[FFmpeg-devel] [PATCH v1] avfilter/vf_framerate: make the scene change detection calculation method consistent with f_select filter

[Marton Balint]

On Thu, 12 Sep 2019, lance.lmwang at gmail.com wrote:

> From: Limin Wang <lance.lmwang at gmail.com>
>
> Have been tested by fate-suite/svq3/Vertical400kbit.sorenson3.mov and the
> testing results are consistent.
>
> If it is acceptable, I want to move get_scene_score() to the public function and
> make the scene change detection calculation method of the relevant module consistent.
>
> Signed-off-by: Limin Wang <lance.lmwang at gmail.com>
> ---
> doc/filters.texi           | 4 ++--
> libavfilter/vf_framerate.c | 6 +++---
> 2 files changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/doc/filters.texi b/doc/filters.texi
> index 9d500e44a9..1c2e86b635 100644
> --- a/doc/filters.texi
> +++ b/doc/filters.texi
> @@ -10626,10 +10626,10 @@ the default is @code{240}.
> 
> @item scene
> Specify the level at which a scene change is detected as a value between
> -0 and 100 to indicate a new scene; a low value reflects a low
> +0 and 1.0 to indicate a new scene; a low value reflects a low

You cannot change the scale of a user facing parameter, that would break
existing command lines.

> probability for the current frame to introduce a new scene, while a higher
> value means the current frame is more likely to be one.
> -The default is @code{8.2}.
> +The default is @code{0.25}.

You should not change the default unless you have a very good reason, as 
it would change existing behaviour.

> 
> @item flags
> Specify flags influencing the filter process.
> diff --git a/libavfilter/vf_framerate.c b/libavfilter/vf_framerate.c
> index 06e463e4d7..def3f19f55 100644
> --- a/libavfilter/vf_framerate.c
> +++ b/libavfilter/vf_framerate.c
> @@ -51,7 +51,7 @@ static const AVOption framerate_options[] = {
>
>     {"interp_start",        "point to start linear interpolation",    OFFSET(interp_start),    AV_OPT_TYPE_INT,      {.i64=15},                 0,       255,     V|F },
>     {"interp_end",          "point to end linear interpolation",      OFFSET(interp_end),      AV_OPT_TYPE_INT,      {.i64=240},                0,       255,     V|F },
> -    {"scene",               "scene change level",                     OFFSET(scene_score),     AV_OPT_TYPE_DOUBLE,   {.dbl=8.2},                0,       INT_MAX, V|F },
> +    {"scene",               "scene change level",                     OFFSET(scene_score),     AV_OPT_TYPE_DOUBLE,   {.dbl=0.25},               0,       1.0,     V|F },
>
>     {"flags",               "set flags",                              OFFSET(flags),           AV_OPT_TYPE_FLAGS,    {.i64=1},                  0,       INT_MAX, V|F, "flags" },
>     {"scene_change_detect", "enable scene change detection",          0,                       AV_OPT_TYPE_CONST,    {.i64=FRAMERATE_FLAG_SCD}, INT_MIN, INT_MAX, V|F, "flags" },
> @@ -77,9 +77,9 @@ static double get_scene_score(AVFilterContext *ctx, AVFrame *crnt, AVFrame *next
>         ff_dlog(ctx, "get_scene_score() process\n");
>         s->sad(crnt->data[0], crnt->linesize[0], next->data[0], next->linesize[0], crnt->width, crnt->height, &sad);
>         emms_c();
> -        mafd = (double)sad * 100.0 / (crnt->width * crnt->height) / (1 << s->bitdepth);
> +        mafd = (double)sad / (crnt->width * crnt->height) / (1ULL << (s->bitdepth - 8));

No, dividing by bitdepth-8 is completely wrong. It is wrong in f_select, 
but we could not change it there as that would have changed the scale of a 
user facing option.

Regards,
Marton


>         diff = fabs(mafd - s->prev_mafd);
> -        ret  = av_clipf(FFMIN(mafd, diff), 0, 100.0);
> +        ret  = av_clipf(FFMIN(mafd, diff) / 100., 0, 1);
>         s->prev_mafd = mafd;
>     }
>     ff_dlog(ctx, "get_scene_score() result is:%f\n", ret);
> -- 
> 2.21.0
>
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request at ffmpeg.org with subject "unsubscribe".