First of all, a big thank you for your interest and efforts in improving cinelerra !
I could build your patched version of cinelerra after applying two other patches from my distribution (Archlinux), which is using gcc-4.5 and libpng-1.4 : http://repos.archlinux.org/wsvn/community/cinelerra-cv/trunk/cinelerra-gcc45.patch http://repos.archlinux.org/wsvn/community/cinelerra-cv/trunk/libpng14.patch The build was successful with included ffmpeg but also with a recent external ffmpeg version (SVN-r23619). The program built with included ffmpeg complained about missing shared libraries. I got it running after : # cd /usr/lib # ln -s libavcodec.so.52 libavcodec-cinelerra.so.52 # ln -s libswscale.so.0 libswscale-cinelerra.so.0 # ln -s libavutil.so.50 libavutil-cinelerra.so.50 # ln -s libavformat.so.52 libavformat-cinelerra.so.52 (But I don't know if it's a good idea.) I tried both versions (included ffmpeg / external ffmpeg) with mpeg2 and AVCHD footage from my camcorder. With mpeg2 no problem, and I had also partial success loading and editing directly AVCHD files (the "original" cinelerra-cv crashed when I tried to load them, so your patched version is much better) : - cinelerra can now open and edit AVCHD files, - but some frames at the beginning and at the end of these clips are not displayed correctly in the viewer/compositor, and also not rendered correctly. It seems that cinelerra with included ffmpeg processes AVCHD slightly better than with external ffmpeg. In both cases the terminal is full of messages like : [swscaler @ 0x7f3657abd590]No accelerated colorspace conversion found from yuv420p to rgb24. [h264 @ 0x7f3625ce6c40]number of reference frames exceeds max (probably corrupt input), discarding one [h264 @ 0x7f3625ce6c40]Cannot parallelize deblocking type 1, decoding such frames in sequential order With external ffmpeg, the first and last frames of a clip are black or look like video noise when they're rendered. With internal ffmpeg, same problem but more frames are displayed/rendered properly. I know AVCHD is not a good format for editing, so problems with it are not really relevant. I just report them as testing results. Regards, Julien ----- Mail Original ----- De: "Monty Montgomery" <[email protected]> À: [email protected] Envoyé: Vendredi 16 Juillet 2010 11h48:50 GMT +01:00 Amsterdam / Berlin / Berne / Rome / Stockholm / Vienne Objet: [CinCV] Updates: ongoing functionality work, personal git repo I've been conyinuing debugging work on the new ffmpeg-based loader, and have branched out into other fixes starting with the audio backend (eg, the latency calculation in the ALSA backend was broken, etc...). I've also needed to fix a few bugs in ffmpeg itself. In order to disseminate the patches a bit more seamlessly, I've put up a copy of my working GIT repository for now: http://git.xiph.org/?p=users/xiphmont/cinelerraCV.git;a=summary The goal is to continue debugging and see this rolled into mainline, so this is not intended to be a permanent source. But it will be around for a while. I've also stashed the individual patches to date at: http://people.xiph.org/~xiphmont/cinelerra/patches/ Monty _______________________________________________ Cinelerra mailing list [email protected] https://init.linpro.no/mailman/skolelinux.no/listinfo/cinelerra _______________________________________________ Cinelerra mailing list [email protected] https://init.linpro.no/mailman/skolelinux.no/listinfo/cinelerra
