Review and Measurements of Schiit Yggdrasil V2 DAC

[SIY]

What controls did you use?

If none...don't be upset by the predictable responses.

You saved me typing.

 

[Alec Kinnear]

What controls did you use?

If none...don't be upset by the predictable responses.

A fair question Woody. I had all three DACs set up on a manual (no electronics) switch. All of the DACs were being fed by the same optical output from a Mac Pro via an optical splitter. Hence same exact input in perfect sync. The signal was amplified by a Topping A50s into the amplifier section of a Yamaha R-S700. I also tried using the built-in preamp functionality on the D50s and the E30. I also had a Loxjie D30 and a Aune X8 through just a little earlier in my quest to find the right DAC/preamp/headphone amp setup for my desk. As well as listening to whole songs, switching between the songs, often I played the music on twenty second loops of particularly revealing sections. Some of the testing was done with headphones (Sennheiser HD598, Denon AHD600, Monoprice M1060) with the A50s as the main headphone amp. Differences are often clearer between DACs via headphones.

What's important here is that there was no lag of recabling or button clicking between listening. I could switch between three DACs in real time. @amirm is right that comparative listening tests must be instant to minimise any psychological colouring. Even five or ten seconds of interruption is enough to introduce uncertainty. Which is why I bought the digital splitters and brought my old manual switcher out of storage and tested a whole bunch of connects to make sure they all transmitted music signal accurately.

I'm able to tell the difference between on good quality, nominally transparent DACs on vocals (which are a huge part of most of the music I listen to, Sandy Denny, Barbara, Mazy Star, Melanie, Melissa Horn, Luna, Bridget St John, Fotheringay, Austra, Françoise Hardy, Lana del Rey as a sample). On instrumental sections, it's harder for me to hear the differences unless a DAC is clearly inferior or out of spec.

I spent days on these tests (while doing other work), blinding them, revisiting them. I was so patient as I was getting results opposite from what I wanted. I wanted to keep a Topping D50s/A50s/P50 stack and dispose of the rest of my gear (Musical Fidelity VDAC II, Musical Fidelity V90, various pre-amps), as I preferred the small form factor and harmonious aesthetics (silver version). The overall quest was to have both headphone and pre-amp volume controls physically on my desk, which ruled out full size components.

All vocals, but in particular female vocals, are recessed in the D50s. Originally I thought it might be the A50s but isolating the D50s from the A50s or putting all the DACs through the A50s revealed that the recessed vocals were specific to the D50s. I could happily enough listen to either a Musical Fidelity V90 or a Topping E30 on a long term basis. The E30 does need a bit of EQ to soften the treble, while the V90 has just a slightly fuller mid-range (electric guitar strumming). The D50s has the most emphatic guitar and bass lines of these three – someone who mainly listens to guitar music and/or instrumental music might like the D50s best.

I don't believe one can hear everything. I was not able to consistently hear a difference between the D50s/A50s stack with or without the P50 linear power supply – even though I wanted to hear such a difference (keeping the complete Topping 50 stack would have suited me best in terms of aesthetics, flexibility, space and budget).

The ferocity and aggression of these true believers (see @VintageFlanker's shrieking below) reminds me less of science and more of the Catholic church's attacks on Galileo and da Vinci for heresy. Galileo and his views on a heliocentric earth (as opposed to flat earth). You may not know that da Vinci was almost tried and murdered for his diet:

Certain researchers claim, however, that they have found extraordinary evidence that some of those beliefs fly in the face of accepted Christianity and that Leonardo da Vinci was in fact an active heretic. Leonardo’s life was in danger in 15th century Italy simply because he was a vegetarian, as shocking that might seem in today’s society.

A little bit of calm and courtesy gentlemen. Until much better tests are conceived and finer measuring instruments are built, there must be a place for both measurements and listening in evaluating audio gear.

 

[Jimbob54]

A fair question Woody. I had all three DACs set up on a manual (no electronics) switch. All of the DACs were being fed by the same optical output from a Mac Pro via an optical splitter. Hence same exact input in perfect sync. The signal was amplified by a Topping A50s into the amplifier section of a Yamaha R-S700. I also tried using the built-in preamp functionality on the D50s and the E30. I also had a Loxjie D30 and a Aune X8 through just a little earlier in my quest to find the right DAC/preamp/headphone amp setup for my desk. As well as listening to whole songs, switching between the songs, often I played the music on twenty second loops of particularly revealing sections. Some of the testing was done with headphones (Sennheiser HD598, Denon AHD600, Monoprice M1060) with the A50s as the main headphone amp. Differences are often clearer between DACs via headphones.

