Schiit Audio Yggdrasil D/A processor Measurements

Sidebar 3: Measurements

I measured the Schiit Audio Yggdrasil with my Audio Precision SYS2722 system (see the January 2008 "As We See It"). As well as the Audio Precision's digital outputs, I used WAV and AIFF test-tone files sourced via USB from my MacBook Pro running on battery power with Pure Music 3.0. Apple's USB Prober utility identified the Schiit DAC as "Schiit Audio Gen 3 USB" from "Schiit," and identified the USB interface device as being from "C-MEDIA ELECTRONICS INC." The USB port operated in the optimal isochronous asynchronous mode. Apple's AudioMIDI utility revealed that, via USB, the Yggdrasil accepted 16-, 24-, and 32-bit integer data sampled at all rates from 44.1 to 192kHz.

The Schiit's maximum output was very slightly higher than the specification, at 4.24V balanced and 2.06V unbalanced; and both sets of outputs preserved absolute polarity, meaning that the XLR jacks are wired with pin 2 hot. The output impedance was a low 196 ohms across the audioband from the balanced outputs, and 180 ohms from the unbalanced outputs, though the latter rose slightly, to 202 ohms, at the bottom of the audioband.

The Yggdrasil's impulse response with data sampled at 44.1kHz is shown in fig.1. Taken from the left channel's output, it reveals the Yggdrasil's reconstruction filter to be a time-symmetrical, finite-impulse-response type, but with more coefficients than is usually seen with this type of filter. This filter produces a very sharp rolloff above the audioband, as evidenced by the red and magenta traces in fig.2 (footnote 1). The blue and cyan traces in fig.2 show the spectrum of the Schiit's output while being fed 24-bit data representing a full-scale 19.1kHz tone. While the image of the tone at 25kHz is suppressed by almost 90dB and harmonic distortion is very low, the noise floor looks much more ragged than I usually see (footnote 2).


Fig.1 Schiit Yggdrasil, impulse response (one sample at 0dBFS, 44.1kHz sampling, 4ms time window).


Fig.2 Schiit Yggdrasil, wideband spectrum of white noise at –4dBFS (left channel red, right magenta) and 19.1kHz tone at 0dBFS (left blue, right cyan), with data sampled at 44.1kHz (20dB/vertical div.).

Fig.3 shows the frequency response at a lower level, –12dBFS, with sample rates of 44.1, 96, and 192kHz. Other than the expected sharp cutoff at half of the two lower sample rates, the responses follow the same basic shape, with a smooth rolloff above the audioband. Note the excellent channel matching in this graph. The Yggdrasil also offered superb channel separation, especially from its balanced output jacks, where the crosstalk in both directions was below –125dB below 2kHz, and still below –110dB at 20kHz. Channel separation from the unbalanced outputs was 6dB or so lower at high frequencies, but the crosstalk was 20dB worse at low frequencies. The Yggdrasil also offered superbly low levels of analog self-noise. Fig.4 shows spectral analysis of the DAC's low-frequency noise floor while it reproduced a full-scale 1kHz tone. The random noise components lie at the level of my analyzer, and the only supply components that can be seen are at 120Hz, at –135dB in the left channel and –130dB in the right. This indicates excellent analog circuit design and circuit-board layout.


Fig.3 Schiit Yggdrasil, frequency response at –12dBFS into 100k ohms with data sampled at: 44.1kHz (left channel cyan, right magenta), 96kHz (left green, right gray), 192kHz (left blue, right red) (0.5dB/vertical div.).


Fig.4 Schiit Yggdrasil, spectrum with noise and spuriae of dithered 24-bit, 1kHz tone at 0dBFS (left channel blue, right red; 20dB/vertical div.).

Because of the Schiit's low level of analog noise, I extended to 160dB the vertical scale for the spectral analysis of its output with a dithered 1kHz tone at –90dBFS (fig.5). The noise floor with 16-bit data (cyan and magenta traces) is actually the dither noise used to encode the data. 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.


