On 3/29/20, Michael Koch <[email protected]> wrote: > Am 29.03.2020 um 18:09 schrieb Zedsquared: >>> It's a decimal input which is then treated as a binary bitmask. So dec >>> 10 becomes bin 1010 i.e. filter the 1st and 3rd planes. >>> >>> Gyan >> Ah! that would explain a lot! >> >> I can confirm I'm still confused, however.. >> >> My understanding is that RGB24 has three planes, red then green then blue. >> From your example it seems the masks act big endian so bit 0x08 >> represents >> the first plane and bit 0x02 the third? >> >> Using planes=10 ( should be Red and Blue) gives a trail effect on the >> blue, >> there is no strong red in that source: >> see http://www.robinbussell.co.uk/mov/greenlagtest11.mp4 >> >> However using planes=4 (should be G plane by my reasoning) seems to have >> no >> effect: >> >> http://www.robinbussell.co.uk/mov/greenlagtest12.mp4 >> >> >> Just for completeness, here is >> >> planes= 2 :http://www.robinbussell.co.uk/mov/greenlagtest13.mp4 .. blue >> trails evident >> >> In the above examples I also turned up the green before lagfun so full >> line >> for last example is: >> >> ffmpeg -i IMG_1685.MOV -q:v 0 -vcodec h264 -acodec aac -strict -2 >> -filter:v >> format=pix_fmts=rgb24,colorlevels=romax=0.6:bomax=0.6,lagfun=decay=0.999:planes=2 >> greenlagtest13.mp4 > > Try this workaround: > > ffmpeg -i IMG_1685.MOV -filter_complex > "format=rgb24,extractplanes=r+g+b[r][g][b];[g]lagfun=decay=0.999[gg];[gg][b][r]mergeplanes=0x001020:gbrp" > -t 5 -y greenlagtest.mp4
You are over complicating things. > > Michael > > _______________________________________________ > ffmpeg-user mailing list > [email protected] > https://ffmpeg.org/mailman/listinfo/ffmpeg-user > > To unsubscribe, visit link above, or email > [email protected] with subject "unsubscribe". _______________________________________________ ffmpeg-user mailing list [email protected] https://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email [email protected] with subject "unsubscribe".
