Schiit Yggdrasil Stereophile Review + Measurements

Discussion in 'Digital: DACs, USB converters, decrapifiers' started by purr1n, Jan 20, 2017.

  1. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Friend MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,686
    Dislikes Received:
    16
    Trophy Points:
    113
    Location:
    Irvine CA
    Whether I agree with Mike's delta-sigma opinon or not, I feel the same way.

    Yggdrasil, based on JA's own measurements, should be in the A+ category.

    LOL! Maybe JA is irked at Mike's opinions about this and that.
     
  2. Grahad

    Grahad Friend

    Friend
    Joined:
    Jan 18, 2017
    Likes Received:
    135
    Dislikes Received:
    0
    Trophy Points:
    83
    Location:
    SGP
    Or maybe he's still pissed by the odd to measure Ragnarok that they didn't tell him anything about. |\/|
     
  3. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Friend MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,686
    Dislikes Received:
    16
    Trophy Points:
    113
    Location:
    Irvine CA
    You mean the bias current control deal?

    The measurement approach inconvenience? Come on. If things work out fine, there is some kool factor to it IMO.
     
    Last edited: Jan 21, 2017
  4. mcduman

    mcduman Rando

    Joined:
    Oct 18, 2015
    Likes Received:
    2
    Dislikes Received:
    0
    Trophy Points:
    3
  5. Perot

    Perot Almost "Made"

    Contributor
    Joined:
    Oct 2, 2015
    Likes Received:
    225
    Dislikes Received:
    12
    Trophy Points:
    43
    Location:
    South Carolina
    Home Page:
  6. landroni

    landroni Friend

    Friend
    Joined:
    Aug 12, 2016
    Likes Received:
    1,292
    Dislikes Received:
    8
    Trophy Points:
    93
    My understanding is that @baldr's performance argument against D/S revolves around "missing codes" (or in layman's terms 'fidelity' or 'accuracy').

    From his earlier takes on the matter:
    The TLDR version of all of this is that you won't see D/S applications in critical D/A tasks like medical imaging or weapons guidance for the simple reason that D/S conversion doesn't guarantee that 'what gets in, gets out', whereas R2R conversion can.
     
    Cryptowolf, dmckean44, OJneg and 2 others like this.
  7. Thad E Ginathom

    Thad E Ginathom Friend

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    6,578
    Dislikes Received:
    11
    Trophy Points:
    113
    Location:
    India
    Good god, I wouldn't waste a wank on those BSers

    ... A Brit.

    :sail:
     
  8. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Friend MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,686
    Dislikes Received:
    16
    Trophy Points:
    113
    Location:
    Irvine CA
    Damit you had to go there. All right, while I respect Mikes heroic efforts in producing decent multi-bit DACs, I don't agree with his opinions there.

    It is apparent to me that Mike has not been involved in the development of a missile or munition guidance system. I will not talk about military stuff. However, DS are not excluded from sensitive applications where "missing codes" would be a liability.

    Precision measurement audio equipment (in to the ones/tens of uV), including the AP, dScope, and Schiits very own Rohde & Schwarz use D/S. What Mike calls "missing codes" should be apparent in characterization.

    "Missing codes" or "math errors" in a delta sigma would exhibit horrid blips and distortion if there were missing codes. It would be obvious in characterization. So I truly don't know what he is talking about.

    Regarding conversion speed and bit resolution.

    If one needs fast conversion speed but limited bit resolution, a precision DAC (such as Schiit's) may be the practical solution. The AD5791BRUZ is actually in the high end of bit resolution for a precision DAC. But it's not necessarily the fastest.

    If one needs relatively slow conversion speed (audio range included) and large bit resolution a DS DAC may be the practical solution (and perhaps the better performing one).

    I seldom see an expensive precision DAC with large number of bits, and low frequency operation, which should exist if precision DACs offered superior performance to DS in applications with these requirements.

    I dunno, but to me it reads like marketing basically telling the consumer to put all their delta sigma DACs in the trash, and buy Schiit's multi-bit awesomeness.

    It's not bad or good.

    Folks can call their products anything they want. Like say, Schiit or Krapp or Kacka or Pizz. But few will be successful in selling them if they describe such products performance as horrid shit and inferior in every way to pretty much everything out there.

