• 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

OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,376
Likes
234,540
Location
Seattle Area
Went to close this tab from my browser and noticed something interesting.

It is said that this graph is from 3 hours and hence not as good as what is to follow:

1531979990744.png


I have drawn the red lines as if you could not tell. :) Those are perceptual masking lines. In the presence of that super loud tone in the middle, you will not hear anything else in the region between the two red lines. So no way any of those spikes are audible.

Now let's look at what is claimed to happen at 408 hours:

1531980271808.png


These spikes in new were not there and are outside of the masking region! So things got worse, not better from audibility point of view.

In addition, those sideband spikes are down -135 dB. Our hearing has a maximum dynamic range of about 116 dB. No way can you hear those spikes even if the main tone was taken away and all you had was distortion.

Of course the whole thing is absurd since music never has full amplitude tone at 11 Khz. Since any jitter is directly proportional to amplitude of our main tone, reducing 11 kHz to actual music levels would push any sidebands into noise floor in a hurry.

Bob Smith was asked about the top graph in our local AES meeting as to whether we hear those sidebands. To my surprise he made the same mistake in this post saying yes. He was asked how. He said if you played a weak signal and then amplified it you would hear it. I like to encourage him to create such a test before claiming so. Because no lossy audio compression would work if that were true!

I always say that knowledge of psychoacoustics is necessary to interpret these graphs. This is a great example of how lack of that results in completely opposite interpretation of what is measured.

Be careful guys. Don't let techies give you false leads in these subjectivist arguments. If Bob thinks there is an audible difference here, he should contact his friends at Schiit, get another Yggdrasil, warm one up and do a double blind test against a cold one. It is the ears that matter, right? Let's have the ears demonstrate what he claims to be true.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,524
Likes
37,057
Look closer. It appears the noise floor dropped 5 db after 408 hours. There are hints of those low level spikes in the 3 hr one. They are uncovered in the quieter 408 hr result.

There are also only one pair sidebands around the central tone by 408 hrs at maybe - 148 db. In the 3 hr plot there are 4 pairs of sidebands. One pair is at maybe - 138 db.

Now I don't think they are audible or important. It is a change and for the better actually.
 
Last edited:

derp1n

Senior Member
Joined
May 28, 2018
Messages
479
Likes
629
Bob Smith should go back to posting subjective reviews about how detailed the farts of viola players are with various DACs. The BS he's posting about resistor warm up now is just mind boggling.
 

svart-hvitt

Major Contributor
Joined
Aug 31, 2017
Messages
2,375
Likes
1,253
Went to close this tab from my browser and noticed something interesting.

It is said that this graph is from 3 hours and hence not as good as what is to follow:

View attachment 14028

I have drawn the red lines as if you could not tell. :) Those are perceptual masking lines. In the presence of that super loud tone in the middle, you will not hear anything else in the region between the two red lines. So no way any of those spikes are audible.

Now let's look at what is claimed to happen at 408 hours:

View attachment 14029

These spikes in new were not there and are outside of the masking region! So things got worse, not better from audibility point of view.

In addition, those sideband spikes are down -135 dB. Our hearing has a maximum dynamic range of about 116 dB. No way can you hear those spikes even if the main tone was taken away and all you had was distortion.

Of course the whole thing is absurd since music never has full amplitude tone at 11 Khz. Since any jitter is directly proportional to amplitude of our main tone, reducing 11 kHz to actual music levels would push any sidebands into noise floor in a hurry.

Bob Smith was asked about the top graph in our local AES meeting as to whether we hear those sidebands. To my surprise he made the same mistake in this post saying yes. He was asked how. He said if you played a weak signal and then amplified it you would hear it. I like to encourage him to create such a test before claiming so. Because no lossy audio compression would work if that were true!

I always say that knowledge of psychoacoustics is necessary to interpret these graphs. This is a great example of how lack of that results in completely opposite interpretation of what is measured.

Be careful guys. Don't let techies give you false leads in these subjectivist arguments. If Bob thinks there is an audible difference here, he should contact his friends at Schiit, get another Yggdrasil, warm one up and do a double blind test against a cold one. It is the ears that matter, right? Let's have the ears demonstrate what he claims to be true.

Nice comment on psychoacoustics. @amirm .

I think more care on this issue, which is sort of bifurcating (audio) science, is good when discussing design choices. Especially in the speaker area, which - regrettably - has little focus on ASR (maybe @amirm needs to do a joint venture with some guy who has access to a measuring chamber...).
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,524
Likes
37,057
Nice comment on psychoacoustics. @amirm .

I think more care on this issue, which is sort of bifurcating (audio) science, is good when discussing design choices. Especially in the speaker area, which - regrettably - has little focus on ASR (maybe @amirm needs to do a joint venture with some guy who has access to a measuring chamber...).

Yes, DACs unless Schiit, have less to do with how a system sounds than nearly anything other than wire. Almost like the old joke about the guy looking for his keys by the lampost even though he lost them in the dark.

Speaker testing is going to be difficult to do much with. Amplifier testing on the other hand would likely be much more useful than DAC testing and could prove to be very beneficial.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,376
Likes
234,540
Location
Seattle Area
Look closer. It appears the noise floor dropped 5 db after 408 hours. There are hints of those low level spikes in the 3 hr one. They are uncovered in the quieter 408 hr result.
The scales are not the same so you can't go by the vertical scale and hence noise level. Here is what 0 dBFS (i.e. top of the graph) for 3 hour test:

1531985791624.png


And here is the "warmed up" 408 hour scale for 0 dBFS:

1531985844690.png


That is a 6 dB difference which is even less than what the noise floor moved by. Strangely the peak signal is the same in both which makes no sense unless levels were changed.

