[FFmpeg-devel] [PATCH v3] area changed: scdet filter

[radu.taraibuta at gmail.com]

> -----Original Message-----
> From: ffmpeg-devel <ffmpeg-devel-bounces at ffmpeg.org> On Behalf Of Tobias
> Rapp
> Sent: joi, 13 iunie 2024 16:18
> To: ffmpeg-devel at ffmpeg.org
> Subject: Re: [FFmpeg-devel] [PATCH v3] area changed: scdet filter
> 
> On 13/06/2024 13:21, radu.taraibuta at gmail.com wrote:
> 
> >> -----Original Message-----
> >> From: ffmpeg-devel<ffmpeg-devel-bounces at ffmpeg.org>  On Behalf Of
> >> Tobias Rapp
> >> Sent: joi, 13 iunie 2024 12:52
> >> To:ffmpeg-devel at ffmpeg.org
> >> Subject: Re: [FFmpeg-devel] [PATCH v3] area changed: scdet filter
> >>
> >> On 12/06/2024 21:51,radu.taraibuta at gmail.com  wrote:
> >>
> >> [...]
> >>
> >>> diff --git a/doc/filters.texi b/doc/filters.texi index
> >>> bfa8ccec8b..53814e003b 100644
> >>> --- a/doc/filters.texi
> >>> +++ b/doc/filters.texi
> >>> @@ -21797,6 +21797,22 @@ Default value is @code{10.}.
> >>>   @item sc_pass, s
> >>>   Set the flag to pass scene change frames to the next filter.
> >>> Default value is @code{0}
> >>>   You can enable it if you want to get snapshot of scene change
> >>> frames only.
> >>> +
> >>> + at item mode
> >>> +Set the scene change detection method. Default value is @code{-1}
> >>> +Available values are:
> >>> +
> >>> + at table @samp
> >>> + at item -1
> >>> +Legacy mode for sum of absolute linear differences. Compare frame
> >>> with previous only and no delay.
> >>> +
> >>> + at item 0
> >>> +Sum of absolute linear differences. Compare frame with both
> >>> +previous
> >>> and next which introduces a 1 frame delay.
> >>> +
> >>> + at item 1
> >>> +Sum of mean of cubic root differences. Compare frame with both
> >>> previous and next which introduces a 1 frame delay.
> >>> +
> >>> + at end table
> >>>   @end table
> >>>
> >>>   @anchor{selectivecolor}
> >> Out of curiosity: How do these three modes roughly compare (CPU)
> >> performance wise?
> > Mode -1 is virtually equal with mode 0 (no frame clone speedup of mode 0
is
> mostly theoretical). Mode 1 is roughly 40% slower, depending on input, it
> could benefit from ASM.
> >
> >> I'd prefer if mode "-1" could be used for the default case, and the
> >> three modes get explicit values 0/1/2.
> > So you want four possible values for mode where the default -1 is the
same as
> one of the others?
> 
> Yes. This pattern is also used for the "shaping" option of the ass filter,
for
> example.It allows users to make it explicit that the filter should choose
a
> default (which may vary between versions).

Scdet filter doesn't choose a mode. The default has to be current "Legacy"
for backward compatibility.
I will renumber modes to 0/1/2 to avoid confusion with an "auto" (-1)
option.

> >> Regarding naming I think that the "Legacy" mode still has its
> >> usecases in the future. Naming it something like "simple" would not
> >> imply that it is deprecated in general, but rather that it has lower
> >> precision (and is possibly faster).
> >>
> > OK, I will rename "Legacy" mode so it doesn't look deprecated to
somebody
> that can't allow 1 frame delay. Otherwise it shouldn't be used.
> 
> Thanks.
> 
> Regards, Tobias
> _______________________________________________
> 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".