    It's obvious Mike believes in his products. And puts a lot of effort in delivering good sounding gear. And indeed, his products don't sound like shit.
     
    Last edited: Jan 22, 2017
    loplop, pedalhead, Cryptowolf and 8 others like this.
  9. anetode

    anetode Moderator

    Staff Member Friend
    Joined:
    Sep 25, 2015
    Likes Received:
    565
    Dislikes Received:
    1
    Trophy Points:
    93
    Location:
    How did I get here
    Home Page:
    First off, you can want 40 bits out of a DAC all you want, you're not getting them. Second, yes, using more DACs will get you more accuracy. But how?

    Well, let's put theory to practice. You're correct that doubling up will lower the overall noise because it is not statistically correlated between each chip. Note however that since the AD5791 is a 20-bit +/-1 LSB voltage output DAC with 10 Vref, then the error when using two of them at FS will be 10V-((10/2^20)/2))V, same as that of a 21 bit DAC with +/- 1 LSB [10V-(10/2^21)V]. By using two DACs per channel you are taking advantage of greater theoretical precision to achieve greater real-world accuracy.

    Because the standard espoused here is one of monotonicity, it is arguable that with certain delta-sigma DACs you can get greater precision because of less statistical variability while not achieving commensurate accuracy in terms of reproducing the true value of the encoded signal.

    Irrespective of that exercise, you will always hit the noisefloor before using up 20 bits in a real recording.
     
    dark_energy likes this.
  10. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Friend MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,686
    Dislikes Received:
    16
    Trophy Points:
    113
    Location:
    Irvine CA
    The deal is, the AD5791 can only take 20 bits of data. After conversion, the problem is that the signal is already quantized to 20-bits on both DACs, and I don't know how to get the missing 3 LSBs back or even 1 LSB back. There could be an argument about improved ENOB. But I just don't see how the ENOB can be greater than 20-bits.

    Say for example you have not a 20-bit DAC but a 1-bit DAC. Add the output of 8 of them, and all of them being driven by the same 1-bit data. I think you are only going to get 1-bit out with a lot less noise.
     
  11. baldr

    baldr Schiit-sterer

    Friend
    Joined:
    Oct 9, 2015
    Likes Received:
    2,151
    Dislikes Received:
    0
    Trophy Points:
    93
    All due respect. I have built at least a dozen or so ds dacs from every manufacturer I could find, including ones I had to sign NDAs to get parts that were very well regarded. In my heart of hearts, I honestly believe the best of them did not sound as good as the worst of my multibit designs. This would be for the purpose of reproducing music which existed in the physical universe as vibrations in the atmosphere as media.

    This is my real perception to which I have a right, just as others have their right to judge their own perceptions. This is not a zero sum game. My being right for me about ds dacs sounding like ass does not deny anybody else their own subjective reality.

    Now for stuff which never existed in the physical universe but only as electrical energy, that's a whole separate deal. There is one area where, rarely, I have no opinions about how to manifest it in an audio chain.
     
  12. Ali-Pacha

    Ali-Pacha Friend

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    261
    Dislikes Received:
    0
    Trophy Points:
    63
    I've never been able to seriously compare my Yggdrasil with TOTL DS DACs, but I did a meaningful experience yesterday with an AMR DP-777, which carries both DS and MB chips (some wolfson for DS and "obsolete" :D Philips for MB, don't remember the refs) with exactly same output circuitry (with tubes, hello Gumbreya :p).

    On a 44.1/16 track, my friend was switching from one to the other chip, and It was instantly obvious which is which : DA has more digital glare that gives some openness to the scene and a reinforced sensation of details (fake ?) and punch, but something more flat and somehow uninvolving in the way it portrays music.

    If you want to perform such test, listen to cymbals / hi hat, you'll immediately know.

    Ali
     
    ohmaigulay, Pyruvate and OJneg like this.
  13. Psalmanazar

    Psalmanazar was: dingleberry; now: chill + interesting

    Slaytanic Cliff Clavin
    Joined:
    Sep 27, 2015
    Likes Received:
    3,248
    Dislikes Received:
    110
    Trophy Points:
    113
    Would the Yggdrasil really want to be a on a list with a few of the recommended "Class A" and "Class B" DACs not worth touching with a ten foot pole? The Mytek 192 and Hugo are godawful while the Vega was mediocre. The list completely ignores great sounding pro gear too like Lynx and Lavry.
     
