Normal
Seek operations will cause a downstream flush, which should clean out all the old data from the decoder/mixer/renderer (effectively resetting the pipeline), so I wouldn't rule out the video decoder misbehaving (especially as the PDVD decoder seems less 'stupid' than many others ). Also once a new video sample has been created and sent to the decoder by TsReader.ax the original data is discarded - TsReader couldn't loop back several seconds (which is what seems to be happening) even if it wanted to - the data isn't available any more.... Knowing if the audio also loops back should tell us quite a lot. It might be interesting to try using UNC paths instead of RTSP and see if the issue still happens. Tony
Seek operations will cause a downstream flush, which should clean out all the old data from the decoder/mixer/renderer (effectively resetting the pipeline), so I wouldn't rule out the video decoder misbehaving (especially as the PDVD decoder seems less 'stupid' than many others ).
Also once a new video sample has been created and sent to the decoder by TsReader.ax the original data is discarded - TsReader couldn't loop back several seconds (which is what seems to be happening) even if it wanted to - the data isn't available any more....
Knowing if the audio also loops back should tell us quite a lot.
It might be interesting to try using UNC paths instead of RTSP and see if the issue still happens.
Tony