[Baldr]

For you, yes and I respect that. For the person who asked the question, obviously they care about the measurements; that is what was quoted as a difference between audio based DACs and otherwise, and that is what the Stereophile conclusion was also about, amongst other things. So why bother is because the question stemmed from a claim of better or worse measurements and the advantages/disadvantages of it vs audio based DACs.

 
Not at all, @Nbe9 called me out personally. It seems he is presenting a case which based on incomplete research, a sample size of one measurer (who had that graph 3db worse than the others), and a sample size of one graph. It seems, based upon his post, that his entire audio worldview is based on harmonic distortion, at the expense of all other measurements he so respects.
 
He, as well as others who listen with their ability to read charts and engineering credentials, real or imagined, may deny themselves the pleasure and emotion this hobby allows conveying the pleasure of music.
 
There are those who prefer sedans with less noise and smoother ride to sports cars, which are genuinely exciting. If one takes the time to read our product descriptions, he will find we plainly state that many of our $100 and up DACs may be all many clients will ever need.
 
Edit:  I was wrong in what I wrote - I meant to say 3db, not 3x.  It is corrected above.  I do not have the energy to debate 1 db.   I have a lot of other fish to fry.

 

[460414]

Originally Posted by theveterans /img/forum/go_quote.gif
 
It's those harmonics that make Yggy sound beautiful

The 3rd harmonic is at ~-93 db, the 2nd + 3rd is ~-91db (in the original image, not the below). Get a sine wave generator (this works fine link) download one of -1db/0db. Set the volume so its loud, then download one of -90db. If you can hear the -90db one at all, it's rather quiet. The relative volume of that is similar to the 2nd plus 3rd harmonic.
The -125db harmonics though... They are approximately 8x quieter to our ears (if they could be heard) than the ~-90db one, for 10db = double/half perceived volume.
But.. The noise floor is in practice at far higher than -145db. If you look at this image:

The magenta line/graph is of a -90db sine with 16 bit data (red is 24 bit). The "real" and perceived volume of the noise floor of the -90db sine should be at -96db for the 1 bit = 6db rule. So the FFT used makes the measured noise floor ~34db lower than it is. Apply this the first graph I posted and the "actual" noise floor is at ~110db, assuming you are feeding it 24 bit data (not to mention the noise floor of the recording). This -110db should sound approximately double as loud as the sum of all the lower level harmonics combined (just under -120db). If you are using 16bit data/recordings then the the lowest possible noise floor is 96db anyway.
 

Originally Posted by Jozurr /img/forum/go_quote.gif

  Actually I don't care either way. Ultimately it is how the device sounds. 
 

For you, yes and I respect that. For the person who asked the question, obviously they care about the measurements; that is what was quoted as a difference between audio based DACs and otherwise, and that is what the Stereophile conclusion was also about, amongst other things. So why bother is because the question stemmed from a claim of better or worse measurements and the advantages/disadvantages of it vs audio based DACs.

This ^
 

Originally Posted by earnmyturns /img/forum/go_quote.gif
 
You might want to explain yourself why you think those measurements are more relevant than these. Anybody with experience measurement-heavy fields knows that complex systems don't have a single figure of merit, and motivated people can find the measurement that fits their story — and that's charitably assuming that the measurements were done well, which is by no means a given even for so-called "professionals." 
 

The linked thread THD measurement he got is ~0.0022% which equals to -93db. Stereophile got ~-91db. In the context of my original post its essentially the same figure, what is the problem?

Originally Posted by buke9 /img/forum/go_quote.gif
 
To me the proof is in the pudding. The sound that goes into my ears matters to me more than a measurement on a graph. It might have some certain distortion at a certain frequency or frequencies but to me it sounds darn good. Have no problem with someone asking questions about it just doesn't concern me much . Any study or measurements have a lot to do with who is doing them and the motivation of those doing it. Not to say there are any reason why those measurements are bad or good just who knows.

See above. I am not saying this DAC sounds bad because of these measurements or whatever. I am just questioning @Baldr's claims.

 
 
Edit:
 
 
Here's some other measurements, all from Stereophile, similar results from SBAF, although more limited.
 
IMD
 
Benchmark DAC 2 HGC (Also look at Auralic Vega and Azur 851D. DAC 2 and Azur 851D are cheaper than the Yggdrasil as well)

 
Yggdrasil

Frequency response
 
DAC 2

Yggdrasil

 
-4dbfs white noise left channel. 0dbfs 20khz tone.
 
DAC2

 
 
Yggdrasil
 

 
Jitter
 
DAC 2

Yggdrasil

 
So, what measurement shows higher accuracy of the Yggdrasil?

 

[460414]

   
Not at all, @Nbe9 called me out personally. It seems he is presenting a case which based on incomplete research, a sample size of one measurer (who had that graph 3x worse than the others), and a sample size of one graph. It seems, based upon his post, that his entire audio worldview is based on harmonic distortion, at the expense of all other measurements he so respects.
 