Fig.5 Schiit Yggdrasil, spectrum with noise and spuriae of dithered 1kHz tone at –90dBFS with: 16-bit data (left channel cyan, right magenta), 24-bit data (left blue, right red) (20dB/vertical div.).

With undithered data and a signal at exactly –90.31dBFS, the Yggdrasil output a superbly symmetrical waveform, with the three DC voltage levels described by the data very well defined and the ringing due to the reconstruction filter clearly visible (fig.6). 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.


Fig.6 Schiit Yggdrasil, waveform of undithered 1kHz sinewave at –90.31dBFS, 16-bit data (left channel blue, right red).


Fig.7 Schiit Yggdrasil, waveform of undithered 1kHz sinewave at –90.31dBFS, 24-bit data (left channel blue, right red).

Distortion levels with high-impedance loads were low, as can be seen in fig.8, where the second and third harmonics lie between –90 and –100dBFS (0.001 and 0.003%) in both channels. However, the Schiit clipped with signals higher than –10dBFS into the punishing 600-ohm load (fig.9). I then tested the Yggdrasil for intermodulation distortion with an equal mix of 19 and 20kHz tones (fig.10), and while the actual intermodulation products were between 90 and 100dB below the signal's peak level, the noise floor again looked ragged, as in fig.2. I suspect that the digital filter begins to overload with full-scale high-frequency tones. As music only very rarely contains such spectral content, perhaps the filter and DSP circuits have been optimized for low-level signals.


Fig.8 Schiit Yggdrasil, spectrum of 50Hz sinewave, DC–1kHz, at 0dBFS into 100k ohms (left channel blue, right red; linear frequency scale).


Fig.9 Schiit Yggdrasil, spectrum of 50Hz sinewave, DC–1kHz, at –10dBFS into 600 ohms (left channel blue, right red; linear frequency scale).


Fig.10 Schiit Yggdrasil, HF intermodulation spectrum, DC–30kHz, 19+20kHz at 0dBFS into 100k ohms, 44.1kHz data (left channel blue, right red; linear frequency scale).

Tested for its rejection of word-clock jitter with a 16-bit J-Test signal, the Schiit Yggdrasil had some problems (fig.11). Although most of the odd-order harmonics of the LSB-level, low-frequency squarewave in the right channel (red trace) are close to the correct level (green line), many components are suppressed, particularly in the left channel (blue), and the two sidebands closest to the spectral spike that represents the 11.025kHz tone are boosted. This behavior was identical with coaxial and optical S/PDIF data and with USB data. With 24-bit data, there were still some very low-level sidebands visible in the left channel with a USB datastream (fig.12, blue trace), but not with S/PDIF data.


Fig.11 Schiit Yggdrasil, high-resolution jitter spectrum of analog output signal, 11.025kHz at –6dBFS, sampled at 44.1kHz with LSB toggled at 229Hz: 16-bit coaxial S/PDIF data (left channel blue, right red). Center frequency of trace, 11.025kHz; frequency range, ±3.5kHz.


Fig.12 Schiit Yggdrasil, high-resolution jitter spectrum of analog output signal, 11.025kHz at –6dBFS, sampled at 44.1kHz with LSB toggled at 229Hz: 24-bit USB data (left channel blue, right red). Center frequency of trace, 11.025kHz; frequency range, ±3.5kHz.

It's difficult to sum up the Schiit Yggdrasil's measured behavior. While the processor's analog circuitry is superbly well designed, its digital circuitry appears to have problems with high-level, high-frequency tones, and with the LSBs of 24-bit data. It's possible, of course, that the former will be rare with music, and that the latter will be obscured by the noise floors of recordings. But it does look as if the digital circuitry is not fully optimized. Hopefully, this could be addressed with a firmware upgrade.—John Atkinson

Footnote 1: My thanks to Jürgen Reis of MBL for suggesting this test to me.

Footnote 2: See, for example, figs. 4 and 5 here.

