I do amateur stereography (converting 2D movies to 3D) using a monocular depth estimation neural network. My pipeline consists of doing a raw rough cut 3D conversion of the entire movie (takes about 4 days of rendering time), screen the movie in my VR theater, note all scenes with artifacts and errors, correct these scenes either manually or with different parameters to the neural network model, and then splice them in using shotcut and finally exporting this on shotcut as a final cut.
I recently completed “The Batman” this way and the results came out excellent. It’s a broadcast quality 3D conversion. The excellent and moody cinematography looks amazing in 3D.
My neural network works best with 16:9 video. In the case of “The Batman” the video was already a 2.35:1 video dropped into a 16:9 frame with black bars added to top and bottom on the bluray itself so I didn’t need to deal with it
However I have other videos that aren’t in 16:9 and need to be manually converted to 16:9 (with black bars). What’s the best way to do this in shotcut?