MQA Review - Technical Analysis

Discussion in 'General Audio Discussion' started by Woland, Apr 15, 2021.

  1. Gazny

    Gazny Almost "Made"

    Contributor
    Joined:
    May 11, 2020
    Likes Received:
    627
    Trophy Points:
    93
    Location:
    SoCal
    Again it seems MQA is a waster mark applied to the bit stream, it seems like some of it is 16 bit data in a 24 bit package(with some noise). I would not recommend Tidal to any NOS analog filtered dac owner.
     
    • Like / Agree Like / Agree x 1
    • List
  2. nishan99

    nishan99 Friend

    Friend
    Joined:
    Dec 1, 2019
    Likes Received:
    710
    Trophy Points:
    93
    Location:
    Layla
    Yeah it's be coming pretty obvious they're shifting to MQA only database to save more space to their servers (saving money $$) and simplifying things on their end by providing one "lossless" file then opting to fuck with it or not.

    Classic greedy and lazy business model.
     
    • Like / Agree Like / Agree x 4
    • List
    Last edited: May 1, 2021
  3. Thad E Ginathom

    Thad E Ginathom Friend

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    9,323
    Trophy Points:
    113
    Location:
    India
    sorry, this is a dumbo question: Wasn't the lossless-compression problem solved already? For many of us FLAC is the default format for storing our digital music.

    Does streaming really need different compression solutions to local storage? I'm guessing that's the thing.
     
  4. Yethal

    Yethal Facebook Friend

    Joined:
    Sep 23, 2018
    Likes Received:
    178
    Trophy Points:
    43
    Location:
    Poland
    It's not just about storage but also bandwidth used.
     
    • Like / Agree Like / Agree x 2
    • List
  5. wormcycle

    wormcycle Friend

    Friend
    Joined:
    Aug 13, 2016
    Likes Received:
    1,514
    Trophy Points:
    113
    Location:
    Toronto, ON, Canada
    Reading the Archimago take down of MQA, and Bob Stuart ever changing claims about solving many problems that they just invented, and biased, at times unhinged (Robert Harley), push for MQA in the audiophile press, I was never inclined to give it any credit but still reserved my judgement if it was just a fraud, until the 38 mins exactly into the @GoldenOne video.
    It shows the example where someone took the MQA distribution file (Master Quality Authenticated:punk:), then blanked the lower eight bits, and the blue light on the MQA DAC was till shining bright.
    They are just crooks. Everything else they are saying becomes irrelevant. That's equivalent to finding an internal combustion engine in Tesla cars.
     
    • Like / Agree Like / Agree x 9
    • List
  6. purr1n

    purr1n Building Magnis part time because it's peaceful.

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    75,112
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Neither storage or bandwidth for music files are issues today.

    One thing we need to remember is that Amir has played this game for much of his career at Microsoft with respect to HD-DVD and the VC-1 codec. People who play these games will hedge their bets. You never know who will win. I don't have a problem with Amir not having a problem with MQA. The thread lockdown at ASR is probably meant more for saving Amir from further grief.

    Atkinson is probably trying to help out a fellow brit, his friend. This is a little bit more disturbing because JA was always supposed to be the more measurement focused objective guy at Stereophile. But we already saw years ago how conclusions of his measurements varied and applied differing standards depending upon the manufacturer relationship with Stereophile, so this stance with MQA, his no-too-subtle pushes for MQA in his gear reviews, isn't surprising.

    Then again... they say follow the money. I wouldn't be surprised if there was money involved here, either directly or indirectly from Bob Stuart to Amir or Atkinson. Not a few beers, but quite a bit too, at least potentially.
     
    • Like / Agree Like / Agree x 5
    • Epic Epic x 3
    • List
    Last edited: May 2, 2021
  7. Hrodulf

    Hrodulf Prohibited from acting as an MOT until year 2050

    Friend
    Joined:
    Sep 28, 2015
    Likes Received:
    3,763
    Trophy Points:
    113
    It's not an issue for listeners, but server spooling is always a large expense position for a small player like Tidal.
     
    • Like / Agree Like / Agree x 2
    • List
  8. crenca

    crenca Friend

    Friend
    Joined:
    May 26, 2017
    Likes Received:
    2,344
    Trophy Points:
    93
    Location:
    Southern New Mexico
    I forget the numbers (Archimago has an analysis somewhere) but MQA is no help here for Tidal given that it's typical file size is no larger than 16/44 PCM, so this is bunk marketing claim. Even for a real HiRes (say, 24/48 and above) like Qobuz it is dubious given the real world numbers, and that's only if you and your customers accept the fiction that MQA is really some species of HiRes...
     