Schiit Audio
24900 Anza Drive, Unit A
Valencia, CA 91355
(323) 230-0079

mink70's picture

This is straight-up poetry:

"Live music may be viewed as a continuously pulsating wavefront. If you hold your hand up, you can almost feel it. Recorded music is a coded narrative simulacrum of that pulsing wavefront. If anything in the recording or playback chain interrupts, bends, truncates, or haphazardly disrupts the original (live) continuity—all the world's smart guys can never restore its hyperfragile relationships of time, frequency, and amplitude. Love, music, and poetry live only in the undamaged continuity of those relationships. (Unlike the stock market or election polling, music is not a stochastic process.)"

allhifi's picture


" ...coded narrative simulacrum .."
" ...pulsing wavefront."
" ..haphazardly disrupts..."
" ....hyperfragile .."

" ..undamaged continuity of those relationships."

It's straight-up alright, and perhaps even poetically-challenged !


allhifi's picture

Hi Kal: This review was indeed informative --two DAC's of opposite ends of the sonic spectrum was my take.
However, from one disc, music file or song to the next, which of the two will resolve each one distinctively --revealing the huge distinctions in recording quality/tonality known to exist within the music itself ?

That, to me, would make my decision (to purchase) much easier.


tonykaz's picture

I've been in Audio since the 1950s, my numerous Schiit pieces of headphone gear are the amongst the best performing Audio Electronics I've ever encountered annnnnnd they're cheap by comparison but not cheaply made or appearing.

Now Schiit is immigrating to the Audiophile world with a range of Preamps and Amps. This will be a breath of fresh air for all of us gripers about the Stratospheric Cost of High-End Audio stuff ( $100,000 Mono Amps, for gods sake ).

Designing is an Art Form, Mr. Stoddard is an Artist. I think he designs the entire product, a global type of designer, he does A Level work.

So, Audiophile World, Brace yourself, your about to get traditional Audio Research levels of performance at NAD 3020 price points! PHEW!

Well Done Schiit, just what we've been waiting for.

Tony in Michigan

ps. I'm a Schiit owner, not affiliated in any way with the Schiit Company, other than being a bit of a Critic

rt66indierock's picture

Forgot the DAC you reviewed. Harmonics wipe a big chunk of high end audio. A banjo gets most of the rest as Art Dudley reported his Capital Audiofest coverage last year. And finally fiddle sawing in Cajun music gets almost everything leftover. It is now pretty easy to choose from what little is left.

You must tell me sometime how you got the last paragraph of the review past the editors. Whatever you did keep doing it.

USAudio's picture

Mike Moffat, the digital designer at Schiit and an innovator in the field, has been doing this a long time. Perhaps JA could reach out to the guys at Schiit for details on why they made the design decisions they did, in particular the use of the 20-bit D/A converters and DSP?

USAudio's picture

Also, I don't see anywhere in the Yggdrasil measurements a mention of its measured resolution like we usually do with digital processors?
Per Schiit's website: "Yggdrasil is the world’s only closed-form multibit DAC, delivering 21 bits of resolution with no guessing anywhere in the digital or analog path." -
Did JA indeed find the Yggdrasil's resolution to be 21 bits?

John Atkinson's picture
USAudio wrote:
I don't see anywhere in the Yggdrasil measurements a mention of its measured resolution like we usually do with digital processors?

I didn't offer that judgment because the Yggdrqasil is not like conventional D/A processors that use 24-bit DC chips. With the latter, the resolution will be limited by the thermal noise of the overall DAC+analog circuitry, which produces a random noisefloor. With the Schiit, however, while the analog noise is very low, as I mentioned in the review, the noisefloor will rise with 24-bit audio due to what appears to be the truncation of the LSBs. As the noisefloor will therefore be related to the encoded signal, my usual estimate of the DAC resolution will be misleading.

John Atkinson
Editor, Stereophile

USAudio's picture

Thanks for the clarification John!

rom661's picture

Prior experience with an earlier product line of his makes me apprehensive.

