[pulseaudio-discuss] Merging soxr

[Andrey Semashev]

On Thursday 08 January 2015 17:58:43 Alexander E. Patrakov wrote:
> 08.01.2015 17:44, Andrey Semashev wrote:
> > On Thursday 08 January 2015 11:29:42 Alexander E. Patrakov wrote:
> >> 08.01.2015 01:52, Andrey Semashev wrote:
> >>>> Also, with PulseAudio forced to 44.1 kHz, FooBar2000 v1.2 (which uses
> >>>> DirectSound and thus, by default, resamples everything to 48 kHz) just
> >>>> plays silence (with a neverending stream of underruns in pulseaudio
> >>>> log)
> >>>> over soxr-vhq and works fine over speex-float-5. soxr-hq and soxr-mq
> >>>> also work fine with wine.
> >>> 
> >>> I didn't quite understand this test, could you elaborate? Are you
> >>> running Windows in a VM here? Which one?
> >> 
> >> That's in wine. FooBar2000 v1.2 (and not any later version) is good for
> >> testing DirectSound-related code paths.
> > 
> > Hmm, I'm having trouble reproducing this (for now I'm trying with my
> > patched PA 4.0 with soxr-vhq). In Foobar2000 1.2.9 on wine 1.6.2 I can
> > only select "Primary Sound Driver" or "Pulseaudio" as the output, in both
> > cases I cannot select the output format - it says that output format will
> > be chosen automatically for the selected device. 'pactl list sink-inputs'
> > says the signal sample rate is 44100 Hz. I tried to add resampler to 48
> > kHz in the Foobar2000 DSP pipeline, but the result is the same. In any
> > case, the sound plays flawlessly.
> > 
> > Are there any specific settings I should do to reproduce the problem?
> 
> Well, I guess this is a left-over from my experimenting with the
> unpatched alsa output module. And your distribution applies the
> non-upstream "pulseaudio output" patch.
> 
> So, please add this registry entry to reproduce the test result:
> 
> [HKEY_CURRENT_USER\Software\Wine\Drivers]
> "Audio"="alsa"
> 
> regedit alsa.reg

No, that still doesn't help. After doing that and rebooting the sound still 
plays fine - I assume, directly through ALSA since 'pactl list sink-inputs' 
displays nothing. I don't see any devices that would route the signal back to 
PA.

> But it still boils down to a low-latency stream.

In that case, I think, there's nothing we can do. The resampler does have 
latency, and we have documented it. Just don't use it if you need a lower 
latency.