DAC Blind Test: Modi2 vs ODAC

Discussion in 'Blind Testing and Psychoacoustics' started by purr1n, Nov 17, 2015.

  1. NoStream

    NoStream Acquaintance

    Joined:
    Sep 27, 2015
    Likes Received:
    58
    Trophy Points:
    18
    Location:
    Oakland, CA
    You're hinting at the fact that the "sabre sound" probably isn't as obvious or objectionable as some people would claim, and I agree, but if you're trying to prove that, you could start with the least stringent test and move your way up. Otherwise you end up with an AB test that doesn't prove anything because of how disorienting and stressful it is on participants.
     
  2. OJneg

    OJneg The Most Insufferable

    Pyrate BWC
    Joined:
    Sep 25, 2015
    Likes Received:
    3,923
    Trophy Points:
    113
    Location:
    Grand Rapids, MI
    You always need a reference point. Or else you're chasing your tail with different recordings and different colorations elsewhere in the chain
     
  3. chakku

    chakku Friend

    Pyrate
    Joined:
    Oct 13, 2015
    Likes Received:
    676
    Trophy Points:
    93
    Location:
    NZ
    Understood, that's fair enough. But yeah @NoStream I think some people do give any DAC using Sabre chips a bit too much flack just for the fact that they use those chips.
     
  4. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    @chakku, why don't you come out and say it instead of beating around the bush: "Marv gives any DAC using Sabre too much flack."

    Of which I will answer: Yes, I do. However, I keep an open mind. I ranked the GOV2 (Sabre) as the DAC with the least amount of digititus among four other combos units in a recent comparison review.

    I think it's more along the lines that I've never heard AKM, BB179x, Cirrus, AD1922, PCM1704/02/63/56 or UltraAnalog implementations have nasty digital treble (Wolfson seems 50/50). Also, for some reason, Chinese companies love the Sabre. I think it's because of the DSD-compliant buzzword. I've noted Chinese companies tend to pick component names with good brand recognition (XMOS, Sabre, etc.), combobulate the parts, and make stuff that looks great on paper. Unfortunately, the results are more often craptastic than not.

    I suggest blind-test DAC training for the Chinese engineers designing this stuff. Maybe they will stop using Sabre after they evaluate DAC chips with their ears under blind testing conditions. Obviously these Chinese audio designers have serious preconceptions / confirmation bias based on Sabre brand name recognition and buzzwords such as "DSD" from the marketing arm of ESS.
     
    Last edited: Nov 20, 2015
  5. Xen

    Xen Friend

    Pyrate
    Joined:
    Sep 30, 2015
    Likes Received:
    207
    Trophy Points:
    43
    Location:
    Houston, TX
    Awesome results!
    So the question is, what kind of measurement(s) will be needed to see a difference between the ODAC and the M2U? Something like Dynamic Range but more granular since you say the ODAC sounds "louder" maybe through "compression"? This would suggest higher DB lows on the ODAC compared to the M2U, kind of like raising the sound floor instead of the noise floor?
     
  6. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    The problem is that the typical measurement techniques (FR, IMD, THD, etc.) employ steady state signals. We'd have to invent new methods / visualizations to measure this, if it's even measurable at all.
     
  7. chakku

    chakku Friend

    Pyrate
    Joined:
    Oct 13, 2015
    Likes Received:
    676
    Trophy Points:
    93
    Location:
    NZ
    @purrin I didn't want to single you out, I've seen others do so before. You are entitled to your opinions and I'm not trying to defend Sabre DACs specifically but notion of more or less disliking a DAC due to the chip used instead of how it is used seems a little strange to me, but then again I am inexperienced. I also imagine it could reflect on others negatively and give them the idea that they should automatically disregard any DAC with a specific chip in it instead of giving it a fair go.

    And yes I wholeheartedly agree that the DSD buzzword is definitely a gimmick used to sell more units. In my eyes so is the whole overkill 32bit/384khz capabilities of some of the new DACs coming to the market.
     
  8. OJneg

    OJneg The Most Insufferable

    Pyrate BWC
    Joined:
    Sep 25, 2015
    Likes Received:
    3,923
    Trophy Points:
    113
    Location:
    Grand Rapids, MI
    When most people are just copying specsheet designs it shouldn't come as a surprise that most DACs that use the same D/A chips sound similar. Yes implementation matters, but the dominating characteristic of any DAC is still the D/A chip. This is old news for people who are used to design.
     
  9. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Yup. That's why I think it's impossible to make a bad sounding AKM DAC.
     
  10. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Naaa, I want Sabre to die. :( Or at least Chinese audio gear designs to make DACs based on something other than the Sabre. Maybe they won't use AKM because they hate Japanese. The world would be a better place if every other DAC wasn't Sabre. Or maybe it would be a good thing if I could get consumers to at least consider something other than Sabre. I'll lay off Sabre once this happens.
     
