Author Topic: bit-perfect output to usb for external dac  (Read 3825 times)

beaker152

  • Newbie
  • *
  • Posts: 5
is it possible to completely bypass the soundcard to send bit-perfect data straight to the usb?

hiccup

  • Sr. Member
  • ****
  • Posts: 7781

beaker152

  • Newbie
  • *
  • Posts: 5
an external dac (topping D50s) that has the rest of its inputs already occupied...

hiccup

  • Sr. Member
  • ****
  • Posts: 7781
Then that should be recognized by Windows (and MusicBee) as a soundcard.
Next, in MusicBee, select your Topping, use Wasapi exclusive mode, and you have bit-perfect output.
(assuming you have disabled any effects etc.)

beaker152

  • Newbie
  • *
  • Posts: 5
cool. 

followip:  is there a way to record the digital output from that config using audacity (or other) so i can
compare the digital output to the flac input?

hiccup

  • Sr. Member
  • ****
  • Posts: 7781
I'm guessing that should be possible, but I have no concrete answer on how to.
It might need using wav and some Asio tools, but I would need to google for that, and I am sure you are able to do that yourself ;-)
(and then perhaps report back on your findings?)

For what it's worth, a while back I had an RME ADI-2 Pro to play with, and that dac has a built in bit-perfect test feature.
I tested it on MusicBee, and it passed all bit-perfect tests for me.

beaker152

  • Newbie
  • *
  • Posts: 5
that's interesting about the RME having a bit-feature...  i was shopping for one a few weeks ago and it seemed it was on backorder everywhere, so i got the D50s as a placeholder.

i haven't yet found out how to capture the digital output stream, but i did do a quick test using the analog output from the dac as input to my pc, recorded a few seconds of a song with audacity, then compared to the flac by aligning the waveforms, inverting, and merging...   most of the result was close to null except near some sharp transients.  i exported the data to csv and used excel to find that there was a temporal shift of 4 samples over the 2 seconds of data and when i manually adjusted for it by removing 4 individual samples from one curve, they overlayed very nicely.

seemed like the bit slips occurred just prior to some sharp transients.
i found it interesting, but can't say if it was due to the dac, my sound card, or audacity...  might record the whole song to see which curve is longer

will keep looking for a way to capture the digital stream...

hiccup

  • Sr. Member
  • ****
  • Posts: 7781
that's interesting about the RME having a bit-feature...  i was shopping for one a few weeks ago and it seemed it was on backorder everywhere, so i got the D50s as a placeholder.

Sorry, drifting a bit off-topic here, but I returned it since I was not completely satisfied with the sound quality, especially using headphones.
In all other aspects, it really is a fantastic device.

Quote
i did do a quick test using the analog output from the dac as input to my pc, recorded a few seconds of a song with audacity, then compared to the flac by aligning the waveforms, inverting, and merging...   most of the result was close to null except near some sharp transients.

Wow. Isn't it impressive how close your analogue recorded stream is to the digital source?

I do respect your quest for perfection, and I am honestly interested to read and learn of your further efforts, but realistically, it will probably not result in anything that will change the sound quality as even Superman's ears could perceive.

beaker152

  • Newbie
  • *
  • Posts: 5
what, you mean those missing few bits won't get me the frequency response equivalent of Salar de Uyuni, the acoustic depth of Valles Marineris, with just a hint of citrus and enough oaky leather for my subwoofer to make my eyeballs wobble?

just kidding--totally not worried.  :)