On Thu, Oct 13, 2016 at 04:41:00PM +0200, Ståle Kristoffersen wrote:
> On 2016-10-12 at 13:57, Carl Eugen Hoyos wrote:
> > 2016-10-11 16:27 GMT+02:00 Michael Niedermayer <mich...@niedermayer.cc>:
> > 
> > > also the img2 demuxer for mjpeg2000 from img2dec.c does not work
> > > for this ?
> > 
> > We separated the jpg from the mjpeg probe, jpg makes sure that the
> > sample starts like a jpg image and contains an end marker and can
> > be decoded, mjpeg probe is less strict but requires more than one
> > frame.
> > Same could be argued for j2k or was our reasoning bad?
> Since nobody chimed in on this, I have created two new patches, one
> containing a separate mjpeg 2000 probe
> (0001-libavf-Auto-detect-mjpeg-2000-in-mpeg-ts2.patch), and one that
> incorporates it into the img2dec jpeg 2000 probe
> (0001-libavf-Make-jpeg-2000-probe-work-for-mjpeg-2000.patch).

> I think it is best to have them separated, as that matches the mjpeg
> behaviour.

have you tested the demuxer ?
i dont think it works outside probing for mpegts
i think for j2k codestream support a jpeg2000 parser is needed first
same as we have for mjpeg

that said i have no preferrance for which way its done but i think
were not all aware that adding a identical mjpeg style demuxer was
only part of the stuff needed ...


Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

No human being will ever know the Truth, for even if they happen to say it
by chance, they would not even known they had done so. -- Xenophanes

Attachment: signature.asc
Description: Digital signature

ffmpeg-devel mailing list

Reply via email to