
Watched some random german comedy sketches here: https://www.youtube.com/watch?v=umjRB5Y_w0c
The automated english subtitles are very good, but don't show who is saying which line (although you can figure it out).
This market will YES if this or similar autogenerated caption videos of the show Pastewka (or similar type of show if that one disappears) have a version of speaker separation which is better than today (where "better than today" is just a stream of text, translated from german to english, with no markers or way to know which speaker said what)
Even if it's just adding a newline to separate speakers, that counts, by deadline, end of june 30 2026
Other ways I can think of doing it:
color to indicate different people's voices
positioning of subtitles below the speaker
actually figuring out the names of characters and using that to distinguish
etc
The point is whether the system will improve re: identifying who is saying what.
Automated dubbing is interesting but let's limit this one to just subtitles.
