This will not work as Whisper generates different subtiltes every time you run it, so there is no one-to-one comparison unless you get the japanese version with no translation and the translated one from the same run.
Sure, it won't be perfect for the whole video but for part of it it can work. For example, on the SSIS-381 test I ran 4 times, the first 4 min 27 sec never changed, it was always identical. So you can run it a few times, find a range that doesn't change using kdiff3 or something similar to compare the files and then get a transcription of that.
Last edited:










