[alsa-devel] Acer Ferrari 5000

Takashi Iwai tiwai at suse.de
Mon Mar 23 18:07:55 CET 2009

At Mon, 23 Mar 2009 09:40:46 -0700,
Russ Dill wrote:
> On Mon, Mar 23, 2009 at 9:36 AM, Takashi Iwai <tiwai at suse.de> wrote:
> > At Mon, 23 Mar 2009 09:27:24 -0700,
> > Russ Dill wrote:
> >>
> >> I'm testing with the acer-aspire quirk set, and things seem to be
> >> working rather well. The headphone plug/unplug detection and speaker
> >> muting actually work :) However, I'm still not getting HDMI audio
> >> working. I'm not that familiar with IEC958, how should I have the
> >> switches set? I currently have both IEC958 and IEC958 Default PCM on.
> >> I've tried with mplayer -ao alsa:device=iec958 <mp3 file>
> >
> > Basically it should be OK.  I guess it's rather an issue of the video
> > driver side.  If the video driver doesn't support HDMI audio output,
> > it doesn't work, of course.
> >
> [ Working on testing patch... ]
> The audio does work over HDMI though, Vista has better support for
> connecting to HDTV and getting the video modes right, so I was
> watching a DVD. Out of curiosity, I checked a box in the audio
> configuration for digital output, and sound started coming out of the
> TV.

Do you mean it working on Linux?  When "IEC958 Default PCM" is set
to on, the analog PCM outputs are also routed to the digital output
automatically.  So you should be hearing the sound played via aplay.

The "iec958" PCM device is for the exclusive digital output, mainly
for the raw digital stream like AC3.  If it's being used with the
analog streams at the same time, you won't be able to open it.

BTW, I found why model=auto doesn't work well on your machine.
It's because (again) of BIOS.  It doesn't set the codec SSID properly
so the driver doesn't accept the HP auto-toggle and other features.

So, in your case, using model=acer-aspire would be likely the best
choice (supposing it matches with most functionality).  If you have no
problem (at least regressions) with this model, I'll add the entry to
point to model=acer-aspire

Anyway, my previous test patch is independent from the model problem.
It's still helpful to know whether it can solve the codec
communication errors with model=auto.



More information about the Alsa-devel mailing list