[pinnahertz]

For those saying MQA is lossy:

You are technically correct, but the only difference is an improvement in sound quality. Let me explain. For every MQA album released, information about the original recording equipment is stored in the file. They do this so they can correct known problems with the equipment when it's processed through an MQA enabled dac. That means the resulting file will technically not be the same as the originally recorded and released song. That technically means it's lossy, by definition, but it's not lossy in the way most people understand the term.

It's wonderful how much it improves the sound quality, especially for older albums. But, you've got to have an MQA enabled dac to really hear an improvement. Otherwise, it'll sound the same as the hd counterpart file with a smaller file size (which is a good thing)

Actually no, that's the MQA coolaid again. There is no information about the recording equipment stored in am MAQ file because that information is not available or even fully known. There is no way to characterize the temporal response of a composite system from a bit of audio recorded with it.

You've also blended two different MQA "features", the "temporal blurring correction" and the lossy codec. They are not the same thing. The "temporal blurring correction" is nonsense because it's not possible to know what to correct for in the entire chain. Their lossy codec is good, but no better than many others.

Any audible improvements in an MQA file are likely the result of remastering, and would be audible equally with any DAC, lossless or mildly lossy codec.

 

[Basshead Paul]

Actually no, that's the MQA coolaid again. There is no information about the recording equipment stored in am MAQ file because that information is not available or even fully known. There is no way to characterize the temporal response of a composite system from a bit of audio recorded with it.

You've also blended two different MQA "features", the "temporal blurring correction" and the lossy codec. They are not the same thing. The "temporal blurring correction" is nonsense because it's not possible to know what to correct for in the entire chain. Their lossy codec is good, but no better than many others.

Any audible improvements in an MQA file are likely the result of remastering, and would be audible equally with any DAC, lossless or mildly lossy codec.

Quick question: have you ever tried listening to MQA files through an MQA enabled dac? I'm not trying to be a smartass, I'm just wondering.

The equipment used for almost all recordings is archived and accessible by contacting the producer. MQA gets the artist, producer, etc to sign off on the MQA album before releasing it, so the information could be obtained at that point.

In any case, I notice that most naysayers have never tried actually listening to MQA files through an mqa enabled dac. The temporal deblurring isn't present unless you listen that way. The audio quality is noticeably better when I listen that way. It's not subtle.

 

[pinnahertz]

Quick question: have you ever tried listening to MQA files through an MQA enabled dac? I'm not trying to be a smartass, I'm just wondering.

Before I answer that quick question, I would ask this two-part question to you:

Part 1: In your comparisons between MQA and non-MQA, were those comparisons done double-blind with all biases accounted for?
Part 2: When making those comparisons, did you confirm that the MQA versions originated from the same masters as the non-MQA versions with no re-mastering or any other differences?

I already know the answer to part 2, because there have been no MQA releases with provenance verified by a third party that confirm this condition. And I strongly suspect I know the answer to Part 1 too.

Now, given the situation that there has been no proof that MQA does anything at all to improve the sound of any recording (and marketing is not proof), then I don't really see why I would invest in any MQA equipment. Proof is easy, and if it's that big an "improvement", it should be really, really easy. Yet, there remains none.

The equipment used for almost all recordings is archived and accessible by contacting the producer. MQA gets the artist, producer, etc to sign off on the MQA album before releasing it, so the information could be obtained at that point.

This issue has been addressed before. It's simply not true that there is equipment info for anything but a very tiny group of recordings. But there's a problem there too. You might generate an equipment list for a recording that might even be complete. Heck, I can do that for any recording I've ever engineered or produced. But what I can't tell you, and what nobody can tell you is how that gear was used or adjusted. Every equalizer has a temporal response, and every adjustable control of every equalizer changes it. The number of passes the signal went through A/D then D/A conversion, including a filter at each step, is unknown, especially with earlier recordings where even though digital recorders were used, mixing, mastering and effects required D/A then A/D who knows how many times and through what filters. Every A/D and D/A has a filter, and they are not all the same. Passing through multiple conversions compounds temporal responses of each. And, speaking as someone in the industry, I can tell you there are no records for this other than distant and fading memories. And we aren't even including the response, both amplitude and temporal, of the monitoring systems used on which the mixes were built in the first place. Why would you want to correct for something that was already considered in the final mix?

