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 Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,960
    Trophy Points:
    113
    Location:
    Irvine CA
    The more bins, the better. Assuming tones. I can't remember how many bins are needed for a given resolution, and for a given reasonable window.

    I think we could get -130 dBFS noise floor with a perfect square wave, which is effectively 1-bit, going at full scale from a "24-bit" DAC into a "24-bit" ADC.

    If the input is a 1 kHz sinusoidal, and we are getting a 1-bit square wave @ the output, all the harmonics are impacting SNR and bit resolution. Noise floor is more of a wideband distortion, if one wants to look at it that way. Harmonic distortion is sort of narrow band noise, if one wants to look at it that way. They both penalize resolution.

    I didn't really read JA's take on the CHORD before. No disrespect intended, but honestly because TL;DR. Will do right now.

    So lets make it concise:

    My AP SYS2722 is awesome-sauce, and the DAVE can talk to it.

    In wild-man mode the CHORD can put out 8.75V balanced at @ "0dB". Above "+3dB" it will run out of juice and f**k up your music (and perhaps your equipment).

    In chicken-shit mode CHORD can put out at most 6.2V balanced (i.e. "-3dB" of wild-man mode).

    Headphone out impedance was less than 0.5 ohms.

    Balanced output was 66 ohms.

    The FIR is very long (it did not fit in my equipment's fur burger), and likely will have a long delay before squirting music. However, this also means ultrasonic shit is gone bye bye.

    All digital filters are linear phase brick-walls, except @ 384kHz which is mediocre. The "HF filter" mode is pretty much redundant.

    Channel separation does not suck.

    There noise floor between 0 and 1 kHz is pretty low. Meaning the whole system rejected 60 Hz power supply noise.

    From 0 to 10 kHz, there was very little distortion using a -90 dBFS dithered 1 kHz tone in both 16-bit and 24-bit modes. This when plugging CHORD through USB and maybe AES (not sure what JA used for fig. 5).

    Note 0 dBFS is proly 8.75 V which maybe a bit hot, and make things look more awesome-sauce than they actually are, but who knows.

    Assuming all these pretty plots were done balanced.

    Distortion (and real bit depth resolution) WILL show it's ugly face above 10 kHz as shown in fig. 7 with the -90 dBFS 16-bit tone. 24-bit mode did not help at all. The "HF filter", as mentioned before, is pretty much redundant and does not do a damned thing to fix the "not shown" 24-bit plot, or the not-so-hot 16-bit picture in fig. 7.

    DAVE does awesome in distortion from 0 to 1 kHz using a 50 Hz tone. You do not care what happens from 1 KHz and above, specially 10 kHz and above *Obi-Wan hand waving*.

    IMD using 19 kHz and 20 kHz, is awesome. A 60 Hz + 7 kHz 4:1 might have been nice. But they are "not shown".

    Jitter seems low.

    CHORD DAVE is beyond reproach 1 kHz and below, perhaps great from 1 kHz to 10 kHz, and "not shown" fugly 10 kHz and above (particularly with undithered tones, but maybe with dithered ones as well).

    +++

    Based on what I read, and on JA's measurements, the CHORD might have and edge on the Yggdrasil from bass to midrange, and sucks ass relatively speaking from midrange to treble. At least that's my read on it.

    In all fairness, JA did seem to run same tests on the Yggdrasil.
     
    Last edited: May 24, 2017
  2. landroni

    landroni Friend

    Pyrate
    Joined:
    Aug 12, 2016
    Likes Received:
    1,164
    Trophy Points:
    93
    This is what struck me most, and gave me an OCD need to put all the relevant statements together to make abundantly clear the BS.

    Chord DAVE:
    "When I changed the bit depth of the incoming data from 16 to 24 with a dithered tone at –90dBFS, the noise floor dropped by 23dB (fig.5), implying resolution close to 20 bits, which is state-of-the-art DAC performance."


