The Great Pulseaudio Mixer Debate: a modest (productive) proposal

Lennart Poettering mzerqung at
Sun Apr 26 15:15:17 UTC 2009

On Sun, 26.04.09 11:46, Kevin Kofler (kevin.kofler at wrote:

> Lennart Poettering wrote:
> > Oh great. This sounds like an invitation to stop working on cleaning
> > up the volume control situation entirely. If we never can get rid of
> > the old cruft and need to prominently feature it in all future release
> > then why even try?
> I think that's the whole point here. We can NEVER get rid of those sliders,
> they all exist for a reason. (Well, almost – I'm sure there's one or the
> other bogus dummy control in some driver, but those are ALSA bugs. Showing
> sliders which actually do something is NOT a bug.)

Uh. We certainly can get rid of them. Many of the more modern HDA
designs got rid of the entire volume control logic and instead
increased the PCM bit depth to 24 bit everywhere of which the usual 4-8
bit headroom can be used for volume control. In ALSA these are all
those devices where the 'softvol' module does volume control.

Gazillions of controls are *not* necessary. Only if you buy a Creative
sound card these days you still get such an immsense number of those
controls. But that's simply because otherwise there is nothing that
would distuingish them from onboard audio. All machines today come
with onboard sound cards whith quality nobody could ever complain

Today we have MMX and SSE and similar CPU extensions. The very reason
we have them is to do signal processing with them. Such as
implementing mixing, volume adjustments, equalizers, and other filters
in them -- in high digital quality. Ironically even Creative sees that
and nowadays a lot of logic is actually in their Windows drivers, not
so much in their sound cards.

The only controls that are really necessary are NOT those which
control signal processing but those which control routing. And of
those the ones which make sense already are (like the spdif controls)
or will be (like input switching) exposed by PA. 

So, you said 'they exist for a reason'. Yes and that reason often has
to do more with marketing than with technical reasons. And since PA is
not a marketing tool I won't buy into those.


Lennart Poettering                        Red Hat, Inc.
lennart [at] poettering [dot] net         ICQ# 11060553           GnuPG 0x1A015CC4

More information about the fedora-devel-list mailing list