I have been toying with The Chronicles Of Myrtana - Archolos audio file mods which are in Russian. Basically it is a bunch of wave files in a pak file and I started looking into AI capabilities.
From a brief look, it appears there is software out there that can take and translate the wav file into a different language using AI.
My question is, time wise, how long is this process usually for the AI roughly?
My thought would be to translate all of the audio files to English, but also train the AI to apply various different voices to each character in the process. Since the files are already properly named and organized and categorized to their character, I think the work would be simply developing a generic voice for each character and then applying the process to each wav files translation.
If this takes too long, I wonder if it would be a faster process to first obtain the scripts for all characters and have them quickly translated in text first, then feed the scripts into each AI voice, save them as the same file names and re-pak the audio into the new translation.
Any thoughts from those of you who have dealt with this?
If the project is going to be a massive one that is extremely time consuming, I will just add it to a later focus, but if it has a reasonable turn over (ie progress is based on how fast I can do the work and not waiting for a program), then it may be something I focus on sooner than later.
AI Audio Language Translation
Ok, it appears that in general, it takes around 24 hours to translate a 1 hour audio file. That may be reasonable depending on how many hours of dialogue there is and AI voice changes appear to be fairly quick (aside from the training).
It may still be easier to translate the scripts first (AI can do this extremely quickly), then have the AI convert it to audio in the voice desired. That would be the fastest I think leaving the bulk of the work with the organization, naming, and repacking the files.
It may still be easier to translate the scripts first (AI can do this extremely quickly), then have the AI convert it to audio in the voice desired. That would be the fastest I think leaving the bulk of the work with the organization, naming, and repacking the files.