Review and Measurements of Schiit Yggdrasil V2 DAC

[Timbo2]

@amirm

Source.



Can you do this as well?

If this outputtable directly from the AP software that’s seems OK. However, if you are asking essentially a “one person” shop to compile a dossier on every measured piece of equipment I don’t think that’s particularly reasonable.

 

[SIY]

If you have a standard set of measurements, you can run in Sequence Mode and print a report. If, like me, you change things up depending on what you're looking for, it's generally done in Bench Mode, and the printouts become a bit more tedious.

Amir has made project files available, which is (IMO) a better way to do things. That way, anyone can play with the data and see what they can extract from it. And of course, it's a plus toward replicability, since anyone with an AP will have all the measurement settings automatically loaded.

 

[bennetng]

For those who interested in dithering, the green line is my $170 soundblaster 24-bit analog loopback (2Vrms unbalanced), red and blue are truncated vs dithered 20-bit unpolluted signal, scanned with a 2 second range and -60dBFS tone. So if people have stuff like what Blumlein 88 have, like a Foucsrite Forte or ADCs with similar quality and adequate input voltage headroom, truncation artifacts can be identified without using ultra expensive analyzers.

 

[rebbiputzmaker]

If you have a standard set of measurements, you can run in Sequence Mode and print a report. If, like me, you change things up depending on what you're looking for, it's generally done in Bench Mode, and the printouts become a bit more tedious.

Amir has made project files available, which is (IMO) a better way to do things. That was, anyone can play with the data and see what they can extract from it. And of course, it's a plus toward replicability, since anyone with an AP will have all the measurement settings automatically loaded.

I guess since they are a manufacturer they are printing out results and the settings used that show an honest representation of their product. They will probably not be looking for different things. A project file is of no real use to a customer.

 

[amirm]

Can you do this as well?

As SIY correctly explained, the Audio Precision software has two operational modes: Bench and Sequence. The results you see from me are currently from Bench mode which doesn't export it all to PDF. I capture the graphs one at a time, annotate them for clarity in Photoshop, and then publish here. Audio Precision calls the Bench mode "R&D" mode.

The sequence mode is for manufacturing automation. You run a set of standardized tests with limits and get pass/fail scores. For factory automation, you know what the device is supposed to do so all the settings can be optimized for that. Again as SIY mentioned, when it comes to testing every device known to man as I do , I don't know the expected outcome. Sometimes a distortion graph limit is 1%, sometimes 10%.

That said, I did start my journey with APx555 trying to create a sequence mode for automation. I had to stop and develop my bench mode process first so that I know precisely what I want to measure. I plan to go back and see if I can now finish the sequence mode.

I provided some data from my unfinished work earlier when someone asked for single number figures of merit, realizing how incomplete my sequence mode still is. It is a struggle between testing more gear and building more fixtures.

Speaking of fixtures, I need to automate my headphone testing loads. Right now changing impedances and graphing is manual step which I hate. I will be using the sequence mode to do this. But I have to first build the hardware load that can be controlled by AP (not hard, just have not had time).

 

[amirm]

If this outputtable directly from the AP software that’s seems OK. However, if you are asking essentially a “one person” shop to compile a dossier on every measured piece of equipment I don’t think that’s particularly reasonable.

Thanks for saying this. I am really not here to document the world. Companies making high-end products need to do this and a) provide it on their website and b) give it to customers. My read of Schiit post is that they are keeping these records to themselves which doesn't do either.

What I like to do is produce enough data for a product to draw informed conclusions. If you want 500 charts, look to Bob Smith (atomicbob) who produces them.

So please put this kind of pressure on the manufacturers. I am so far behind in measuring gear that I need to spend what time I have left on those than producing more and more documentation.

 

[Candlesticks]

Thanks for saying this. I am really not here to document the world. Companies making high-end products need to do this and a) provide it on their website and b) give it to customers. My read of Schiit post is that they are keeping these records to themselves which doesn't do either.

