>>> As far as I'm aware, gstreamer sits higher in the
>>> stack than PulseAudio, so GNOME distros that use PulseAudio
>>> still have to use gstreamer as well.
>>
>> True, but Gstremer can also directly control audio elements. It seems
>> abundantly clear Pulse Audio is being developed only for ALSA and
>> Linux and the developer has no intentions to develop for anything else.
>> With a few improvements in GStreamer's mixer panel and audio device
>> configurator, you can provide a better usability experience for
>> Solaris than Pulse Audio on ALSA.
>
> More importantly, I think Apple and Microsoft both have demonstrated the
> importance of having a well-integrated, supported, and high-performance
> audio solution as part of the operating system instead of relying on
> third-parties to add it.

Doesn't matter whether it's a third party or not in this case, IMO. It's 
how it works in the open source world, I guess. The deal with ALSA and 
PulseAudio is the same as with audio in the newer Windows releases. For 
some reason, some folks figured they needed to hammer an userland mixer 
between the applications and the kernel.

In the Windows case, it's the userland audio stack hammered between the 
applications and Kernel Streaming. KS worked perfectly fine, but only 
represented as much outputs as the hardware offers (usually just one, 
cards like the SBLive and Audigy had multiple KS outputs that were mixed 
down in hardware). The same deal as with ALSA, IIRC. Windows had a 
kernel mixer on top of KS, which was replaced by userland audio for 
"stability and quality reasons" in Vista and beyond. Not sure what 
that's supposed to mean, since I had yet to see a box go down due to 
kmixer. Apparently just improving the kmixer wasn't an option.

Anyway, the issue I have with PulseAudio is that it's not as portable as 
they claim, as someone else said, just a Linux solution. Practically, 
just another hack to fix ALSA, as the creator dismisses anything other 
as Linux as crap, anyway.

Network audio is a niche gimmick that introduces unnecessary complexity. 
Any other additional features, like the directional audio thing (quasi 
just audio effects), should be implemented in userland an optional 
interface library (something like liboss.so) since they're on a per-app 
basis usually.

Never understood how Pulseaudio gained popularity so quickly anyway. I 
just remember how the guy was working at Fedora, pushed it out in a 
release and suddenly everyone was picking it up for no good reason (IMO).

-mg

Reply via email to