Who knows if the rest of the FFT depth and averaging are the same.

Notice that the rest of the text in that log window is different too which indicates different tests, not the same script.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,524
Likes
37,057
The scales are not the same so you can't go by the vertical scale and hence noise level. Here is what 0 dBFS (i.e. top of the graph) for 3 hour test:

View attachment 14030

And here is the "warmed up" 408 hour scale for 0 dBFS:

View attachment 14031

That is a 6 dB difference which is even less than what the noise floor moved by. Strangely the peak signal is the same in both which makes no sense unless levels were changed.

Who knows if the rest of the FFT depth and averaging are the same.

Notice that the rest of the text in that log window is different too which indicates different tests, not the same script.
So maybe delete all posts related to that? They simply aren't comparable in any useful sense.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,376
Likes
234,540
Location
Seattle Area
So maybe delete all posts related to that? They simply aren't comparable in any useful sense.
He was asked about this and this is his answer:

1532043277874.png


Can you believe that? Two completely different outputs (balanced vs unbalanced) are tested and he just brushes it aside after claiming the only variable was warm-up time.

1532043397381.png


Oh yeh? How about the comment about lower noise floor??? And how has he verified that the jitter on Yggdarsil doesn't vary on one output vs the other when the noise floor changes this way?
 

Thomas savage

Grand Contributor
The Watchman
Forum Donor
Joined
Feb 24, 2016
Messages
10,260
Likes
16,298
Location
uk, taunton
He was asked about this and this is his answer:

View attachment 14054

Can you believe that? Two completely different outputs (balanced vs unbalanced) are tested and he just brushes it aside after claiming the only variable was warm-up time.

View attachment 14055

Oh yeh? How about the comment about lower noise floor??? And how has he verified that the jitter on Yggdarsil doesn't vary on one output vs the other when the noise floor changes this way?
You know things have gone fruity when the schiit promoter is trying so hard to show their DAC’s have variable performance over time .

Only in the Audiophile world would this happen, they have lost their minds in their collective desperation to afffrim subjectivist ideas that ultimately underpin the whole company it seems.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,376
Likes
234,540
Location
Seattle Area
You know things have gone fruity when the schiit promoter is trying so hard to show their DAC’s have variable performance over time .
That's a fascinating point. As engineers we get paid to design products with stable performance, not one that keeps changing its mind over time.

Fascinating to watch Bob Smith give totally incorrect explanations of why this can be so. "Oh, look at this resistor value changing over time." Now multiply that by 1000 parameters in the DAC and tell me why they would all directionally go toward better sound, rather than worse or no change.
 

Palladium

Addicted to Fun and Learning
Joined
Aug 4, 2017
Messages
631
Likes
769
My hearing also changes over the course of a day too and I don't need to pay $2500 for that. ;)

It's fascinating really just how many mental hoops the believers can jump through to defend this company when one can spend exactly $0 on add-on DACs and can still easily win them in audio quality by default, much less in performance/$ terms.
 
Last edited:

Bhargu

Member
Joined
Jun 5, 2018
Messages
83
Likes
14
I can't understand why people are fine with equipment that supposedly takes weeks to achieve thermal stability. With such a big frame of reference, people will even accept it if someone says it gets infinitely better at infinite time. A fact which should be regarded as faulty engineering is being heralded as a miracle of engineering. If the whole world was this forgiving, engineering would have been the easiest job around.

I have come across class-A amps that took 10-20 minutes to warm up from cold (temperature controlled room at 16 degree Celsius, around 60F). Are there a lot of brands claiming 1+ hours for warmup? @amirm , how long did you warm up those Benchmark DACs and Oppo DAC before you took the measurements?
 

maxxevv

Major Contributor
Joined
Apr 12, 2018
Messages
1,872
Likes
1,964
Any properly designed modern electronic device should not vary in performance if there are no drastic temperature changes. As long as the temperature has stabilised in use, there is no good reason for performance to vary.
 

Grave

Senior Member
Joined
Jun 30, 2018
Messages
382
Likes
204
Must warm up Schitt for at least 100 hours, after 1000 hours the sound really opens up.
 

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,967
Likes
2,523
Location
Iasi, RO
I've seen people on Head-Fi and SBAF that are keeping powered ON their Schiit devices based on subjective results posted by others. By powered ON I mean 24/7.

Technically speaking, I'm not sure why someone will actually recommend hours of warm-up or continuous running, after all in 20 minutes most semiconductors and capacitors should be pretty warm, so DC-output gets stable enough, thermal agitation too and resistors gets warmed-up so background noise increases a bit etc.

Speaking about 24/7 run, what if a bypass capacitor fails and blows up (thantalum maybe)? Or if the PSU breaks somehow and gets on fire while everyone's at the office. :(
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
I can only relate my experience. I used to have A TagMcLaren AV processor. In my stupidity I kept it on 24/7. Never heard a jot of difference in sound because of it, however after 5 years it started generating dsp start up error codes. After a good poke around I found a lot of power supply ripple and several of the power supply caps had failed (not leaking just low value). The psu area was quite hot and they had simply lifed even though they were high temp rated.

Replaced them, kept it off when not in use and was trouble free for another 7 years before selling it.

So my advice, don't do it. Waste of electricity to boot.
 
Last edited:
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,376
Likes
234,540
Location
Seattle Area
The Yggdrasil runs pretty warm too so you definitely reduce its life by keeping it on all the time.

I think the whole affair has to do with elasticity and variability of hearing. Same gear can some days sound dull, and others, much better. This can also be because the listener is playing different material. The owner hears the "good" sound after a few days and assumes it is because of being left on for days. Which of course it is not.
 
Top Bottom