What's important here is that there was no lag of recabling or button clicking between listening. I could switch between three DACs in real time. @amirm is right that comparative listening tests must be instant to minimise any psychological colouring. Even five or ten seconds of interruption is enough to introduce uncertainty. Which is why I bought the digital splitters and brought my old manual switcher out of storage and tested a whole bunch of connects to make sure they all transmitted music signal accurately.

I'm able to tell the difference between on good quality, nominally transparent DACs on vocals (which are a huge part of most of the music I listen to, Sandy Denny, Barbara, Mazy Star, Melanie, Melissa Horn, Luna, Bridget St John, Fotheringay, Austra, Françoise Hardy, Lana del Rey as a sample). On instrumental sections, it's harder for me to hear the differences unless a DAC is clearly inferior or out of spec.

I spent days on these tests (while doing other work), blinding them, revisiting them. I was so patient as I was getting results opposite from what I wanted. I wanted to keep a Topping D50s/A50s/P50 stack and dispose of the rest of my gear (Musical Fidelity VDAC II, Musical Fidelity V90, various pre-amps), as I preferred the small form factor and harmonious aesthetics (silver version). The overall quest was to have both headphone and pre-amp volume controls physically on my desk, which ruled out full size components.

All vocals, but in particular female vocals, are recessed in the D50s. Originally I thought it might be the A50s but isolating the D50s from the A50s or putting all the DACs through the A50s revealed that the recessed vocals were specific to the D50s. I could happily enough listen to either a Musical Fidelity V90 or a Topping E30 on a long term basis. The E30 does need a bit of EQ to soften the treble, while the V90 has just a slightly fuller mid-range (electric guitar strumming). The D50s has the most emphatic guitar and bass lines of these three – someone who mainly listens to guitar music and/or instrumental music might like the D50s best.

I don't believe one can hear everything. I was not able to consistently hear a difference between the D50s/A50s stack with or without the P50 linear power supply – even though I wanted to hear such a difference (keeping the complete Topping 50 stack would have suited me best in terms of aesthetics, flexibility, space and budget).

The ferocity and aggression of these true believers (see @VintageFlanker's shrieking below) reminds me less of science and more of the Catholic church's attacks on Galileo and da Vinci for heresy. Galileo and his views on a heliocentric earth (as opposed to flat earth). You may not know that da Vinci was almost tried and murdered for his diet:


A little bit of calm and courtesy gentlemen. Until much better tests are conceived and finer measuring instruments are built, there must be a place for both measurements and listening in evaluating audio gear.

How did you match the levels?

Edit, and did you select the same filter types?

 

[BDWoody]

I spent days on these tests (while doing other work), blinding them, revisiting them.

Thank you.
How did you match levels, and what was your blinding process?

 

[Alec Kinnear]

How do you know Audirvana isn't bit perfect?

Audirvana is the only player which does not sound identical out of the Swinsian, Calibri, Decibel, Fidelia, Audirvana. All on OS X of course, all supposed to be bit-perfect. Let's take the Audirvana conversation over here.

 

[DSJR]

Recessed vocals but no obvious response change - inverted absolute phase on one dac I wonder? Some systems, rooms and listeners appear sensitive to it with some music recordings..

 

[Alec Kinnear]

How did you match the levels?

All the DACs were going through a single pre-amp (using different pre-amps would be unfair).

The levels could not all be matched. For the D50s vs the other DACs with pre-amp functionality I was able to match levels by running the D50s at -2 dB for instance and the Aune X8 at full output. Levels were matched on 1 kHz signal tones. For the Musical Fidelity V90, a small manual adjustment of the volume of the A50s had to be made as the V90 runs one dB louder. The output is 2.1 VRMS I believe instead of 2 VRMS. I may be getting the measurement wrong. Yes, the slight volume difference did interfere with testing and made it much more difficult.

@BDWoody No need for thanks. I did not do these tests to please the Inquisitors of AudioScienceReview.com but to put together some accurate information for myself and to share with others. My intention was to spend my money wisely. The NAD C372 had been an excellent choice fifteen years ago and the Musical Fidelity VDAC II and V90 were very good choices in 2011 and 2014. On the other hand, the Yamaha R-S700 had not been a success (amplifier section is not bad, pre-amp is awful). I'm a relative newcomer here as I found @amirm's objective tests very useful when researching DACs and pre-amps.