What I like to do is produce enough data for a product to draw informed conclusions. If you want 500 charts, look to Bob Smith (atomicbob) who produces them.

So please put this kind of pressure on the manufacturers. I am so far behind in measuring gear that I need to spend what time I have left on those than producing more and more documentation.

No you put pressure on everyone for sake of the reviewers credibility, the manufacturers ability to resolve a problem and ultimately the consumers benefit. Knowing what was done is fundamental to any kind of technical support in replicating a problem.

I can right now level the accusation against you, which many already have, which is that you are incompetent or deliberately botching measurements. I don't believe that to be true but if all it takes is an automatically generated PDF report and an upload of a project to show transparently what was done then I have to question why that is suddenly such a difficult task. Anyone would be right to be suspicious of someone who won't be transparent with detailed test methodology.

 

[amirm]

I can right now level the accusation against you, which many already have, which is that you are incompetent or deliberately botching measurements.

I am OK with this kind of accusation. Honestly I am. I rather put the work toward testing new gear than constantly addressing this. If people don't trust my measurements or my competence, that is cool. Let's move on. There are plenty of people who find value in the work and my job is to serve them by testing more equipment, than producing more documentation.

 

[SIY]

I can right now level the accusation against you, which many already have, which is that you are incompetent or deliberately botching measurements. I don't believe that to be true but if all it takes is an automatically generated PDF report and an upload of a project to show transparently what was done then I have to question why that is suddenly such a difficult task. Anyone would be right to be suspicious of someone who won't be transparent with what is being done.

That is utter nonsense. The project files, which Amir is sharing, have FAR more info in them on the test than a pdf, which is easy to manipulate. This is completely transparent for anyone who is technically competent.

 

[Candlesticks]

I am OK with this kind of accusation. Honestly I am. I rather put the work toward testing new gear than constantly addressing this. If people don't trust my measurements or my competence, that is cool. Let's move on. There are plenty of people who find value in the work and my job is to serve them by testing more equipment, than producing more documentation.

The purpose of documentation is replication. I don't trust you as I wouldn't trust anyone else. It's because we all make mistakes and that is why we document everything. If creating that documentation is not a huge time endeavor then anyone would be right to question why someone wouldn't do it. After all it allows others to check for mistakes.

I assume everyone here believes that every product should be tested fairly and accurately regardless of who makes it. The only reasons why I think someone wouldn't want to provide transparency is because they are worried others might see their inexperience (ego) or they are deliberately not measuring a product accurately. None of this is related to friends, enemies or any stupid objective/subjective "war". This relates to the only thing that matters which is accurate product measurements for consumers.

That is utter nonsense. The project files, which Amir is sharing, have FAR more info in them on the test than a pdf, which is easy to manipulate. This is completely transparent for anyone who is technically competent.

The only thing amirm uploaded was a 80kB .approxj file. The benefit of the PDF report is that anyone can read it and both the project file and PDF report reinforce each other. This will make trouble shooting much faster in the future because people don't have to make countless forum posts or emails to replicate test methodology.

 

[SIY]

The only thing amirm uploaded was a 80kB .approxj file. The benefit of the PDF report is that anyone can read it and both the project file and PDF report reinforce each other. This will make trouble shooting much faster in the future because people don't have to make countless forum posts or emails to replicate test methodology.

"Project file on request." I requested it, he sent it, it was solid. One email, not "countless." Transparent and far more useful than a pdf. But then, I don't have a particular agenda.

If you're dying for a pdf, get the project file and make whatever pdfs you like.

 

[amirm]

The purpose of documentation is replication. I don't trust you as I wouldn't trust anyone else. It's because we all make mistakes and that is why we document everything. If creating that documentation is not a huge time endeavor then anyone would be right to question why someone wouldn't do it. After all it allows others to check for mistakes.

I am giving you exact bit for bit graph captures of AP software. Having the same graph in PDF form doesn't help you any more. See my dashboard views and look at the bottom for test conditions. I don't know why think those screenshots are not trustworthy but some text pdf file is.

