[FFmpeg-devel] [PATCH] Set padding_bug_score to zero at the start of each frame
michaelni at gmx.at
Tue Mar 4 03:03:06 CET 2014
On Mon, Mar 03, 2014 at 04:36:37PM -0500, gregory.wolfe at kodakalaris.com wrote:
> Greetings! At Kodak Alaris (we're the Kodak photo kiosk people) we are
> using FFmpeg to extract frames from customer video files so that they can
> be included in print products (e.g., photo books, collages, calendars).
> I have just completed updating the version of FFmpeg we are using from
> ffmpeg-0.5 (circa 2009) to the latest development branch. This is the
> of two changes I've made. I have read and attempted to follow all the
> recommendations for submitting patches, but please advise if I missed
> The full patch is in two patch files (sorry) because I'm somewhat new at
> using git (we use ClearCase). I've also attached before/after command
> line output from ffmpeg. A test video file that can be used to generate
> the before/after output, H263PaddingBugFixVID_6647.MOV, has been
> uploaded to the FFmpeg FTP server.
> FYI, I am using the MinGW32 build environment on Windows 7.
> Patched file: libavcodec/h263dec.c
> Description (from commit message):
> Set padding_bug_score to zero at the start of each frame.
> If the padding_bug_score is not zeroed out at the start of each frame,
> negative increments can accumulate. This can result in the padding bug
> fix-up code not running, and you get an error condition from
> decode_slice() with the message "slice end not reached but screenspace
> end ...".
very nice bug report
issue fixed differently
(zeroing the score might cause some sideeffects for some mpeg4 files)
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
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 198 bytes
Desc: Digital signature
More information about the ffmpeg-devel