Yes, it's possible, but that's unrelated to the Nim language.
You have several solutions:
* The Shazam/OpenSubtitles way: you fingerprint the audio of the movie or the
file, recognize the movie and retrieve the proper subs and/or audio from a
database.
* The "this-could-be-a-big-business" way: you create a software that does
real-time audio translation.
You sell it to the governments in South America and in the UN, put all
translators out of business and you forget about Nim ;).
* Google search there is a spanish app that retrieve spanish audio for
current airing US movies, seems very hacky though.
Anyway to be serious, you can certainly write it in Nim but you have the
following issues:
* Analyzing audio to extract sentences
* understanding sentences
* traanslating them
For audio, look into "Automatic speech recognition" For sentences look into
"Natural language Understanding" For translations look into "Neural Machine
Translation"
Combine all and you're done:
* This is at the very bleeding edge of deep learning
* Most of the tools are in Python today or proprietary
* I suggest you start playing with deep learning for audio first, for example
using Mozilla or Baidu Deep Speech datasets
