[FFmpeg-devel] [PATCH] avcodec/vvcdec: fix boundary strength when IBC involved
Nuo Mi
nuomi2021 at gmail.com
Sat Mar 2 16:04:40 EET 2024
The following cases should set bs to 1:
If the prediction modes are not the same.
If both prediction modes are MODE_IBC, but the motion vector delta is larger than 8 of 1/16 pixels.
see 8.8.3.5
How to reproduce it:
vvencapp -i sintel_trailer_2k_1080p24.y4m --preset fast --additional "IBC=1" -o sintel.266
ffmpeg -i sintel.266 -f md5 -
md5 will mismatch
Found-by: 6ws at https://github.com/ffvvc/FFmpeg/issues/187#issuecomment-1962842135
---
libavcodec/vvc/vvc_filter.c | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/libavcodec/vvc/vvc_filter.c b/libavcodec/vvc/vvc_filter.c
index 379d90d02b..dded447bfa 100644
--- a/libavcodec/vvc/vvc_filter.c
+++ b/libavcodec/vvc/vvc_filter.c
@@ -309,6 +309,10 @@ static int boundary_strength(const VVCLocalContext *lc, const MvField *curr, con
const RefPicList *neigh_rpl)
{
RefPicList *rpl = lc->sc->rpl;
+
+ if (curr->pred_flag == PF_IBC)
+ return FFABS(neigh->mv[0].x - curr->mv[0].x) >= 8 || FFABS(neigh->mv[0].y - curr->mv[0].y) >= 8;
+
if (curr->pred_flag == PF_BI && neigh->pred_flag == PF_BI) {
// same L0 and L1
if (rpl[0].list[curr->ref_idx[0]] == neigh_rpl[0].list[neigh->ref_idx[0]] &&
@@ -497,6 +501,7 @@ static av_always_inline int deblock_bs(const VVCLocalContext *lc,
const int cb_p = (y_p >> log2_min_cb_size) * min_cb_width + (x_p >> log2_min_cb_size);
const int cb_q = (y_q >> log2_min_cb_size) * min_cb_width + (x_q >> log2_min_cb_size);
const uint8_t intra = fc->tab.cpm[chroma][cb_p] == MODE_INTRA || fc->tab.cpm[chroma][cb_q] == MODE_INTRA;
+ const uint8_t same_mode = fc->tab.cpm[chroma][cb_p] == fc->tab.cpm[chroma][cb_q];
if (pcmf)
return 0;
@@ -517,6 +522,9 @@ static av_always_inline int deblock_bs(const VVCLocalContext *lc,
if ((off_to_cb && ((off_to_cb % 8) || !has_sub_block)))
return 0; // inside a cu, not aligned to 8 or with no subblocks
+ if (!same_mode)
+ return 1;
+
return boundary_strength(lc, mvf_q, mvf_p, rpl_p);
}
--
2.25.1
More information about the ffmpeg-devel
mailing list