• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Review and Measurements of Schiit Yggdrasil V2 DAC

zalive

Active Member
Joined
Sep 4, 2018
Messages
263
Likes
38
This does not mean people won't "like" the device, and I have no objection to people preferring audible distortion over transparency.

Just to be more complex, we're possibly (or probably) talking about personal preferrence of a specific distortion profile.

But the whole idea about 'ok, I've done three objective tests and this is what faithfully translates to reality as we hear it' is possibly flawed. I've just spent a nice evening with a local guru (audio equipment designer, actually), so I specifically asked them: how does digital ringing (he explained to me it's acually pre-ringing and post-ringing in stake) show in measurements. And he said it's about time domain (though not necessarily jitter) difference and that it would show in a spectral analysis - don't ask me further as I myself would not know about this to offer an explanation, and I didn't ask further). And that the time difference would not show in THD or IMD measurements. So...if he's right then the objective set of measurements is not enough for a full legit evaluation of what we hear. His thesis is that time domain is actually the most important in perceiving reality as our hearing mechanism is particularly sensitive to it, so natural time domain relations must be preserved for the high fidelity.

All this makes it possibly more complex, and perhaps beyond possible to describe equipment behavior with three different measurements model. Which would possibly mean: objective as done here/today is not objective enough as it doesn't cover all the aspects that we do hear.
 
Last edited:

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
@Ray2k, glad you're not offended.

If I asked you if you preferred Bach to Schubert, I would do so presuming you could tell the difference between the two.

But we have here two electronic devices. Apart from the fact that they are not artists, it's important to note that the science - not just the measurements of these two units, but actually a large body of research that spans electronics, psychoacoustics and psychology - suggests that there is either no audible difference between them, or only a slight audible difference.

This is why I'm curious to see if you can discern the difference once the psychological variables have been removed.

I'm not suggesting you're more prone to the influence of the psychological variables than anyone else here btw. We simply all are. It is human.

Anyway, I know you're not going to do it and I respect that.

My other question is why you'd want to try a more expensive DAC than the Benchmark or Schiit? Why do you think that a more expensive DAC will sound better?

And final question, do you think the Schiit sounds better to you because it's more transparent than the Benchmark?
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
Just to be more complex, we're possibly (or probably) talking to personal preferrence of a specific distortion profile.

But the whole idea about 'ok, I've done three objective tests and this is what faithfully translates to reality as we hear it' is possibly flawed. I've just spent a nice evening with a local guru (audio equipment designer, actually), so I specifically asked them: how does digital ringing (he explained to me it's acually pre-ringing and post-ringing in stake) show in measurements. And he said it's about time domain (though not necessarily jitter) difference and that it would show in a spectral analysis - don't ask me further as I myself would not know about this to offer an explanation, and I didn't ask further). And that the time difference would not show in THD or IMD measurements. So...if he's right then the objective set of measurements is not enough for a full legit evaluation of what we hear. His thesis is that time domain is actually the most important in perceiving reality as our hearing mechanism is particularly sensitive to it, so natural time domain relations must be preserved for the high fidelity.

All this makes it possibly more complex, and perhaps beyond possible to describe equipment behavior with three different measurements model. Which would possibly mean: objective as done here/today is not objective enough as it doesn't cover all the aspects that we do hear.

@DonH56 is our resident expert on these matters but I think I know enough to give a brief reply to this one.

Yes, your friend is correct in one sense, most DACs produce pre and post ringing, and this does not show up on the measurements that Amir takes here.

However, he's incorrect in a few important ways.

Firstly, Amir's measurements focus on the key performance aspects and (for reasons I'm about to explain) an impulse response which would show up that ringing is not one of them.

Secondly, although your friend is correct that digital ringing occurs in almost every DAC, in any half-decent DAC (or just about any DAC running at a higher res sample rate) this ringing occurs only above 20KHz, which is above the range of hearing for most humans. So yes, it happens, but no, you can't hear it. And I don't mean you can't hear it in the way you can't hear 0.001% distortion. I mean you can't hear it in the sense that your ears do not even hear anything in the frequency range in which it occurs.

Thirdly, his thesis that the time domain is actually most important is a popular myth, but it has been debunked many times by controlled studies, which as in all other areas of audio have sought to establish thresholds of audibility. It turns out that the time domain is one area in which humans are have an incredibly low level of sensitivity - at least compared to the common myths.

But the whole idea about 'ok, I've done three objective tests and this is what faithfully translates to reality as we hear it' is possibly flawed.

