[FFmpeg-devel] checked region version in dvbsubdec.c

Michael Niedermayer michaelni at gmx.at
Mon Apr 28 22:09:18 CEST 2014


On Mon, Apr 28, 2014 at 02:03:04PM +0530, anshul wrote:
> On 04/22/2014 07:09 PM, Michael Niedermayer wrote:
> >On Tue, Apr 15, 2014 at 06:37:38PM +0530, anshul wrote:
> >>Hi
> >>
> >>I referred ETSI EN 300 743 V1.3.1 (2006-11) for dvb subtitle.
> >>
> >>This document they say that ...
> >>   region_version_number: This indicates the version of this region.
> >>    The version number is incremented (modulo 16) if one or more of
> >>     the following conditions is true :
> >>      -> the region_fill_flag is set;
> >>      -> the regions CLUT family has been modified;
> >>      -> the region has a non-zero length object list.
> >>
> >>so I have changed code to not check anything if version number are same.
> >The version could also match when streams are concatenated
> >so i would suggest that more is checked than the 4 bit version
> >
> >[...]
> >
> >
> >
> >_______________________________________________
> >ffmpeg-devel mailing list
> >ffmpeg-devel at ffmpeg.org
> >http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
> 
> Hi
> 
> Sorry for too late reply,
> 
> version of two stream can only match if Page composition segment of
> Second stream
> has not been received yet  and end of display set segment of first
> stream is not received.

why ?


> 
> even if version match that does mean we have not received page
> composition segment
> of second stream yet, so this region segment is invalid if it comes
> before page composition segment.
> 
> Since Region segment to parsed are mention in page composition
> segment, there would be a problem
> if version does not match of second stream and first stream.
> 

> concatenation should be considered as special behavior and all
> segment information of previous
> component should be reset.
> 
> I don't know where to start for that, may be some streams where you
> get problem while concatenation
> could help me.

i dont see anything that would reset things on seeking
theres no fush() in dvbsubdec

so if iam not missing anything then maybe adding such a flush function
would be a start


[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

The bravest are surely those who have the clearest vision
of what is before them, glory and danger alike, and yet
notwithstanding go out to meet it. -- Thucydides
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20140428/0c59bc6d/attachment.asc>


More information about the ffmpeg-devel mailing list