[Stil]

  Anyone compared the Schiit Gungnir to the Meridian Director DAC? Same price.

I'm interested in this too.  

 

[Xymordos]

Perhaps your dislike towards the 9018 chips is due to the distortion that were tested with AP and confirmed with ESS. Between 1khz and 5khz, with signal between 0db and 60db, there is nonlinear distortion. It looks inaudible, but some people might believe otherwise.

Spoiler: Warning: Spoiler!

Teal is -35dB input signal, and Blue is -36dB input signal

 

[purrin]

Wow, if that for is real, the most scary part is not the distortion per se, but the completely different characteristics of the distortion between -35db and -36db.

 

[Sapientiam]

Teal is -35dB input signal, and Blue is -36dB input signal

 
Very useful data - thanks! I've been harping on about noise modulation in S-D DACs for a while now, very nice to see another smoking gun. The plot of the Weiss Medea+ showed a kink in the THD+N vs level plot around this very level.
 
Are you sure you got the colours the right way around though? From eyeballing the plot, the blue looks to have a slightly higher level than the cyan. I'd also expect the noise to be triggered at a specific absolute signal level - the lower amplitude sine does not pass through the critical point around -35dB so doesn't trigger the noise bursts.
 
Remember that all FFTs only show the average noise - a shift of 20dB in the 2kHz region looks big but given the generated noise is only happening over a relatively small part of the sinewave, the subjective effect is worse than the plot seems to suggest.

 

[yfei]

1) Why it looks like the peak frequency of the 2 curves are different?   not exactly matching 1kHz.
2) the distortion is below -130~140db, can it really cause audible differences?
 
 
Quote:

Perhaps your dislike towards the 9018 chips is due to the distortion that were tested with AP and confirmed with ESS. Between 1khz and 5khz, with signal between 0db and 60db, there is nonlinear distortion. It looks inaudible, but some people might believe otherwise.

Spoiler: Warning: Spoiler!

Teal is -35dB input signal, and Blue is -36dB input signal

 

[Sapientiam]

the distortion is below -130~140db, can it really cause audible differences?

 
Interpreting FFTs is not that straightforward.. A casual look at the plot shows noise 'floor' around -150dB. But remember that's not the total noise (integrated noise from 20-20kHz) rather each FFT 'point' shows you the noise present in a very narrow bandwidth. You have to add all those points up to get what's called the 'integrated noise'.
 
Eyeballing the place on the plot where the frequency scale is most expanded (bottom left), note that the 'steps' occur at roughly 3Hz intervals. This is a clue as to the effective noise bandwidth for each data point. It's only a good guide though if no averaging has been done - the AP can be set to average 10s or even 100s of waveforms.

 

[purrin]

   
Thanks for the support. I don't own the TEAC but am very close to buying one based on so many positive reviews.

 
This week, a friend and I made some visits to a few other friends (who typically carry a few DACs here and there at their places of work). Here are some pictures of the TEAC in various states of use disuse.
 

(The Airmotiv speakers were actually liked.)
 

 
 It's probably a great DAC if you haven't heard anything better.

 

[Xymordos]

I did mention that it might not be audible, since many of the big DAC companies use it. I know that one of the Taiwan developers managed to test this on his AP and decided not to use it just in case, since there was no fix (confirmed by ESS). John Westlake (designer of MDAC) also said that the 9018 was the most buggy chip he used, and he'll never use it again because of how frustrating it was.

 

[bowtung]

Just wondering does anyone know how the fostex hp-a8 compares to the gungir. I noticed that they have the same akm 4399 chip. Of course implementation matters, but it would be interesting to know how these different akm dacs sound? 
 
There are a lot of info and comparison on 9018 dacs, but not much on the akm 4399 ones.
 
Thanks

 

[Andrew_WOT]

  I did mention that it might not be audible, since many of the big DAC companies use it. I know that one of the Taiwan developers managed to test this on his AP and decided not to use it just in case, since there was no fix (confirmed by ESS). John Westlake (designer of MDAC) also said that the 9018 was the most buggy chip he used, and he'll never use it again because of how frustrating it was.

There is an interesting site of the Russian DAC designer/developer who considers delta-sigma (ESS) a top of the line chips today, and he used to build multibit DACs before. 
 http://goo.gl/RNp4xB
There is also a pdf describing his 5 years adventure dealing with ESS to make it sound to its best. It's in Russian, but you can probably run it via some google translate.
 
