[FFmpeg-user] Avid DNxHD QTs dropped/doubled frame problem

[Jesse]

Hi, all.  I've run in to an issue when importing DNxHD Quicktime 
movies made with ffmpeg in to Avid Media Composer.
The source material are image sequences rendered at 24 FPS.

The output format is DNxHD 36 Mb/s, 23.976 FPS.

I am doing my encoding with ffmpeg SVN-r20372 packaged with Fedora 
Core 12.

All output movies from ffmpeg play back with the correct number of 
frames in mplayer on Linux, and in Quicktime on a Mac.

However, when the movies are imported in to an Avid Media Composer 
project set to DNxHD36, 23.976 FPS, movies with frame lengths ending 
in 0, 3, and 6 (e.g.  10, 13, 16, 20, 23, ...), have a doubled first 
frame, and a missing last frame.

All DNxHD36 Quicktimes created with Quicktime on a Mac maintain the 
correct order and number of frames when imported in to the Avid, which 
is why I think my problem originates with ffmpeg.


This is an example of my ffmpeg command on Linux:

ffmpeg -f image2 -r 24 -i ./1-20/test.%04d.tif -vcodec dnxhd -b 36Mb -r 23.976 -s 1920x1080 ./test.1-20.mov


This is the complete terminal output of the run.

FFmpeg version SVN-r20372, Copyright (c) 2000-2009 Fabrice Bellard, et al.
  built on Nov  7 2009 10:57:27 with gcc 4.4.2 20091027 (Red Hat 4.4.2-7)
  configuration: --prefix=/usr --bindir=/usr/bin --datadir=/usr/share/ffmpeg --incdir=/usr/include/ffmpeg --libdir=/usr/lib64 --mandir=/usr/share/man --arch=x86_64 --extra-cflags='-O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector --param=ssp-buffer-size=4 -m64 -mtune=generic' --extra-version=rpmfusion --enable-bzlib --enable-libdc1394 --enable-libdirac --enable-libfaad --enable-libgsm --enable-libmp3lame --enable-libopenjpeg --enable-libschroedinger --enable-libspeex --enable-libtheora --enable-libvorbis --enable-libx264 --enable-libxvid --enable-x11grab --enable-avfilter --enable-avfilter-lavf --enable-postproc --enable-pthreads --disable-static --enable-shared --enable-gpl --disable-debug --disable-stripping --shlibdir=/usr/lib64 --enable-runtime-cpudetect
  libavutil     50. 3. 0 / 50. 3. 0
  libavcodec    52.37. 1 / 52.37. 1
  libavformat   52.39. 2 / 52.39. 2
  libavdevice   52. 2. 0 / 52. 2. 0
  libavfilter    1. 4. 1 /  1. 4. 1
  libswscale     0. 7. 1 /  0. 7. 1
  libpostproc   51. 2. 0 / 51. 2. 0
[image2 @ 0x9735e0]MAX_READ_SIZE:5000000 reached
Input #0, image2, from '1-20/test.%04d.tif':
  Duration: 00:00:00.83, start: 0.000000, bitrate: N/A
    Stream #0.0: Video: tiff, rgb24, 1920x1080, 24 tbr, 24 tbn, 24 tbc
Output #0, mov, to './test.1-20.mov':
    Stream #0.0: Video: dnxhd, yuv422p, 1920x1080, q=2-31, 36000 kb/s, 2997 tbn, 23.98 tbc
Stream mapping:
  Stream #0.0 -> #0.0
Press [q] to stop encoding
frame=   20 fps=  8 q=1.0 Lsize=    3681kB time=0.83 bitrate=36148.6kbits/s    
video:3680kB audio:0kB global headers:0kB muxing overhead 0.024573%


Does anyone see anything conspicuously wrong with my command line that 
would cause the behavior I described above?   Is it normal to see 
"2997 tbn, 23.98 tbc" when writing DNxHD Quicktimes?

Thanks for reading.

-Jesse