USAudio's picture

I think you'll find with a little web searching and on that Schiit customer support is very well regarded. Anecdotally, and this isn't probably typical, but I once sent a question to Schiit technical support on a Sunday afternoon and got a response within the hour! Mike Moffat's partner, Jason Stoddard, is heavily involved in communicating with the audio community, particularly through

rom661's picture

As I said, this was a different company. Hope your experience is good.

USAudio's picture

"As I said, this was a different company."

"Not Schitt"
It's spelled "Schiit", but I guess that is based on your perspective! ;-)

allhifi's picture

RE: " "Not Schitt" It's spelled "Schiit", but I guess that is based on your perspective!"

Seriously? Who gives a shit.


SteveG's picture

As JA knows, the Schiit employs a 20 or 21-bit DAC. While the DAC may be able to accept a 24-bit signal, it cannot know what to do with the LSB of a 24-bit signal. Thus, his tests reveal neither that the hardware is defective nor that there is something amiss in the software.

John Atkinson's picture
SteveG wrote:
As JA knows, the Schiit employs a 20 or 21-bit DAC. While the DAC may be able to accept a 24-bit signal, it cannot know what to do with the LSB of a 24-bit signal.

My comments were not odd. When you have 24-bit data but 20-bit DACs, you need to dither those data to match the DAC. Otherwise, simply chopping off the 4 LSBs, called "truncation," reintroduces quantizing distortion. Schiit's Jason Stoddard has subsequently said that the Yggdrasil "rounds" 24-bit data but my measurements suggest that the LSBs of 24-bit data are simply truncated.

John Atkinson
Editor, Stereophile

gevorg's picture

>>When you have 24-bit data but 20-bit DACs...

Aren't most, if not all, DACs cannot truly resolve beyond 20-21 bits anyway, including those on Stereophile Class A list? So what difference would "rounding" or "truncation" make with the last ~4 bits of thermal noise/etc? Why should this matter for the audiophile, mastering engineer, critical listener, etc?

John Atkinson's picture
gevorg wrote:
Aren't most, if not all, DACs cannot truly resolve beyond 20-21 bits anyway, including those on Stereophile Class A list? So what difference would "rounding" or "truncation" make with the last ~4 bits of thermal noise/etc?

As I said but you appeared to ignore, truncation of the LSBs in 24-bit data reintroduces quantization noise/distortion. Research by people like Bob Katz has shown that this will be audible even with real-world DACs.

gevorg wrote:
Why should this matter for the audiophile, mastering engineer, critical listener, etc?

Now you could say that you prefer the sound of truncation against that of redithering and I have no argument with that opinion. But if it costs the engineer nothing other than a little bit more silicon real estate to do it right, why not? The end result is reduced–bit-depth data presented to the DAC that will never have audible consequences whereas simply truncating the data will produce artifacts that will be audible, perhaps not with every recording but definitely with some.

John Atkinson
Editor, Stereophile

scottbuzby's picture


Schiit Audio has made significant changes to Yggdrasil. Please re-measure a current production unit. I think you will find it measures much better than the original release.


bapcha's picture

Truncation and rounding create identical amount of rms noise. This is very basic. Claiming that rounding is better than truncation is incorrect - especially since you referred to qualtizing distortion (same thing as noise). Yes, you can check me on that. I studied under Dr. Henry Samueli, UCLA, Broadcom (one of the best known DSP guys on earth)

LarryMagoo's picture


It's a shame you cannot judge the Yggy with your ears as well. Because if you did, it would not have ended up in the Class B....its really that simple!

All any of has to "sell" is credibility. You just lost yours!


allhifi's picture

Wonderful comparison; Mytek "Brooklyn" and "Yggdrasil":

A wonderful contrast of two DAC's. A modern take on A Tale of Two ...." !

It's clear the "Brooklyn" incorportaes the latest understanding and technology of modern DAC design. I suppose the "Yggy" appeals to those who cannot break the bond of analog desires.

The superiority of Brooklyn's resolution shupld be commended --as is their MQA adoption. I envision a superior linear P/S on the Mytek wold elevate into the super-DAC territory (or perhaps the Manhatten 2) ?