Regardless, as I said and will repeat, I am fine with people having doubts. It is considerable amount of work to conduct and publish these reviews. They are provided for free at no cost to you all, and as a hobby for me. The moment you increase the drudgery on my end where it stops being fun, that will be the end of my publishing these reviews.

So there is no currency here to cash to force me to become the documentation shop for the world. It is not like any other reviewer anywhere is providing what you are asking me anyway.

 

[Candlesticks]

"Project file on request." I requested it, he sent it, it was solid. One email, not "countless." Transparent and far more useful than a pdf. But then, I don't have a particular agenda.

If you're dying for a pdf, get the project file and make whatever pdfs you like.

Upload it somewhere please.

 

[SIY]

This is Amir's file. You're welcome to contact him and request it, just like I did. Then have fun!

 

[Thomas savage]

I am giving you exact bit for bit graph captures of AP software. Having the same graph in PDF form doesn't help you any more. See my dashboard views and look at the bottom for test conditions. I don't know why think those screenshots are not trustworthy but some text pdf file is.

Regardless, as I said and will repeat, I am fine with people having doubts. It is considerable amount of work to conduct and publish these reviews. They are provided for free at no cost to you all, and as a hobby for me. The moment you increase the drudgery on my end where it stops being fun, that will be the end of my publishing these reviews.

So there is no currency here to cash to force me to become the documentation shop for the world. It is not like any other reviewer anywhere is providing what you are asking me anyway.

Weeding for PDF’s ? , I wonder if @Candlesticks would be amenable to this.

 

[Candlesticks]

This is Amir's file. You're welcome to contact him and request it, just like I did. Then have fun!

@amirm can you upload the project file you sent to SIY please.

 

[Deleted member 65]

Isn't it time to close the Schiit book for good now.
ASR has written many chapters all with the same ending.
I propose we see the Yggdrasil V2 thread as the Schiit epilogue in ASR ... ;-)

 

[Thomas savage]

@amirm can you upload the project file you sent to SIY please.

 

[έχω δίκιο]

Isn't it time to close the Schiit book for good now.
ASR has written many chapters all with the same ending.
I propose we see the Yggdrasil V2 thread as the Schiit epilogue in ASR ... ;-)

I agree! It's time to move on from this Schiit show and get back to testing and discussing other products. Every time a Schiit Audio product is measured, it performs poorly and there is a Schiitstorm™ of comments with demands for more graphs, more files, and more measurements.

Amirm summed it up well when he wrote "If your ear doesn't hear these problems, consider yourself blessed and enjoy your unit. But please don't dispute the measurement data. They provide very high confidence results given the multiple times I have tested the model."

​

 

[amirm]

I have been meaning to comment on this but keep forgetting. Besides Jude, Bob Smith (atomicbob) has been posting measurements. Wanted to point this flaw in his jitter measurement methodology which he keeps using on everything he tests. Here it is on Yggy V2:

If you can read that, you have eaten far more carrots than I have . So let's pull that into photoshop and making more legible:

As we see, his graph starts at 7.5 kHz and ends at 14.5 kHz. Why? We need to see full bandwidth of audio.

As I noted there, cutting off low frequencies means we don't see power supply contributions although we can infer it from the sidebands around his main tone as I show in my measurements:

I routinely see power supply issues in DACs I measure and not showing the spectrum down to 20 Hz is a definite mistake.

While there is not much of interest in higher frequencies here, nevertheless the graph needs to show that so that we can tell. Many DACs have garbage at higher frequencies induced by front panel displays, other periodic events running inside the DAC, etc.

When told about this issue, he posts more charts addressing something different and finishes with this:

Well, the coverage of jitter measurements in AES17 are quite brief. Instead reader is sent to another AES standard:

In there, we find this:

So clearly they are interested to see jitter contributions from far wider bandwidth than Bob uses.

Yes, JA at stereophile also makes the same mistake.

Summary
Bob's Jitter measurements are incomplete. They show the least interesting range of frequencies which is around our perceptual masking tone. Jitter spectrum should include full audible bandwidth.