Thanks to all of you for your replies ! This is becoming interesting.
Yes,
@maverickronin got my point. I'd be happy to spend a lot of time (on an old PC, let's not forget that part) if I was certain of the result. Plus it's not even my laptop, it's my GF's, and she's currently writing a professional thesis on it. So it's not that I can't have it for some hours or even a day. It's that I can't risk that something goes wrong at some point (even if it's unlikely) that would affect her work afterwards (i.e. Office not working, Internet not working, etc.). I'll be talking with her in the next few days and we'll see if we can find a solution.
@RayDunzl , your message is really helpful !
It proves that somehow on Windows 10,
A4A can actually detect Topping DACs (at least a D10, but it should be the same with a E30). That would definitely solve my problem. So unless A4A developer answers my emails, maybe upgrading to Win 10 is the next best solution. We'll see.
This inherently doesn't work without the syncing between the two dacs. The XMOS interface is asynchronous. You can force it to work for a while and maybe you are lucky the oscillators are closely matched then you may not see issues. But this should require a real multichannel dac with dsp.
Thanks for chiming in John
The sync issue has been talked several times already :
- The two headphones are independent (it's two people listening to the same music at the same time). So if one headphone happened to be slightly our of sync vs. the other, nobody would ever notice unless that drift became close to a second or so, which is unlikely.
- Pressing "stop" in the player stops the audio feed and resets audio drift to 0, right ? So simply pressing "stop" after each song would do the trick.
- What would be the audio drift for a single typical song ? I have no precise data, but somebody on ASR suggested that average audio drift between two DACs could be close to 2 samples per second (maybe you have more accurate data). That means 480 samples for a typical 4-minute song. It's smaller than most audio buffers (A4A buffer is 1024 samples, foobar2000 buffer can be as big as the whole song).
- Even if the audio delay became bigger than A4A buffer, what would happen exactly in worst case scenario ? An occasional audio glitch ? That would be a really small price to pay IMHO.
Let me know if I'm wrong of course