Michael Niedermayer wrote: > Michael Niedermayer <michae...@gmx.at> added the comment: > > On Thu, Mar 19, 2009 at 09:22:53PM +0000, Baptiste Coudurier wrote: >> Michael Niedermayer wrote: >>> Michael Niedermayer <michae...@gmx.at> added the comment: >>> >>> On Sun, Mar 15, 2009 at 10:01:49PM +0000, Baptiste Coudurier wrote: >>>> On 3/9/2009 1:35 PM, Baptiste Coudurier wrote: >>>>> New submission from Baptiste Coudurier <baptiste.coudur...@gmail.com>: >>>>> >>>>> It seems FFmpeg mpeg2 decoder skips the 2 first B frame in this sample: >>>>> skip_first_b_frames.m2v in incoming >>>>> >>>>> mpeg2dec does decode correctly the 2 first B frames. >>>>> >>>>> I believe these 2 frames does not contain both predictions while still >>>>> marked as B. >>>>> >>>> Anything I can do to help ? I've tried commenting some code to force >>>> decoding without any success. >>> To be honest, i would prefer not to support decoding these special case >>> b frames. >>> >>> Some reasons are >>> 1. the flags might be wrong (damage, buggy encoder, malicious user) >>> so we either need to check for each MB if the pointers are valid or >>> allocate a dummy frame (later would be prefered) >> allocating dummy last picture works. >> See attached patch, I does not seem ugly, does it ? > > it does it work when B frames had pointers to the previous frame? > i mean, i dont see a check for these flags ...
Sorry, I don't understand, you mean it does look ugly ? Since last_picture is now allocated, even if the B frame had pointers to the previous, would it still cause a problem ? Which flags should I check for ? -- Baptiste COUDURIER GnuPG Key Id: 0x5C1ABAAA Key fingerprint 8D77134D20CC9220201FC5DB0AC9325C5C1ABAAA checking for life_signs in -lkenny... no FFmpeg maintainer http://www.ffmpeg.org