The entire scheme is bollox. And, as with every aspect of MQA, there has been no verified third party proof. Why? We can't get our mitts on the encoding process! So we can't test the whole MQA chain for audibility.

In any case, I notice that most naysayers have never tried actually listening to MQA files through an mqa enabled dac. The temporal deblurring isn't present unless you listen that way.

Sorry to inform, but temporal deblurring isn't present that way either because the degree required is unknown. In fact, "de-blurring" is a term made up by MQA that exists nowhere else in the audio industry. And it's well coined for the purpose, incorporating a strong negative connotation without any actual substantiation like a well constructed technical paper, for instance.

The audio quality is noticeably better when I listen that way. It's not subtle.

I recognize you think you hear a difference. I recognized there may actually be a difference. There's no way I, you, or anyone so far can prove it has anything to do with any specific MQA process. But until someone can pull of a controlled ABX/DBT and show a reliable audible difference (we can discuss later if it's an improvement or not), there's no reason to attribute any audible differences to any part of MQA.

Now to answer your question, I don't own any MQA enabled gear. The MQA enabled auditions I've heard have been inconclusive.

 

[castleofargh]

For those saying MQA is lossy:

You are technically correct, but the only difference is an improvement in sound quality. Let me explain. For every MQA album released, information about the original recording equipment is stored in the file. They do this so they can correct known problems with the equipment when it's processed through an MQA enabled dac. That means the resulting file will technically not be the same as the originally recorded and released song. That technically means it's lossy, by definition, but it's not lossy in the way most people understand the term.

It's wonderful how much it improves the sound quality, especially for older albums. But, you've got to have an MQA enabled dac to really hear an improvement. Otherwise, it'll sound the same as the hd counterpart file with a smaller file size (which is a good thing)

you're talking about an album made almost from scratch processed for that issue before it was mixed or anything, as that's the only moment where such treatment could be done effectively. so first you're wrong about where this is applied if applied at all. the original tracks(tracks, not songs) would need to be processed one by one, then mixed, mastered and released as MQA. everything done long before it reaches the DAC. but do you see the amount of technical and artistic work required? if that's how all MQA file was born, the actual catalog would hardly be reaching even a hundred songs right now. not the 30000 or more claimed by Warner alone. they went from discussing the thing to having that many files in almost no time. which makes it obvious how they batch processed the all thing.
I'm far from being the all seeing eye, but I have seen no evidence that this was done, only that it's one of the multitude of stuff they said they could at some point decide to do if their dreams of imposing MQA into every part of the music industry was to come true(plz no!). I've seen them announcing like 2 such albums so far, recorded for MQA from scratch, but that's also different as they might just encode the way they like instead of trying to correct another DAC using a usual band limiting they call wrong. so don't expect your MQA library to be made that way. it's not.


as for lossy format, it has to do strictly with the encoding/decoding fo the format. if you take a signal, convert it to MQA and then extract it back, you will not get the original signal. => it's a lossy format.
and that lossy effect has nothing to do with anti ringing filters or correcting self proclaimed flaws in the production process, it's purely due to the method to store part of the information for the ultrasonic content.
it's a definition, in no way does it quantify fidelity of a playback system or even the amount of loss within the format conversion. you can take a 24/192 file, turn it into 8bit/22khz PCM if you like, then encode it into FLAC and send that after extraction to the DAC using asio or wasapi, run a nonsense low pass filter in the DAC and gazillion factors of oversampling, cripple it with noise in the amp, drown anything below -50dB with THD and alter the signature by up to 20dB with the headphone, and still say that you're using a bit perfect system and a lossless format. ^_^ that would technically be right.
people decide to associate a heavy load of psychological ideas to certain terms, but those terms define very specific areas in a very specific way. they're not "feel good" words. I personally dislike or disagree with much of what MQA stands for, but that lossy format thing isn't one of them.

