1. WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Archimago's MQA listening test results

Discussion in 'Psychoacoustics: Science of How We Hear' started by Blumlein 88, Sep 23, 2017.

  1. Fitzcaraldo215

    Fitzcaraldo215 Addicted to Fun and Learning

    Joined:
    Mar 4, 2016
    Messages:
    774
    Likes Received:
    159
    I think I have answered that. If they did it themselves or commissioned it, it would be seen as tainting the results. For all I know, they or their licensees have already done this, and I would be surprised if they had not already have done it. But, the firestorm of resistance it would receive upon publication might not make business or PR sense. It must be as independent as possible.

    Tobacco money or Rupert are distracting red herrings that contribute nothing to the debate. I don't like tobacco money either. But, IMHO, it weakens any arguments you might have against the technology itself by playing that card, unless you believe in simple guilt by association. Who is to know whether Rupert is an active or a passive investor, unless you have been in the board room? Yet, you assume he is "in control".
     
  2. Fitzcaraldo215

    Fitzcaraldo215 Addicted to Fun and Learning

    Joined:
    Mar 4, 2016
    Messages:
    774
    Likes Received:
    159
    It is? Possibly so, but where has this been objectively established beyond doubt? The Archimago tests avoid the question.
     
  3. svart-hvitt

    svart-hvitt Active Member

    Joined:
    Aug 31, 2017
    Messages:
    160
    Likes Received:
    22
    Anthony Edward Rupert is an MQA director. This is hands-on investing by the Ruperts. Owners are in control.

    If you mean that ownership control, director position are a red herring in understanding the business logic of a corporation, I’m afraid you have little experience in said matters.

    I think it’s relevant to judge an investor’s ethics. More and more people put weight on social responsibility these days and BAT is a continuous offender of social responsibilty. The Ruperts could have sold their BAT stake long ago but are still leveraging on the dividends from BAT.

    What concerns the technology, the only blind test I know of says MQA and hi-res is a 50/50 proposition. Like CD vs hi-res, which you already knew.

    What concerns MQA’s claim that MQA is informationally better than the original master (cfr. 2L’s/Morten Lindberg’s listening note and the claims of the press), it’s a violation of kindergarten-level logic.

    What concerns the euphonic value of MQA vs other formats; pick a number!

    EDIT: You are of the opinion that the Ruperts and their money is a distraction and that focus should be on the technology. I am afraid you are being fooled or fool yourself. MQA is not about technology at all. MQA is in fact a distraction. Focus should be on the potential profit of the MQA scheme.
     
    Last edited: Sep 30, 2017
  4. Fitzcaraldo215

    Fitzcaraldo215 Addicted to Fun and Learning

    Joined:
    Mar 4, 2016
    Messages:
    774
    Likes Received:
    159
    Look, I have been on the boards of privately held investment entities. Ownership of an investment interest does not nessarily equal control of day to day decisions, particularly when it comes to the esoterica of technologies not understood by the average Joe.

    Yes, I feel your pain about this investment by an evil tobacco empire. But, it proves absolutely nothing about what is going on with MQA, unless you can site some evidence other than that they are investors.
     
  5. Blumlein 88

    Blumlein 88 Major Contributor

    Joined:
    Feb 23, 2016
    Messages:
    1,942
    Likes Received:
    472
    So the following was posted by your evil twin:

    Exactly, and all music being composed entirely of transient signals means that music features many many such transition zones.

    That seems the key audiophile question. What's your evidence for asserting that we cannot hear it? Meridian has pointed to academic studies indicating that we can hear it. I've come to no conclusion myself.

    The evidence you cannot hear it is it happens at a frequency we cannot hear. While you said you had reached no conclusion you did address the hearing of it.
     
  6. svart-hvitt

    svart-hvitt Active Member

    Joined:
    Aug 31, 2017
    Messages:
    160
    Likes Received:
    22
    I agree about the opacity of MQA as a business; we need to use publicly available information and common sense to make up our minds.

    What concerns MQA as a consumer-oriented technology, it’s a non-event.
     
  7. Fitzcaraldo215

    Fitzcaraldo215 Addicted to Fun and Learning

    Joined:
    Mar 4, 2016
    Messages:
    774
    Likes Received:
    159
    If it is a non-event, why get so whipped up about it?
     
