duncanthrax / scream

Virtual network sound card for Microsoft Windows
Microsoft Public License
1.69k stars 142 forks source link

Native data flow #187

Open Traktorist3d opened 1 year ago

Traktorist3d commented 1 year ago

Hello, thank you very much for such a useful thing as Scream. Tell me, is it possible and how to make all the data go to the receiver without oversampling, but as it is. If I listen to 16 bits 44.100, so that it goes to the receiver, if I turn on 96k 24 bits, send such data to the receiver without changes.

duncanthrax commented 1 year ago

Hmm not sure what you mean, because that's exactly what the Scream driver does. PCM data coming from the Windows sound subsystem is dumped onto the network without any conversion.

However, receiver sound stacks might resample to match their own outputs.

Traktorist3d commented 1 year ago

I'm sorry, but I checked again, the sampling rate is really the same as the original signal, it's very good. But why is 32Bit always on the receiver, no matter what is supplied, maybe it's the Pure receiver ... Scream

duncanthrax commented 1 year ago

Ah yes. I do remember that Windows actually dumps 4 bytes (32bit) per 24bit sample. Probably for alignment purposes. Scream will then signal 32bits fixed point and the receivers dump it as such. It should not make a difference. I am not sure if Windows does this in all cases, it may depend on the source material being played, maybe.