Blinding. I would move around the switch not looking at, go and do some other things and listen for awhile and then give an evaluation to what I was listening to. My accuracy with the D50s was upwards of 80%. My accuracy between the E30 and V90 was lower (levels didn't matter here as they were extended listening tests and not swapping between DACs). I found non-blinded tests where I focused on twenty seconds of music with particular vocals or instrumental sounds more useful for isolating particular sonic ticks in either stereo separation (soundstage) and frequency emphasis. The important part is to switch instantly and to listen to exactly the same section of music. Switching mid-song between DACs while moving forward can be deceiving as there are nuances of performance even on chorus sections which cloud one's ability to distinguish sonically between DACs.

All of this is fairly subtle stuff – which is why it mostly doesn't show up in objective measurements. On the other hand, like you, I spend many thousands of hours listening to music, and in may case, particularly vocal music. I'd like to enjoy the most suitable (to me) playback of this music within my budget. I don't say most accurate, as all of the DACs mentioned here reach the audible limits of accuracy in terms of measurement. Accuracy was a pre-requisite. Absolute accuracy is an illusion, as all speakers and all headphones colour music even if an absolutely neutral DAC or amplification could be found (Benchmark products are pretty close). Since there will be some tonality in whatever we choose, it should be a tonality which suits the individual listener.

 

[Jimbob54]

All the DACs were going through a single pre-amp (using different pre-amps would be unfair).

The levels could not all be matched. For the D50s vs the other DACs with pre-amp functionality I was able to match levels by running the D50s at -2 dB for instance and the Aune X8 at full output. Levels were matched on 1 kHz signal tones. For the Musical Fidelity V90, a small manual adjustment of the volume of the A50s had to be made as the V90 runs one dB louder. The output is 2.1 VRMS I believe instead of 2 VRMS. I may be getting the measurement wrong. Yes, the slight volume difference did interfere with testing and made it much more difficult.

@BDWoody No need for thanks. I did not do these tests to please the Inquisitors of AudioScienceReview.com but to put together some accurate information for myself and to share with others. My intention was to spend my money wisely. The NAD C372 had been an excellent choice fifteen years ago and the Musical Fidelity VDAC II and V90 were very good choices in 2011 and 2014. I'm a relative newcomer here as I found @amirm's objective tests very useful when researching DACs and pre-amps.

Blinding. I would move around the switch not looking at, go and do some other things and listen for awhile and then give an evaluation to what I was listening to. My accuracy with the D50s was upwards of 80%. My accuracy between the E30 and V90 was lower (levels didn't matter here as they were extended listening tests and not swapping between DACs). I found non-blinded tests where I focused on twenty seconds of music with particular vocals or instrumental sounds more useful for isolating particular sonic ticks in either stereo separation (soundstage) and frequency emphasis. The important part is to switch instantly and to listen to exactly the same section of music. Switching mid-song between DACs while moving forward can be deceiving as there are nuances of performance even on chorus sections which cloud one's ability to distinguish sonically between DACs.

All of this is fairly subtle stuff – which is why it mostly doesn't show up in objective measurements. On the other hand, like you, I spend many thousands of hours listening to music, and in may case, particularly vocal music. I'd like to enjoy the most suitable (to me) playback of this music within my budget. I don't say most accurate, as all of the DACs mentioned here reach the audible limits of accuracy in terms of measurement. Accuracy was a pre-requisite. Absolute accuracy is an illusion, as all speakers and all headphones colour music even if an absolutely neutral DAC or amplification could be found (Benchmark products are pretty close). Since there will be some tonality in whatever we choose, it should be a tonality which suits the individual listener.

Sorry, I edited my question to add in how did you deal with the filters to ensure comparability (I imagine some can't vary the filter).

 

[Alec Kinnear]

Recessed vocals but no obvious response change - inverted absolute phase on one dac I wonder? Some systems, rooms and listeners appear sensitive to it with some music recordings..

Perhaps. Is it possible for a switching power supply like the one which comes with the D50s to have inverted phase? The D50s sounded the same when plugged into the P50 power supply as well.

Sorry, I edited my question to add in how did you deal with the filters to ensure comparability (I imagine some can't vary the filter).

When I did filter tests, I found the differences except on DSD music (which suffers from what should be inaudible high frequency noise) to be so subtle as to be unimportant for a preference for one DAC or another. When doing the filter tests though I was not hunting for a filter I loathed but to ensure I wasn't listening to a filter which had a negative effect on the music. My own hearing fades out above 13 kHz these days so filter testing which probably explains why the low pass filters on these DACs didn't make much difference to my listening. I'm an imperfect instrument and the manufacturer won't send me replacement parts as I'm outside of warranty period and that even in Europe he has no legal obligation after seven years to provide any parts at all.