  8. svart-hvitt

    svart-hvitt Active Member

    Joined:
    Aug 31, 2017
    Messages:
    160
    Likes Received:
    22
    It’s a technological non-event. It’s the smoke and mirrors on the business side that’s intriguing.
     
    Last edited: Sep 30, 2017
  9. Ken Newton

    Ken Newton Active Member

    Joined:
    Mar 29, 2016
    Messages:
    165
    Likes Received:
    33
    I primarily addressed the transient response implications of an MQA channel. I raised the question of whether or not transient spreading might be subjectively audible with music as a secondary point at best. Even though that is, of course, the end question of the discussion. As you say yourself, I explicitly stated that I have reached no conclusion myself, so, no evil twin is required to confuse that simple distinction, unless it's your own?

    Meanwhile, I'm curious as to whether you perceive 192kHz PCM as sounding better than RBCD, and if so, why do you suppose that is? According to your view, it certainly couldn't be due to any ultrasonic frequency content, whether program related or brickwall FIR filter ringing related as these should be inaudible in themselves, yes? What do you see as the possible explanations?
     
    Last edited: Oct 1, 2017
  10. Blumlein 88

    Blumlein 88 Major Contributor

    Joined:
    Feb 23, 2016
    Messages:
    1,942
    Likes Received:
    472
    Don't hear the 192 as different when the same source or master is used. In some cases recordings I have made so I know what happened to the signal. Don't hear any benefit above 48 khz. I am old enough I am sure my hearing stops before 20 khz.
     
  11. Fitzcaraldo215

    Fitzcaraldo215 Addicted to Fun and Learning

    Joined:
    Mar 4, 2016
    Messages:
    774
    Likes Received:
    159
    I am an old fart no doubt with diminished hearing. However, I do hear a small, subtle but noticeable and preferable improvement in hi rez with increasingly diminished returns above 88/96k. There is no slam dunk, obvious improvement from hi rez, however. The Joshua Reiss meta analysis paper of a few years ago provides an excellent summary of the empirical evidence. Some test subjects did hear it, some did not, and some of the past listening experiments were seriously flawed. But, overall, there was a slight, but persistent preference for hi rez, which was noticeably stronger if the obviously flawed testing (like Meyer Moran) was disregarded.

    I also think that comparisons of hi rez vs RBCD resolution remasters of analog recordings are particularly unrevealing of any differences, as are uprezzed recordings from native RBCD recordings. Native hi rez recordings vs. downsamplings of same to RBCD resolution are the only true test, in my opinion.

    Friends and I had reached the same conclusions and preferences numerous times in informal testing. Yes, it is theoretically possible that a different master was used in some cases. But, that is highly unlikely with the small label classical music we used from labels committed to recording and mastering natively in hi rez. And, it is highly unlikely that all of the samples we compared used different masters. We have compared downloads via PC as well as the RBCD layer vs. the DSD stereo layer of SACDs converted to 88k PCM using various Universal disc players. We have not repeated such comparisons in recent years, because we felt we had heard enough often enough to reach a firm enough conclusion.

    As always, YMMV.
     
    Ken Newton likes this.
  12. krabapple

    krabapple Active Member

    Joined:
    Apr 15, 2016
    Messages:
    124
    Likes Received:
    17
    It's not at all clear that all the papers remaining in the analysis weren't 'flawed' (or that M&M's paper was as 'flawed' as Reiss claims -- e.g., I have zero patience for the goalpost-moving argument 'but they didn't use pure hi rez recordings', because neither did audiophiles who gushed about the 'sound' of SACDs and DVDAs back in the day).

    So here's where we are. It takes a meta-analysis of several decades of papers of disparate methods, gear, test signals, and subject training level to (supposedly) reveal... a 'slight but persistent preference'. The mountain labors mightily and brings forth (maybe) a mouse. It shouldn't reassure *anyone* about their ability to 'hear' hi rez. And yet, human nature being what it is, a typical 'audiophile' will be so reassured.

    Well, of course, as without objective data, who knows what 'remastering' moves may have been performed? Yet that's never stopped enthuisiasts from claiming not only 1) the hi rez release sounds better and 2) it's because of hi rez.

    Which is what some 'hi rez' releases are (and again, that doesn't stop the praise from gushing forth from the usual parties)

    Yes, as you just said up there...."comparisons of hi rez vs RBCD resolution remasters of analog recordings are particularly unrevealing ".

    You could compare the objective data about the different versions, for a clue. Mastering differences tend not to be invisible.

