These Japanese scientists are simply amazing, always up to something completely new in terms of scientific or technical discoveries: on Reuters today, one of them reports that he reconstructed Mona Lisa's voice (also Leonardo Da Vinci's). You need some Japanese language skills (familiarity with internet would also do, however, as it worked for me) to know how to get around on that site, but you can listen to both Mona Lisa and Leonardo, talking in Italian (what else?!) here. Apparently Dr. Suzuki is an expert in the area of human voice reconstruction based on facial traits etc. (he claims he can achieve 90% accuracy in most cases), hence let's give him some credibility, although in this case he bases himself on paintings (however, not any paintings, but Da Vinci paintings...). But it is definitely quite some news to start the day!
2 comments:
I've heard that the accuracy is around 50% because it is based only on external physical characteristics. But it's cool anyways:)
I would just do a clear experiment, sent an accurate photo of mine to the guy and tell him to reconstruct my voice just to see exactly to what extent he's able to do it. Probably however he is faced with too much demand in this sense already :-)
Another thing would be, unless he did it already, to reconstruct voices of the current (or not only current) Presidents and PM's for most developed countries, where we definitely have enough recordings of speeches by them. It's just that these ideas are so simple that for sure he must have done them already- then it would be just a question of finding out if he's got a site with such samples.
Post a Comment