ubuntu-audio-dev team mailing list archive

["David Henningsson" <launchpad.web@xxxxxxxxxxxxxx>]

> On Fri, Jan 22, 2010 at 4:04 PM, David Henningsson
> <launchpad.web@xxxxxxxxxxxxxx> wrote:
>> * Alsa-plugins-pulse have been buggy for a long time. I've traced down
>> one important bug (LP: #485488) but there's still the startup race /
>> obsolete underrun thing to fix (LP: #464008). I've asked twice on the
>> pulse mailinglist for comments/solutions but haven't got an answer, so
>> probably they're as unsure as I am of the best way to fix/workaround
>> it...
>
> We should test Jaroslav's duplicate-wakeups fix first.

I just don't see how they are related, IIRC that fix was only related to
hw devices...? And now, I'm trying to look it up, but now I can't find
that recent commit in alsa-lib either, has it been rebased somehow?

> I'll devote some cycles to the latter bug report this coming week.

Okay, see
https://bugs.launchpad.net/ubuntu/+source/alsa-plugins/+bug/464008/comments/4
for some thoughts about it.

>> * I'm curiously observing the Jack-in-main process. Will Jack reach main
>> in time to be properly tested with Lucid? There are quite a few
>> applications in main (not only pulseaudio) with a Jack backend, will
>> they be recompiled and tested with Jack support in time for Lucid?
> The jack-in-main bug report has been filed (#510481); Luke and I have
> both worked on and reviewed it, so now it's awaiting approval from
> Martin.
>
> Xine and PulseAudio at least will gain this backend immediately for
> testing.

Looks good, I think there was at least one more, but I can't remember
which one...

>> * The logout sound. Still not working for me. It's not that having a
>> logout sound is a very important feature to me, it's more like this
>> long-standing bug is saying to us that we can't even get the simple
>> things right... (Even though fixing it turns out to be harder than it
>> seems at first glance.)
>
> There are reports that the symptom can't be reproduced with a new
> user. I've never been able to reproduce it with a fresh user; YMMV.

Even if it isn't reproducable with a fresh user, there is still the issue
of figuring out how a user becomes "unfresh".

>> * Another process locking the sound card: Daniel has sorted out the
>> slmodemd problem. Timidity is still unsolved; I got a report that the
>> solution suggested does not work. It has been discussed at the
>> pulseaudio mailinglist but no good solution has come up.
>
> Should we be considering a timidity->fluidsynth migration, then?

Heh, good question. Since I'm a part of the developer team upstream my
view might be a bit biased, but it also means we might get things fixed
quicker. :-) A few points to consider:

* FluidSynth renders sf2 files better than timidity, but OTOH it does not
read .pat (aka freepats) at all. Hopefully that isn't such an issue and
after all, fluid-soundfont sounds better than freepats anyway.

* There is no FluidSynth switch do make it become a daemon currently, and
we also need to make startup scripts as outlined here:
http://fluidsynth.resonance.org/trac/ticket/66

* http://fluidsynth.resonance.org/trac/ticket/67 should be a nice-to-have
in this use case, but then AFAIK Timidity has the same problem so
switching shouldn't change things for the worse.

>> * Advanced mixer (volume control): A lot of people still need alsamixer
>> and that is not going away for Lucid, I'm afraid. Gnome's switch to
>> pulseaudio causes a regression in that regard. I wish there was an
>> "advanced" button in g-v-c which brought up the original ALSA mixer.
> I don't know if an extra button in the GUI will do much; there are
> certainly UI arguments either way. Perhaps we should ask the Ubuntu
> Desktop team to evaluate reenabling the non-pulse mixer perspective in
> a separate binary package for 10.04.

If people install that package, will they lose the pulse-g-v-c or will
they have both (latter preferred)?

>> * Cards needing non-free firmware: An interesting one I added recently.
>> I've been talking to the maintainer of linux-firmware and hopefully most
>> of alsa-firmware will go into linux-firmware soon (some of it is already
>> there). Since Ubuntu swallows linux-firmware without much hesitation,
>> that'll reduce the need for alsa-firmware.
>
> We need to be clear that the alsa-source binary package (used with
> module-assistant) generated by alsa-driver has non-Free drivers
> stripped:
>   * Removed because of DFSG violations:
>     - alsa-kernel/pci/cs46xx/cs46xx_image.h
>     - alsa-kernel/pci/cs46xx/imgs
>     - alsa-kernel/pci/korg1212/korg1212-firmware.h
>     - alsa-kernel/pci/maestro3.c
>     - alsa-kernel/pci/ymfpci/ymfpci_image.h

And interesting enough, those four are already in linux-firmware. And they
didn't become more free by moving there, yet linux-firmware is in Ubuntu's
main. And then there are seven firmwares in alsa-firmware that seem free
enough to me, but they're banished from Ubuntu. Makes sense? Not to me.

>> I think it works pretty well with the apport-collect information, it
>> gives enough information to give a diagnosis in many cases. However, we
>> also have the answers on launchpad which don't get the attention they
>> deserve. We need the apport-collect information there as well, and I
>> also think we should set ubuntu audio team as answer contacts for the
>> packages it receives bug notifications for. What do you think?
>
> I think the ubuntu-audio-dev team may be more appropriate, as it's a
> more focused group of individuals.

The ubuntu-audio team is currently mostly a bug mail subscription, but I
think that the same people that would be interested in having audio bug
mail would also be interested in answering audio questions as well. Also,
there are more than the four of us that know how to change the output
profile (or whatever) in g-v-c.

// David