I have to sit down and seriously listen to this new album and compare it to the old one ........ I will try to do that within couple of days ......... I may still need to listen to few more different albums for me to make any useful judgement ........ Even if I hear improvement in the sound in the new Radka album, that may be due to re-mastering, not necessarily due to MQA, just like you mentioned ........ That is the reason why I have to listen to more albums for me to make any valid judgements ........

Bogolu Haranath's picture

Looks like (or sounds like) "audio nirvana" was promised and then delivered by MQA ........ at least for some people .........

rt66indierock's picture

I’m a member of what John Atkinson calls the “We oppose MQA fraternity.” In fact John specifically mentioned me in a February 25th post on Computer Audiophile. I’d like to catch you up on few things.

First Jim Austin reported on May 19th in the MQA is Vaporware thread that “the listening tests at McGill failed to establish a difference between MQA and PCM at the same rate.” Another set of data supporting that MQA in itself doesn’t improve sound quality.

Member Archimago created an intermediate phase filter that might have better performance than linear and minimum phase filters. Julius O. Smith III of Stanford University said something similar and qualifies it as well “in some cases.” I wrote early in 2017 that I wanted control of the filters. The research done in 2018 seems to support that one filter may not be optimal. One reason I oppose a format like MQA that limits my choices.

Jim writes about Post Shannon sampling and Bob Stuart confirms in his way that MQA is the first use in an audio but neither provides a satisfactory reason why it would be considered better than conventional sampling. Member mansr provides a reason why it wouldn’t be better. “The Shannon-Nyquist sampling theorem provides a sufficient condition for fixed-interval sampling to fully capture a signal and enable subsequent reconstruction. Later research has defined other conditions allowing certain signals to be accurately captured without fulfilling the Shannon-Nyquist criterion. A search for terms like sparse signal, sparse sampling, compressed sensing, and finite rate of innovation will turn up hundreds of papers spanning decades.

None of this is new. The reason it hasn't been applied to audio is that it is unnecessary. An audio signal has such a low bandwidth to begin with that the Shannon-Nyquist requirement is easily met. The data rates involved also pose no problems for processing, transmission, or storage systems. Even if some form of sparse sampling of audio could cut the data rate in half, say, there are good reasons not to do this outside very specific applications. Traditional sampling produces a signal that is easy to process in a multitude of ways (think of all the operations a DAW can do). That isn't necessarily true of sparse sampling. Why should we complicate everything only for the sake of a data rate reduction we don't need? I can't think of a single reason.”

My own research confirms sparse sampling would complicate processing.

dalethorn's picture

From what you say here and all I've read here and elsewhere, the only valid pro-MQA arguments would be 1) The positive effect of MQA's proprietary time-domain filters, if they're actually positive, and 2) It "sounds better" to certain experts, if that's actually the case.

As of today, it appears that the technical tests published at CA which ostensibly show MQA to be detrimental to music, haven't scared off enough record companies and customers to slow down or halt the adoption of MQA. I'll take a shot in the dark on that one and say that the anti-MQA folk haven't reached audiophiles on the level they'd need to, to get a groundswell of opposition going in the wider audiophile community.

On #2, one person noted above a couple of MQA-encoded tracks that allegedly show where the MQA process screwed up and introduced artifacts that shouldn't be there. Maybe if some person or persons build a list of such tracks that can be downloaded (not streamed) in MQA and non-MQA formats for comparison, and those faults can be noted and verified by anyone, then perhaps we can get that groundswell going.

It seems odd to me that with the whole world of open digital music under a possible threat from MQA, that such a list verified by many audiophiles doesn't already exist.

spacehound's picture

1)You comment that if Craven in no longer involved it will be a problem for MQA.
Yeah. If, as it appears, the MQA outfit depends on the expertise of one man (Stuart is just the PR man, he has no software skills) it could be the end for MQA. So let's hope it's true.

2) I'm no fan of Conaker and his CA site either. The number of people who actually post rather than just 'lurk' is tiny and has been the same small set of elderly guys for years. And his attempts to employ 'staff' on any kind of regular basis (as Stereophile does) have all failed rapidly. Read into that what you will.

