Thoughts on a bunch of DACs (and why delta-sigma kinda sucks, just to get you to think about stuff)

[purrin]

No full access to prototype, but I finally did get to hear the same unit as the one at RMAF. The same one Maxvla, Misterrogers, and a few others heard. Take everything with a grain of salt because obviously the Yggy I heard isn't the final production.
 

• Neutral sounding, not bass light, not warm, not cold, not rolled, not laid back. Maybe slightly "aggressive" a la Gungnir. It's all relative because I feel Gungnir tonality is fine. So I think folks expecting that warm, soft, and murky PCM1704, syrupy super smooth UltraAnalog, or laid-back and rolled Metrum sound, or ever-present tonally dense "NOS" sound could be disappointed. Schiit doesn't necessarily have a house sound. But with the exception of some of their first generation stuff, they don't make gear that sounds overly polite or forgiving, fuzzy, rolled, laid back, with blunted attacks, or overly warm or bassy. The warmest amp in existing lineup is the Lyr 2, and even then, it's just a touch. 
• The Yggy is a fricking resolution monster. It will scrap every single bit from the PCM data. It's unbelievable in this regard. I've heard a lot of resolving DACs, many that I haven't even ranked, and the Yggy beats them all by a good margin. Makes me wonder if all other DACs really aren't effectively 16-bits / 24-bits, but more like 14. I mean, there's just all this schiit flying everywhere. It was information overload for a while - just not used to it - especially into headphones which I haven't been using as much as speakers lately.
• There was no added glare, grain, sigma-delta hashy treble garbage, yet at the same time, it didn't impart a smoothness over everything like how the SFD-1 tends to do. Some recordings sounded pretty rough, rougher than usual, e.g. the more recent Pixies MSFLs which tend toward bright. Yggy won't remove glare or grain present in the recordings. Pristine recordings sounded absolutely wonderful, especially those recorded and mixed in the analog domain from years past. You can tell which recordings were mixed or mastered in the analog domain and which ones were mixed or mastered in the digital domain. I think given more time to fully warm up, the Yggy would have smoothed out even more and eased up on some of that aggression, as this was the case with the Theta Gen Va where I felt about four hours was about right. I only had about two to three hours on and off with the Yggy. Most of the time was spent waiting for it to warm up.
• In comparison to the Gen Va which I had on hand for a direct comparison, the Yggy started to sound eerily similar to it after it warmed up for about an hour or two. Really no surprise. Same texture, twangs, growls, roars. The kind of sounds that makes the hairs on your back stand up. Same super precise nuanced taut bass that stops on a dime. The differences, as already alluded to above, were more detail and resolution - just on another level; and clearer with more impact than the Gen Va, which already is super hard hitting.
• Yggy had same holographic 3d soundstage as Gen Va. As already indicated in posts elsewhere, I got a exact sense of the size of the church, the ceiling, the walls, in that Cowboy Junkies album, from the reverb and ambient cues. And you get this great "sense of space" as I call it, from only the ambient noise too. When music is playing, the imaging fakes you out and makes you look left, right, down, especially when you are using open headphones and get confused between ambient sounds and sounds from the recording. To me that's a good thing.

 
Yggy it is very much a modern DAC in that it is super resolving with unadulterated attacks. In the sense that it is a vintage 90s DAC, there is no sigma-delta treble hashy raspy stuff and a lot of harmonic texture.
 
Did I want to steal it and take it home with me? Most certainly yes. I miss it already.

 

[zachchen1996]

  The Yggy is a fricking resolution monster. It will scrap every single bit from the PCM data. It's unbelievable in this regard. I've heard a lot of resolving DACs, many that I haven't even ranked, and the Yggy beats them all by a good margin. Makes me wonder if all other DACs really aren't effectively 16-bits / 24-bits, but more like 14. I mean, there's just all this schiit flying everywhere. It was information overload for a while - just not used to it - especially into headphones which I haven't been using as much as speakers lately.

 

 

[Baldr]

Admittedly off topic in a DAC thread:
 
REVEAL:  The square chip is an ADSP-21478

 

[NoPants]

thanks for the reveal, now I can sleep soundly at night

 

[jacal01]

But a great reveal nonetheless. 
 
Yeah, those do look like AD DAC chips, now that you mention it.  Makes sense, considering that pretty much all of TI's R2R DAC chips are now defunct, except for multiplying DACs and the PCM1704, which we know you don't like.  AD1851R-J work for you?  

