[FFmpeg-devel] [PATCH] Enable PAFF decoding
Neil Brown
neilb
Fri Oct 12 04:52:30 CEST 2007
On Friday October 12, uoti.urpala at pp1.inet.fi wrote:
>
> On Fri, 2007-10-12 at 12:05 +1000, Neil Brown wrote:
> > On Friday October 12, michaelni at gmx.at wrote:
> > > ive not looked at the stream at all, but from what you posted these b frames
> > > have timestamps which lay prior to the first IDR picture if that is the case
> > > the sane solution is to drop them
> >
> > In my 11 second clip there are two IDR frames. Each is followed (in
> > decode order) by two B frames (pairs of B fields) which should be
> > displayed (based on picture order count) before the IDR frame.
> > (poc numbers for fields are, in decode order,
> > 0, 1, -4, -3, -2, -1, 4, 5, 2, 3, ...
> > I P B B B B P P B B
> > )
> >
> > If we dropped all such frames, it would be sad - even if it would be sane.
>
> Is this the start of the stream or not? Usually such B frames would
> depend on earlier frames (before -4) and so would be dropped _if_
> decoding starts from frame 0 here, but not if we get here after decoding
> earlier parts and have frames buffered from there.
This is the poc/type sequence both at the start of the stream and
from each internal IDR field (sample size = 2).
i.e. the sequence occurs multiple times in the stream.
I cannot see anything in the spec which says anything about where
referenced pictures should be in display order - only that P and B
slices can reference 1 or 2 (respectively) earlier pictures in
decoding order.
Also, as '0' is an IDR, field -4 could not possibly reference anything
before field '0' and so must only reference fields 0 and 1.
The spec explicitly allows negative poc numbers (section 8.2.1 sets
the range for poc between -2^31 and 2^31-1).
The implication of a negative poc number combined with the
requirements that an IDR must have a poc of 0 (same section) is that
the first negative poc must be decoded with reference only to pictures
that are later in the display order (... unless the first negative poc
is an I picture, which would be pointless).
NeilBrown
More information about the ffmpeg-devel
mailing list