Review and Measurements of Schiit Yggdrasil V2 DAC

[gvl]

Is it possible to overlay, approximately, the profile of the filter on the AtomicBob's graph to better understand the content that is filtered out in calculations? Or is it too narrow for display purposes?

 

[Kyle / MrHeeHo]

Amir I'd like to thank you for yout attempts at making measurements more understandable to layman, and I think it's relevant to say that here because neither Jude nor Atomicbob attempt to do so.

Forgive me if this has been brought up and I missed it but are there industry standard settings for some of these tests, and if not does there need to be? I'd assume yes to the latter.

Edit: Is there any value in showing the sine wave the AP is inputing compared to the sine wave the DAC is outputting?

 

[Blumlein 88]

I was asked to comment on a specific "dynamic range" measurement by Bob Smith (AtomicBob). Specifically, he shows the FFT spectrum of a -60 dB tone and on it, declares a signal to noise ratio of 122 dB. As usual, his charts are impossible to read. So please allow me to annotate them such:
View attachment 13771

As you see, his "FFT meters" are declaring that there is 121+ dB of Signal to noise ratio.

That data directly conflicts what the FFT is actually showing. The mains hum alone is enough to make that difference around 65 dB, not 122 dB. Add up all the other distortions and noise and there is no way we have 122 dB of proper dynamic range.

And no, you can't look at the noise floor of the FFT and measure that difference. FFT noise floor gets artificially lowered based on its parameters (called "FFT gain"). But even if we did, that noise floor is at -160 db so subtracting our -60 dB signal from it, we get 100 dB, not 122.

So what is going on? First, let's look at the same measurement using my Audio Precision analyzer of the same -60 dB signal:

View attachment 13772

We see the same mains contribution at 120 Hz and bunch of harmonic distortion. The dedicated meter in Audio Precision is reporting about 60 dB of dynamic range above our noise and distortion which matches more or less the manual math I performed on AtomicBob's graph. And we can with eye confirm the same. Starting with -60 dB signal, the noise floor would have to be -180 dB for us to get the math and no way can we get there.

The key thing for now is that both of our FFT measurements are producing essentially identical results. So the issue is not the device being tested but what the meters on AtomicBob's graph is saying.

Alas, despite all the shouting that goes on on accuracy and documentation of measurements, we have none here from AtomicBob. The meter says: "USER: DAC SNR Residual Async." Good luck trying to find out what that means.

Fortunately I have used the Prism Sound analyzer and still have the software. So I went in there and found this custom script to make measurements from FFT. This is what it looks like when not minimized as he has done:
View attachment 13773

I know, I know, still makes no sense. But stay with me. What this is trying to do is filter out the tone at 1000 Hz ("band reject"). As with any filter, the bandwidth matters. Here, we are interested in taking out our main tone and look at what is left as our distortion+noise power. Unfortunately, the filter used here by default is improper. It has a wide bandwidth of 1/3 octave instead of just a hertz or two to take out the 1 kHz tone.

Prism Sound help file explains the motivation and problems with it:

View attachment 13774

Yup, the 1/3 octave filtering is there to emulate old analog THD+N meters! It was hard to filter sharply as we can today with our digital signal processing (and much better analog ones too). As they say, using of this method results in "residual components .... to be underestimated." And underestimated it is and hence the reason he is showing much better values than it should.

Lesson here is that custom scripts for making measurements in analyzers need to read and understood. And results confirmed to make sure it passes the smell test. Clearly an FFT that shows noise components just -65 dB below the signal can't have any useful figure out merit of 120+ dB.

Summary
My measurements/FFT spectrum of a -60 dB tone essentially matches AtomicBob's data. That results in SINAD (signal over noise+distortion power) of just 60 dB. The meter used in AtomicBob's graph to derive the Signal to Noise ratio is simply wrong and not configured correctly. We can easily confirm this on his graph as I have shown.

Let me know if you have any questions.