    • Like / Agree Like / Agree x 2
    • List
    Last edited: May 2, 2021
  9. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Friend MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    10,035
    Trophy Points:
    113
    Location:
    Irvine CA
    Are storage and bandwidth issues not already addressed by mature and available audio coding standards (some perhaps free)? Is Tidal unable to make use of such wide arrange of opinions? Not sure I follow?
     
    • Like / Agree Like / Agree x 1
    • List
  10. Merrick

    Merrick A lidless ear

    Friend
    Joined:
    Jan 6, 2016
    Likes Received:
    7,755
    Trophy Points:
    113
    Location:
    Portland, OR
    High levels of FLAC compression should save more space than an MQA encode, at the cost of more processing to decompress. But then who knows how much processing MQA unfolds take? Tidal might have the worst of all possible worlds right now.
     
  11. Hrodulf

    Hrodulf Prohibited from acting as an MOT until year 2050

    Friend
    Joined:
    Sep 28, 2015
    Likes Received:
    3,763
    Trophy Points:
    113
    If Tidal wants to offer 24/192 or higher, then mqa just needs to be smaller than FLAC of the same resolution. Doesn't it achieve that?

    Edit: here's what deeper FLAC compression gets you.

    SmartSelect_20210503-015239_Chrome.jpg

    Not too impressive.
     
    Last edited: May 2, 2021
  12. Merrick

    Merrick A lidless ear

    Friend
    Joined:
    Jan 6, 2016
    Likes Received:
    7,755
    Trophy Points:
    113
    Location:
    Portland, OR
    Do we even know what a corresponding MQA 24/192 equivalent file is in size, and the processing it takes to unfold it though? Without that it’s just guesswork.
     
  13. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Friend MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    10,035
    Trophy Points:
    113
    Location:
    Irvine CA
    And how are these apples to apples?

    Are you saying that MQA is lossless?

    EDIT: As far as I know MQA is an ad-hoc lossy approach (a very poor one from what I've gathered). So not sure why you are comparing against a codec with real lossless compression.
     
    Last edited: May 2, 2021
  14. crenca

    crenca Friend

    Friend
    Joined:
    May 26, 2017
    Likes Received:
    2,344
    Trophy Points:
    93
    Location:
    Southern New Mexico
    No. The audio band (so up to 48khz samples) is non compressed PCM, though we don't know what DSP is applied. The ultrasonics above 24 khz up to 48khz (so up to 96khz samples) is lossy compressed into the LSBits. After that, it's all upsampling (in the so called "rendering" phase) and marketing speak - there is no actual content (lossy or otherwise) though MQA wants you to believe there is. Every claim above 24/96 (such as 24/192) is fiction.

    On top of this, MQA does not compress as well (due to whatever proprietary compression is going on in the LSB) as regular PCM when put in the misnamed FLAC (which is a data compression scheme, not an actual audio sampling/digitizing algorithm), so real 24/96khz source PCM is barely smaller when run through MQA and then FLACed than the equivalent PCM source.
     
    • Like / Agree Like / Agree x 1
    • List
  15. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Friend MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    10,035
    Trophy Points:
    113
    Location:
    Irvine CA
    It's not an honest PCM even in the audio band.

    https://arstechnica.com/gadgets/201...hing-you-need-to-know-about-high-res-audio/4/

    For 24/192 original recording:

    1) Decimated to 24/48 using a poor IIR filter with aliasing problems (right from the start, this broke any lossless claims.)

    2) An optional second filter to reduce ringing on the aliased and lossy 24/48 file.

    3) The 24/48 file gets noise shaped to 16/96.

    4) The 16/96 gets split into two bands: 0-24kHz and 24-48kHz (by which filter again?) - BTW, the 20 bit claim they have from the noise shaper in the 0-24kHz range probably crashed and burned after this step which seems to map back to 16/48 anyway.

    5) 16 bit 24-48 kHz band (noise now from the noise shaper) gets lossy compressed to 8 bits.

    6) A new 24/48 file is created by concatenating the 0-24 kHz 16/48 with the 24-48 kHz 8/48 (not sure if the intention is to add this back to support the 20 bit claim, but this is questionable because of the filtering and compression process).

    A "legacy" DAC sees a processed 24/48 where the 8 LSBs are basically noise in best case scenario, since there is some talk about further restricting playback to as low as 13-bits.

    Even if you were to unravel this mess. This is not lossless by any stretch.

    We seem to be down to heavily aliased and noisy 16/48 PCM using an original 24/192. Unless MQA wants to claim credit for the noise shaper 8 bit compressed quantization noise in the 8 LSBs of the 24/48 PCM.

    How is this better than straight 24/48 or even properly processed 16/48 PCM? What are the benefits again?

    So @Hrodulf, this is not apples to apples when compared to 24/192 FLAC. I'm not even sure why anybody would pick this over a straight 24/48 PCM stream. Folks can do 24/48 FLAC, can't they? Are we saying this is more efficient than that? Better quality perhaps?
     
