Review and Measurements of Schiit Yggdrasil V2 DAC

[ElNino]

This one? https://www.audiosciencereview.com/forum/index.php?threads/review-battle-of-schiit-audio-dacs.5487/

In the review there it looks pretty bad. Not audio-gd bad but close... Besides, all the new version 'B' does is actual dithering from 24-bit to 16-bit for better linearity. The other characteristics have not changed.

That's actually not too bad as far as Schiit goes. No one's posted a Bifrost 2 multitone, so it's hard to compare. The Schiit AP measurements for the revision B Modi Multibit are here: https://www.schiit.com/public/upload/PDF/Schiit DAC APx555 Standard Test Suite_ Modi Multibit.pdf The 1kHz FFT is actually remarkably clean for a 16-bit converter, apart from the prominent second harmonic. The 3rd harmonic is 100dB down and there is no other prominent harmonic, apart from the usual rise in high frequencies that's visible on all R2R devices.

 

[ElNino]

You're right btw, Bifrost 2 IMD is pretty bad. That is unfortunate, since it seems to measure better distortion-wise than their other multibits. Odd.

Yes, plus you should check out the 19+20kHz IMD measurement posted on SBAF. The 1kHz difference tone is only something like -63dB down in one channel.

 

[Veri]

That's actually not too bad as far as Schiit goes. No one's posted a Bifrost 2 multitone, so it's hard to compare. The Schiit AP measurements for the revision B Modi Multibit are here: https://www.schiit.com/public/upload/PDF/Schiit DAC APx555 Standard Test Suite_ Modi Multibit.pdf The 1kHz FFT is actually remarkably clean for a 16-bit converter, apart from the prominent second harmonic. The 3rd harmonic is 100dB down and there is no other prominent harmonic, apart from the usual rise in high frequencies that's visible on all R2R devices.

I disagree. That multitone does not look good and borders at audible distortion in music. I would not recommend modi MB to anyone for its price.
Also, your criticism for IMD is actually worse in the modi multibit.

Yes, plus you should check out the 19+20kHz IMD measurement posted on SBAF. The 1kHz difference tone is only something like -63dB down in one channel.

Intriguing. Could perhaps be a measurement artifact? Not very encouraging in any case.

This is only solidifying my stance on never buying any multibit schiit
I stand by modi + magni heresy being their best product out right now. Bifrost2/asgard would not get close to its performance :/

 

[ElNino]

I disagree. That multitone does not look good and borders at audible distortion in music. I would not recommend modi MB to anyone for its price.
Also, your criticism for IMD is actually worse in the modi multibit.

I think the Modi Multibit CCIF graph is overall slightly better. At least it dips to around -85dB at some point, whereas the Bifrost 2 CCIF never falls below -78dB.

Given the poor 19+20kHz measurement for the Bifrost 2, and generally higher levels of IMD, my theory is that the Bifrost 2 multitone will be awful, once someone gets around to measuring it. The 1kHz FFT of the Modi Multibit is also quite a bit cleaner than the 1kHz FFT of the Bifrost 2, apart from the 2nd harmonic. The Bifrost 2 has a spray of higher order harmonics that are higher in level than the Modi Multibit.

 

[ElNino]

BTW, the weirdest device for IMD is the Gungnir Multibit. If you look at the CCIF charts in the AP measurements Schiit posted for the Gungnir, one channel has excellent performance (easily better than the Modi Multibit and Yggdrasil) and the other channel has poor performance. It's like one channel is good and one is broken.

 

[Veri]

Given the poor 19+20kHz measurement for the Bifrost 2, and generally higher levels of IMD, my theory is that the Bifrost 2 multitone will be awful, once someone gets around to measuring it.

One member here can't tell it apart from an apple usb/lightning dongle
I think it would sound 'fine'/indiscernible.

 

[majingotan]

Surprisingly with all these horrible IMD and multitone tests, I still couldn’t discern a difference from a sighted A/B volume matched a borderline mediocre measuring Chord Mojo and a decently measuring Apple lightning dongle to my horribly measuring Bifrost 2. Though I would think that maybe if I listen at over 105 dB SPL, I might finally hear the distortions from the Bifrost 2. I did try going 105 dB SPL but the HS7 starts showing signs of elevated distortion since it’s not meant to be played that loud. It only has ~50 W RMS total (the 95W power claim is actually peak power and not RMS)

 

[ousi]

I bought and returned multiple Schiit DACs since Modi 2 days including Yggdrasil ("V1") (Bifrost, Modi, Modi2, Modi 3, Modi Multi-bit, Gungnir MB, Gungnir stayed for a while but still sold. Yggdrasil V1.... then I just gave up). The only one my ears can tolerate for a while is the Gungnir (not Multibit). All others will make my ears hurt, especially with hot and "badly" mastered tracks as a result of loudness war, and surprisingly on the Multi-bit implementation. It's interesting people said it's "warmer" but I just found it fuzzier with less resolution, and some strange "noise" that my ears are not happy about. The only multi-bit DAC that works well for me was the now defunct Monarchy Audio DAC-24 using dual PCM63. If anyone wants those DAC chips I know Mr. Poon and he's willing to sell the lot. Guaranteened it's not ruler flat as it has a tube buffer, but the subjective listening experience it's the best DAC I ever had ever... too bad it doesn't decode DSD which I have a tons of...

