| Message ID | 20220923183640.8314-11-vr_qemu@t-online.de |
|---|---|
| State | New |
| Headers | show |
| Series | audio: misc. improvements and bug fixes | expand |
On Fri, Sep 23, 2022 at 10:48 PM Volker Rümelin <vr_qemu@t-online.de> wrote: > The calculation of the buffer size needed to store audio samples > after resampling is wrong for audio recording. For audio recording > sw->ratio is calculated as > > sw->ratio = frontend sample rate / backend sample rate. > > From this follows > > frontend samples = frontend sample rate / backend sample rate > * backend samples > frontend samples = sw->ratio * backend samples > > In 2 of 3 places in the audio recording code where sw->ratio > is used in a calculation to get the number of frontend frames, > the calculation is wrong. Fix this. The 3rd formula in > audio_pcm_sw_read() is correct. > > Resolves: https://gitlab.com/qemu-project/qemu/-/issues/71 > Signed-off-by: Volker Rümelin <vr_qemu@t-online.de> > Would you mind adding the test to qtest? lgtm Acked-by: Marc-André Lureau <marcandre.lureau@redhat.com> > --- > audio/audio.c | 2 +- > audio/audio_template.h | 4 ++++ > 2 files changed, 5 insertions(+), 1 deletion(-) > > diff --git a/audio/audio.c b/audio/audio.c > index ba0c62b120..60c7472d37 100644 > --- a/audio/audio.c > +++ b/audio/audio.c > @@ -995,7 +995,7 @@ void AUD_set_active_in (SWVoiceIn *sw, int on) > */ > static size_t audio_frontend_frames_in(SWVoiceIn *sw, size_t frames_in) > { > - return ((int64_t)frames_in << 32) / sw->ratio; > + return (int64_t)frames_in * sw->ratio >> 32; > } > > static size_t audio_get_avail (SWVoiceIn *sw) > diff --git a/audio/audio_template.h b/audio/audio_template.h > index 7192b19e73..6a0337ac6b 100644 > --- a/audio/audio_template.h > +++ b/audio/audio_template.h > @@ -112,7 +112,11 @@ static int glue (audio_pcm_sw_alloc_resources_, TYPE) > (SW *sw) > return 0; > } > > +#ifdef DAC > samples = ((int64_t) sw->HWBUF->size << 32) / sw->ratio; > +#else > + samples = (int64_t)sw->HWBUF->size * sw->ratio >> 32; > +#endif > > sw->buf = audio_calloc(__func__, samples, sizeof(struct st_sample)); > if (!sw->buf) { > -- > 2.35.3 > > >
Am 27.09.22 um 13:54 schrieb Marc-André Lureau: > > On Fri, Sep 23, 2022 at 10:48 PM Volker Rümelin <vr_qemu@t-online.de> > wrote: > > The calculation of the buffer size needed to store audio samples > after resampling is wrong for audio recording. For audio recording > sw->ratio is calculated as > > sw->ratio = frontend sample rate / backend sample rate. > > >From this follows > > frontend samples = frontend sample rate / backend sample rate > * backend samples > frontend samples = sw->ratio * backend samples > > In 2 of 3 places in the audio recording code where sw->ratio > is used in a calculation to get the number of frontend frames, > the calculation is wrong. Fix this. The 3rd formula in > audio_pcm_sw_read() is correct. > > Resolves: https://gitlab.com/qemu-project/qemu/-/issues/71 > Signed-off-by: Volker Rümelin <vr_qemu@t-online.de> > > > Would you mind adding the test to qtest? > > lgtm > Acked-by: Marc-André Lureau <marcandre.lureau@redhat.com> > Hi Marc-André, I will give it a try. But it will be a separate patch, because the test from issue #71 now checks for the error at https://lists.nongnu.org/archive/html/qemu-devel/2022-09/msg02347.html and not the one from issue #71. With best regards, Volker
diff --git a/audio/audio.c b/audio/audio.c index ba0c62b120..60c7472d37 100644 --- a/audio/audio.c +++ b/audio/audio.c @@ -995,7 +995,7 @@ void AUD_set_active_in (SWVoiceIn *sw, int on) */ static size_t audio_frontend_frames_in(SWVoiceIn *sw, size_t frames_in) { - return ((int64_t)frames_in << 32) / sw->ratio; + return (int64_t)frames_in * sw->ratio >> 32; } static size_t audio_get_avail (SWVoiceIn *sw) diff --git a/audio/audio_template.h b/audio/audio_template.h index 7192b19e73..6a0337ac6b 100644 --- a/audio/audio_template.h +++ b/audio/audio_template.h @@ -112,7 +112,11 @@ static int glue (audio_pcm_sw_alloc_resources_, TYPE) (SW *sw) return 0; } +#ifdef DAC samples = ((int64_t) sw->HWBUF->size << 32) / sw->ratio; +#else + samples = (int64_t)sw->HWBUF->size * sw->ratio >> 32; +#endif sw->buf = audio_calloc(__func__, samples, sizeof(struct st_sample)); if (!sw->buf) {
The calculation of the buffer size needed to store audio samples after resampling is wrong for audio recording. For audio recording sw->ratio is calculated as sw->ratio = frontend sample rate / backend sample rate. From this follows frontend samples = frontend sample rate / backend sample rate * backend samples frontend samples = sw->ratio * backend samples In 2 of 3 places in the audio recording code where sw->ratio is used in a calculation to get the number of frontend frames, the calculation is wrong. Fix this. The 3rd formula in audio_pcm_sw_read() is correct. Resolves: https://gitlab.com/qemu-project/qemu/-/issues/71 Signed-off-by: Volker Rümelin <vr_qemu@t-online.de> --- audio/audio.c | 2 +- audio/audio_template.h | 4 ++++ 2 files changed, 5 insertions(+), 1 deletion(-)