[evil-zen]

I have a M-Audio Revolution 7.1 soundcard outputting a digital signal to my Ack dAck 1. Using foobar2000 to play udial via the soundcard is perfect with no weird sounds at all at default settings.

But if i tried to output it to my DAC, it is not good with various type of weird sounds depending on the settings I used. I tried fiddling with the 3 resamplers(SSRC,PPHS and SRC), output data format, dithering, output(ASIO,waveout,directsound,KS) but most of them doesn't work well.

The only combination that works perfectly is with DSP Resampler SRC (Super Rabbit Code) with Sinc Fastest quality at 88.2 or 96khz.

Is there a reason for this? I sure hope my DAC isn't resampling since it is a non-oversampling unit. I hope I can get a more technical answer from the likes of Glassman, Wodgy or jeifmeister.

Thanks!

 

[lini]

Well, it's a filterless design, isn't it? Welcome to hf artifacts, then...

Greetings from Hannover!

Manfred / lini

 

[sgrossklass]

For some fun, try routing the DAC's output to the sound card's input, adjust levels to obtain best SNR (wave slider all up, rec level depending on what's needed) and run some tests in RMAA at various sample rates (and 24 or 32 bits). The Revo has a fairly good ADC so should not bottleneck things all too much. If the DAC measures all clean, then it's probably high-freq intermod in the amp - one reason why no sane engineer would ever even consider a NOS filterless design. With such a design, you'll get lots of aliasing beyond fs/2 (see sampling theorem), which would in theory be inaudible given a perfect, zero-distortion amp which possibly serves as a lowpass beyond 20 kHz as well (they all do, more or less). (If you got a spare, could you send it my way?

) Anyway, it's interesting that there are no audio performance specs whatsoever for the dACK!. Only the output level is given, ~0.7Vrms. Something you can only sell to audiophools.

 

[evil-zen]

Quote:

Originally Posted by sgrossklass If the DAC measures all clean, then it's probably high-freq intermod in the amp - one reason why no sane engineer would ever even consider a NOS filterless design. With such a design, you'll get lots of aliasing beyond fs/2 (see sampling theorem), which would in theory be inaudible given a perfect, zero-distortion amp which possibly serves as a lowpass beyond 20 kHz as well (they all do, more or less).

Well actually I've measured my DAC using RMAA last year. Seems perfectly fine to me with no HF artifacts. In fact I measured my Gilmore V2 amp too and it is almost the same as the above with a poorer stereo crosstalk probably due to the pot issue.

What I am hearing is sort of an alien sound not unlike what AC97 soundcards produce with no resampling not HF artifacts.

Actually two other people also showed the same result with SRC Sinc Fastest except they are not using a DAC.

 

[sgrossklass]

Quote:

Originally Posted by evil-zen Well actually I've measured my DAC using RMAA last year. Seems perfectly fine to me with no HF artifacts.

To me the graphs look quite different. There actually is quite a bit of high-freq stuff in there, the THD measurement alone shows a nice zoo of high-order harmonics as well as anharmonics (albeit few of them reach more than -90 dB), and with the IMD measurement you can see a whole lot of IM mud above 600...700 Hz instead of a clean background (the distribution of harmonics for the bass tone also is quite strange - 2nd -110 dB, 3rd -105 dB, 4th -100 dB, 5th below noise again). The frequency response drops off from a few kHz onwards already and reaches -.5 dB at ~12 kHz, probably the result of a low-order lowpass to keep at least the most nasty aliasing dreck away. Overall, this looks a bit like resampling à la Creative, not quite as bad and not quite the same but certainly not exactly pretty.

 

[evil-zen]

bump