[Audio-video] 2011-04-20 video could use OpenTLD/FaceDetector source code to automatically re-edit/remove camera movement, and crop podium and ceiling

[David Marceau]

On 11-04-22 02:48 PM, Jason Self wrote:
> The recording uses H.264 for video. Rather than convert that one to a 
> different 
> format I contaced Philip Peng to ask if they could create one from the 
> original 
> recording instead so that we can have better quality.
> 
> 
> 
> _______________________________________________
> Audio-video mailing list
> address@hidden
> https://lists.gnu.org/mailman/listinfo/audio-video

The video capture in .mkv was excellent resolution quality, but with the
camera constantly moving the point of interest (Mr. Stallman's head
speaking) that was annoying.

The actual video size could have been cropped down to the point of
interest on a video encoding, but because of the camera movement, that's
hard to do.  It would be interesting to see a piece of software like
OpenTLD and FaceDetector be applied to do just that on a frame by frame
basis in order to crop out the podium and the ceiling and constantly
keep Mr. Stalling's face even with all the camera movement.

Cheers,
David Marceau