 

[ousi]

Interesting discussion. Honestly, if anyone ever dipped a toe into DSP will know that "sound stage" or perceived width and depth of the audio presented can be manipulated with subtle phase change, crosstalk and amplitude of different frequency. The latest batch of "good sounding" home speakers all does that, and at least the Sonos Arc soundbar (which I have) does the same too to make it an illusion of how wide the sound is. For headphone, one example is the Macintosh HxR which did some processing in the amount of cross-talk and phase delay to make the illusion of listening to a pair of speakers instead of headphones. I had an MHA150 before, didn't find it very exciting though.

One thing that Armin's measurements didn't cover is the amount of phase shift for different frequencies. L/R Crosstalk vs frequency was actually covered somehow with the L/R separation. Frequency response is always covered.

If one wants different sound character, we can just "tune" it post DAC. The DAC (or sources) should be accurate and as noise free as possible so that you can effectively apply EQ or tune by different equipment (be it the pre amp, power amp, "cables", speakers, headphones, etc) downstream. Just my $0.02.

 

[DSJR]

Perhaps. Is it possible for a switching power supply like the one which comes with the D50s to have inverted phase? The D50s sounded the same when plugged into the P50 power supply as well.

Apologies, I was referring to the audio output being inverted absolute phase. I've no idea but it's all I could think of as with 'eyes switched off,' there really shouldn't be a difference if levels are matched and you don't know which is which (old ground here, sorry)

 

[Jimbob54]

Interesting discussion. Honestly, if anyone ever dipped a toe into DSP will know that "sound stage" or perceived width and depth of the audio presented can be manipulated with subtle phase change, crosstalk and amplitude of different frequency. The latest batch of "good sounding" home speakers all does that, and at least the Sonos Arc soundbar (which I have) does the same too to make it an illusion of how wide the sound is. For headphone, one example is the Macintosh HxR which did some processing in the amount of cross-talk and phase delay to make the illusion of listening to a pair of speakers instead of headphones. I had an MHA150 before, didn't find it very exciting though.

One thing that Armin's measurements didn't cover is the amount of phase shift for different frequencies. L/R Crosstalk vs frequency was actually covered somehow with the L/R separation. Frequency response is always covered.

If one wants different sound character, we can just "tune" it post DAC. The DAC (or sources) should be accurate and as noise free as possible so that you can effectively apply EQ or tune by different equipment (be it the pre amp, power amp, "cables", speakers, headphones, etc) downstream. Just my $0.02.

Why not eq /dsp at source (pre dac)? Though I don't think you're ruling that out, just wanting the hardware boxes to be as neutral as possible?

 

[ousi]

Why not eq /dsp at source (pre dac)? Though I don't think you're ruling that out, just wanting the hardware boxes to be as neutral as possible?

Fixed EQ at source gives you less flexibility. As we all know audiophile taste changes It also depends on what you called source. Some DACs and even Roon allow you to apply DSP to do EQ.
What I meant is that the conversion should be as neutral as possible upstream so that you have a clean slate to add the taste you want - be it just some EQ or room correction DSP

 

[Jimbob54]

EQ at source gives you less flexibility. As we all know audiophile taste changes It also depends on what you called source. Some DACs and even Roon allow you to apply DSP to do EQ

I'm not sure we are on the same page. Eq in or just after the player software is infinitely more flexible. Roon for eg. I call the source whatever is playing the file/disc. Eq as early up the stream as possible in the digital domain.

 

[daftcombo]

I have one here and it offers a much smaller soundstage than a Topping E30 and an NAD C165BEE. So all your measurements really don't help in this case. The limited set of measurements made at AudioScienceReview don't even begin to characterise musical sound.

Did you do a level-matched ABX or is it just in your mind?

 

[ousi]

I'm not sure we are on the same page. Eq in or just after the player software is infinitely more flexible. Roon for eg. I call the source whatever is playing the file/disc. Eq as early up the stream as possible in the digital domain.

I meant "Fixed EQ" as in DACs with built-in "sound signature" which you cannot change, or those so-called remastered discs that has certain frequency ranged turned up higher, or overall compressed in general. Imagine applying EQ on Roon and then you go through some DAC that adds its own sound signature. The source I meant is the source device that converts digital signal to analog signal.

Depends on implementation of the EQ, sometimes I prefer to do it in the analog domain. There are some professional processors that does the EQ in DSP without taking into account of the extra headroom needed, and caused hard clipping. Roon seems to have implemented it correctly so I do use it sometimes for my bedroom setup.

 

[rkbates]

The worst part, @VintageFlanker is the poster (@hudo) is right about the RME ADI-2 DAC FS. I have one here and it offers a much smaller soundstage than a Topping E30 and an NAD C165BEE. So all your measurements really don't help in this case. The limited set of measurements made at AudioScienceReview don't even begin to characterise musical sound. The only thing the set of measurements made here can help with is identifying poor gear. Identifying poor/flawed gear is no small achievement, but it's only a start.

