Julf
Facts are our friends
Yes, that would be ideal. Unfortunately MQA is not making their proprietary encoder, or any details about it, available unless you sign a NDA.
MQA 6 months later- have thoughts evolved?
- Thread starter bec143
- Start date
Yes, that would be ideal. Unfortunately MQA is not making their proprietary encoder, or any details about it, available unless you sign a NDA.
DNM
pfm Member
And it would appear that for some who did sign the NDA and had a good look under the bonnet, they regret doing so because they're now gagged from voicing their honest opinion about the methodology.
camverton
pfm Member
Yes, I was referring to the testing method used by Archimago, which was described as reasonable. For coming to any worthwhile conclusion I dont think reasonable is sufficient.
Perhaps we are at the stage where it is sensible to reserve judgement
This is why I started the thread, not to se if anyone else had read the Archimago review, (which I found to be quite good), but to see if people who had listened to MQA for the past 6 month have evolved an opinion over time that differs from their early thoughts or the consensus at the time.
I have listened to a lot of MQA content, but casually at work and only when I happen to hit on something MQA on Tidal. I certainly can't listen critically this way (iMac/E2/Audioengine desktops, although it can actually sound way better than you might think! However, there are some MQA albums that sound radically different - When I get back to these and do some careful listening at home, I think it's all, or nearly all, due to mastering. Things just sound way too different to be anything else.
SO I don't think that I have experienced any convincing advantages of MQA thus far
DNM
pfm Member
Sorry yes, NDA/underlying tech being separate to Archimago's testing methodology.
https://docs.google.com/spreadsheet...-Oo2MjIa3orv9DKZfwiRQKmTAA/edit#gid=945476039 - this is a dynamic list, regularly updated.
MQA also claims to reverse problems introduced in the original analogue recording process by identifying microphone and reel-to-reel artefacts and reversing them. So it is not just applied to A2D processes. How it works I cannot explain technically but I can imagine logically. It does require some accurate metadata (what equipment was used in which session, etc.) for sure.
Your analogy is flawed but, if you believe the claims for the technology, then yes. If MQA applies multiple correction curves at the remastering process to correct for mikes, tape records and A2D processors, then applies further correction curves to correct for the D2A characteristics then it is, in effect, unscrambling eggs.
I wonder if Devialet are challenged to the same extent when they claim to correct for the playback speakers...
DNM
pfm Member
Hmm, think about it. MQA applying correction for microphones and ADC's used in the 1970's onwards vs Devialet applying correction to contemporary speakers that they themselves have actually measured in the lab today!
Does MQA really have any idea which Mic's and ADC's were used for every recording they've MQA'd? Of course they don't, its bollox. If you enjoy the lossy distortion effects and marketing story then good for you, but don't kid yourself or others that its correcting the timing anomalies in the recordings.
Speakers and their interaction with the room have a far greater impact on what MQA claims to solve than MQA itself. Just look at the measured impulse responses in Archimago's test. Active speaker crossovers and driver time alignment, DRC, not to mention passive room treatments, have a far greater impact on correcting impulse response than MQA. If you think Archimago's measurements are bogus then complain on his website. No one appears to have done so BTW.
camverton
pfm Member
There were a few people who challenged Archimago on his website and had a very interesting technical discussion amidst various somewhat infantile insults from a few supporters of Archmago and detractors of MQA.
That Bob Stuart has been heavily involved in developing MQA leads me to at least keep an open mind. As I understand it the lossy part of MQA is above 22kHz, so probably not important.
What I can't understand is why there is so much hostility to MQA. In theory, at least, it sounds like the analogue/digital haters dream, assuming that one accepts that timing is one of things that makes digital so bad compared with analogue - and that means pretty old recordings.
My own comparisons of CD standard versus hi-res downloads have not shown any benefit in me paying extra for high res, although comparing the files did show that there were actual differences even if I couldn't discern them audibly. As with so many of these discussions it is all a moot point, given the enormous orders of magnitude more that choice of speaker and room make.
It is that, that makes me wonder if MQA is worthwhile for me, although I can see a definite advantage for those streaming in an area with poor broadband speeds.
That Bob Stuart has been heavily involved in developing MQA leads me to at least keep an open mind. As I understand it the lossy part of MQA is above 22kHz, so probably not important.