I'm not sure what this is a reference to. Could you explain?
 

zalive

Active Member
Joined
Sep 4, 2018
Messages
263
Likes
38
@DonH56 is our resident expert on these matters but I think I know enough to give a brief reply to this one.

Yes, your friend is correct in one sense, most DACs produce pre and post ringing, and this does not show up on the measurements that Amir takes here.

However, he's incorrect in a few important ways.

Firstly, Amir's measurements focus on the key performance aspects and (for reasons I'm about to explain) an impulse response which would show up that ringing is not one of them.

Secondly, although your friend is correct that digital ringing occurs in almost every DAC, in any half-decent DAC (or just about any DAC running at a higher res sample rate) this ringing occurs only above 20KHz, which is above the range of hearing for most humans. So yes, it happens, but no, you can't hear it. And I don't mean you can't hear it in the way you can't hear 0.001% distortion. I mean you can't hear it in the sense that your ears do not even hear anything in the frequency range in which it occurs.

Thirdly, his thesis that the time domain is actually most important is a popular myth, but it has been debunked many times by controlled studies, which as in all other areas of audio have sought to establish thresholds of audibility. It turns out that the time domain is one area in which humans are have an incredibly low level of sensitivity - at least compared to the common myths.

I'm not sure what this is a reference to. Could you explain?

I can't represent or defend him, I'm not sure what's the source of his claims related to time domain. His standpoint is though (that I know, I specifically asked it) that you hear differently, say, a 10 kHz square waveform from a sinuus 10 kHz square waveform, Because of Fourier analysis related to harmonics consisting a waveform, that would imply we do hear frequencies beyond 20 kHz, through it's influence on waveforms within a hearing spectrum. I don't know whether scientific tests were performed on this. whether we hear a difference between different waveforms and a same frequency.

Anyway, I'm strongly for double blind listening tests. There's no way objectivists can prove a certain set of measurements represents what we hear in a blind test unless it's actually confirmed in blind test. Without this it's still an interpretation and a theory without a final proof.
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
I can't represent or defend him, I'm not sure what's the source of his claims related to time domain. His standpoint is though (that I know, I specifically asked it) that you hear differently, say, a 10 kHz square waveform from a sinuus 10 kHz square waveform, Because of Fourier analysis related to harmonics consisting a waveform, that would imply we do hear frequencies beyond 20 kHz, through it's influence on waveforms within a hearing spectrum. I don't know whether scientific tests were performed on this. whether we hear a difference between different waveforms and a same frequency.

I'm not sure exactly what he's talking about, and I realise you're in the tough position of trying to relay arguments that deal with an area you're new to. Perhaps he's talking about intermodulation distortion, i.e. frequencies above the audible range intermodulating with frequencies in the audible range to create intermodulation harmonics also within the audible range. This is a real thing, but it is a form of distortion happens only when there is a flaw in the playback system.

Anyway, you can actually test your friend's claim for yourself, there's a website that allows you to generate different kinds of waveforms at different frequencies and listen to them. Select say 10KHz or 15KHz and flick back and forth between a sine wave (pure tone) and a square wave.

But be warned, due to some weird quirk of the website, the squarewave is about 1dB louder than the sinewave (and the others are inexplicably quieter). You'll definitely be able to hear this level difference. But (and I know this is not strictly speaking fully possible), try to ignore the level difference and concentrate on trying to hear a tonal difference.

If you want to avoid the 1dB level difference and do the experiment in a properly controlled way, unfortunately you'll have to record the waveforms in something like Audacity, then use software to properly level match them, then play them back with Foobar and an ABX comparator plugin (under which circumstances you unfortunately won't be able to hear the difference at all unless you have hearing that extends higher than average in frequency).

But also, just think it through first. If you can't hear a fundamental tone at 22KHz, how could you hear a harmonic at the same frequency? What is your ear doing to the inaudible tone, just because there is now an audible tone present? But by all means try it if you don't believe me.

Also, if you do this, please make sure that your sample rate is set to at least 48KHz, but preferably higher. It needs to slightly more than double whatever the highest frequency is that you want it to reproduce.

Anyway, I'm strongly for double blind listening tests. There's no way objectivists can prove a certain set of measurements represents what we hear in a blind test unless it's actually confirmed in blind test. Without this it's still an interpretation and a theory without a final proof.

Nice :) But... I think I disagree slightly again, if I've understood you correctly.

