kernel/5137: Strange soundbug on SiS 7012 cards

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

kernel/5137: Strange soundbug on SiS 7012 cards

Jonathan Schleifer
>Number:         5137
>Category:       kernel
>Synopsis:       Sound is sometimes strange, occours on all SiS 7012
>Confidential:   yes
>Severity:       serious
>Priority:       medium
>Responsible:    bugs
>State:          open
>Quarter:        
>Keywords:      
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sun May 28 13:40:01 GMT 2006
>Closed-Date:
>Last-Modified:
>Originator:     Jonathan Schleifer
>Release:        All from 3.5 (maybe even before) up to -current
>Organization:
net
>Environment:
       
        System      : OpenBSD 3.9
        Architecture: OpenBSD.i386
        Machine     : i386
>Description:
        It's very difficult to describe this problem, but I'll try my
        best. When I play some audiofiles, for example in XMMS, they
        sound really strange, just like if there's a buffer underrun.
        But this bug doesn't always occour. If it occours, seeking in
        the current track fixes it in XMMS. But after a few tracks, I
        have the same bug again. It's the same if I don't have the bug
        and seek a few times, then suddenly the bug appears sometimes. I
        already had a look at the code, but didn't see something
        obvious that could be the reason. It's the same with mplayer:
        If I use the sun audio plugin, it's just like in XMMS. Seeking
        fixes it if it's broken and sometimes breaks it if it's
        working. Using the SDL audio plugin, I can seek as long as I
        want if the playback started without the bug. But if the
        playback started with the bug, I can seek as long as I want and
        the bug stays. So I guess that it has something to do with the
        initialization of the device, because XMMS and the sun audio
        plugin for mplayer reopen the device for seeking, while the SDL
        audio plugin doesn't. It's a chance about 15% that the bug
        occours when an application opens the device, so it's really
        annoying and making audio unusable for me. I tested this on 2
        machines, on my laptop and my desktop. They both have a SiS
        7012, but the chip differs a bit: The one in the laptop needs
        to be initialized a bit different than the one in the desktop
        and the one in the desktop has no volume for PCM (only 0 and
        100). That's the only difference as far as I know, so it aren't
        exactly the same chips. The bug occours on both, so it seems to
        be driver related. Additionally, everything gets played too
        fast on the desktop. It seems like nothing gets converted from
        44100 Hz to 48000 Hz on the desktop, while on the laptop
        everything gets converted.
>How-To-Repeat:
        Open the audio device with your favourite application and do
        some playback.
>Fix:
        I don't know, but it could be a problem with a buffer. I
        already looked at the code, but didn't see anything.


>Release-Note:
>Audit-Trail:
>Unformatted:
 >cards I own Severity: serious

Reply | Threaded
Open this post in threaded view
|

Re: kernel/5137: Strange soundbug on SiS 7012 cards

Jonathan Schleifer
Jonathan Schleifer <[hidden email]> wrote:

> >Release:        All from 3.5 (maybe even before) up to -current

I just updated to the newest snapshot today, the problem is still
there. It's very annoying if you have a playlist in XMMS and every few
songs you have to restart the current song / wind in the song to have
the sound normal again.
Is there nobody else using a SiS 7012? Does nobody have any idea what
it could be?

--
Jonathan

Reply | Threaded
Open this post in threaded view
|

Re: kernel/5137: Strange soundbug on SiS 7012 cards

Can Erkin Acar
On Thu, Jul 13, 2006 at 07:43:22PM +0200, Jonathan Schleifer wrote:

> Jonathan Schleifer <[hidden email]> wrote:
>
> > >Release:        All from 3.5 (maybe even before) up to -current
>
> I just updated to the newest snapshot today, the problem is still
> there. It's very annoying if you have a playlist in XMMS and every few
> songs you have to restart the current song / wind in the song to have
> the sound normal again.
> Is there nobody else using a SiS 7012? Does nobody have any idea what
> it could be?

Just for the record, I have the same problem with a machine here (amd64)
Will try to look into it.

Reply | Threaded
Open this post in threaded view
|

Re: kernel/5137: Strange soundbug on SiS 7012 cards

Jonathan Schleifer
In reply to this post by Jonathan Schleifer
Now that's really funny:
I'm an audacious developer and ported it to OpenBSD. While I was at it,
I got the OSS audio plugin running on OpenBSD. And guess what happened:
With the OSS audio plugin, I had no sound bugs at all. I'll compile
mplayer with OSS support now and test that. I guess mplayer will work
without that soundbug then, too.

So it seems like it's a bug with the sun audio system + the auich
driver. I don't know much about the sun audio system, so when someone
has an idea how to fix it, I would really appreciate it.

Anyway, why do all apps come WITHOUT an OSS plugin when installed from
ports? That XMMS comes without OSS is a thing I can fully understand,
since it's OSS plugin is horribly broken - that broken, that it won't
run on OpenBSD at all (Audacious had the same plugin, but I fixed it).
But why all the other apps, like mplayer, too?

--
Jonathan