[FFmpeg-devel] avformat/mov: improve qt metadata reading and writing

[David Murmann]

Hi all,

this has been brought up before, the MOV muxer/demuxer currently does not support all
possible datatypes that are allowed in keys/mdta style metadata. This is specified here:
https://developer.apple.com/library/mac/documentation/QuickTime/QTFF/Metadata/Metadata.html#//apple_ref/doc/uid/TP40000939-CH1-SW35

At the moment only data types 1 (UTF-8) and 23 (Float32) work. I have a patch that adds
support for 21 and 22, as I have found those in actual files produced by ARRI cameras.

I have a second patch that adds a flag to the muxer to output keys/mdta style metadata,
to be able to keep the metadata while transcoding camera files.

This patch tries to do something similar, but is mine is shorter (though not as complete)
so hopefully easy to review:
https://lists.ffmpeg.org/pipermail/ffmpeg-devel/2015-September/178347.html
(I actually found this only after writing my patches.)

Patchset is in the next mails.

till then, David.