I think you're talking about the kind of double blind test where we just take one piece of equipment out of the chain and replace it with another, is that right?

IMO, these tests are of value only if we are trying to establish whether we can discern a difference between two devices, or doing market research.

If you want to validate a set of measurements, you can't do it by blind testing with the piece of gear measured, there are just too many variables being changed.

That's why audibility studies use the same equipment throughout the whole chain and change only one variable at a time (e.g. noise, distortion, phase, etc. etc.). This allows audibility thresholds to be determined (obviously with a margin of error, as in all science). These thresholds can then in turn be used to interpret the measurements of a specific device.

It's an imperfect process, but it's far, far more precise and valid than just blind testing a piece of gear and making assumptions about how the sound relates to its specific measurements. That's an interesting thing to do, but only slightly less of a stab in the dark than sighted testing.
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,201
Likes
16,983
Location
Riverview FL
Anyway, you can actually test your friend's claim for yourself, there's a website that allows you to generate different kinds of waveforms at different frequencies and listen to them. Select say 10KHz or 15KHz and flick back and forth between a sine wave (pure tone) and a square wave.

Well... For grins or flames:

Assuming a digital source, and depending on the sample rate of the "square"...

Here's a -6dB 12khz "square" at 48kHz sample rate, and that square resampled to 384kHz (my approximation of what the reconstruction filter would do with it during conversion to analog). It don't be looking very square...

1536190678028.png
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
Well... For grins or comment:

Assuming a digital source, and depending on the sample rate of the "square"...

Here's a -6dB 12khz "square" at 48kHz sample rate, and that sqaure resampled to 384kHz (my approximation of what the reconstruction filter would do with it during conversion to analog). It don't be looking very square...

View attachment 15378

Sure, the "square wave" is going to look very different at different sample rates. But these differences won't be audible in respect of harmonics falling above the limits of human hearing :)
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,201
Likes
16,983
Location
Riverview FL
More grins and flames invited:

Gets more interesting at other frequencies... I have no explanation for this mess!

1536191800358.png
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,370
Likes
234,436
Location
Seattle Area
OK, I am back. Tired from driving all day but I am back. :)

When I received the Yggdrasil DAC for this review, I happen to read owner's comments that he had bought a Topping DX7s and did not like it as much. Searching, here is one of his posts on another forum:

1536194741240.png


So in coordinating the shipping of the unit to me, I also asked for some specific music tracks where his observations would be most obvious. This is what he wrote to me:

---
There's Muddy Water's Folk Singer where its fairly noticeable with his voice. I have that as FLAC now, but used to listen to it on Tidal (Master version).
Candido Camero's Wa Wa Wa, only have that as FLAC (from Chesky's headphone demonstration disc).

--

I happen to have the high-resolution downloads of Muddy Water's Folk Singer and Wa Wa Wa. And since I also had owner's Yggdrasil and Topping DX7s, it made for a very nice situation where I could try to replicate his results.

The output levels of the two DACs is every so slightly different but was not enough to need equalization. I connected both DACs via USB, synchronized them using Roon,, and performed AB using my Stax headphone amplifier and its Omega headphone.

I can say with very high confidence that there was no difference at all when playing the two. Mind you, I would "perceive" differences at times but it would not at be consistent. What I thought was Topping or Yggy would vary in some instances. In others it was hopeless to tell the difference.

No way, no how was there a difference is soundstage, depth, etc. I focused very hard on this and the image would be at the same precise place in both DACs.

There is no doubt in my mind that under blind listening tests, all the advocates of Yggdrasil would fail such a comparison. I am happy to donate $500 to a charity of their choice if they can demonstrate otherwise.

Again, this test was done with the owner's DAC, selection of music, and comparison to the same alternate DAC.

Until such time that people make controlled tests as I did, their subjective data unfortunately is not useful for anything. Our memory of audio fidelity is extremely poor. No way can you remember the nuances of music being heard in one device, and then minutes, days or weeks later try to compare it to another. Doing so as subjective audiophiles do just leaves the astray.
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,201
Likes
16,983
Location
Riverview FL
Are you using Square Waves, no alias to generate these? If not try that.

No, forgot about that option since I first saw it mentioned recently...

Let's see:

1536195923770.png
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
I can say with very high confidence that there was no difference at all when playing the two. Mind you, I would "perceive" differences at times but it would not at be consistent. What I thought was Topping or Yggy would vary in some instances. In others it was hopeless to tell the difference.