3) However, Archimago has produced reports on all sorts of audio stuff for years and nobody has been able to disprove his measurements on any of it.
What is more, MANY, now including myself (and I published my real name on Stereophile some months ago) have been able to duplicate his MQA results and such repeatability by anyone who tries is proof that Archimago's results are valid.

4) Nobody I know has ever heard of Tidal or MQA though most have heard of Spotify. Some even use that.
So after several years of trying MQA's impact is still near zero.

5) Surely nobody believes that the record labels have actually paid substantial amounts in advance for an MQA license?
Remember, MQA is the 'supplicant' here, trying to sell stuff to the record labels, who are pretty 'hard boiled'. So they probably have paid nothing and are just 'testing the waters' by putting their existing catalogues through an 'instant MQA sausage machine' (thus vanishes the MQA claimed 'Authentication') and seeing what the takeup is.
Take Tidal for instance. Only 0.02 percent of its catalogue is MQA. And it's decreasing, as Tidal releases far more non-MQA material than it does MQA stuff.
And it's all from those record labels. So they can't be releasing much new MQA material either.

Bogolu Haranath's picture

Even if Craven has left the company, it may not be a big deal for the company ............ Stuart has made enough money, he can easily hire enough software engineers to keep MQA going ......... Stuart probably also part owner of the patents ....... Also, some big (and rich) companies like Sony are known share holders of MQA ........... Whether we like it or not, big record labels are gonna shove MQA down our ears ..........

dalethorn's picture

I would worry more if Craven is no longer involved in maintaining the software. Partly that's because the software needs (or is going to need) a lot of work by someone who 1) Knows the existing code well, and 2) Has the genius (whatever) to maintain and revise it consistent with its original premises. The other part is because Craven's code affects millions of people and might eventually affect a million treasured recordings, and having him remain involved could make him more reachable by various bodies of inquiry, should that be necessary.

spacehound's picture

His absence, if true, might make MQA go away. That's good.

As for those affected, tough.
It won't be me, that's for sure. For every MQA recording newly issued, there are vast numbers which aren't, even from those labels who claim to have MQA'd their existing catalogue, which, anyway, remains in its original form. None of them are going to be fool enough to discard that constant and regular money source.

So its presence isn't increasing, it's already declining as a percentage of recording sold and/or streamed.

And the world's biggest market, the EU, already has 100 Mbps internet as an 'average' and it's getting faster by the month (mine is now 200+ 24/7 and I didn't ask for it, it just 'arrived' at no cost increase). Even 100 Mbps is about 10 simultaneous 192/24 streams. So the stated original purpose of MQA which was 'As good as CD but a lower bit rate', has already vanished.

Bogolu Haranath's picture

Is MQA "stairway to heaven" or "highway to hell"? .............. That seems to be question ........

dalethorn's picture

If we're reaching for absolutes we're sure to be disappointed. Any kind of engineering involves compromises - lots of compromises is the usual case. But this doesn't look like compromise as much as it looks like strong-arming. The ball is in your court.

Bogolu Haranath's picture

I am not sure what the microphones are recording is the "absolute sound" or not ........ We are listening to what the microphone(s)'s interpretation of the sound is .........

spacehound's picture

Despite what the 'golden ears' would have us believe, human hearing only goes up to about 20KHz maximum. And it isn't 'flat'. It is way down at even 10KHz - you don't need a high frequency response to hear an approaching tiger, which is what hearing is all about :-)

And as for the much vaunted 'time domain' that the MQA people constantly waffle on about, it's just the reciprocal of the frequency domain.
EG; 20KHz upper limit means a rise time from 'noise level' to maximum of 0.00005 seconds. Which is very easily achieved in electronics, microphones, and speakers.

In practice, for audio, the time domain is useful only for measuring the noise spectrum and 'analogue' distortion levels. Both of which are near enough inaudible unless something is 'broken'.

Thus MQA's comments about 'time domain' are pure BS..

