I hope I do not get bashed for such a stupid question, but I'm trying to assess how far are we of a implementation of what I'm about to tell you.
To give a little context, I have this friend that is blind. He loves Naruto (an anime/TV Show), but he can't watch it/hear it anymore because all episodes have not been dubbed yet, and as been that way for years.
I was thinking how could I help him to finish it, and the first thought on my mind was to find fan-made subtitles and put a text-to-speech system on top of it. But this is too basic, because it will lack all emotions expressed by the voice actor, and all actors will have the same voice. I know that there are many other problems like voice-character lip-syncing due to the translation, but I think this is optional for now, and for a blind person is less important.
The question is, is there any way I can feed a algorithm a voice record and somehow it will keep the original tone/voice when the text-to-speech system runs? Any type of feedback about this is welcome, and any other possible problems I'm not seeing.