No way, no how was there a difference is soundstage, depth, etc. I focused very hard on this and the image would be at the same precise place in both DACs.

There is no doubt in my mind that under blind listening tests, all the advocates of Yggdrasil would fail such a comparison. I am happy to donate $500 to a charity of their choice if they can demonstrate otherwise.

Again, this test was done with the owner's DAC, selection of music, and comparison to the same alternate DAC.

Until such time that people make controlled tests as I did, their subjective data unfortunately is not useful for anything. Our memory of audio fidelity is extremely poor. No way can you remember the nuances of music being heard in one device, and then minutes, days or weeks later try to compare it to another. Doing so as subjective audiophiles do just leaves the astray.

Of course there was no difference. But no matter how many times you say this there will always be people telling you that as the measurements of those 2 DACs differs so much the difference must be heard, right? :p
 

zalive

Active Member
Joined
Sep 4, 2018
Messages
263
Likes
38
Why would you ears be more pained in a blind comparison than in a sighted comparison? Aren't they doing exactly the same thing in both cases?

Anyway don't worry, I can see this is not going to be a rational discussion... Best of luck to you and I'm very happy you are enjoying your Yggy. Enjoyment is the aim of the game at the end of the day :)

From his description he sounds as he had confidence in both products to perform well. But wanted to pick a device which performs better between the two. He sounds pretty disappointed in Mutec, as its sound was nowhere near what he expected to hear from it. This is what sometimes happen. You actually have a bias towards hearing a good sound, but then you simply...don't hear what you expected to hear.

I never listened to Mutec myself (anything from them actually). But I've read more than one opinion describing its sound not being natural, so it doesn't surprise me if someone dislikes this sound. Actually there are listeners who are not biased with seing great measurements, or content after they hear ultra resolution (details), but instead they listen to the music/sound as a whole, and naturality of presentation is highly valued by such.
 

zalive

Active Member
Joined
Sep 4, 2018
Messages
263
Likes
38
I'm not sure exactly what he's talking about, and I realise you're in the tough position of trying to relay arguments that deal with an area you're new to. Perhaps he's talking about intermodulation distortion, i.e. frequencies above the audible range intermodulating with frequencies in the audible range to create intermodulation harmonics also within the audible range. This is a real thing, but it is a form of distortion happens only when there is a flaw in the playback system.

I'm in no position here. In a way it's a loss that people like him are not personally as available to discuss those topics. It would be an interesting discussion for sure. One more thing I recalled from what he said. He said that tech measurements are performed by designers as a tell tale sign to see whether they did everything right and sound. Because in case they messed up something it should show on measurements. Also he said he personally used in practice various tests from simple to much more complex to induce further pressure on the circuitry if this would cause measurements to show more. But he says, the set of test which is performed in reality, and a listening experience, are completely two different worlds. I know he collaborates with more than one local experts as they help each other tweak and improve equipment they design...and after measurements, it's their listening what they use to confirm the quality, or quality of their tweaks, whether they improved sound or not. He says measurements they perform (and to my understanding they do measure a lot, to an extent which would here likely be found more than sufficient) are not enough to guarantee a great sound, so they can't avoid listening when developing a device. However, as from his position, he himself would never introduce/allow more distortion on purpose. He's personally even not fond of tubes because he finds them demanding when it comes to developing a neutral presentation.
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
From his description he sounds as he had confidence in both products to perform well. But wanted to pick a device which performs better between the two. He sounds pretty disappointed in Mutec, as its sound was nowhere near what he expected to hear from it. This is what sometimes happen. You actually have a bias towards hearing a good sound, but then you simply...don't hear what you expected to hear.

I never listened to Mutec myself (anything from them actually). But I've read more than one opinion describing its sound not being natural, so it doesn't surprise me if someone dislikes this sound. Actually there are listeners who are not biased with seing great measurements, or content after they hear ultra resolution (details), but instead they listen to the music/sound as a whole, and naturality of presentation is highly valued by such.

As the name of this forum implies this may not be the best place to promote subjective impressions. ;)
 

zalive

Active Member
Joined
Sep 4, 2018
Messages
263
Likes
38
Nice :) But... I think I disagree slightly again, if I've understood you correctly.

I think you're talking about the kind of double blind test where we just take one piece of equipment out of the chain and replace it with another, is that right?

IMO, these tests are of value only if we are trying to establish whether we can discern a difference between two devices, or doing market research.

If you want to validate a set of measurements, you can't do it by blind testing with the piece of gear measured, there are just too many variables being changed.

