On 6/22/17 2:52 PM, wm4 wrote: > Affects in particular DXVA2/D3D11VA, which asks for 128 pixels > alignment. The AVHWFramesContext will return AVFrames with that larger > size set. We need to crop it back in the decoder. > --- > libavcodec/hevc_refs.c | 3 +++ > 1 file changed, 3 insertions(+) > > diff --git a/libavcodec/hevc_refs.c b/libavcodec/hevc_refs.c > index 0c1918793e..d21cd982df 100644 > --- a/libavcodec/hevc_refs.c > +++ b/libavcodec/hevc_refs.c > @@ -93,6 +93,9 @@ static HEVCFrame *alloc_frame(HEVCContext *s) > if (ret < 0) > return NULL; > > + frame->tf.f->width = FFMIN(frame->tf.f->width, s->avctx->width); > + frame->tf.f->height = FFMIN(frame->tf.f->height, s->avctx->height); > + > frame->rpl_buf = av_buffer_allocz(s->pkt.nb_nals * > sizeof(RefPicListTab)); > if (!frame->rpl_buf) > goto fail; >
fate-hevc-conformance-RAP_B_Bossen_1 changes its hash because of it. _______________________________________________ libav-devel mailing list libav-devel@libav.org https://lists.libav.org/mailman/listinfo/libav-devel