cross-posted from: https://lemmy.dbzer0.com/post/43980617
And are translations always done based on the native language or do they translate from e.g. the English subtitle to another language? Asking because this definitely feels like something they would skimp out on if they could.
Most of us can’t read as quickly as we can absorb spoken language.
I learned this directly when I decided to create English subtitles for French films and TV as an exercise when I was trying to get from B2 to C1 in French. It was a good exercise, but the result was unusable because the text often goes by too quickly to read.
That’s when I understood that it really is an art.
The subtitle artist must make descriptions work, punch lines land, and reproduce dialog with the correct gravitas. And they have to do it while cutting 50 percent or more of a meaningful, culturally-grounded translation.
It doesn’t seem like the kind of thing an id-less ml model could ever do.
That assertion is invalid, at least in English. Speech is ~130 wpm 1 and reading is >200 wpm 2
And subtitles going by too quickly is a symptom of speech speed vs read speed, not audio comprehension rate.
I guess it depends on the subtitle type? I know some are just the spoken audio and some are descriptions of side conversations/surrounding noises as well. That said I’m one of those that reads faster than hearing but that’s due to bad hearing and my brain trying to figure out what it heard is slower than reading/hearing combined. Or even sometimes the subtitles are a bit fast so I’m still about even.