    And only very recently, Yggdrasil:
    "However, with 24-bit data (blue, red), while the noise floor lies at or below –160dBFS, a regular series of distortion components can be seen, in which the third, fifth, seventh, or ninth harmonics are highest in level. This will be due to the Yggdrasil's use of 20-bit D/A converters; the bottom four bits with 24-bit data will be truncated. "

    "With undithered 24-bit data at the same level (fig.7), although the overall shape of the reconstructed sinewave is good, you can see significant errors at the signal's zero-crossing points. Again, this will be due to the design choice to use 20-bit converters. "

    And of course the infamous "obsolete":
    [​IMG]

    And then later on still on Facebook:
    "We don't have a "social media guy" - I, John Atkinson, post to Facebook that which I think interesting and relevant. The datasheet for the DAC chip clearly states that it is a 20-bit part and my measurements in the review clearly reveal the truncation of 24-bit data."

    And of course "sub-optimal engineering" and the "sound of truncation":
    [​IMG]


    So, John Atkinson, does all this make Chord DAVE obsolete in the context of 24-bit audio and sup-optimally engineered? Or is it instead that Schiit Yggdrasil is state-of-the-art DAC performance?
     
  3. Darren G

    Darren G Friend

    Pyrate
    Joined:
    Apr 4, 2017
    Likes Received:
    589
    Trophy Points:
    93
    Location:
    Pennsylvania, PA
    I remind myself how many of the live music events I've heard that have been amplified. How much stress/effort goes into verifying the mics, amps, and speakers have been measured to the least significant bits of accuracy, to measure the noise, to measure the distortion? Measurements are fine, they have a place, but even home gear that measures well can end up sounding dry, boring, repetitive, harsh, have just one tiny attribute that is annoying, etc. What is different is that if that home gear has annoying traits (vs pleasant), we'll be living with it day after day. Live events are short duration and easy to overlook that if we had to listen to that gear daily, annoyances would eventually stand out.

    Besides all of that good luck finding a transducer that is purely flat, free of all coloration. Again measurements are fine, but I get more out of reading other's impressions, and in particular, how components pair over the long run. Magazines are very limited in how many cross gear impressions can be tested, for how long, and yes, magazine writers are hardly free of personal bias.
     
  4. Larry Megugorac

    Larry Megugorac Craps on Filipino accents to ease inner poverty

    Joined:
    Jul 27, 2016
    Likes Received:
    31
    Trophy Points:
    18
    Location:
    Brentwood
    Home Page:
    Yea it's worse than that...because JA said, in that same Chord review....Marvey should have added this...."implying resolution close to 20 bits, which is "State of the Art DAC performance".....This guy talks out both sides of his famous mouth!!!
     
  5. powermatic

    powermatic Friend

    Pyrate
    Joined:
    Sep 16, 2016
    Likes Received:
    690
    Trophy Points:
    93
    Someone should engage JA himself concerning this-he's a constant presence on 'Critic's Corner' at Audio Asylum, and seems always at the ready to take on the tough issues. Here, for instance, he helps a loyal subscriber solve The Case Of The Purloined Periodical.

    https://www.audioasylum.com/forums/critics/messages/8/86341.html
     
  6. Maxx134

    Maxx134 Dunning–Kruger effect poster boy

    Pyrate
    Joined:
    Oct 27, 2015
    Likes Received:
    818
    Trophy Points:
    93
    Location:
    NYC
    All this makes me curious has anyone here heard both dacs side by side?

    I have only heard them separately so my impressions would be skewed.

    *edit* - yeah I can see why I received dislike, as my post not helpfull.
    Ok then,

    What I meant about skewed,
    is that my impression will be more subjective , than over detailed comments and accuracy.

    So far my impression of the Dave, is of most dacs of this type,
    Which to me sound very clean and shiny, yet realistic.
    One thing I am fairly sure is that I perceived less width with the Dave than Yggdrasil.
    Yggdrasil wider soundstage.
     
    Last edited: May 24, 2017
  7. 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
    Really besides the point. DAVE vs. Yggdrasil discussions can be noted in Torq's DAC thread.

    It's not a question what is the better DAC, it's a question of Atkinson's double standards (20-bit is bad for Schiit, but 20-bit is state-of-the-art for CHORD or MSB) and the "noisy" 24-bit -90db plot from the DAVE which he so conveniently happened to NOT show.
     
  8. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,960
    Trophy Points:
    113
    Location:
    Irvine CA
    Yeah. Saw that. That was BS.

