[Bug 28723] Sound stutter in Rage when emulated windows version is set to "Windows 7" (XAudio2 -> mmdevapi sound output path)

[wine-bugs at winehq.org]

http://bugs.winehq.org/show_bug.cgi?id=28723

--- Comment #16 from Andrew Eikum <aeikum at codeweavers.com> 2011-11-02 08:47:38 CDT ---
Created attachment 37257
  --> http://bugs.winehq.org/attachment.cgi?id=37257
winealsa.drv: Don't set ALSA period time

Thanks for the nice logs and updates, Alexey. Sorry it's taken me so long to
get to this bug.

I'll try to describe what I think is happening based on the log in comment 3.
RAGE's audio thread runs in a loop polling IAudioClient::GetCurrentPadding().
When the current padding value is <= 441, it writes another packet (441 frames)
of data. But, I think ALSA's period size is being set to something about the
same size, so when GCP() is called, its value changes in chunks of
approximately 441 frames. So you see patterns like this (trimmed for clarity):

007d:trace:alsa:AudioClient_GetCurrentPadding PCM state: 3, avail: 15927, pad:
441
007d:trace:alsa:AudioRenderClient_GetBuffer (0x48a9e70)->(441, 0x352de924)
007d:trace:alsa:AudioClient_GetCurrentPadding PCM state: 3, avail: 15927, pad:
441
007d:trace:alsa:AudioRenderClient_ReleaseBuffer (0x48a9e70)->(441, 0)
007d:trace:alsa:AudioClient_GetCurrentPadding PCM state: 3, avail: 15927, pad:
882
007d:trace:alsa:AudioClient_GetCurrentPadding PCM state: 3, avail: 15927, pad:
882
007d:trace:alsa:AudioClient_GetCurrentPadding PCM state: 3, avail: 15905, pad:
463
007d:trace:alsa:AudioClient_GetCurrentPadding PCM state: 3, avail: 16281, pad:
87
007d:trace:alsa:AudioRenderClient_GetBuffer (0x48a9e70)->(441, 0x352de924)
007d:trace:alsa:AudioClient_GetCurrentPadding PCM state: 4, avail: -32, pad: 0
007d:trace:alsa:AudioRenderClient_ReleaseBuffer (0x48a9e70)->(441, 0)

Notice the large skips in GCP() return values. When RAGE finally sees the <=441
value, ALSA will have run out of data by the time the GetBuffer() call
completes.

I'm attaching a patch here which I hope fixes this. It just doesn't set the
ALSA period time at all, letting ALSA figure out what it wants to use. I'm not
sure about this in combination with fixes for Bug 28282, but it would be useful
to know if this patch works for you.

-- 
Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email
Do not reply to this email, post in Bugzilla using the
above URL to reply.
------- You are receiving this mail because: -------
You are watching all bug changes.