If this project is so time-consuming and tedious, why not recruit and train a few people to help? Would it be possible to cut the movie into bite-sized chunks, distribute those to volunteers, and let them do the easy parts? That's how a lot of book preservation projects work. The whole thing gets scanned and OCRed, then different editors make multiple passes over the text to ensure it matches the scanned page. Many hands make light work, many eyes make bugs shallow, etc. Or are the software requirements prohibitive?
(No, I'm not offering to help, but that's just because I'm moving and changing jobs. I may be more helpful in the future.)