    Last edited: Jan 22, 2017
    Garns, Aklegal and LauSing2 like this.
  14. Priidik

    Priidik Friend

    Friend BWC
    Joined:
    Sep 27, 2015
    Likes Received:
    1,460
    Dislikes Received:
    2
    Trophy Points:
    93
    Location:
    Estonia
    Thanks, I spilled coffee on my laptop. Really should learn not to read your posts with coffee on hand.
    Even a lowly Metrum Quad (worst multibit dac I have heard), while severely flawed in many respects do not sin like every S-D dac I have heard.
    Some things seem to be above known measurements and specs, looks like it applies to dacs as well as amps and transducers.
     
    Aklegal and landroni like this.
  15. gixxerwimp

    gixxerwimp Professional tricycle rider

    Friend
    Joined:
    Sep 30, 2015
    Likes Received:
    2,368
    Dislikes Received:
    13
    Trophy Points:
    113
    Location:
    small island claimed by China
    Here's the reason the Yggdrasil isn't on the A+ list.
     
    Psalmanazar, Dino, JK47 and 2 others like this.
  16. landroni

    landroni Friend

    Friend
    Joined:
    Aug 12, 2016
    Likes Received:
    1,292
    Dislikes Received:
    8
    Trophy Points:
    93
    Yggdrasil handles 24/96 PCM gracefully. If JA bitches about 24 vs 20 bits, then he should as well subject say the Vega to the same rigueur, as it won't handle 24 bits natively as it uses, essentially, 1-bit switches (however many of them working in unison).
     
    Psalmanazar and Dino like this.
  17. gixxerwimp

    gixxerwimp Professional tricycle rider

    Friend
    Joined:
    Sep 30, 2015
    Likes Received:
    2,368
    Dislikes Received:
    13
    Trophy Points:
    113
    Location:
    small island claimed by China
    Touché! Just got back from the gym so the endorphins seem to have affected my reading comprehension o_O
     
    landroni likes this.
  18. Kattefjaes

    Kattefjaes Mostly Harmless

    Friend
    Joined:
    Sep 5, 2016
    Likes Received:
    5,120
    Dislikes Received:
    41
    Trophy Points:
    113
    Location:
    London, UK
    Having made my way through this thread, I mostly want to see how Stereophile write up the Edition 6, since they seem to like living very moist.
     
    Merrick, Psalmanazar, Ash1412 and 3 others like this.
  19. Vorlon

    Vorlon self-important, pompous ass

    Friend Executed
    Joined:
    Jan 3, 2017
    Likes Received:
    403
    Dislikes Received:
    14
    Trophy Points:
    63
    Location:
    Finland
    Also looking forward to the potential manufacturers comments section, but it seems this thread has since the beginning moved more into a discussion of the classic D/S vs. R2R theme. In that context I have to say I find the brutal honesty @ baldr refreshing. To me it´s respectable that companies can be this open about their conclusions no matter in which direction they go to to (Benchmark is refreshing for the same reason). Can´t see much of a future for D/S Schiit products in the higher price range though...

    My 25 cents: Having heard a few high end D/S and R2R designs, I have to say I can´t agree with the conclusion either. In the designs I´ve heard they both do some things better than the other and neither is generally superior. The problem with comparing the two is that in almost all cases our perception is colored by the fact that R2R designs are generally of the higher price range while most D/S designs are cheap. This can (not saying it automatically does) create a cognitive bias towards R2R. Before the Modi Multibit there basically hasn´t been a mainstream cheap R2R (haven´t heard it so no idea how it sounds) on the market.
     
    Psalmanazar and ultrabike like this.
  20. OJneg

    OJneg The Most Insufferable

    Friend BWC Executed
    Joined:
    Sep 25, 2015
    Likes Received:
    3,748
    Dislikes Received:
    36
    Trophy Points:
    113
    Location:
    Grand Rapids
    Great compilation. This post might need to be pinned.
     

Share This Page