[Pkg-alsa-devel] Bug#503839: Bug#503839: /usr/lib/alsa-lib/libasound_module_pcm_vdownmix.so: vdownmix from 5.1 does not incorporate LFE channel

[Josh Triplett]

On Mon, Sep 07, 2009 at 09:36:56PM +0200, Apelete Seketeli wrote:
> > With your setup, does "speaker-test -c 6 -t wav" generate sound for all
> > six channels?
> 
> Ok, so I don't get the .1 output either, but is it relevant here ? The
> speaker test doesn't output it, but any 5.1 video just plays fine, 6
> channels and all. At lest I assume, since I don't feel like there's
> any sound missing while watching...

It does make a difference; whether you can easily notice depends on the
soundtrack.  Some soundtracks mix more or less onto the LFE, and more
importantly off of other channels.  Lacking the LFE for some soundtracks
can make some sounds softer to the point of almost becoming inaudible,
while leaving other sounds nearly unchanged.

For an analogous situation, consider what would happen if you took
stereo audio, dropped the right channel entirely, and mixed the left
channel onto both.  Often you wouldn't notice, if you listened to
something mostly balanced between channels; on the other hand, sounds
balanced mostly to the right channel would become very quiet.  The same
thing can happen with the LFE to some extent, depending on how any given
sound got mixed into the channels.  Particularly with the LFE, you
probably wouldn't notice any sound go missing (because sounds probably
won't appear only on the LFE channel), but the relative
volumes of sounds may change.

> > It certainly *shouldn't* depend on the sound card.  Mixing 5.1 to 2.0 in
> > software should work anywhere, and I have working 2.0.  Of course, that
> > doesn't mean it *doesn't* depend on the sound card, just that it
> > *shouldn't*. :)
> 
> Well, here is what I get on an X60:
> 
> 
> ~$ speaker-test -c 6 -t wav
> 
> speaker-test 1.0.20
> 
> Playback device is default
> Stream parameters are 48000Hz, S16_LE, 6 channels
> WAV file(s)
> Rate set to 48000Hz (requested 48000Hz)
> Buffer size range from 64 to 16384
> Period size range from 32 to 8192
> Using max buffer size 16384
> Periods = 4
> was set period_size = 4096
> was set buffer_size = 16384
>  0 - Front Left
>  4 - Center
>  1 - Front Right
>  3 - Rear Right
>  2 - Rear Left
>  5 - LFE
> Time per period = 8,401077

Go figure.  So, it looks like you somehow have a "surround51" device to
begin with, since you use that as a target for vdownmix without defining
it.  The presence of this device may also depend on the "model"
mechanism in snd-hda-intel.

In any case, if you don't get the LFE channel either, it sounds like
vdownmix may just not handle it correctly.

> I'll be glad to further test too, since there's obviously something
> strange here.

Indeed.  Thanks.

- Josh Triplett