 

[gregorio]

[1] For every MQA album released, information about the original recording equipment is stored in the file. [2] They do this so they can correct known problems with the equipment when it's processed through an MQA enabled dac. [3] That means the resulting file will technically not be the same as the originally recorded and released song. That technically means it's lossy, by definition, but it's not lossy in the way most people understand the term.
[4] The equipment used for almost all recordings is archived and accessible by contacting the producer. [5] MQA gets the artist, producer, etc to sign off on the MQA album before releasing it, so the information could be obtained at that point.

1. No it's not!
2. Which is completely irrelevant because the known problems of the equipment is completely insignificant compared to how that equipment was actually used. Take for example microphones, which MQA states it can correct for: The timing imperfections or "temporal blurring" of the mic itself is irrelevant, how the mics are positioned relative to the source and the other mics makes hundreds of times more difference than the timing imperfection of the mics themselves and more than a hundred times more difference than anything the ADC or DAC do to the timing! When it comes to mixing, if I hear a timing imperfection I will choose to correct it manually or deliberately leave it as it is because I find it artistically more pleasing, the very last thing I want is for a codec to come along and apply a correction to something that I've already corrected or specifically do not want corrected!!!
3. No! By definition and taking the previous point into consideration, what that technically means is low/poor fidelity, it has nothing to do with being lossy. MQA is lossy for an entirely different reason, because it looses data during encoding which cannot be perfectly recovered upon decoding.
4. Now you're just making stuff up! The equipment used is hardly ever archived and on those rare occasions when it is, the producer wouldn't have that information, the various studios involved would. However, as mentioned, the equipment is irrelevant, how it was used is what's important and that is NOT archived, MQA couldn't do anything with that data even if it were and even if MQA did have that data and were able to use it for correction, that's just about the very last thing I as an engineer would want! This also highlights another MQA lie, that MQA allows you to hear it how the studio/artists heard/intended it. What I and the artists heard in the studio was the mix which included all these supposed timing errors (temporal blurring), if MQA comes along and supposedly corrects all those supposed timing errors then by definition it is deliberately is not reproducing what I and the artists heard in the studio. These two claims are mutually exclusive, at least one of them MUST be a lie!
5. No, that's just something else you've made up (or been mislead to believe). MQA just gets the recording right's holder to sign off on the album, which in the vast majority of cases would be the record label. The artists, producer and engineers are not even consulted, let alone have any say in the matter and they wouldn't have the relevant information anyway.

You are falling for the typical audiophile marketing ploy; focus on an irrelevant detail, make it seem important while ignoring the elephant in the room and then promise to fix that irrelevant detail. It's like promising to fix a paint imperfection on the trunk of a car which is only visible under laboratory lighting and ignoring the fact that the front of the car is completely wrecked from a head on collision with a truck!

G

 

[old tech]

And we aren't even including the response, both amplitude and temporal, of the monitoring systems used on which the mixes were built in the first place. Why would you want to correct for something that was already considered in the final mix.

Ahh, temporal blurring. A marketing concept that has no basis in audio engineering. A term designed for the gullible section of the audiophile world who have no understanding of music recording, distribution or playback. Not the first time a science sounding marketing idea has been used for this purpose and it wont be the last either.

 

[old tech]

Quick question: have you ever tried listening to MQA files through an MQA enabled dac? I'm not trying to be a smartass, I'm just wondering.

The equipment used for almost all recordings is archived and accessible by contacting the producer. MQA gets the artist, producer, etc to sign off on the MQA album before releasing it, so the information could be obtained at that point.