But back to the $2K price-point, if Schitt can offer a decent linear P/S, how is it that Mytek cannot do the same for the "Brookyln" ?

Hmmm, my vote would be the Brooklyn". Let's move forward ...


hollowman's picture

I was just looking thru two early-90s Stereophile issues (Dec. 1992; Mar. 1993) ... not a single analog/phono review or article (tho' I did see an Acoustic Sounds ad with LPs and equip.). Some tube gear. And MANY multibit DACs or CDPs -- reviews, articles, ads, classifieds.
I have a late-1980s aftermarket-modified Philips CDP, with the classic TDA1541A and SAA7220 (4x DF). And it's better sounding than all more-modern DACs in my collection (that feature delta-sigma).
Indeed, those early-90s Stereophiles were full of glowing reviews of Arcams and MSBs and Naims ... all using that classic Philips platform.

My suggestion for JA and other Stereophile reviewers is to go into their attic or basement storage ... and dig out these gems. And use THEM as a basis of comparisons against Chord, Border Patrol, Benchmark, etc.

Oh ... JA ... please webify that 1995 Fabio article. He had VERY high-end $$$ system based on a Krell CDP. No phono, no tubes.

allhifi's picture

Herb's: Hmmm. Very odd, you start the review by saying:

" .. From the start, I enjoyed playing CDs. But compared to LPs, digital sounded deficient in contrast and conspicuously artificial, in a plastic-bottles-in-the-ocean kind of way. It felt distant and mechanical." (Sure did)

"Nonetheless, my brain readily adapted to its shortcomings."

Is that a joke? Of course not, so it's therefore most telling; Your 'brain' adapted to its short-comings ? WTF. Those "shortcomings" (back in the 1980's) were so severe/nasty that no normal "brain" could adapt to is shortcomings.

Yet, your comments become even more ridiculous, when you say (35-years later):

" My original Yggdrasil made music in a fun, highly articulate way, but its empty spaces were filled with a fine, vibrating, subliminal grain. That vibrating haze might have originated in my CD transport, my computer, or my brain—but with the Analog 2 upgrade, it was now completely gone. Was the cause the Gen 5 USB upgrade? No matter. Through the Yggdrasil Analog 2, backgrounds were now more transparent and "black" than before. Which, to my delight, allowed new sonic delicacies to emerge."

OMG, LOL; In other words, 35-years prior to Schiit's DAC/Yggy, you were happier than a swine in dung (your "brain" adapted to 1982's horrific CD sound), and then, near 40-years later you're taken/impressed by the second Generation of a $2K DAC because you can now:

" to my delight, allowed new sonic delicacies to emerge."

AND (the best part), you now have the "sensitivities" to identify and enjoy :

A)".Analog 2, backgrounds were now more transparent and "black" than before."

B) " ...Which, to my delight, allowed new sonic delicacies to emerge."

C) " ... I'm mesmerized by the tiniest subtleties of the master's touch. My mind follows and sees the butterfly-like tactility of every keystroke."

D) "The Analog 2 upgrade retained all of the original Yggdrasil's virtues, especially its force and drive, but with a new spiderweb delicacy, more lower-octave density, and, most important, a more complete transparency.

I'm near ready to throw-up:

" ..spider-web delicacy ...low octave density ...complete transparency"

Yet, back in 1982, again, using your words:

" ..FROM THE START, I enjoyed playing CDs ..."

An oxymoron if there ever was one; if one loves/understands/feels music, there is not a lobotomy in hell that could have fixed our brains to make sense of that atrocity of a (SQ) technology called the Compact Disc, back in 1982.

In fact, Herby, that was pretty much the status quo "sound" for the next 20 (CD) years as well -SHIT.

But, here you are, extolling the (rightfully) intricate details, subtleties (virtues) of digital replay NEARLY 40-YEARS after you admittedly enjoyed back in 82' !

A fake, and fraud.

peter jasz