Problem with Soundblaster Live and Wine

List overview All Threads

newer

older

bug in sound support?

Re: Donations - RFC

Constantin Bergemann

21 Jun 2004 21 Jun '04

7:50 p.m.

Hi,

I have a problem with my Soundblaster Live and Wine. When running Windows Applications, they often set the mixer level for the PCM channel to 100%. However, the Soundblaster Live under Linux tends to sound a bit dirty at these levels, you have to use values under 80%.

Is there anything I can do to tell wine to let this value unchanged? I am using ALSA and the OSS module of wine. Is anyone even aware of this problem?

Thanks, Tony.

Show replies by date

Francois Gouget

22 Jun 22 Jun

6:04 p.m.

Hi,

Constantin Bergemann wrote:

...

Hi,

I have a problem with my Soundblaster Live and Wine. When running Windows Applications, they often set the mixer level for the PCM channel to 100%. However, the Soundblaster Live under Linux tends to sound a bit dirty at these levels, you have to use values under 80%.

Is there anything I can do to tell wine to let this value unchanged? I am using ALSA and the OSS module of wine. Is anyone even aware of this problem?

This is most likely caused by DirectSound. Currently calling DirectSoundCreate() has the side-effect of setting the sound volume to 100%. In fact anything that causes the creating of a primary buffer does that so that some applications repeteadly reset the volume to 100% which is very annoying.

Based on my test it's also wrong. On Windows starting a DirectSound application does not change the volume. One way to verify this is with the DirectSound regression test. Do:

set WINETEST_INTERACTIVE=1 dsound_crosstest.exe dsound

This should play the test tones at the current volume. When the same test is run in Wine it first sets the volume to the max and then plays the tones.

The attached hack should fix this problem but probably not in the right way. I'm not sure of the exact Windows behavior. I believe it just initializes the DirectSound volume to whatever is the current volume. That means we should query the current volume from wineoss and set it that way. We should really add a trace in the dsound test to make it possible to test that theory.

-- Francois Gouget fgouget@codeweavers.com Index: dlls/dsound/primary.c =================================================================== RCS file: /var/cvs/wine/dlls/dsound/primary.c,v retrieving revision 1.23 diff -u -r1.23 primary.c --- a/dlls/dsound/primary.c 13 Feb 2004 20:22:02 -0000 1.23 +++ b/dlls/dsound/primary.c 20 May 2004 14:28:55 -0000 @@ -74,8 +74,6 @@ HRESULT err = DS_OK; TRACE("(%p)\n",This); - DSOUND_RecalcVolPan(&(This->volpan)); - /* are we using waveOut stuff? */ if (!This->driver) { LPBYTE newbuf; @@ -135,11 +133,6 @@ } if ((err == DS_OK) && (merr != DS_OK)) err = merr; - - if (!err) { - DWORD vol = (This->volpan.dwTotalLeftAmpFactor & 0xffff) | (This->volpan.dwTotalRightAmpFactor << 16); - err = mmErr(waveOutSetVolume(This->hwo, vol)); - } } else { if (!This->hwbuf) { err = IDsDriver_CreateSoundBuffer(This->driver,&(This->wfx), @@ -154,7 +147,6 @@ if (dsound->state == STATE_PLAYING) dsound->state = STATE_STARTING; else if (dsound->state == STATE_STOPPING) dsound->state = STATE_STOPPED; } - err = IDsDriverBuffer_SetVolumePan(This->hwbuf, &(This->volpan)); } return err;

7518

Age (days ago)

7519

Last active (days ago)

wine-devel@winehq.org

1 comments

2 participants

tags (0)

participants (2)

Constantin Bergemann
Francois Gouget