On 29/09/17 02:58, Jun Zhao wrote: > From e2a7cce88d2a47c7e598b59d24258fea8d809c22 Mon Sep 17 00:00:00 2001 > From: Jun Zhao <jun.z...@intel.com> > Date: Thu, 21 Sep 2017 02:41:29 -0400 > Subject: [PATCH V2 1/2] ffmpeg: re-enable hwaccel_lax_profile_check use > hwaccel_flags. > > re-enable hwaccel_lax_profile_check option use hwaccel_flags. > > Signed-off-by: Jun Zhao <jun.z...@intel.com> > --- > ffmpeg.c | 3 +++ > 1 file changed, 3 insertions(+) > > diff --git a/ffmpeg.c b/ffmpeg.c > index 1d248bc269..4f2b9d69d9 100644 > --- a/ffmpeg.c > +++ b/ffmpeg.c > @@ -2859,6 +2859,9 @@ static enum AVPixelFormat get_format(AVCodecContext *s, > const enum AVPixelFormat > > ist->active_hwaccel_id = hwaccel->id; > ist->hwaccel_pix_fmt = *p; > + > + if (hwaccel_lax_profile_check) > + s->hwaccel_flags |= AV_HWACCEL_FLAG_ALLOW_PROFILE_MISMATCH; > break; > } > > -- > 2.11.0 >
I'm inclined to say we should just remove the ad-hoc global lax_profile_check option (which was previously implemented only in ffmpeg.c code for VAAPI) and add a per-stream -hwaccel_flags option to set AVCodecContext.hwaccel_flags directly instead. That can contain "ignore_level" and "allow_profile_mismatch" together, and maybe future flags if appropriate. What do you think? - Mark _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
