Cool. Seems like we're getting closer to figuring this thing out...
Two more tests to try: 1. Try with the dummy video driver to see if there are still scratchy noises. import os os.environ['SDL_VIDEODRIVER'] = 'dummy' This will tell us if a video driver is having an effect. 2. try pygame program with a 4096 sized buffer. This will let us know that a 4096 sized buffer magically works better :) cu, On Wed, Mar 26, 2008 at 12:45 PM, Brian Fisher <[EMAIL PROTECTED]> wrote: > the fact that playmus SDL_Init(SDL_INIT_AUDIO) is interesting - I > wonder if having a window or not affects whether playmus can use the > dsound backend? How would one figure out what audio backend SDL chose? > > > > On Tue, Mar 25, 2008 at 5:06 PM, Lenard Lindstrom <[EMAIL PROTECTED]> wrote: > > Two things I noted with playmus. First it initializes SDL with > > SDL_Init(SDL_INIT_AUDIO). Second it starts playback with > > Mix_FadeInMusic(music,looping,2000). It lacks a feature to start > > playback in the middle of a song unfortunately. > > > > > > Lenard > > > > > > > > > > René Dudfield wrote: > > > Hello, > > > > > > Yeah, it would be very useful if you could try the playmus program to > > > try and reproduce your problem. This will help with a bug report to > > > SDL. > > > > > > It would also help if we could summarise all your testing to make a > > > better SDL bug report. eg. > > > I tried this, then this, then this, then this, etc. > > > > > > > > > > > > Note that Bo is using windows XP. > > > > > > I don't think it's a bug in directx, but a change in the way threads > > > work from win98 to winxp - exposing a bug in the sdl code. > > > > > > That sound code in SDL was mostly written in the time of win9x, and > win2k. > > > > > > I think the change in the video driver has also changed the behaviour > > > of the thread code for the sound. Since using windib for video > > > changes the amount of cpu time used in the sound threads and also the > > > properties of the video thread. > > > > > > > > > > > >