[FFmpeg-user] Audio decoding and MXFcontainer problem with DNxHD MXF

tim nicholson nichot20 at yahoo.com
Wed May 21 08:14:46 CEST 2014


On 07/05/14 14:25, Christoph Gerstbauer wrote:
> Hello,
> 
> I have a audio decoding/MXFcontainer problems with a MXF Op1a file which
> was encoded by Adobe Premiere (Videocodec DNxHD, AudioCodec pcm_s24le).
> 
> When I open the file with ffprobe it shows me following warning messages:
> 1.) [mxf @ 0000000002a58080] Field dominance 0 support is not implemented
> 2.) [mxf @ 0000000002a58080] OPAtom misinterpreted as OP1a?KLV for edit
> unit 1 extending into next edit unit is not implemented
> 
> When I transcode it (video and audio) I get an additional error at
> decoding the audio:
> 3.) [pcm_s24le @ 00000000003b8d20] Invalid PCM packet, data has size 4
> but at least a size of 24 was expected
> Error while decoding stream #0:1: Invalid data found when processing input
> 
> In the final target file, the audio error is hearable between
> 00:00:05.000 - END OF FILE. Sounds like clipping.
> When I decode the file with FFmbc, there is no problem with the audio
> (but an other warning:
> [dnxhd @ 00000000003893E0] frame size changed: 1920x540 -> 1920x1080)
------------------------------------------------------------------/
normal for progressive
------------------------------------------------------/
normal for interlaced

I wonder if there is mixed interlaced/progressive that is upsetting
things with the frame size change not being detected.

>
> For error 1.) (Field order)
> I checked with mxfdump: there is no KLV field order information (like "[
> k = FieldDominance 32.12, l = 1 (0001) ] 0  01") in the MXF Container
> itself. That would fit. -> "Field dominance 0", but I am not shure if
> ffmpeg reads out this information from the MXF, does it?
[..]

-- 
Tim


More information about the ffmpeg-user mailing list