I already got an excellently working Win10 Gaming VM. Since I own headphones of higher quality I improved my Pulseaudio and Alsa settings on my Arch Linux to get the best out of it, and therefore had to tinker with my QEMU settings of the VM to match frequency and so on.
I now got it working… though am confused.
My host’s pulseaudio settings/changes:
default-sample-format = float32le
default-sample-rate = 48000
alternate-sample-rate = 44100
default-sample-channels = 2
default-fragments = 2
resample-method = soxr-vhq
enable-lfe-remixing = no
high-priority = yes
nice-level = -11
realtime-priority = 9
daemonize = no
Guest’s qemu xml file:
<domain xmlns:qemu="http://libvirt.org/schemas/domain/qemu/1.0" type="kvm">
[...]
<sound model="ich9">
<address type="pci" domain="0x0000" bus="0x00" slot="0x1b" function="0x0"/>
</sound>
[...]
<qemu:commandline>
<qemu:arg value="-audiodev"/>
<qemu:arg value="id=pa,driver=pa,in.fixed-settings=on,in.frequency=48000,in.format=s32,in.channels=2,server=/run/user/1001/pulse/native"/>
</qemu:commandline>
</domain>
The audio-passthrough from the Windows 10 guest machine works flawlessly again. What I just wonder is why my guest machine still only seems to support a 16bit bitrate. The virtual hda soundcard should technically support more (at least the real one does, right?), and as far as I understood I absolutely have to make sure everything in the chain from my VM to my real sound output device works within the same specifications (32bit 48.000hz, in this case). But I couldn’t find any settings I didn’t already set within the VM’s xml config file.
So, am I just very lucky to have qemu properly forwarding the sound without cracks and distortions, or is it just the way qemu works, not supporting / showing >16bit audio?