I don't really have any suggestions for improvement, other than further refinement of the lipsync algorithms. Expanded audio support (compressed audio files, MP3, etc.) would be nice, but it isn't a big issue. I can always convert. It's better for the audio to be uncompressed until the final cut, anyway.