I know I have media that is not in a language that I understand. Without using existing metadata, is there a way to scan the audio track of a video file to determine what language is being spoken? I’ve tried a few things myself, using FFMpeg and some others, but they seem to rely on metadata.
you can create a custom filter…
if you lets say speak French, Spanish and English that filter could look like this:
Audio Language does not match French
Audio Language does not match Spanish
Audio Language does not match English
As the custom filter combines the criteria with a logical AND, this will show all the movies with different language only audio tracks.
The same way you can filter for files with audio tracks you don’t need and remove them (e.g. using MKVToolNix).
oh… if that’s already part of the metadata you want to avoid – no… I’m not aware of any other options as that would require an app to recognize spoken words and identify the spoken language
I know you posted in server-linux category but here is a script for Windows that will find all media with undetermined language and then set it to English, though you could change it to find any language you want and ex out the “set to Eng” part aswell… Works on MKV files. I’m sure it can be modified for linux use.