Hi, I merged more stuff into the CVS HEAD. You can now enjoy all choppy and crappy wav/au/aiff/etc. (libsndfile) and Ogg/Vorbis (libvorbisfile) sound.
Try './utils/pplay xyz.ogg' To play video and seperate audio, look at 'examples/mix_audio.veml' I can probably hack the player to play the sound non-choppy, but the Right Thing is to toss the player into the junk bin and write a half decent player class with a seperate audio thread. Oh, then there is a filter plugin which can do audio resampling (libsamplerate). As soon as it is working, it would be possible to insert the external audio track into the raw DV file. (This is already possible if the external audio has one of the few support sample rates in DV) The other option is to keep the original samplerate and put it into an seperate audio track in the .avi file. Actually I would rather use .ogg instead of .avi, but both should be working in a while. We need to think about how to modify VEML to handle all this. Probably by using as much SMIL as possible. It is straight forward to replace <scenelist> with <seq> which directly maps to the SeqContainer class in piave and <scene> with <par> I guess, but there are some issues with effects. The <par> tag does not directly map to any piave class since it allows many inputs as far as I see it. A piave BNode does only have two tracks and has also different boundary conditions. Of course my definition makes more sense ;-) but since I agree to the Charles' and Dan's arguments, I will propose something that is close to SMIL. That way we can replace my simple parser with Dan's full SMIL timing parser in the future. Do you think you could support the mixing/replacing of audio in kdenlive any time soon, or is this far away? I guess this would require too many changes, doesn't it? Cheers, Rolf BTW, did I mention that pgrab is in the CVS HEAD? You can use it to capture DV from your DV-Cam if it's supported by avc1395. -- contacts: www.physi.uni-heidelberg.de/~dubitzky