I think what atomicbob is doing is what is in AES17. The protocol for dynamic range is to measure with a -60 db 1 khz signal. Notch it out. Yes how the notch is done may make a little difference. Get a noise level reading for that bandwidth 22khz, and then take the difference between -60 db FS and the noise floor, add 60 db to it for dynamic range. So what he shows is what would be called something like 125 db dynamic range( well 121 or 122 db from what is shown). And that is why AES17 dynamic range and SNR are not the same thing. Some DACs modulate the noise floor with signal. But the dynamic range from max to low level signals would be this larger number because as signal level drops the noise floor may drop with it some few decibels. While a max signal vs the underlying noise floor might be a few decibels less due to the noise floor modulation.

 

[mindbomb]

Yea, I think the idea with that measurement is to estimate the 0dbfs noise performance, and not state the dynamic range at -60db.

 

[Blumlein 88]

I think I'm more aligned with Rob Watts on measuring linearity. You can have a DAC of excellent linearity poorly applied. You can test to see if the DAC is putting out the proper levels even if it is being obscured by noise from a poor implementation. Or you can see the DAC simply isn't linear regardless of implementation. For one thing measurements of the Yggy indicate even with filtering isn't capable of super excellent linearity.

I get the idea you don't get to listen to the filtered linearity, but we apply tests all the time to see how something is functioning for which that could be said. It is one reason I like SFDR (spurious free dynamic range). That is where you take your max level and see what the highest level of any aberration for any reason is. You know you have this much clean dynamic range. Whatever spurious signal is highest sets that. It could be jitter sidebands, 60 hz noise, other noise, idle tones anything sticking up from the noise floor. This is what you get to hear as the clean range of sound.

 

[amirm]

Yea, I think the idea with that measurement is to estimate the 0dbfs noise performance, and not state the dynamic range at -60db.

Ah, OK that makes sense.

Here is AES17 canned measurement in Audio Precision then:

So about 5 to 7 dB worse than what he is reporting.

 

[amirm]

To be clear, I am not a fan of such dynamic range measurements when they ignore the distortions that are there. The SINAD includes their effects and is a much more proper measure of true dynamic range of the device. From my original review:

The 86 dB is far more representative of what the DAC is doing than 115 dB I just showed. Audibility of distortion spikes is a lot higher than random noise floor so we better show dynamic range relative to the sum of distortion+noise than noise alone.

 

[Blumlein 88]

To be clear, I am not a fan of such dynamic range measurements when they ignore the distortions that are there. The SINAD includes their effects and is a much more proper measure of true dynamic range of the device. From my original review:

The 86 dB is far more representative of what the DAC is doing than 115 dB I just showed. Audibility of distortion spikes is a lot higher than random noise floor so we better show dynamic range relative to the sum of distortion+noise than noise alone.

I agree, but that is a different measurement. When signal levels drop, you do have a lower noise/distortion floor. So it isn't completely without validity. It also is deceptive if you think you have more than 115 db dynamic range at full signal. So there is two different measurements. You've measured gear that would be near 115 db in both the AES dynamic range measure, a full scale SNR measure, and SINAD. Obviously that is the better piece of gear.

 

[SIY]

The nice thing about modern test gear is that it's only a couple of mouse clicks to get both.

 

[Thomas savage]

The nice thing about modern test gear is that it's only a couple of mouse clicks to get both.

That’s what I’d like to see, if for no other reason it gives us something to talk about lol

 

[amirm]

Edit: Is there any value in showing the sine wave the AP is inputing compared to the sine wave the DAC is outputting?

Yes, it is actually a good suggestion. Here is what the digital waveform and its spectrum looks like at -120 dBFS and what comes out of Schiit Yggdrasil DAC:

The digital sine wave looks a bit noisy because we have added "dither" to it to make it render at such a low amplitude at 24 bits. That results in highly clean spectrum on top right.

In contrast, what is coming out of Schiit DAC below is clearly steppy which means it has truncated digital samples. That truncation causes all the distortion spikes on the right (since the waveform is square, we get odd harmonics of our original tone at 1 kHz). We also have power supply noise on the left.

