[FFmpeg-devel] [PATCH] libavfilter: add photosensitivity filter

[Paul B Mahol]

On 7/13/19, Michael Niedermayer <michael at niedermayer.cc> wrote:
> On Sat, Jul 13, 2019 at 09:32:38AM +0200, Paul B Mahol wrote:
>> On 7/13/19, Michael Niedermayer <michael at niedermayer.cc> wrote:
>> > On Fri, Jul 12, 2019 at 11:26:34AM +0200, Paul B Mahol wrote:
>> >> From: Vladimir Panteleev <git at thecybershadow.net>
>> >>
>> >> Signed-off-by: Paul B Mahol <onemda at gmail.com>
>> >> ---
>> >>  doc/filters.texi                  |  20 ++
>> >>  libavfilter/Makefile              |   1 +
>> >>  libavfilter/allfilters.c          |   1 +
>> >>  libavfilter/vf_photosensitivity.c | 336
>> >> ++++++++++++++++++++++++++++++
>> >>  4 files changed, 358 insertions(+)
>> >>  create mode 100644 libavfilter/vf_photosensitivity.c
>> >
>> > I support this filter or some variant of it to be applied
>> >
>> > some area of work is that it seems to work less well at higher
>> > framerates
>> > for example:
>> > -vf  fps=120,photosensitivity Denno.ogg
>> > or 240fps
>> >
>>
>> Please refrain from commenting for stuff you know very little or nothing.
>> Thanks.
>>
>> Your comments are not valuable at all.
>
> i normally just ignore your insults but here i think i should not
>
> if the same video at 120fps fails to be filtered as well as it is at 30fps
> there is a problem. Because again its the same data, its not played faster
> it just internally has more frames.
> A viewer seeing a frame for 33ms cannot know if thats 1 frame at 30fps
> or 4 times the same frame at 120fps.
> So if one is filtered stronger than the other they cannot both be optimal
>
> About understanding, i looked at the paper referenced by compn, and the
> code
> of the filter. That is not much indeed, but i doubt the other people here
> have
> better knowledge about the subject. If some expert is around here who knows
> this stuff. His/her oppinion and input would no doubt be very valuable

I'm expert to this very subject.

I'm stating that filter is good at reducing/removing abrupt changes in
color/brightness for typical videos.

On another hand, you are pointing to irrelevant corner cases.
Besides user can increase number of frames used in calculation of changes.

What I will not ever tolerate is your belittling other peoples work.