Bug 530049 - MMX volume adjustment code is broken
Summary: MMX volume adjustment code is broken
Keywords:
Status: CLOSED UPSTREAM
Alias: None
Product: Fedora
Classification: Fedora
Component: pulseaudio
Version: rawhide
Hardware: All
OS: Linux
low
medium
Target Milestone: ---
Assignee: Lennart Poettering
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
: 533001 (view as bug list)
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2009-10-21 10:59 UTC by Michael Schwendt
Modified: 2009-11-23 20:25 UTC (History)
6 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2009-10-31 01:59:57 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
cat /proc/cpuinfo (437 bytes, text/plain)
2009-10-21 15:48 UTC, Michael Schwendt
no flags Details

Description Michael Schwendt 2009-10-21 10:59:09 UTC
$ play somefile.ogg

Then open "Sound Preferences" to alter the "sox" application's volume level. For anything lower than 100%, sound gets distored heavily.

Also reproducible with other audio players.

$ rpm -q pulseaudio alsa-plugins-pulseaudio
pulseaudio-0.9.19-1.fc12.i686
alsa-plugins-pulseaudio-1.0.21-2.fc12.i686

Comment 1 Lennart Poettering 2009-10-21 15:29:23 UTC
Please provide the contents of /proc/cpuinfo!

Comment 2 Michael Schwendt 2009-10-21 15:48:27 UTC
Created attachment 365538 [details]
cat /proc/cpuinfo

Comment 3 Michael Schwendt 2009-10-21 16:06:51 UTC
If I let two audio apps play, distortion goes away.

Comment 4 Michael Schwendt 2009-10-21 16:45:04 UTC
No, I was mistaken. It's only less distortion with two audio apps playing - or sometimes it seems fix itself magically. Still something's wrong with the mixing. When either app finishes, heavy distortion returns most of the time. There's an additional burst of noise (approx. 500 ms to less than 1 s) when "play" starts and stops and the volume of a 2nd app is reduced.

Comment 5 Lennart Poettering 2009-10-23 18:47:07 UTC
OK, this seems to be a problem with the MMX volume adjustment code, which is only enabled when SSE2 is not available.

Comment 6 Lennart Poettering 2009-10-23 18:49:00 UTC
A temporary workaround should be to set PULSE_NO_SIMD=1 when running PA. Could you please verify that?

Comment 7 Michael Schwendt 2009-10-24 10:06:36 UTC
Confirmed.

Comment 8 Lennart Poettering 2009-10-30 04:15:41 UTC
Now fixed upstream, will prepare an update shortly.

Comment 9 Joachim Backes 2009-11-04 17:17:26 UTC
*** Bug 533001 has been marked as a duplicate of this bug. ***


Note You need to log in before you can comment on or make changes to this bug.