On Tue, May 28, 2013 at 12:32:13PM +0300, Martin Storsjö wrote: > On Tue, 28 May 2013, Kostya Shishkov wrote: > > >On Tue, May 28, 2013 at 12:21:44PM +0300, Martin Storsjö wrote: > >>On Tue, 28 May 2013, Kostya Shishkov wrote: > >> > >>>On Tue, May 28, 2013 at 11:25:58AM +0300, Martin Storsjö wrote: > >>>>From: Michael Niedermayer <[email protected]> > >>>> > >>>>The existing implementation had little to do with VC1. > >>> > >>>http://26-26-54.hardwarebug.org/33 > >>> > >>>>This could be implemented by adjusting the reference frames > >>>>ithemselfs but that would make frame multi-threading difficult. > >>>>--- > >>>>libavcodec/vc1.c | 91 > >>>>++++++++++++++++++++++++++++++++++++++++----------- > >>>>libavcodec/vc1.h | 8 +++-- > >>>>libavcodec/vc1dec.c | 49 ++++++++++++++++++--------- > >>>>3 files changed, 111 insertions(+), 37 deletions(-) > >>> > >>>In general looks WTFy. Maybe (I cannot say even with 80% certainity) it > >>>tries > >>>to deal with the fact there's intensity compensation defined in 8.3.8 and > >>>10.3.8 and it mandates that intensity-compensated reference frame is used > >>>for > >>>the subsequent fields/B-frames too but that's not obvious from the code or > >>>commit > >>>message. > >> > >>Ok, so what would be the right course of action? Patches 2-3 that > >>you ok'd depend pretty much on this change. Should I try to split > >>this patch into smaller pieces to get the parts that patches 2-3 > >>depend on separated, or would this one just be less WTFy with a > >>seriously amended commit message? > > > >The latter would do for me. > > Right, so e.g.: > > ---8<--- > vc1dec: Redesign the intensity compensation > > Use the intensity-compensated reference frame for subsequent > fields/B-frames.
Since we currently don't change the reference frame we have to maintain lookup tables for intensity compensation in the following dependent frames. _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