  11. Xen

    Xen Friend

    Pyrate
    Joined:
    Sep 30, 2015
    Likes Received:
    207
    Trophy Points:
    43
    Location:
    Houston, TX
    Since you were talking about the sound itself and not an illusion created by brain (e.g. soundstage), I would think it should be measurable. The fun and the pain comes from figuring out those new methods. :sail:

    EDIT: Hmm... I remember atomicbob talking about L/R matching so maybe even replication of soundstage can be measure, but would seem ore difficult as it is dependent on small differences in timing of L/R information.
     
  12. chakku

    chakku Friend

    Pyrate
    Joined:
    Oct 13, 2015
    Likes Received:
    676
    Trophy Points:
    93
    Location:
    NZ
    Excuse me while I go cry in a corner as my quad-Sabre Benchmark DAC journeys to my doorstep. :oops:
     
  13. Psalmanazar

    Psalmanazar Most improved member; A+

    Pyrate Slaytanic Cliff Clavin
    Joined:
    Sep 27, 2015
    Likes Received:
    5,345
    Trophy Points:
    113
    @purrin
    What API did you use? I did a comparison myself tonight with both outputting to the Magni and O2 attempting to level match with my phone. I've found the M2U on Schiit's ASIO drivers to just crush the ODAC that has to operate on DirectSound or WASAPI Push to not pop and stutter. With DirectSound, they're more comparable.
     
  14. paranoidroid

    paranoidroid Friend

    Pyrate
    Joined:
    Sep 30, 2015
    Likes Received:
    180
    Trophy Points:
    43
    Location:
    San Francisco
    That's really interesting. Based on my limited understanding it seems like if it goes through DirectSound it can get dithered and also goes through the Windows mixer - so not bit perfect output? I wonder how much of the AB'd differences was from that alone. Anyone know how OSX handles things? When I did my AB testing it was whatever default drivers OSX used for the Schiit & ODAC.
     
  15. SSL

    SSL Friend

    Pyrate
    Joined:
    Nov 12, 2015
    Likes Received:
    1,239
    Trophy Points:
    93
    Relevant: http://archimago.blogspot.com/2013/06/measurements-part-i-bit-perfect.html (see the conclusion)
    Also: http://archimago.blogspot.com/2015/11/measurements-windows-10-audio-stack.html

    To summarize, DS up-sampling is suspect, but otherwise should perform similarly to bit-perfect if set to 24-bits and the same sampling rate as the source material.
     
  16. chakku

    chakku Friend

    Pyrate
    Joined:
    Oct 13, 2015
    Likes Received:
    676
    Trophy Points:
    93
    Location:
    NZ
    Yup, this is why I use a resampler to make all my audio match my output sample rate (Though it it set not to touch anything already at the same sample rate). SoX VHQ linear phase is very good.
     
  17. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    I used WASAPI.
     
  18. Psalmanazar

    Psalmanazar Most improved member; A+

    Pyrate Slaytanic Cliff Clavin
    Joined:
    Sep 27, 2015
    Likes Received:
    5,345
    Trophy Points:
    113
    It is 100% instantly obvious with instrument separation, spacing, and clarity even if you set the output sample rate to 44.1 kHz. A blind test would be pointless. The DirectSound resampler sucks compared to something like SoX. It's also mixing the output from other programs (including the OS) into the exported PCM sample stream which is a big problem.
    @SSL that rolloff...
    I'm not sure the ODAC rev A supports WASAPI well at all due to its Tenor TE7022 USB controller. The Rev B (a completely different DAC) uses the SaviAudio SA9023 which has ASIO drivers but I'm not sure how good the implementation would be compared to the CMedia controllers and drivers Schiit DACs use.
     
  19. Thad E Ginathom

    Thad E Ginathom Friend

    Pyrate
    Joined:
    Sep 27, 2015
    Likes Received:
    14,130
    Trophy Points:
    113
    Location:
    India
    Weird... Windows does not even support the essential USB audio protocol, leading to the requirement for proprietary drivers and does not [?] provide a straightforward unmixed, un-resampled audio path: Why do people continue to use it?
     
  20. SKiring

    SKiring Friend

    Pyrate
    Joined:
    Oct 3, 2015
    Likes Received:
    203
    Trophy Points:
    43
    Location:
    Rotterdam
    Well it's mostly easy usage (read: being used to it) and Windows does not resample if the sample used is the same as the output chosen in the options. My advise (and this is a recent discovery) is to keep it on 16/44 default since nearly all audio (streaming audio services, movies, games etc.) are in basic 16/44. Then use WASAPI or ASIO and bypass Windows when using any digital files. Essentially this would cause the least amount of "ïnterference".

    With that out of the way, I tried to A/B multiple outputs ranging from 16/44 to 32/96 and in all honesty I can't hear an outright difference or it's so small that in casual listening I wouldn't be able to hear it. We're passed Windows NT 5.x kernel, the very reason ASIO was made is to get rid of the horrible resampling of NT 5.x. Newer DirectSound does this so fast that the differences are absolutely minimal. Still I use the above methodology because if nothing else, it soothes my mind.
     

Share This Page