Bogolu Haranath's picture

The "time domain" response may be the "impulse response" of the microphone diaphragm transducer ........ This may be what the MQA is talking about ........ The suboptimal impulse response of the microphones could be causing the suboptimal time domain response ........... This could be a type of distortion which may be showing up in the recording .......... This could be another form of IMD originating from the microphone .......... If you are interested, you can see some of my other comments in this forum above ......... I mentioned about Ehrlund microphones ........ You can visit Ehrlund microphone website for more detailed explanation ........ They call this "ringing", like ringing of a "round" bell .......... This type of "ringing" could be within the audible range ............

spacehound's picture

...everything you read in adverts,

Bogolu Haranath's picture

I am glad you did not use "approaching tiger" and MQA in the same sentence :-) .............. By the way I said IMD not WMD ......... There is a real big difference between the two :-) ............. Difference between MQA and WMD? ....... I am not sure :-) ............... Many "nothing burgers" have come and gone before in the audio world ......... We have to wait and see what happens with MQA .......... It may turn out to be "the emperor has no clothes" ............ MQA may be, again may be, trying to solve the problem with out knowing what is causing the problem ........ the transducer diaphragm of the microphone ........

Bogolu Haranath's picture

I downloaded an album by Eva Marklund ..... Album name "Stralande" ....... I got that after a review on a website ....... They say all the mics used were Ehrlund triangular mics .......... The singer I think is Swedish and she sings in Swedish ........ The drum sound, sounds excellent to me ...... Also, the vocals are very open and sound excellent to me as well ....... I downloaded it from iTunes ........ 256 Kbps, no MQA mentioned ........ Check it out, if you want to ......... Let me know if you like it .......... I used two headphones Lcd-x and Lcd-Mx4 for listening ...........

dalethorn's picture

The iTunes release date is 2014 and the low prices (plus lack of "Mastered for iTunes") indicate a 256 kbps-only recording. So that won't be good enough to validate the quality of microphones.

Bogolu Haranath's picture

iTunes allows sampling for couple of minutes of each track ........ You can check out vocals and drum sound ........ They say the mics were starting to be available since 2012, at least in Sweden .......... The company is a Swedish company ........... You can check out streaming websites like Tidal as well ........

dalethorn's picture

I sampled her album on Amazon, iTunes, and CDBaby. There definitely is not enough resolution to validate a microphone's quality in this recording, even if it were on a CD. The recording engineers would need to remaster this album and release it in higher resolution.

Bogolu Haranath's picture

Did you check hi-res download websites? ...........

dalethorn's picture

Yes - none are available in the U.S. But as I noted, I think the recording itself is very shy on featuring percussion detail and impacts, or whatever else would show off a mic's ability to capture detail and impact with great clarity. In the U.S., her recording is far enough outside of the mainstream that I doubt we'd ever see a remaster.

Bogolu Haranath's picture

Also, another album by Dirty Loops "Loopified" is also supposed to be made using Ehrlund mics, I downloaded from iTunes ........ It is kind of hard rock album ......... If you like, you can check that one too ........ Marklund released another album "Brinnande" the same year 2014 with similar type of music, vocals and drums .............

dalethorn's picture

In the 1997-98 MLB seasons, Mark McGuire and Sammy Sosa were beating the crap out of Babe Ruth's home run record (or Roger Maris' asterisked record), and you could hear McGuire saying "Unbelievable" every five minutes on the news those years. In a dissimilar way, the Dirty Loops album is also unbelievable. To that end, I downloaded one track "It hurts", and indeed it does.

Watching the kbps meter on Foobar2000, the highest value I saw was 343 kbps, so I assume the container is somewhere above 320 kbps, which when converted to WAV format is suitable for evaluating the gear you mentioned.

BTW, my Vox player showed 271 kbps with no variation as the track played, which is a clue that the meter you are using is also frozen on the first value it reads, so that you're not able to see the maximum container rate for the VBR (variable bitrate) track you're testing.

The quality of the percussion I hear is OK, but typical for the other EDM and hard-rock tracks I have. Not being familiar with enough of their music compared to other similar artists, I can't say anything else.

Bogolu Haranath's picture

Are you listening to this new music with Ehrlund mics on the headphones? Are you using any EQ? Are you running this new music through EQ? If so, I would suggest listening with and without EQ, if that is possible ..........

dalethorn's picture

Using without EQ, with the iTunes player and Vox player, from my Macbook Pro into the DragonFly Red DAC and then into the Senal 1200 headphone. Sometimes I add an amp to the DragonFly if I need more power or definition.

