[alsa-devel] HDA: Problem with sound becoming corrupted

[Colin Guthrie]

'Twas brillig, and Jean-Yves Avenard at 10/03/11 02:41 did gyre and gimble:
> Hi
> 
> On 10 March 2011 02:10, Colin Guthrie <gmane at colin.guthr.ie> wrote:
>> The problem is that the sound suddenly goes nuts and starts sounding
>> awful. It's like some buffer pointer is just a little bit out for some
>> reason.
>>
>> I can fix the issue by closing and reopening the PCM device (as I use
>> PA, this is just a simple "pasuspender echo" away).
> 
> 
>>
>> The problem is very easy to reproduce - it happens about 20+ times a day
>> to me - to the extent that I have added a panel launcher with the above fix.
> 
> I have to raise my hand on this one too... It has puzzled me for a few
> months now.
> I had thought it was related to the new video/audio card I'm using a
> GT220 , but this is about the same time I upgraded alsa.
> 
> Every once in a while, the audio becomes slightly distorted. Voices
> sounds like they are gargling with mouthwash while talking.
> 
> Interestingly, it sounds the same for AC3 or DTS passthrough, it's
> slightly corrupted, but not enough to break the AC3 stream from being
> usable by the amp.
> 
> However, closing the PCM device and re-opening it doesn't fix the
> issue for me ; I need to reboot.
> 
> This is seen with MythTV using ALSA directly (I wrote the audio code in MythTV)
> 
> Few MythTV users have reported it, I can't explain what is causing it..

[Resend as I used the wrong email address the first time! Things have
changed since the first one tho' Jean-Yves!]

Sounds like it could be related indeed.

I've been trying an older kernel today: 2.6.36.2 with:
 Advanced Linux Sound Architecture Driver Version 1.0.23.

It worked well initially, but sadly I did get two occurrences of the
problem so far (both specifically triggered rather than happening
organically). This seems to tie in with my feeling that the problem
started off as a nuisance but has gotten significantly more problematic
with more recent kernels....

With a more recent kernel I can trigger it just changing the volume
rapidly (which triggers a "bloop" noise). With this older kernel I still
used this to trigger the problems, but it takes a while (it feels to
take a lot longer than the newer kernel). I cannot discount this being a
subjective comparison however.

I'll look further.

Col

-- 

Colin Guthrie
gmane(at)colin.guthr.ie
http://colin.guthr.ie/

Day Job:
  Tribalogic Limited [http://www.tribalogic.net/]
Open Source:
  Mageia Contributor [http://www.mageia.org/]
  PulseAudio Hacker [http://www.pulseaudio.org/]
  Trac Hacker [http://trac.edgewall.org/]