[FFmpeg-user] Setting option video_size / progressive / interlaced

[Carl Eugen Hoyos]

billyx <billyx <at> gmx.at> writes:

> The command line I think should be correct is: 
> ffmpeg.exe -f h264 -r 12 -s 704x576 -i
> ch00000000000001-130930-095853-105853-01p101000000.264 
> -vcodec copy -acodec copy test.mp4

Difficult to answer...

You hopefully should not have to specify "-f h264", 
FFmpeg should auto-detect it, if it fails there may 
be a bug (sample welcome). You cannot specify a 
size for h264 streams, so please remove -s 704x576.
You actually can specify an input frame rate, but 
this is only a (particularly ugly) hack around a 
bug in FFmpeg, I suggest you remove "-r 12" until 
the speed is the only thing wrong about the output 
file.

-acodec copy makes no sense if your input is raw 
h264.
I wonder why you are using -vcodec copy, I suspect 
what you actually want is to reencode the input 
file...

But the real question is:
Why do you get so many errors while decoding?
Is the file broken (because it comes from an 
transmission that is known not to be error- 
free)? Or is this an intentionally broken 
file from a camera (cctv or similar)? In that 
case please provide a sample and please tell 
us about the camera (type).

Re-reading your command line, I realize that this 
might be the stream of a mxf file.
Is FFmpeg unable to read the mxf file? If yes, a 
sample would be very welcome!
(mxf as input should fix the problem you are 
trying to solve with -r 12.)

If all you need is to scale, please use the scale 
filter, if you want merge two fields that are 
encoded after each other, please see the filter 
documentation, one of the interlace filter should 
support it.

Carl Eugen