What I can't understand is why there is so much hostility to MQA. In theory, at least, it sounds like the analogue/digital haters dream, assuming that one accepts that timing is one of things that makes digital so bad compared with analogue - and that means pretty old recordings.
My own comparisons of CD standard versus hi-res downloads have not shown any benefit in me paying extra for high res, although comparing the files did show that there were actual differences even if I couldn't discern them audibly. As with so many of these discussions it is all a moot point, given the enormous orders of magnitude more that choice of speaker and room make.
It is that, that makes me wonder if MQA is worthwhile for me, although I can see a definite advantage for those streaming in an area with poor broadband speeds.
davidsrsb
pfm Member
Devialet are working with the speaker manufacturer and with actual examples of the speaker. DSP correcting the speaker amplitude and phase anomalies is well proven.
MQA can only really hope to correct for ADC errors on the capture of the master tape. The construction of that master tape with many microphones, sometimes in more than one studio, complex effects and unknown post editing is too far removed and rarely documented to be practical.
I asked for a list of albums recorded and mastered in MQA. The vast majority of the albums on your list were recorded and mastered before MQA was invented, many indeed were recorded before digital audio was invented. (cf Bill Evans, Charlie Mingus, Cream ..).
So, how about a list of albums recorded and mastered in MQA.
Objective answer:
As I note in this thread above, no they don't; but they claim that when they do have those data, they can, the result is termed a "white glove" remaster. Where they do not have the full metadata they say they make informed guesses. Likewise on playback, the first unfold extracts the corrected source while the second unfold corrects for the downstream DAC where known, otherwise again an informed guess is made. I am not trying to kid anyone, I am merely repeating the claims and descriptions of the technology and how it works which I have read from multiple sources and discussed with people close to the technology.
Subjective answer:
I have listened at length to many different MQA recordings via Tidal on high quality systems. Yes I have enjoyed the results, yes they can sound very different (for whatever reason), and generally I find them more detailed and more musical then CD or non-MQA streamed versions. I also recently enjoyed immensely the distortion effects of a Prima Luna valve amplifier through Spendor SP1 speakers fed with vinyl by a Heybrook TT1.
Werner
pfm Member
1.Shady claims.
2.Based on bad science.
3.Technically not even necessary.
4.Obsoletes existing DACs and/or media players, creating landfill.
5.Incompatible with existing DSP, creating landfill.
6.A grab to control all, from music to replay gear, and extract money at all stages (*).
7.Opposing the recent trend of open source standards.
8.Internally contains DRM mechanisms that could be activated in the future.
Note how this list does not contain 'sound quality'. That is because MQA's sound quality does not matter. It could be magnificent (**), MQA would still be bad for all.
(* https://s3-eu-west-1.amazonaws.com/...yv0B+lkEvRUb7WH2zB4fJVvUNO3o6YpGT+iEo1e6vzwU=, section Business Review)
(** It isn't.)
Werner
pfm Member
It has been established so far that the second unfold really is just upsampling, and that the filter used for this task is dictated by the music, and not by the DAC downstream. In other words, so far there has been no evidence of correction for the replay DAC, quite the contrary.
camverton
pfm Member
Thanks Werner, an interesting list!
I’m glad the industry didn’t take too much notice as I would still be listening to my classical music with the numerous frustrations of vinyl rather than digital...
As for sound quality, the range of anecdotal experience is too great to be able to draw any conclusion. I guess it’s something we’ll all have to judge for ourselves; pity my explorer2 DAC died. Is there a list of MQA enabled DACs anywhere?
Edit: 21 manufacturers according to the MQA website, including some very well known names; I thought someone here claimed that manufacturers didn’t want to know. If anything, this thread is encouraging me to try this for myself. Although, as I mentioned earlier I am very happy with the setup I have which includes many CDs ripped to FLAC. Decisions, decisions.
I’m glad the industry didn’t take too much notice as I would still be listening to my classical music with the numerous frustrations of vinyl rather than digital...
As for sound quality, the range of anecdotal experience is too great to be able to draw any conclusion. I guess it’s something we’ll all have to judge for ourselves; pity my explorer2 DAC died. Is there a list of MQA enabled DACs anywhere?