Basically what's needed is a recommendation for a specific track and time within that track, to identify a particular sound. The alternative would be to compare 2 different masters, if they were done differently and sounded different.

Bogolu Haranath's picture

Ok ,,,,,,,, Thanks for the info ........... I am using my Mac Pro also ........ I have Dragonfly Red also ........ I am using the new Hugo2 with Audeze 'phones, for this new music ......... I got the new Rogue RH-5 headphone amp which was very favorably reviewed in Stereophile ........ You know tubes add a little bit of "magic" to the sound ........ I am gonna try the Rogue sometime this weekend with all the 'phones including the Utopia and Lcd-4 ......... You are right ....... We need more examples and more music for making any valid judgements .......... By the way, did you read the new Keith Howard article about measurements in Innerfidelity? ........ I am all excited about the improved measurements, he is gonna make ..........

dalethorn's picture

I have not - thanks for that tip. I believe that many if not most of the people who occasionally read Tyll's posts missed a lot of his points, especially how even the flagship headphones vary from sample to sample.

The thing Tyll was good at was describing the actual sound of each headphone, compared to neutral. You had to read between the lines of course, because he didn't like to be brutal in most of his reviews unless it was Ultrasone or some similar thing.

The thing that is MOST needed in headphone measurements is to normalize the high frequency part of those curves, because of the ragged lines that represent "ear pinnae" response. It's absurd - if the headphone sounds neutral, it probably is neutral, and those ragged lines are not only misleading, but they disguise a lot of bad headphones. It's not at all difficult to flatten those lines - just average the response of 30 or so flagship headphones and then report how each individual headphone varies from that.

As far as the dirty loops recording or any of the others is concerned, it might be a real challenge to get some good test tracks, here in the U.S. at least, since the technicians already are set with their favorite brands of recording gear. I usually get mine at sweetwater in Fort Wayne IN.

dalethorn's picture

BTW, I did go back to the "It hurts" track *with* EQ this time, using the Shinola On-Ear headphone, and it sounds better with a corrected/neutral response. Much better, with good impacts.

I checked the Innerfidelity article, and I tend to agree that the direction it's going to go will be away from a hardcore site like Head-Case, and it'll become more inviting to a wider range of audiophiles who aren't into the Tyll thing. It sounds like a good plan.

Bogolu Haranath's picture

Glad you did try with EQ ....... Try it with Focal Elear ........ When you get the Focal Clear, I think you can get even better sound ......... You can try the Eva Marklund, "Stralande" as well, may be couple of tracks ......... She has a very good voice and the drums sound very good as well ........ About Innerfidelity - The new measurements should help better analysis and understanding about the sound of the 'phones .........

dalethorn's picture

The Focal Clear would sound better out of the box than the Shinola, but once EQ'd, the Shinola detail is about as good. If you really want to hear the ultimate in upper harmonics detail with a headphone, try the Sennheiser HD800 with a decent tube/valve amp, and play Chesky's song Girl From Guatemala. At the 3:00 point is the most outrageously sparkling burst of treble percussion you'll likely ever hear. A solid state Lehmann Black Cube amp also does very well with this and the HD800.

Bogolu Haranath's picture

For the price of Focal Utopia, you can get both Focal Clear and Rogue RH-5 ......... $4000 ........ Now, that combination is what I would call "audio utopia" .......... You can use AQ-DF Red or Hugo2 to connect between the laptop/desktop and RH-5 ...........You can use a portable DAP/phone as well ......You can always use EQ to further enhance the sound, of course ......... HR would probably agree with me about Clear and RH-5 combo ......... He is listening (or listened) to that combo ...........

dalethorn's picture

I read that article, but did not relate to the RH5 as I usually relate to the more compact (and often battery-powered) headphone amps. A few years ago I got a DAC plus headphone amp in two boxes, but both of those ran off of "wall-wart" AC-DC transformers. So I rigged up batteries with the correct voltage and tested the DAC and amp that way compared to running off of AC mains current feeding the AC-DC adapters.

I got lower noise and cleaner sound running from the batteries, and I finally figured there were two main factors in a user's satisfaction when running amps from the mains - one would be the quality of one's internal power supplies and isolation, and the other the quality of the mains current. I rarely trust the former, and never the latter which is almost always bad.

Pages

X