    But I'm guessing 'informal' means, 'not controlled', and thus, scientifically, 'can't be used to draw conclusions'?

    Yours might too, if you applied DBT methods. Or maybe considered, 'would Joshua Reiss include my results in a metadata analysis?"
     
    Blumlein 88 likes this.
  13. Blumlein 88

    Blumlein 88 Major Contributor

    Joined:
    Feb 23, 2016
    Messages:
    1,942
    Likes Received:
    472
    You have to be very careful. Some high quality minimalist online files available for comparison didn't account for the fact many downsamplers drop the level .2 or .25 db when downsampling. That is just enough to make one sound better under careful listening. That is assuming the downsampling was of good quality otherwise. Which is a second case of confusing effects when comparing one sample rate to another. Unless they use Izotope or Sox most DAWs don't have top quality downsampling algorithms.

    There are a handful of cases where testing indicated a genuine difference vs 44 khz that look pretty compelling. All had four things in common. They parallel recorded with everything the same except sample rate. They used selected listeners most of which were under 30 years old. They used exceptionally high quality recording gear. They didn't use a large number of microphones to record with.

    Even in the above cases results were border line for 5% confidence levels. None would have reached 3 sigma levels of confidence. So there might, maybe be something small to it with excellence all the way around. I agree with krabapple. The differences are very small if they exist at all. Probably don't once you pass 30 years old, and probably are swamped greatly by what happens in mixing/mastering for 99% of available music.

    I am all for quality, but there are a myriad of easier things to do making more difference than what 96 khz vs 44 khz likely provides.
     
  14. Fitzcaraldo215

    Fitzcaraldo215 Addicted to Fun and Learning

    Joined:
    Mar 4, 2016
    Messages:
    774
    Likes Received:
    159
    Happy listening to whatever it is you enjoy most, old buddy.
     
  15. Jakob1863

    Jakob1863 Active Member

    Joined:
    Jul 21, 2016
    Messages:
    238
    Likes Received:
    46
    Location:
    Germany
    In the context of the assertions M&M published within their article it is a serious argument while pointing to other people´s (maybe) wrongdoing in anecdotical listening is not. I hope we can agree on that..
    Btw, M&M confirmed those "audiophiles" claims of better sound quality based on the same anecdotical process (means without supplying data from controlled experiments), the only difference was that "audiophiles" attributed the better sound quality to a " hi rez" virtue while M&M didn´t.

    I´m not sure about your argument, but it seems to be based on a misunderstanding about what a systematic review/meta-analysis is able to provide. One of the difficulties within the audio field are often experiments with small sample sizes and a general lack of replication (which is often true too in other fields). Small samples sizes tend to lead to underpowered experiments which means that real effects have to be very large to be detected.
    Given the heterogenity of the data/experiments the result isn´t that surprising and the call for further experiments is quite common overall for meta-analysises.
    Reiss provided some usefull hints for the experimental setup that should be considered when doing additional studies.

    Yeah, enthusiasts do such things; people who believe the opposite aren´t shy in their claims too. :)
    People interested should evaluate for themselves if there is something in it that matters and that they should imo still do so, even if there are results from controlled listening experiments that "hi rez" provides better quality (because of the hi rez property).
     
  16. Fitzcaraldo215

    Fitzcaraldo215 Addicted to Fun and Learning

    Joined:
    Mar 4, 2016
    Messages:
    774
    Likes Received:
    159
    Level matching might be an issue in most any comparison. Actually, there is a bigger level difference if one compares the hi rez layer of an SACD, which is DSD, to the RBCD layer, which is PCM. There is always a substantial level difference involving those different formats on a hybrid SACD.

    I would not swear that my own attempts at level matching were definitive and scientifically unassailable. However, it was necessary to apply level match at the playback end as carefully as possible. I do have a test SACD, from Pentatone as I recall, that has test tones, which was useful in attempting the level match between DSD converted to 88k/24 PCM vs. converted to 44k/16. I think the best way was to adjust levels using the same 1k Hz test tone file converted at different sampling rates by JRiver and to go on from there comparing the two down conversions at different resolutions with music played by JRiver from the hard drive.

    This also avoided the DSD vs. PCM can of worms in the comparison, as well as differences in mastered level on the disc between RBCD and hi rez, as you described. Also, our most useful comparisons were sourced from the same file on hard drive at the different sample rates, etc., and they did not involve the mechanism in an optical disc player, separate lasers for different layers, etc. And, of course we compared stereo vs. stereo, not using the Mch program on the disc which is our primary interest.

