Pulse audio

[Daniel Chen]

On Thu, Oct 8, 2009 at 4:23 PM, Martin Olsson <mnemo at minimum.se> wrote:
> Sound was broken for me in all releases before hardy and then
> in hardy it worked _perfectly_ with Skype, Flash etc. Sound

Again, just because it worked for you does not mean that it wasn't
broken. If you care to look outside your hardware and your
configuration, you'll see it clearly.

> notice; ALSA did everything I ever wanted from it. Anyway,
> then came jaunty and it was broken again. Now in karmic
> alpha I got audio back but I got these extremely load sparks
> and cracks which give me a really poor audio experience.

I'm not surprised that ALSA "did everything [you] ever wanted from
it." However, it does not change that there were latent bugs. Clearly
they existed despite ALSA working for you.

As for Karmic, are you using the ubuntu-audio-dev PPA? If not, you should.

> By now it's obvious that karmic as well with line up along
> with the releases that did not reach back up to the level
> where ALSA was for me. I honestly wonder, when will it stop?

Pretty clearly, you can continue to use Hardy. When you choose to test
a development release that becomes a stable release, you choose to
test an entirely different stack.

Developments are quick in the audio world. You may not follow the git
commits and the breakages - and you shouldn't be required to - but
just because things appear to continue to be broken doesn't mean no
one cares or no one is working to fix the regressions.

> I have a _lot_ of respect for the work that the Ubuntu audio
> team (and Lennart) is doing but the TB decision to accept
> PA into Ubuntu was a _BIG_ mistake. The appropriate action
> would have been to talk some sense into upstream. If you

How do you intend to "talk some sense into upstream"? Upstream,
presuming you mean PulseAudio, is _one_ project. Its success in any
distribution depends on perfect alignment of the layers beneath it:
linux, alsa-lib(, and to some extent, alsa-plugins). Every current
desktop Linux distribution ships a different combination of stack
components. Drilling down, even the linux configurations are
different. Even the compiler flags are different. A more persuasive
test is to take the precise Fedora 12 configuration of PulseAudio and
demonstrate that it works remarkably better in Ubuntu Karmic than
Ubuntu Karmic's.

In other words, you need to maintain the precise configuration across
the board before you can really say something is broken and thus needs
to be beaten into upstream.

> I was very glad that Canonical posted a job listing for
> "Desktop Architect – Sound Experience" recently, clearly
> someone is noticing this and pulling the right strings.

My understanding is that it is more a UI position.

> I do definitely think following upstream is the only sensible
> thing to do but not to follow them into an 18 month walk in the
> valley of death without water.

I don't think the situation is nearly as bleak as you paint it here,
but I caution you to consider the myriad hardware combinations that
wreak havoc on the default PA configuration.

In other words, it may suck for "you", but it sucks a whole lot worse
for people in the trenches, because there are thousands of "you" with
"your" craptastic hardware, and the people in the trenches have to
balance thousands of configurations.

-Dan