He, as well as others who listen with their ability to read charts and engineering credentials, real or imagined, may deny themselves the pleasure and emotion this hobby allows conveying the pleasure of music.
 
There are those who prefer sedans with less noise and smoother ride to sports cars, which are genuinely exciting. If one takes the time to read our product descriptions, he will find we plainly state that many of our $100 and up DACs may be all many clients will ever need.

 
To address the sample size:

 

  You might want to explain yourself why you think those measurements are more relevant than these. Anybody with experience measurement-heavy fields knows that complex systems don't have a single figure of merit, and motivated people can find the measurement that fits their story — and that's charitably assuming that the measurements were done well, which is by no means a given even for so-called "professionals." 

The linked thread THD measurement he got is ~0.0022% which equals to -93db. Stereophile got ~-91db. In the context of my original post its essentially the same figure, what is the problem?

 
But to the rest of the post.. Ok right, please give me some actual evidence to your original claim.
Also: where are these measurements of 3x lower distortion?

 

[buke9]

The 3rd harmonic is at ~-93 db, the 2nd + 3rd is ~-91db. Get a sine wave generator ([COLOR=000000]this works fine[/COLOR]) download one of -1db/0db. Set the volume so its loud, then download one of -90db. If you can hear the -90db one at all, it's rather quiet. The relative volume of that is similar to the 2nd plus 3rd harmonic.
The -125db harmonics though... They are approximately 8x quieter to our ears (if they could be heard) than the ~-90db one, for 10db = double/half perceived volume.
But.. The noise floor is in practice at far higher than -145db. If you look at this image:


The magenta line/graph is of a -90db sine with 16 bit data (red is 24 bit). The "real" perceived  volume of the noise floor of the -90db sine should be at -96db for the 1 bit = 6db rule. So the FFT used makes the measured noise floor ~34db lower than it is. Apply this the first graph I posted and the "actual" noise floor is at ~110db, assuming you are feeding it 24 bit data (not to mention the noise floor of the recording). This -110db should sound approximately double as loud as the sum of all the lower level harmonics combined (just under -120db). If you are using 16bit data then the the lowest possible noise floor is 96db anyway.

This ^

The linked thread THD measurement he got is ~0.0022% which equals to -93db. Stereophile got ~-91db. In the context of my original post its essentially the same figure, what is the problem?

Like I said I have no problems with one questioning something that someone else has claimed I'm all for transparency.

 

[wink]

When it comes to "good" measurements, "Oh look how this $2.5K DAC measures so much better than expensive DACs" and when the measurements are "bad" then "if you can't hear it why bother?". How convenient.

Actually, it's all the plankton swimming around waiting to manifest themselves in a display of high energy melifluous musicality...... 

 

[460414]

@Baldr please just give me some evidence to back up your claim of superior accuracy of this DAC vs audio DACs in audio applications.

 

[noshortcuts]

We all have plenty of evidence ... it seems to me ... of a troll in our midst. Are you, nbe9, a customer? What is your investment in any of this Schiit beyond a hidden agenda? How are you honest and well intended while hiding behind the data of others? Why do you ask this question, really? Why not ask respectfully? What's the rush for an answer? ... Had you shown investment in audio, respect of engineers, honest intent, and patience in waiting for answers ... then no doubt you would have received equally thougtfull and respectful and honest response. I know better than to feed a troll and yet, tonight, I wonder if it isn't possible to ask why troll here?

 

[manpowre]

I think this thread went way out of bonds with calling people out personally. I would like to add a few bullet points about yggdrasil both on negative and positive way. First when I was looking for the DAC that I wanted to spend hours with every day for my headphone setup which was at that point LCD3 with Oppo HA-1 I reviewed Hegel HD25, HD30, Oppo HA.1 dac, and a few different 100 dollar dacs I cant even remember their name. The superior DAC was definetely the HD30, but it was so crisp and so hard in its attacks that I couldnt imagine using it in a headphone setup. Im sure HD30 was made for floor standing speakers where there is a high end equalizer in between in the signal path to adjust the room. But HD25 and HD30 was very silent in 24 bit mode. It has a noise floor out of this world, dead silent.. HD25 as more forgiving DAC, easyer to live with in Headphone setup, but due to an issue with USB input I couldnt live with it (loose USB input that made signal drop).
 
So when I reviewed Yggdrasil, I knew of its 21 bit locked mode where the noise floor wont be as deep as the Hegel products, but I was looking for the musicality for my LCD3, and at the point in time I got the Yggdrasil I owned LCD4, and got to do a real good review before purchase. 
 
For me what made Yggdrasil special is the musicality of the DAC and its consistency in reproducing 99% of the music I listen to to a satisfying level where I can use it day out day in hours every day and still enjoy it.
 
Yes, I do miss the possibility of 24 bit and the possibility to purchase high resolution material and push it through for all the damn details. Ive even considered a few times to get a used HD20 just for that for the fun of my hobby, or even looking at Oppo's new DAC with a sabre chipset to get that crisp high resolution for those special recordings. Or even if the day comes issues with Focal Utopia is solved to get a bricasti and connect Utopia directly to it.
 
