Future of automatic translation: where I want it to go

I’m thrilled about Michael Schumacher’s comeback to Formula 1, and past week I was watching all news about this subject. Today I’ve found some video clips, on Youtube of course, about his first practice yesterday on Ferrari’s Mugello racing circuit. There was also an interview, but the problem was all videos were spoken in German language, which needless to say, I don’t master at all.

Automatic translation

At this point come to me the idea of automatic translation for audio, not just for text. I think this needs to be developed in two stages: one is the speech recognition feature, that transforms the original voice into text and then a text to speech technology that, in this case, will analyze Schumacher’s voice, extract enough samples which are then used to reconstruct its speech in the destination language of the translation. An easy to implement solution would be to display captions on the video.

I know this is practically impossible now, but that’s a direction I want translation services to take in the future. Once the language barrier will no longer exist, Internet penetration will get close to that 100 % perfect rate (given that everyone could afford a PC and an Internet connection).

Maybe the future will suprize me, as it did in so many occasions.