    LOL! While JA made a big deal of the similar 24-bit -90dBFS plot from the Yggdrasil where he writes "With undithered 24-bit data at the same level (fig.7), although the overall shape of the reconstructed sinewave is good, you can see significant errors at the signal's zero-crossing points. Again, this will be due to the design choice to use 20-bit converters."

    One more thing that I think was not very fair in general when doing measurements:

    For Yggdrasil the 16-bit -90dBFS ranged from -0.2 mV to 0.2 mV. I think this corresponds roughly to a 0 dBFS at 4.4721 Vrms. Yggdrasil max output is 4 Vrms (balanced).
    For DAVE the 16-bit -90dBFS ranged from -1.5mV to 1.5mV. I think this corresponds roughly to a 0 dBFS at 33.5 Vrms. WTF?! I thought DAVE clipped at 12.3 Vrms!

    With that 0 dBFS reference on the DAVE, distortion and noise is likely going to look a proly much better than it should relative to other measured equipment. If I'm not mistaken, that 16-bit -90dBFS measurement was run too hot to DAVE's advantage. And it was still horrid! In the 24-bit case so much, that the characterization was completely omitted from the review!
     
  9. AllanMarcus

    AllanMarcus Friend

    Pyrate
    Joined:
    Oct 23, 2015
    Likes Received:
    2,969
    Trophy Points:
    113
    Location:
    Los Alamos, NM
    Home Page:
  10. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,960
    Trophy Points:
    113
    Location:
    Irvine CA
    I think that means JA is not a fan of R-2R audio DACs in general. Just like Mike Moffat is not a fan of Delta Sigma DACs in general.

    It's sort of a preference I guess. I actually like Delta Sigma more in general. I know other folks like R-2R more in general. But, that is no reason to use double standards when reviewing or measuring equipment.
     
    Last edited: May 24, 2017
  11. landroni

    landroni Friend

    Pyrate
    Joined:
    Aug 12, 2016
    Likes Received:
    1,164
    Trophy Points:
    93
    What are the chances that we could get a DAVE in your or @atomicbob's hands for testing measurements in more controlled and less financially invested conditions? Your running comments on JA's measurements (thanks!) make it seem as if he took a very skewed approach in appraising the technical characteristics of this device.

    I imagine JA has no problems with R2R implementations like MSB, when the devices come with the appropriately ludicrous price-tags.
     
  12. Maxx134

    Maxx134 Dunning–Kruger effect poster boy

    Pyrate
    Joined:
    Oct 27, 2015
    Likes Received:
    818
    Trophy Points:
    93
    Location:
    NYC
    Thanks for answering my post question ,
    so I could have been steered to proper thread and post any comments over there...

    Instead of getting an infantile "dislike" by "winders", which I have no reasons so incorrectly assumed was for lack of info.

    Another example of loose usage on dislike button which reduces thread posts in general, because it's just easier to press that button and leave others wondering..
    :rolleyes:
     
  13. Maxx134

    Maxx134 Dunning–Kruger effect poster boy

    Pyrate
    Joined:
    Oct 27, 2015
    Likes Received:
    818
    Trophy Points:
    93
    Location:
    NYC
    Interesting to see if anyone brings this topic up over there,,
    Thanks for link
     
  14. landroni

    landroni Friend

    Pyrate
    Joined:
    Aug 12, 2016
    Likes Received:
    1,164
    Trophy Points:
    93
    Absolutely not the place, but this is an important point: Friends who enjoy wielding their Dislike powers should always state their reasons, in a separate post or by liking the post of someone who already explained the issues. This applies equally whether the dislike concerns randos or friends.
     
  15. Torq

    Torq MOT: Headphone.com

    Pyrate
    Joined:
    Jan 15, 2016
    Likes Received:
    8,193
    Trophy Points:
    113
    Location:
    Seattle, WA
    DAVE is a bit of an enigma. It was, initially, extremely impressive. Some of that was down the nature of the thing, and some was, I suspect, the fact I'd gone through so many, very expensive, DACs to that point - none of which lived up to Yggdrasil's performance (for me), that it was perhaps a more impactful event than I thought.

    The problem is that every time I listen to DAVE I like it less than I did before.