But for now, its damn musical, and thats what we always should look for, and its bang for the bucks even considered that the noise level isnt as low as I wish the Yggy should be.
 
Yggdrasil is fast quick and high detailed for the type of music I like. Alot of drums, brass, cymbals and horns with guitars.
 
All the measurements in the world cant decide if a DAC is musical or not. I proved that with HD30 Im sure the Oppo HA-1 DAC section might give good measurements, but it was a lighter sounding dac than the AKM chips from Hegel and the Yggdrasil that for me didnt make sense (got listening fatigue with the HA-1 dac).
 
I wouldnt say Yggdrasil is superior, but for the cost of it, it gives high level of musicality within its 21 bit mode, precise and accurate to a high end level, and within that domain its competing with higher priced DAC's, but with less functionality as multibit possibility.
 
The funny thing is, if I sell my gear used, I could buy a Bricasti M1 + the Focal Utopia for the money and go upper High End. But Im afraid I would miss the LCD4 and the musicality I experience atm with the Yggdrasil and V281 amp.

 

[landroni]

But to the rest of the post.. Ok right, please give me some actual evidence to your original claim.

 

  @Baldr please just give me some evidence to back up your claim of superior accuracy of this DAC vs audio DACs in audio applications.

There is an English word for this...

 
EDIT: Post working links.

 

[Noldir]

For people using USB to AES converters, do you guys think the USB cable still makes a difference? What AES cable are you using? Have you noticed any difference in AES ables? I remember someone recommended a glass toslink cable (Cant remember if it was Mike from Schiit but I recall reading something like that) over plastic based toslink cables. I'm wondering if there's any recommendation/opinion like that for AES cables.

I know the folks at Blue jeans cables have written about the plastic toslink cables:

Optical Digital Audio Cable: Mitsubishi Eska POF
When we have a choice, we prefer to run digital audio in coax; it's more robust over distance, and the cable is interchangeable with cable used for certain other applications (e.g., composite video). However, an increasing number of devices are coming onto the market with digital audio available only in optical form, following the TOSlink standard. For these applications, we build our optical cables using the finest high-performance Plastic Optical Fiber (POF), Mitsubishi's ESKA Fiber. While POF is in general rather lossy stuff compared to glass optical fiber, we prefer it for optical digital audio use because it's much more physically durable and because its aperture matches the spec for optical digital audio use, unlike glass fiber which is too small and must be used in bundles. Our fiber is encased first in a tough cladding layer and then, for added durability, a flexible outer PVC jacket similar in texture to the PVC on some of our high-flex Belden cables (e.g. Belden 1505F). In our own usage, we've tested these cables at lengths up to 50 feet and found them to perform perfectly even at those extended distances.

 

[460414]

The above post (post 3981) is directed to the people talking about the musicality and perceived sound of that DAC. I am not questioning whether this DAC sounds good or not, and I am certainly not saying that this DAC sounds bad because of these measurements of that "x dac sounds better because it has lower distortion" or anything of the sort.
I am simply looking for said explanation.

Correct me if I am wrong, I do not possess a high level of knowledge of DACs or claim to have one:
A DAC with 100% accuracy: whatever you give it digitally, it gives you a perfect representation of it in analogue. This would by definition gives you zero distortion. Altering this output would introduce distortion. Hence: an alterations to the output, is inaccuracy. If the Yggdrasil was more accurate than other DACs with audio intended DAC chips, it would have lower distortion. A 0dbfs sine wave is excellent for showing inaccuracies as it goes progressively from zero to full amplitude showing any non linearities and inaccuracies.

I am not saying that anyone here made a bad choice in buying this DAC or questioning the perceived sound quality or even perceived sound accuracy. If I were to make a recommendation for a DAC at this price point based on listening reviews/impressions of others this DAC would probably be at the top of my list.

 

[earnmyturns]

This "24 bit" silliness is getting old. AFAIK, no recording studio can capture those low-order bits accurately. 24- (let alone 32-) bit encoding is just a design and marketing convenience, not a real reflection of what what can be recorded and reproduced by realistic transducers. One of these days, if we are not careful, people will be arguing for the importance of going below the noise floor of the Big Bang.

 

[Baldr]

This "24 bit" silliness is getting old. AFAIK, no recording studio can capture those low-order bits accurately. 24- (let alone 32-) bit encoding is just a design and marketing convenience, not a real reflection of what what can be recorded and reproduced by realistic transducers. One of these days, if we are not careful, people will be arguing for the importance of going below the noise floor of the Big Bang.

Wow! Someone who keeps my hours. Well done!

 

[earnmyturns]

Wow! Someone who keeps my hours. Well done!

Not normally At LHR waiting for my SFO flight.

 

[landroni]

Wow! Someone who keeps my hours. Well done!

In Europe we're very much awake. In fact, it's just about the right time for lunch...