 

[amirm]

The nice thing about modern test gear is that it's only a couple of mouse clicks to get both.

As it turns out and I have shown, the difference between the two methods is extremely small. After all, even I filter down -50 dB. The big difference in published reports is due to different units being tested. Clearly what Jude/AtomicBob have is different than what customers have bought. They are either hand-picked or have corrected firmware/design.

 

[Thomas savage]

It’s nice to read through these open and honest discussions of our measurements , free from flame throwers, egotism and the like.

 

[amirm]

It’s nice to read through these open and honest discussions of our measurements , free from flame throwers, egotism and the like.

Yeh but I have a boat to wash, repair and clean to put in the water! I should have finished it all two weeks ago but instead, I am still making Schiit measurements...

 

[Thomas savage]

As it turns out and I have shown, the difference between the two methods is extremely small. After all, even I filter down -50 dB. The big difference in published reports is due to different units being tested. Clearly what Jude/AtomicBob have is different than what customers have bought. They are either hand-picked or have corrected firmware/design.

This is the real story here imo, being obfuscated (possibly intentionally) by ‘measurement wars’.

 

[mindbomb]

Here is AES17 canned measurement in Audio Precision then:

wait, is that unbalanced? because it is a few db worse than the previous -60db sinad measurement you showed if you were to have just added 60db to it. (the 59.940 and 57.168)

edit - difference almost certainly from the weighing filter used

 

[Blumlein 88]

Yes, it is actually a good suggestion. Here is what the digital waveform and its spectrum looks like at -120 dBFS and what comes out of Schiit Yggdrasil DAC:

View attachment 13777

The digital sine wave looks a bit noisy because we have added "dither" to it to make it render at such a low amplitude at 24 bits. That results in highly clean spectrum on top right.

In contrast, what is coming out of Schiit DAC below is clearly steppy which means it has truncated digital samples. That truncation causes all the distortion spikes on the right (since the waveform is square, we get odd harmonics of our original tone at 1 kHz). We also have power supply noise on the left.

I really don't see anything different at all. Done in my best imitation of Dolores on Westworld.

 

[gvl]

Yes, it is actually a good suggestion. Here is what the digital waveform and its spectrum looks like at -120 dBFS and what comes out of Schiit Yggdrasil DAC:

View attachment 13777

The digital sine wave looks a bit noisy because we have added "dither" to it to make it render at such a low amplitude at 24 bits. That results in highly clean spectrum on top right.

In contrast, what is coming out of Schiit DAC below is clearly steppy which means it has truncated digital samples. That truncation causes all the distortion spikes on the right (since the waveform is square, we get odd harmonics of our original tone at 1 kHz). We also have power supply noise on the left.

Hmm, wouldn't it be steppy anyway even if they dithered internally to 20 bits? It's just flipping the LSB. Probably it would look less like a square wave with a little plateu at the 0 crossing resulting in a slightly cleaner spectrum. No? It is a 20 bit DAC after all, you can't expect it to produce the AP waveform in the time domain.

 

[amirm]

Hmm, wouldn't it be steppy anyway even if they dithered internally to 20 bits? It's just flipping the LSB. Probably it would look less like a square wave with a little plateu at the 0 crossing resulting in a slightly cleaner spectrum. No? It is a 20 bit DAC after all, you can't expect it to produce the AP waveform in the time domain.

No. The purpose of dither is to avoid those clear steps. Here is the AP digital signal at 20 bits with dither:

Unfortunately you can't see the image in real-time and see the random dither being applied. But take note of the variability regardless relative to the simple square wave shown by Yggdrasil. It is that randomness that provides the necessary magic to produce signals below 20 bits.

See the clear difference in the spectrum on the right. 100% clean sine wave at 1 Khz and no distortion. Compare that with the Yggdrasil producing all of those odd harmonics.

 

[gvl]

Interesting, thank you. Does this look anything like sine-wave on the DAC output? The 20-bit dithered signal that is.