    A couple of months ago I had been contemplating buying one of these, used, from a UK dealer, that would have come in at the equivalent of $6,500 (they're substantially cheaper in the UK to start with). As of the last time I listened, I think my buy-in price would be down around $3,500. Though to be fair, some of that is down to me absolutely hating the look and operational ergonomics of the thing and the fact that, at that price, if I decided not to keep it it would be easy to unload.

    The next time I have an opportunity to borrow one I'll see about getting it to @atomicbob if he's interested in measuring it.
     
  16. Larry Megugorac

    Larry Megugorac Craps on Filipino accents to ease inner poverty

    Joined:
    Jul 27, 2016
    Likes Received:
    31
    Trophy Points:
    18
    Location:
    Brentwood
    Home Page:
    What does impress about Schiit Audio is that their products offer real value and performance!! Something that cannot be said about Chord's Dave.
     
  17. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,960
    Trophy Points:
    113
    Location:
    Irvine CA
    I think @atomicbob is your best bet, because he has better measurement equipment than I do.

    However, Those -90 dBFS numbers for the DAVE did not look to hot, so possible I might be able to do something there. @atomicbob already has the Yggdrasil I think, so he would only need the DAVE. I don't have neither unfortunately.

    Like @Torq said, we'll see if @atomicbob has the bandwidth to put the DAVE through the test bench.

    There appears to be some of that indeed. But based on the response he gave to @AllanMarcus, it seemed as if he felt R2R was not the way to go, because current R2Rs covering the audio range seldom go beyond 20-bit.

    But there could be much more to it. I don't know.
     
  18. 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
    That statement doesn't make much sense. It would seem that thermal noise is just as much as a limiting factor in most cases. I don't exactly see the noise floor on his other DAC measurements hitting below -154db (which should indicate a 20-bit DAC) very often.
     
  19. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,960
    Trophy Points:
    113
    Location:
    Irvine CA
    It depends on what dB is referenced to. I may go all technical here, but I feel it may shed some light.

    1) Over a band of 20 KHz, thermal noise is about -131 dBm (dB referenced to milliwatts). This is not to be confused with SNR. We only know noise floor power, and have no idea about the signal power. Therefore we have no idea about number of bits with just the noise floor.

    2) When people talk about dBFS, we now have some idea of maximum signal power which is base lined to 0 dBFS. Furthermore, 0 dBFS for an audio DAC typically is about 4.0 Vrms balanced and 2.0 Vrms unbalanced. That's how Schiit did it with the Yggdrasil.

    Putting 1) and 2) together can give us an idea of the maximum possible number of bits before hitting the noise floor, assuming say 600 ohms:

    P(signal) = (4.0 Vrms)^2/(600.0) = 26.7 mW => 14.25 dBm

    This gives us a range of 145.25 dB above the noise floor (14.25 + 131) which is roughly 24 bits.

    And that is probably why 24-bits is such a magic number. Again, assuming 600 ohm load.

    This is the number to hit, but hard to do because distortion and non-idealities kick in. Not only that, if the load goes north of 600 ohms, then the number of achievable bits goes down. If it goes south then in theory one can achieve more bits, but the load is difficult and many devices will distort as the load decreases (in some cases due to current demands)

    Now, DAVE did not seem to put 0 dBFS at 4.0 Vrms, but at 33.5 Vrms (a voltage at which the DAVE clips). This give us 1870 mW with 600 ohms => 32.7 dBm, and therefore a range of 163.7 dB which is roughly 27.3 bits (which they can't achieve because again, they clip). They achieve this by running the DAC hot and maybe even clipping the MSBs. But even w/o the clipping issues, what standard amp is going to take 33.5 Vrms? Hell, what average Joe's headphone does? And they clip at way bellow that voltage? So this IMO is cooking the numbers by design.
     
    Last edited: May 25, 2017
  20. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    18,652
    Trophy Points:
    113
    Location:
    On planet
    If a DAVE makes its way to my lab I will squeeze it in, lab schedule is booked with a ton of dialogue editing and acoustic measurements, all on the usual corp ASAP turnaround schedule. DAVE brings to mind that old Cheech & Chong bit, Dave's not here.
     

Share This Page