Reimar Döffinger <[email protected]> added the comment: On Tue, Jul 27, 2010 at 02:45:06PM +0000, P R wrote: > (gdb) r -threads 1 -f rawvideo -pix_fmt uyvy422 -s 720x576 -i > /var/www/transcodetests/video.yuv -r 25 -f s16le -ac 2 -ar 48000 -i > /var/www/transcodetests/audio.wav -vcodec dvvideo -s 720x576 -pix_fmt yuv420p > -aspect 16:9 -acodec pcm_s16le -ac 2 -ar 48000 > /var/www/transcodetests/muxed.dv > -vcodec libx264 -vpre placebo -b 256k -bt 256k -s 512x288 -aspect 512:288 -ab > 96k -ac 2 -f ipod /var/www/transcodetests/muxed.m4v
Please always provide a minimal test-case. This one contains a lot of stuff completely irrelevant to the issue. Below is enough to reproduce fmpeg_g -f rawvideo -s 720x576 -i /dev/zero test2.avi -s 300x400 test.avi Note that AFAIK we do not have even a single test case in our regression tests that tests encoding to multiple output files. I suspect that as usual the cause is the libavfilter addition. I haven't tested if it works any better using the scale filter explicitly instead of -s > GDB disassembly around PC: > (gdb) disass $pc-32 That's only the disassembly _before_ PC, which is rather pointless. You forgot the $pc+32 part. ________________________________________________ FFmpeg issue tracker <[email protected]> <https://roundup.ffmpeg.org/issue2129> ________________________________________________