Edit: 21 manufacturers according to the MQA website, including some very well known names; I thought someone here claimed that manufacturers didn’t want to know. If anything, this thread is encouraging me to try this for myself. Although, as I mentioned earlier I am very happy with the setup I have which includes many CDs ripped to FLAC. Decisions, decisions.
MQA attempts (but according to most people fails) to deliver hi-res music in the bandwidth of red book. If you put 24/96 in you don't get the same out. Let's suppose you think there's an advantage to a hi-res master. Why not just buy it? And if you like classical music notice that Qobuz now supports hi-Res streaming up to 24/96, so you could stream from a vast catalogue of hi-res classical masters for a modest monthly subscription. No magic DAC or vaguely unspecified jiggery-pokery.
MQA is dead in the water. It is a bad solution to a non-existent problem.
camverton
pfm Member
Good points AndyU. As it happens I am not convinced by the need for hi-res based upon my limited tests. The “timing” issue sounds interesting and maybe worth following up. As for streaming from Qobuz I am considering a move to a more rural area where broadband speeds are still not up to the task and of course the Qobuz catalogue has many gaping holes; Hyperion for one.
If 21 DAC manufacturers are including it in their products then I suspect it is, for better or worse, far from dead in the water.
If 21 DAC manufacturers are including it in their products then I suspect it is, for better or worse, far from dead in the water.
Jim Audiomisc
pfm Member
It might have been possible to initially take a more positive view of MQA if they hadn't gone down the same line as HDCD. i.e. Publishing some 'generalised' details in patents so they can commercially control and gain income while *symultaneously* keeping key details of the implimentation secret. An attempt to have the bun *and* the sixpence.
Add that into the list Werner gives and you can see why mere engineers are wary. What you see may not be what you get. HDCD also made some claims about fiddling about with the reconstruction filter - details vague. Then after enough years had passed it became open and clear that in reality this simply wasn't done. Plus the HDCD process *degrades* non-HDCD replay. The 'peak shaping' applied by HDCD will alter the shapes of waveform peaks when relayed without HDCD decording, making them sound different to an unencoded plain LPCM version.
The above is a real PITA when you come across recordings that are *only* available as HDCD - even when not identified as such at the point when you buy them. Some have even lost their key ID so can't be decoded and played correctly on an HDCD player. Example of the way something like this can give the biz another way to mess up.
And has been demonstated, you can get much the same reductions in HQ audio file sizes and stream rates *without* MQA or any secret or proprietary methods, and also avoiding some of the problems indicated by the MQA patents. So there is no need for MQA.
The idea that MQA can routinely 'correct' for the entire kit and mic setups of ancient recordings strikes as being a fantasy. Might be possible for a few special cases. But in general, a fantasy. The require info simply won't be available, even if the recordings contain any HF anyway. Many old studio mics have a response that is heading south before you get to 20kHz. And heaven knows what the Dolby A systems will be doing in many cases. To just give two examples.
Add that into the list Werner gives and you can see why mere engineers are wary. What you see may not be what you get. HDCD also made some claims about fiddling about with the reconstruction filter - details vague. Then after enough years had passed it became open and clear that in reality this simply wasn't done. Plus the HDCD process *degrades* non-HDCD replay. The 'peak shaping' applied by HDCD will alter the shapes of waveform peaks when relayed without HDCD decording, making them sound different to an unencoded plain LPCM version.
The above is a real PITA when you come across recordings that are *only* available as HDCD - even when not identified as such at the point when you buy them. Some have even lost their key ID so can't be decoded and played correctly on an HDCD player. Example of the way something like this can give the biz another way to mess up.
And has been demonstated, you can get much the same reductions in HQ audio file sizes and stream rates *without* MQA or any secret or proprietary methods, and also avoiding some of the problems indicated by the MQA patents. So there is no need for MQA.
The idea that MQA can routinely 'correct' for the entire kit and mic setups of ancient recordings strikes as being a fantasy. Might be possible for a few special cases. But in general, a fantasy. The require info simply won't be available, even if the recordings contain any HF anyway. Many old studio mics have a response that is heading south before you get to 20kHz. And heaven knows what the Dolby A systems will be doing in many cases. To just give two examples.