    • Like / Agree Like / Agree x 10
    • List
    Last edited: May 7, 2021
  16. Thad E Ginathom

    Thad E Ginathom Friend

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    9,323
    Trophy Points:
    113
    Location:
    India
    This is what I thought. But I have not kept up with either the computer-techie or the audo-techie scene for quite a while. All my (mostly ancient) music collection is stored as FLAC, except for portable devices, for which I use OGG. OGG was the latest thing on my audio-compression horizon. These days, due to eyes working at least a bit better than ears, I am more interested in when jpeg gets replaced with something better for pictures.

    Lossless is lossless. By definition it is as good as it gets. One can only improve on processing time and compression amount. Neither is of any consequence to the listener unless they have huge libraries and little cash to spend on hardware. For streaming, someone might bring up the word latency: latency is of no consequence whatsoever to the listener (another audiophool/marketing myth) unless or until there is a noticeable time lag between pressing play/pause and something happening. Otherwise, whether the music takes 5ms or 5 minutes to reach you, it is still the same music.

    I can see how these things do matter, and matter financially, at the server/spooling/streaming end of streamed music.

    Looks like they fell for the marketing and brought badly.

    Looks like BS wants to be the Gates of music. Back when I was an IT manager, I had a director who thought he was in charge. The reality (which he could not see) was that every desk had a Windows machine, not because the IT department wanted it (I hated it!), but because the workers wanted it*. It was there not because MS had sold it to us, but because MS had sold it to them, who then demanded it of us. Cloud computing followed similar marketing lines.

    Monsanto/etc want money from every mouthful of food. MS wanted money from every personal computer. Now BS wants money from every DAC, from every earful of music, and thus he has to get the public to demand it and the manufacturers to play along. And the audio world is notoriously gullible.



    *Off-topic PS. Yes, of course a PC was a huge step up from a dumb terminal. But Windows? It took two of us to run the system before, and five after. Well, I got promoted accordingly. Not so bad. Did good for my career, but bad for my life of having nothing much to do except Unix, which I actually enjoyed.

    On-topic PPS. I think that maybe audiophile history took a wrong turn obsessing on lossless and bit-perfect. Maybe. And hey, probably it's all the same in the end. Except for companies who can sell their magic sauce to that obsession, whether it is or not. An yes, me too: even now, I wouldn't accept a lossy file if lossless is available. Even though the difference is certainly beyond my hearing.
     
    • Like / Agree Like / Agree x 1
    • Epic Epic x 1
    • List
    Last edited: May 3, 2021
  17. Hrodulf

    Hrodulf Prohibited from acting as an MOT until year 2050

    Friend
    Joined:
    Sep 28, 2015
    Likes Received:
    3,763
    Trophy Points:
    113
    I'm not debating that MQA is an audibly dubious compression method. My inquiry was whether it makes sense for server spooling reduction reasons to implement MQA vs. properly compressed PCM, information retention effectiveness notwithstanding.

    If it can deliver something that lights up hi-rez indicators on one's DAC, while staying barely larger than 16/44 FLAC, then commercially it might seem like a win. Also Bob Stuart assures that it's totally the same thing and might even be better by their de-blurring magic. And let's not forget that Meridian has been pretty good at pumping people up to generate demand, so even with no filesize reduction, implementing MQA might make sense as a marketing coop campaign.
     
    • Like / Agree Like / Agree x 1
    • List
  18. Metro

    Metro Friend

    Friend
    Joined:
    Dec 27, 2016
    Likes Received:
    1,324
    Trophy Points:
    93
    Location:
    San Francisco
    Annnd... the new MQA thread also shut down, by the same moderator.

    Screen Shot 2021-05-05 at 10.05.05 PM.png
     
    • Like / Agree Like / Agree x 2
    • List
  19. Gazny

    Gazny Almost "Made"

    Contributor
    Joined:
    May 11, 2020
    Likes Received:
    627
    Trophy Points:
    93
    Location:
    SoCal
    what a great signature for a mod :rolleyes:
     
  20. nishan99

    nishan99 Friend

    Friend
    Joined:
    Dec 1, 2019
    Likes Received:
    710
    Trophy Points:
    93
    Location:
    Layla
    Yet another pathetic pawn sacrifice from MQA I guess.






    In summary Hans is trying to reestablish authority about who Bob Stuart is and simply suggesting we just have to believe papa Bob and use our ears to decide for ourselves, totally dodging the points @GoldenOne made about the the false claims of MQA and the BS marketing surrounding it.

    Funny how things turn subjective when they fail miserably at their objective claims and goals, it's if SMSL asked us to use our ears when their product is not up to the review samples they sent to promote how objectively superior the product is.
     
    • Like / Agree Like / Agree x 9
    • List
    Last edited: May 7, 2021

Share This Page