 

[magiccabbage]

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

• The Yggy is a fricking resolution monster. It will scrap every single bit from the PCM data. It's unbelievable in this regard. I've heard a lot of resolving DACs, many that I haven't even ranked, and the Yggy beats them all by a good margin. Makes me wonder if all other DACs really aren't effectively 16-bits / 24-bits, but more like 14. I mean, there's just all this schiit flying everywhere. It was information overload for a while - just not used to it - especially into headphones which I haven't been using as much as speakers lately

 
Great to here. I cant wait to hear it with my DNA Stratus which should arrive very soon

 

[twinkle]

  Admittedly off topic in a DAC thread:
 
REVEAL:  The square chip is an ADSP-21478

Ah-ha! you were right Jacal01 about it being a dsp chip

Sharky!
 
Can I say I'm relieved it's not the ESS that was shown?
Now that cat needs help methinks (I'm talking about zachchen's cat).

 

[jacal01]

I think I'm right about the AD1851 DAC chip, too, altho the PCM63 was a dual channel DAC  (ala AD1866).
 
Those TI chips from before are all monotonic multiplying R2R DACs, and while their specs are fairly rigorous, I never could find evidence that they're suitable in an audio application.
 
Aaand to be snarky (vs. sharky), you notice that purrin's DAC like/dislike pie chart comparisons are all categorized by their implemented DAC chips, not by their digital filter processor chips. 

 

[StefanJK]

  I think I'm right about the AD1851 DAC chip.
 
 

 AD1851 being 18-bit isn't a problem?  PCM63 is 20-bit.

 

[jacal01]

No, I suspect that 16 bit resolution is going to be the best we'll get from the audio multi-bit R2R DAC chips still in production.  The 18 bit actually refers to the AD1861 chip, which is now obsoleted.  It appears to me that the different bit resolutions mainly manifest in their respective signal-to-noise ratio specs, altho the THD+N of the 24 bit PCM1704 is actually more like 17 bit, with the 16 bit R2Rs probably closer to 14 bit.
 
However, we do know that there is more to DAC sonics than mere numerical bit resolution, as evidenced by Mike Moffat's scathing denunciation of the PCM1704 relative the the PCM63, and hopefully now these AD1851/1866 R2R DAC chips.  Much obviously is in the circuit implementation, with the Schiit bit perfect programmed DSP digital filtering pulling whatever remaining however niggly DAC resolution (and linearity) shortcoming chestnuts out of the fire.  That's the plan, anyway.

 

[StefanJK]

  No, I suspect that 16 bit resolution is going to be the best we'll get from the audio multi-bit R2R DAC chips still in production.  The 18 bit actually refers to the AD1861 chip, which is now obsoleted.  It appears to me that the different bit resolutions mainly manifest in their respective signal-to-noise ratio specs, altho the THD+N of the 24 bit PCM1704 is actually more like 17 bit, with the 16 bit R2Rs probably closer to 14 bit.
 
However, we do know that there is more to DAC sonics than mere numerical bit resolution, as evidenced by Mike Moffat's scathing denunciation of the PCM1704 relative the the PCM63, and hopefully now these AD1851/1866 R2R DAC chips.  Much obviously is in the circuit implementation, with the Schiit bit perfect programmed DSP digital filtering pulling whatever remaining however niggly DAC resolution (and linearity) shortcoming chestnuts out of the fire.  That's the plan, anyway.

Sorry, went for the AD1861 specs.  As you write, AD1851 is 16 bit.  I suspect Schiit is going with at least 18 bit, more likely something 20 bit like the AD 5791 (with some preprocessing hacks?).  But I'm just guessing without any field relevant expertise.  I'd bet you a dollar, maybe ten after a bit more googling.

 

[Jason Stoddard]

  No, I suspect that 16 bit resolution is going to be the best we'll get from the audio multi-bit R2R DAC chips still in production.  The 18 bit actually refers to the AD1861 chip, which is now obsoleted.  It appears to me that the different bit resolutions mainly manifest in their respective signal-to-noise ratio specs, altho the THD+N of the 24 bit PCM1704 is actually more like 17 bit, with the 16 bit R2Rs probably closer to 14 bit.
 
However, we do know that there is more to DAC sonics than mere numerical bit resolution, as evidenced by Mike Moffat's scathing denunciation of the PCM1704 relative the the PCM63, and hopefully now these AD1851/1866 R2R DAC chips.  Much obviously is in the circuit implementation, with the Schiit bit perfect programmed DSP digital filtering pulling whatever remaining however niggly DAC resolution (and linearity) shortcoming chestnuts out of the fire.  That's the plan, anyway.

 
That's one plan. Not our plan. 

 

[NoPants]

you're not going to pull chesnuts? now I don't know what to believe

 

[jacal01]

The DAC chips used, or the DSP filter function?

 

[StefanJK]

There goes the easily winnable leg of that bet with jacal01 I proposed.