[alsa-devel] Bug in setting channels, related to recent ELD/EDID changes?

[VDR User]

On Tue, Jan 4, 2011 at 10:46 AM, James Courtier-Dutton
<james.dutton at gmail.com> wrote:
>>> A/V receivers showing TV information is indeed something that wasn't
>>> thought about :(
>>
>> This has certainly caused problems.  Anyone thought of a way to
>> resolve it, or better yet reversed the patch that created the
>> situation?
>
> I think this is related to general hot-plug support in ALSA.
> If you plug in a USB speakers, ALSA should detect new speakers and
> have a way of informing applications about it.
> When you unplug USB speakers, ALSA should detect the lack of speakers
> and have a way of informing application about it.
> In the same way the if you power on a HDMI amp, instead of just two
> speakers (in the TV), we now have 5.1 speakers.
> ALSA should have a way to inform applications of the changes.
>
> I do not think ALSA alone can handle these hot plug/unplug events. It
> has to be applications that are made to understand these events and
> adjust accordingly.
> Before now, ALSA has had to be told manually how many speakers a user
> has attached, E.g. Stereo, 5.1 etc.
> With modern sound devices, this information is available automatically
> from the hardware and changes without notice.
>
> So, I propose that an API should be added to achieve this.
> So, instead of the application setting the speaker arrangement, the
> speaker arrangement is set in ALSA, and the application requests
> information about it.
> When the application writes sound samples to the sound card and the
> speaker arrangement has changed, ALSA should use a specific error
> number reported back so the application knows that it should re-check
> the speaker-arrangement API and re-open the sound card with the new
> details.

That sounds like a great solution and hopefully it be considered
when/if steps are taken to fix this problem.  Unfortunately thus far
it seems the preference is to ignore it.  That being the case, will
someone please post a patch that reverses the change that caused this
problem in the first place?

Thanks,
Derek