In the case of the RME ADI-2 DAC FS it does have other attributes which help make up for the tight soundstage. On the sound state the RME ADI-2 DAC FS offers almost zero hiss and an excellent parametric equaliser. On the feature side, good headphone amplifier with separate EQ, tone and volume controls;

Based on your obsession with measurements and persistently rude posts, you are slowly convincing me that neither music nor the musical interest you. Since you like schooling people, allow me to return the favour. As Stereophile understood, measurements are a good starting point to reveal hidden flaws and recognise exceptional performance. What matters though is how audio equipment reproduces music, which is why the measurements follow one or two listening tests. Measurements are just a starting point.

In the particular case of this Schiit DAC, more than the flawed measurements what disturbs me is the poor grounding and inadequate safety features, particularly at the price point. On the other hand, when touching the case of the Topping D50s, A50s and P50 all give off 5v and 12v shocks (at that level it isn't a hard shock but an uncomfortable feeling like crawling on one's skin) as they are poorly grounded. The properly engineered for safety RME ADI-2 DAC FS does not do so. A device which doesn't send electricity through my body every time I touch it matters to me.

You'll have to create some new tests if you wish to characterise musical sound with measurements only. A good starting point would be a test which differentiates between a wide and open soundstage and a more intimate closed one. When I've tried to measure such differences with a UMIK-1 coming out of speakers, it's impossible to differentiate most DACs, even when the differences are easily audible. Here's a comparison of the Topping E30, Topping D50s and Musical Fidelity V90.

View attachment 121964

Yet the sonic differences are clear when listening to them (vocals are recessed in D50s, treble is a little brighter in E30, widest soundstage, vocals are smoothest with least grain on V90).

Amir measures for stereo separation but that doesn't seem to have a direct co-relation to the feeling of spaciousness or soundstage. Another hitch is that Amir uses Audirvana as his player – Audirvana output is not bit perfect. Any measurement of stereo separation that Amir passes through Audirvana is worthless in any case, as Audirvana sounds better precisely because of a stereo separation filter (there's also a very slight boost in treble around 8 kHz which adds presence).

In Amir's own words, the whole point is the music:



In any case, a little courtesy and mutual respect would go a long way to making AudioScienceReview a more agreeable place where both music and measurements productively co-exist.

Just trying to understand the discussion (and learn a bit maybe) - are you saying these measurements with the UMIK-1 don't distinguish the DACs? At 400 Hz I see a 5 dB difference, other places I see 0 dB. This is definitely audible, so assuming nothing else has changed (mike position etc) then the difference must be down to the DACs and yet the DACs definitely don't have a 5dB difference in frequency response. So what is causing the 5 dB difference? In the interests of disclosure, I have yet to see any evidence that measurements don't tell the full story, I do however see errors in measurement methodology, technique and interpretation (and made lots of them myself).

 

[Kw6]

I never heard the Schiit Yiggy A2 but is it true the soundstage is more 3D and has many layers to it as reviews say? Or is it marketing BS? Thanks!

 

[majingotan]

Spatial properties such as soundstage and the illusion of a 3D presentation, layering etc. are from the music source and NOT the DAC (unless the DAC is severely broken like NOS filterless DACs where FR is NOT flat frequency). Yggdrasil has flat FR on balanced signal path so it should not add its own effect to the sound.

Quoting Amir from his post here: https://www.audiosciencereview.com/...cal-utopia-review-headphone.22103/post-734010

On the question of spatial qualities, this is *hugely* dependent on content. I have collected a library of content for headphone testing which shows this effect. Without it, it is easy to conclude that there is no such thing.

Also, frequency response in 1 to 3 kHz is a large determinant of this effect. Take any headphone you have and boost this region and listen for spatial qualities. Likely you hear the effect (ignore the tonality difference -- just listen for externalization of the sound).

 

[Kw6]

Spatial properties such as soundstage and the illusion of a 3D presentation, layering etc. are from the music source and NOT the DAC (unless the DAC is severely broken like NOS filterless DACs where FR is NOT flat frequency). Yggdrasil has flat FR on balanced signal path so it should not add its own effect to the sound.

Quoting Amir from his post here: https://www.audiosciencereview.com/...cal-utopia-review-headphone.22103/post-734010

Ok thanks for clearing that up! I don't know why reviewers say that. One person said the Yiggy was superior in layered depth in comparison to Chord Hugo 2 which was comparatively 2 dimensional!