    But, my purpose was not writing a scientific paper, and neither my friends nor I were attempting to prove for posterity that hi rez is definitively superior. We just wanted to get some idea about the hi rez thing early in our embrace of SACD. And, FWIW, we think we heard something small but positive repeated on many discs, blind, but not double-blind. No doubt our methods could be picked apart with hypothetical or even real criticisms, but it was good enough for us.

    I do not disagree that many things, like speakers, room correction, recording/mastering, etc. make a potentially bigger difference than does hi rez. But, if you already have good equipment, have corrected your room and you play good recordings, there is some upside to hi rez. In our case, it does not matter much, because we are all dedicated Mch listeners. And, the major source of Mch music is still SACD, which always uses hi rez for the 5.0/1 layer. But, same goes for BD and even the relatively few Mch downloads out there, which seem invariably in at least 48k/24 bit.

    And, we are all well over 30, but I never did think the audibility of ultrasonic frequencies in hi rez had much to do with its benefits, except indirectly via enabling filtering at much higher ultrasonic frequencies further from the audible band.
     
  17. Cosmik

    Cosmik Addicted to Fun and Learning

    Joined:
    Apr 24, 2016
    Messages:
    984
    Likes Received:
    251
    Location:
    UK
    Not reading all this with great detail, but Amir the other day showed a hi res file with a constant ultrasonic tone in it - that would not be present in the 44 kHz version. Presumably this could distort back into the audible band..? I think he said it could be a nearby computer monitor, so something 'real' and not just a sampling artefact or whatever. This would be presumably picked up in a parallel recording situation unless this was checked for too..? Even if there was no obvious tone, just ultrasonic electronic 'hash' might do the same thing..?
     
  18. krabapple

    krabapple Active Member

    Joined:
    Apr 15, 2016
    Messages:
    124
    Likes Received:
    17
    I remember you from hydrogenaudio, where you made yourdisapproval of M&M clear. I'm glad we can at least agree on the wrongheadedness of this particular critique of M&M.

    Well, not quite. It is indeed the flawed tendency of audiophiles to attribute a difference they hear (real or not) to 'hi rez'. M&M , quite correctly, posit mastering differences as a possible real reason differences are heard when consumes compare SACDs/DVDAs to CDs. But *their* protocol simply converted hi rez recordings to standard rez in real time, with the listeners given considerable latitude as to preferred listening conditions, and what it 'confirmed' was that under those conditions, evidence of difference was not forthcoming. Indeed , of all the tests Reiss reviews, M&M's comes closest to replicating how audiophiles 'normally' compare releases -- while removing the typically glaring issues of remastering and 'sightedness'.


    I'm familiar with the purpose of meta-analyses, thanks (for me, it's biomedical research). I'm not sure you realize the impetus of this particular analysis. There are commercial interests vested in something, anything that gives them a scientific basis, no matter how thin, to tout 'hi rez' delivery formats as a virtual *necessity* for a high quality consumer audio experience. And there's a troubling claim of 'importance' for its results in Reiss's press release statements. When actually, at very best, if one accepts all of Reiss's choices for data inclusion and exclusion, it's evidence for a small population of listeners in the skinny part of the positive distribution tail who are hearing 'something', under highly optimized conditions. And yes, more research is needed. More hype is not.


    A pity Reiss didn't say that, too.
     
    Last edited: Oct 16, 2017 at 7:42 PM
  19. Blumlein 88

    Blumlein 88 Major Contributor

    Joined:
    Feb 23, 2016
    Messages:
    1,942
    Likes Received:
    472
    Yes the situation you describe could occur. I think two of those I was referring to did make sure their playback had bandwidth above 20 kHz and no significant IMD was occurring.
     
  20. Darwin

    Darwin New Member

    Joined:
    Oct 11, 2017
    Messages:
    4
    Likes Received:
    0
    Like most of Archimago's articles he pretty obviously starts with a premise then finds a way to prove it. Lots of charts and data which are mostly smoke and mirrors.
    There was nothing remotely scientific about his article.
    Anyone who has a high school level understanding of the scientific method can see that.
    I guess I'm biased because I like MQA especially since it seems to have the best chance of succeeding in a streaming format.
     

Share This Page