 
Mostly the difficulty was poor documentation and confusing spec and as a result most manufacturers implementing inferior sounding solutions based on simplest functions of the chip.
Interesting that AURALiC mentioned that very problem when designing their Vega, so they had to work very closely with chip manufacturer.

 

[aamefford]

@purrin and "minions" - reading the first post review section, you guys make a really good case for buying/building a gamma2 or a gugnir, depending on how much under $1000 one wants to spend, and calling it a day.  That is unless one just wants to spend in the land of what I would term "crazy money" for marginal diminishing returns.  Fair comment?  To quantify - the Gugnir is already past at my "crazy money" threshold, but the Gugnir / Mjolnir combo is on my wishlist.
 
Currently thinking I may pick up a gamma2 to stand in for my CEntrance HIFI-M8, which is doing double duty as it's own portable rig and as the DAC to my Burson Soloist.  What I'm really looking for, is a dac to pair with the burson, that will be the best spending of $250 to $500.
 
I am pretty solidly seated in "upper Mid-Fi" with my gear and budget.  I'm trying to maximize my gear within said budget...

 

[zerodeefex]

  @purrin and "minions" - reading the first post review section, you guys make a really good case for buying/building a gamma2 or a gugnir, depending on how much under $1000 one wants to spend, and calling it a day.  That is unless one just wants to spend in the land of what I would term "crazy money" for marginal diminishing returns.  Fair comment?  To quantify - the Gugnir is already past at my "crazy money" threshold, but the Gugnir / Mjolnir combo is on my wishlist.
 
Currently thinking I may pick up a gamma2 to stand in for my CEntrance HIFI-M8, which is doing double duty as it's own portable rig and as the DAC to my Burson Soloist.  What I'm really looking for, is a dac to pair with the burson, that will be the best spending of $250 to $500.
 
I am pretty solidly seated in "upper Mid-Fi" with my gear and budget.  I'm trying to maximize my gear within said budget...

 
commissioning a gamma2 will be 300-350.
 
Sell the M8 for $600, making your total budget $1100.  Get a used Gungnir for $600 (seems to be the going rate) + the USB gen2 install yourself for $100 + Leckerton UHA760 as your new portable.
The gamma2 is great, but not that resolving. I think you'd be happer in the long run going straight to the Gungnir.

 

[EraserXIV]

I was quite surprised how good the gamma 2 was while I had it. I didn't even realize how good it was until I sold it to "upgrade" I have since been through 2-3 DACs in the $500+ range and haven't found something that has impressed me enough to justify the extra cost. I'm still searching and have since increased my budget to the $1000 range.

 

[GoldfishX]

Well, I can confirm what was said earlier...The Gungnir and the HD800 really do not seem to like each other. Too bright, too aggressive, was not a pleasant experience, even with some light EQing in the high mid/treble bands. My Peachtree Dac-It crushed it pretty good with the HD800's, using the exact same source and the Lyr with the same set of Siemens tubes. Very transparent and musical pairing, it passed pretty much everything I threw at it, from Missing Persons to Dokken to Mozart to the NES Brinstar theme from Metroid.
 
Gungnir worked pretty good with both the Mad Dogs and HE-500, but was still different in both cases. The HE-500 was especially strange, I got a really dark sound with really rolled off treble (yes, I reset the EQ from the HD800). Was nice, but sounded unnatural. On the Mad Dogs, I had to EQ some of the treble down (the bass was perfect). Never had to do that when it was paired with the Dac-It.
 
At this point, I'm highly considering running the Dac-It in my home set-up and either selling the Gungnir or using it at work (I'd pair it with the Mad Dogs and the Magni in my cubicle). Not being so good with the HD800 just really sucks. Only thing I can think is maybe try a warmer tube in the Lyr, like a Mullard to tone things down. I've only had the Dac-It for about 2 weeks and it's passed every test I've thrown at it. I'm amazed, to be honest, as I previously had the impression it was an entry-level DAC, nothing too special. I'm disappointed the Gungnir, the Schiit flagship, didn't blow it out of the water (not even close).
 
Both DAC's were done using coaxial connections to my DX50.

 

[olor1n]

Lol purrin. Was it my whinging that bumped the M51 to class D?
 
Are there any ninjas that can lend you the NAD? It'd be interesting to see what the OR5 can do for it.