Back to the topic... People said it's my system matching (well single-driver full range speaker they said; but it also happened for my floorstanders like Sonus Faber Olympic II which I sold and my current Paradigm Premier 800F) but for a DAC which is a source it should be as accurate as possible. Leave it for the pre and power to deal with the "tone" or "color". I do enjoy a good bit of distortion from some good tubes (not the "new" ones) from time to time. But if the source is not clean, it will just be worse down the chain...

 

[manisandher]

It seems I ruffled some feathers over at Audiophile Style...

I posted @amirm 's 1 kHz measurements of the Yggdrasil on a thread about Schiit's new Unison USB board. Chris Connaker deleted my posts, citing "debunked measurements" as one of the reasons. And yet, Schiit's own measurements (posted on their site) are just as damning as Amir's:

The only explanation I can think of is that Chris is looking out for his sponsor. I'm open to another explanation though, if anyone has one...

Mani.

 

[amirm]

I'm open to another explanation though, if anyone has one...

He doesn't understand the measurements and so goes by what people says. And yes, he likes to tilt in the direction that helps his goal.

 

[manisandher]

He doesn't understand the measurements and so goes by what people says. And yes, he likes to tilt in the direction that helps his goal.

Here are Schiit's own measurements vs. yours:

I'm assuming that the difference in noise floor is due to a difference in bin size for each plot. But why does yours show significantly more noise than theirs?

Mani.

 

[Veri]

Here are Schiit's own measurements vs. yours:

View attachment 47843

I'm assuming that the difference in noise floor is due to a difference in bin size for each plot. But why does yours show significantly more noise than theirs?

FFT size ? Besides the abundancy of 'spikes' these measurements are still remarkably alike..

 

[manisandher]

Besides the abundancy of 'spikes' these measurements are still remarkably alike..

Yep. And yet Chris Connaker is saying that Amir's measurements have been "debunked".

I'd still like to know why there are 'spikes' on one and not the other though.

Mani.

 

[Soniclife]

Yep. And yet Chris Connaker is saying that Amir's measurements have been "debunked".

I'd still like to know why there are 'spikes' on one and not the other though.

Mani.

Was this the measurement from before they changed the firmware to not truncate to 20 bits?

 

[amirm]

Yep. And yet Chris Connaker is saying that Amir's measurements have been "debunked".

I'd still like to know why there are 'spikes' on one and not the other though.

Mani.

Last I looked they used toslink input and not usb.

 

[manisandher]

Last I looked they used toslink input and not usb.

Nah, that can't possibly be correct. Surely they would have used the USB input, seeing as Moffat had apparently "solved" USB back in 2017 with his Gen 5 USB board.

But just to be on the safe side, his new Unison board really, really "solves" USB... until the next one comes out in a couple of years time, of course .

A total joke.

 

[TGB]

I'm doing a side by side comparison between Schiit Yggdrasil v2 and Matrix Audio Element X. I'm kind of undecided, yes - the Element X is cleaner, but it sounds more "digital" and flat. I'm borrowing the Element X for some days to compare, goal is to see if I can replace Moon 430HAD, Schiit Yggdrasil v2 and Moon MiND2 Network Player with the Element X (doing this switch will leave my bank account filled). I like the all in one box with Element X, but again - I'm leaning against it sounding flatter and more boring, hence there are more in this world than measurements.

Defective display on the Element X, it is the hardly used demo model for the Norwegian distributor.

Edit:
I have turned of the Yggdrasil v2 when not using it; else I will potentially fry the Element X

 

[SIY]

...hence there are more in this world than measurements.

Yes indeed, like actual controlled listening tests (level-matched and ears-only), which help prevent nonsensical claims about the "sound" from being posted.

 

[Jimster480]

Flat?
If you mean that the Yggdrasil sounds worse because of distortion then yes. Listen for microdetails.

There is no such thing as "sounding digital" unless you have copious amounts of distortion. You should do ABX blind testing. If you are using sighted bias to confirm what you believe before you listen then you won't get anywhere.

 

[TGB]

Flat?
If you mean that the Yggdrasil sounds worse because of distortion then yes. Listen for microdetails.

There is no such thing as "sounding digital" unless you have copious amounts of distortion. You should do ABX blind testing. If you are using sighted bias to confirm what you believe before you listen then you won't get anywhere.

Technically the Element X is better, better at microdetails - but it sounds slightly flat and boring (just came from a live-event; the Yggdrasil v2 reproduced that feeling much closer. I have not decided yet, but sigma-delta DACs have a different sound than multibit (ladder) DACs. My input was that there is more in the world of audio than perfect measurements. Currently I need more time to decide (I love more money on my account which makes me think of element X). Based on the measurements of the element X I expected this to be easy.....

Edit:
After some hours of listening to the element X I though my mind were set on selling the Yggdrasil v2. Then I fired up the Yggdrasil...