Colin Guthrie wrote: > CJ van den Berg wrote: >> I don't know about SDL, but ALSA definitely has the same problem on >> Debian. It seems, from reading the plug-in code, that the ALSA plug-in >> drops the connection on every buffer under run. I don't know what the >> purpose of that is supposed to be. It doesn't make much sense to me. >> >> As a workaround, you can raise the minimum period size in the ALSA >> plug-in. That prevents the stuttering in Skype and Ekiga, but doesn't >> really fix the problem. It only makes it less likely. >> >> The strange thing is though that Skype still creates thousands of >> connections even with a larger period size (and no stuttering), so there >> must be other reasons for the plug-in to drop the connection too, or I'm >> missing something somewhere. Hmm, come to think of it, perhaps I only >> raised the period size for the sink and not the source... > > Thanks for the info CJ, I'll maybe have to learn some slightly in depth > pulse/sound stuff at some point and actually look into the API calls in > SDL and see if I can come up with a little patch for it. The default buffer attributes seem to be a little ill thought out. I've just reverted them to the defaults and it seems to work OK for the test cases I tried although I'm waiting for more feedback from others. If anyone with more knowledge than me cares to work out better defaults I'm happy to try them. Patch: http://svn.mandriva.com/cgi-bin/viewvc.cgi/packages/cooker/SDL12/current/SOURCES/SDL-1.2.13-pulseaudio-default-buffering.patch?view=markup Source to which it applies: http://www.libsdl.org/cgi/viewvc.cgi/branches/SDL-1.2/src/audio/pulse/SDL_pulseaudio.c?view=markup Col