That's why audibility studies use the same equipment throughout the whole chain and change only one variable at a time (e.g. noise, distortion, phase, etc. etc.). This allows audibility thresholds to be determined (obviously with a margin of error, as in all science). These thresholds can then in turn be used to interpret the measurements of a specific device.

It's an imperfect process, but it's far, far more precise and valid than just blind testing a piece of gear and making assumptions about how the sound relates to its specific measurements. That's an interesting thing to do, but only slightly less of a stab in the dark than sighted testing.

I mean blind A-B comparison done with high scientific standards which objectivist would find sufficient. Whatever this includes. Of course you need to keep other links in a musical chain the same for the valid comparison, and use an environment which makes sense, but this is understood.

Suggested test would be interesting because it's where objectivists and subjectivists opinion differ - however objectivists say the difference might not be audible, so they predict two possible outcomes, possibly inclining to no decision which is better...right? While subjectivist tell Yggrdrasil is clearly better. The outcome decides who is right. Wouldn't each side want their proof on this? And the best...result would be objective, no matter which it would be.
 

zalive

Active Member
Joined
Sep 4, 2018
Messages
263
Likes
38
As the name of this forum implies this may not be the best place to promote subjective impressions. ;)

Only I don't see anybody promoting anything. However if there's rule here that it's forbidden to express any subjective impression, let me know.
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
I mean blind A-B comparison done with high scientific standards which objectivist would find sufficient. Whatever this includes. Of course you need to keep other links in a musical chain the same for the valid comparison, and use an environment which makes sense, but this is understood.

Suggested test would be interesting because it's where objectivists and subjectivists opinion differ - however objectivists say the difference might not be audible, so they predict two possible outcomes, possibly inclining to no decision which is better...right? While subjectivist tell Yggrdrasil is clearly better. The outcome decides who is right. Wouldn't each side want their proof on this? And the best...result would be objective, no matter which it would be.

I don't think anyone's against a blind test. I'm interested in knowing whether people can discern any difference. I'm also mildly interested in which one people prefer (if there is an audible difference), but knowing the answer to this question doesn't tell us that much really. There's plenty of evidence some or many people prefer an audibly distorted signal, and if the Schiit is not transparent, preferring it would be in keeping with this.

But I think the more likely outcome is that there's no audible difference.

What the results certainly won't have anything to do with is pre-/post-ringing or harmonics above the audible spectrum.
 
Last edited:

zalive

Active Member
Joined
Sep 4, 2018
Messages
263
Likes
38
I don't think anyone's against a blind test. I'm interested in knowing whether people can discern any difference. I'm also mildly interested in which one people prefer (if there is an audible difference), but knowing the answer to this question doesn't tell us that much really. There's plenty of evidence some or many people prefer an audibly distorted signal, and if the Schiit is not transparent, preferring it would be in keeping with this.

But I think the more likely outcome is that there's no audible difference.

Quite possible. Outside of inaudible, blind test is not perfect, because listeners involved might not be perfectly relaxed in a situation in which they feel their competence to hear a difference will be judged/on the line. It's not the same as doing a blind test for your self at home. Also, I was told my many different audiophile, as a general advice, that I should assess any equipment in my home environment. Because this is environment I know well, devices in system are mostly ones I know well too, so it's easier to pay attention to difference. Might be even more reasons. But still it's the best we have.

I myself have to admit that difference between quality DACs are minor. Sometimes there's something specific which will bother you and prevent you from enjoying the music, however, in a completely non-blind environment, what I experience are nuances. Bit of difference in dynamics, definition of a certain part of frequency spectrum, bit of difference in perception of tonal neutrality/character, bit more or less resolution, bit more or less feeling of 'musicality' - like level of unpleasant distortions being lower, so sound involves you in music more. All mentioned is not on the side of big differences. I myself am not sure whether I'd recognize a device between couple similar sounding in a blind test. I'd like to do it once for fun. The reason why I never did it is fuss/effort...I obviously need assistance.

What the results certainly won't have anything to do with is pre-/post-ringing or harmonics above the audible spectrum.

I have this local expert and an audio designer telling me otherwise. And as for certainty of scientific findings, I'm alwasy on a skeptic side. I've seen scientific findings change over time. New study shows different, or differentiate findings better and sheds a new light. Also, there's an important difference between 'we didn't find relevance' and 'relevance doesn't exist'. So pardon me if I cannot accept this as a fact.
 
Top Bottom