In any case, I notice that most naysayers have never tried actually listening to MQA files through an mqa enabled dac. The temporal deblurring isn't present unless you listen that way. The audio quality is noticeably better when I listen that way. It's not subtle.

Perhaps you should consider participating in this blind test.
https://archimago.blogspot.com.au/2017/07/internet-blind-test-mqa-core-decoding.html

 

[goodvibes]

Well, you can objectively see the MQA file isn't bit perfect. You are arguing perception is not valid then offering a perception test. The test is fine at face value but why FLAC and not the wav file that the MQA was made from? Even the lossless compression from the download sites could be questioned. The reason some companies don't offer MQA is because they have gone to Meridian with their own master files, tested and noticed loss.

To me it's not about whether MQA is as good or better than the uncompressed formats. It isn't. It's whether it's good enough to benefit services like Tidal which is bit rate limited. Problem is that it's much ado for something like that and raising the price of any dedicated kit that uses it. That and Meridian is demanding a look at proprietary aspects of competitor's equipment that would use it. Completely unnecessary.

I have never gotten the best result for evaluating anything via a PC and USB DAC. Fine if that's what you got and want to listen to music which is what it's all about but it's a weak link in any subjective comparison. That's subjective, LOL.

 

[audiobill]

Peter from Lumin recently compiled this list of Tidal MQA albums, currently showing over 3300 titles.

 

[old tech]

^Just on the issue of the marketing term "temporal smears", my broad understanding of this is that MQA claims to reduce or eliminate timing errors in AD/DA conversions in the production chain.

As these timing errors are so small, even with early 80s convertors and beyond the capability of our ears to hear it what is the big deal? But a bigger issue is that most of the MQA offerings as listed in the link from the member in the post above this, are recordings made on analog tape recorders. With the article below as an example (see under "The Reel Thing"), it is well understood that timing errors from the tape recorder, the tape and the tape to tape head contact is by orders of magnitude far greater than from any ADC or DAC. So how does MQA deal with those far more significant temporal smears?

https://www.soundonsound.com/techniques/analogue-warmth

 

[pinnahertz]

The reason some companies don't offer MQA is because they have gone to Meridian with their own master files, tested and noticed loss.

Do you have proof of this statement, a quote, or a link to a "story" somewhere?

 

[pinnahertz]

Peter from Lumin recently compiled this list of Tidal MQA albums, currently showing over 3300 titles.

Ah, so they're closing in on something like <.5% then.

Noise in the stats.

 

[audiobill]

Ah, so they're closing in on something like <.5% then.

Noise in the stats.

My reason for sharing that list was simply to show that there are significantly more MQA albums on Tidal than the several hundred currently shown in their Masters tab. Sorry the conversion isn't happening faster but at least there's some progress being made and I thought it would be of interest to readers of this thread. Happy listening to all in whatever format you prefer, it's all about the music!

 

[Basshead Paul]

Perhaps you should consider participating in this blind test.
https://archimago.blogspot.com.au/2017/07/internet-blind-test-mqa-core-decoding.html

It isn't possible to do a blind test of MQA that way. Core decoding is only possible if you're listening through an MQA dac.

As a result, that test is only comparing MQA files which have been unfolded once, through the software decoding part. The files won't sound much different than normal hi res files at that point. They'll just be smaller in file size, which can still be useful.

 

[castleofargh]

Do you have proof of this statement, a quote, or a link to a "story" somewhere?

I guess he's refering to the Highresaudio website who played turncoat a few times. they supported MQA, then said they would stop because after some personal tests they noticed variations and said they wouldn't support something that isn't lossless. and now I believe they're supporting MQA again.
so maybe not the most convincing voice in the crowd.





to my mischievous crew of Sound Science: it's an appreciation thread! try to limit your answers to correcting technical mistakes, and leave the anti MQA arguments outside of it. it's annoying when people turn a sound science post into an appreciation thread, let's not do the opposite here.