andrey@beyond.cs.caltech.edu (Andre T. Yew) (08/31/90)
Thanks to all of the people who responded. Most suggestions for flushing the pipeline were generally the same -- give the pipeline something to do which on completion would guarantee that the pipeline is empty. Suggestions ranged from using gsync(), swapbuffer(), finish() ... to readrect() and feedback. I ended up calling finish() first then gsync() (call me paranoid :-). Apparently, that helped my problem, but did not fully solve it. That is, I'm still experiencing some jitter in my recordings periodically and in the same locations, and I doubt it's the recorder's fault. I don't suppose the 'Pipeline has any round-off problems? Oh well, back to the drawing board. Thanks again, everyone. -- Andre Yew andrey@through.cs.caltech.edu (131.215.128.1)