[FFmpeg-user] Efficiently doing thousands of edits?

[Michael Shaffer]

I'm pretty sure you could use Python and OpenCV to create a solution..
Basically..
For each frame of video,
1) Create a subset image of the the lower left corner using a specific
bounding box.
2) Calculate an average of the black value of the subset to determine if is
a black image or not.
3) If the black value is normal, store that subset as the current "good"
image
4) If the black value shows it is actually black, don't change the "good"
image
5) Overlay the "good" image to the current frame.
6) Go to the next video frame



On Wed, Apr 10, 2019 at 5:06 PM Carl Zwanzig <cpz at tuunq.com> wrote:

> On 4/9/2019 4:17 PM, John Hawkinson wrote:
> > What would you recommend for video editing tools that use ffmpeg's
> libraries?
>
> I suspect that most of the opensource editing projects use them. I've
> dabbled with ShotCut, and it seemed OK. I also suspect that they're not
> going be much of a help (need EDLs and all that).
>
> An option, perhaps not very good, is to split the existing video into one
> file per quadrant, process the one with the dropouts, then recombine into
> a
> single video. If you have the space, the intermediate files could be
> uncompressed so as to lessen quality loss along the way.
>
> Later,
>
> z!
>
> _______________________________________________
> ffmpeg-user mailing list
> ffmpeg-user at ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-user
>
> To unsubscribe, visit link above, or email
> ffmpeg-user-request at ffmpeg.org with subject "unsubscribe".