On Tue, May 28, 2013 at 11:26:05AM +0300, Martin Storsjö wrote:
> From: Michael Niedermayer <[email protected]>
>
> This fixes several chroma artifacts in several videos.
> ---
> Should this be squashed into the previous commit, or keep
> them separate (where the previous refactors and adds new possible
> parameters while this one actually takes them into use)?
Squash - if it introduces new parameters it should use them properly from the
start.
> ---
> libavcodec/vc1dec.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/libavcodec/vc1dec.c b/libavcodec/vc1dec.c
> index d956f7f..426bc7a 100644
> --- a/libavcodec/vc1dec.c
> +++ b/libavcodec/vc1dec.c
> @@ -4532,6 +4532,7 @@ static int vc1_decode_b_mb_intfr(VC1Context *v)
> vc1_mc_4mv_luma(v, i, 1, 1);
> }
> vc1_mc_4mv_chroma4(v, 0, 0, 0);
> + vc1_mc_4mv_chroma4(v, 1, 1, 1);
> } else {
> vc1_mc_1mv(v, 0);
> vc1_interp_mc(v);
> @@ -4551,6 +4552,7 @@ static int vc1_decode_b_mb_intfr(VC1Context *v)
> }
>
> vc1_mc_4mv_chroma4(v, 0, 0, 0);
> + vc1_mc_4mv_chroma4(v, 1, 1, 1);
> } else if (bmvtype == BMV_TYPE_INTERPOLATED) {
> mvbp = v->twomvbp;
> dmv_x = dmv_y = 0;
> @@ -4598,7 +4600,7 @@ static int vc1_decode_b_mb_intfr(VC1Context *v)
> vc1_mc_4mv_luma(v, 1, dir, 0);
> vc1_mc_4mv_luma(v, 2, dir2, 0);
> vc1_mc_4mv_luma(v, 3, dir2, 0);
> - vc1_mc_4mv_chroma4(v, 0, 0, 0);
> + vc1_mc_4mv_chroma4(v, dir, dir2, 0);
> } else {
> dir = bmvtype == BMV_TYPE_BACKWARD;
>
> --
probably OK
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel
