Voice of the Mona Lisa constructed

[ecoli]

Japanese researchers were able to create, by measuring the face and hands of Da Vinci's Mona Lisa, a model of what her voice would sound like. The model was made by physically duplicating what her skull may have been like, and should reproduce her voice within about 90% accuracy.

 

Creation of such "voice prints" is usually used in police investigation work.

 

http://today.reuters.com/tv/videoSto...RelatedVideo-3

[Martin]

Does the reconstructed version speak Italian or Japanese?

 

 

wow

 

she speaks modern Italian with a delicate little lisp. it is a beautiful reconstruction.

 

(SOUNDBITE) (Italian) MONA LISA SAYING:

 

"My name is Mona Lisa. My real identity is wrapped in mystery. Some people say I am Mary Magdalene, others Lady Gioconda, or Isabella D'Este, Da Vinci's mother or even some say I am Da Vinci himself. However the only thing I can say, is that I am the most loved woman in the world and the woman with the most mysterious smile.''

 

It reminds me of Pygmalion (My Fair Lady) where the scientist they fool at the party says "She must be a foreigner. You only speak that perfectly if you have been TAUGHT english."

this woman is an Italian Eliza Dolittle.

[ecoli]

yes it is Martin... but is it technically a reconstruction? Because, it might be possibly that the Mona lisa was never a real person. so, it might be better off being called a 'construction.'

 

She does speak Italian, the researchers brought in an Italian to get the accent right.

[Genecks]

I don't buy it.

Use it on Leonardo's "The Last Supper" or a photograph of Abraham Lincoln.

[ecoli]

I don't buy it.

Use it on Leonardo's "The Last Supper" or a photograph of Abraham Lincoln.

 

What's that supposed to mean?

[Martin]

I think there are probably a lot of photo of Abe Lincoln and if you reconstructed his skull and neck from those you might get a fairly accurate anatomical model.

 

 

I think the people who do this must have tried it out on living subjects to see how close they could come. It would be easy to do a blind test of the technology. Take my picture from various angles and reconstruct my voice (not ever having heard it) and see how close you actually come.

 

Tests measuring the validity of this technique must be published somewhere. To the extent that it has been proven to WORK which I dont know, then it might actually give some of voice quality of Lincoln.

 

======================

 

Obviously the Davinci demo is PUBLICITY by the japanese expert. As such it is charming and does not pretend to scientific value.

 

I dont think anyone knows who was the model of Davinci MonaLisa painting. If he used a model, then, if he did an accurate painting of that model, then the voice reconstruction could be said to be an attempt to capture voice quality of THAT MODEL whoever she was.

 

It is not important.

=======================

 

GenX says try Davinci LAST SUPPER which has a full face picture of Jesus of Nazareth. However Davinci did not use Jesus of Nazareth as his model. He would have just used some guy who had time to sit.

If he used a model when he painted Last Supper one could try to reconstruct the voice of that model, but it would not be so interesting.

 

I think it would be more amusing to use familiar portraits of FAMOUS POPES or Queen Elizabeth or William Shakespeare or some of Titian's courtesans where you have historical documents of who the person actually was and what they did and in some cases quotes of what they said.

[gcol]

Synthesise a voice based on a composite of all pictures and representations of God. Make a talking book of the bible, sell it as "The word of God by God".

 

Make someone a helluva load of money. And the publicity, wow!

[JesuBungle]

But couldn't this process be quite unreliable? I mean, who's never met a woman that sounded like a man or the other way around. I just think that there's a lot more factors that go into a person's voice than just looks.

[ecoli]

But couldn't this process be quite unreliable? I mean, who's never met a woman that sounded like a man or the other way around. I just think that there's a lot more factors that go into a person's voice than just looks.

 

yes, it's more then external looks. The used her external physical appearance to estimate the size/shape of her skull. Which would give information about how sound would resonate, etc.

[olifhar]

Japanese researchers were able to create' date=' by measuring the face and hands of Da Vinci's Mona Lisa, a model of what her voice would sound like. The model was made by physically duplicating what her skull may have been like, and should reproduce her voice within about 90% accuracy.

 

Creation of such "voice prints" is usually used in police investigation work.

 

http://today.reuters.com/tv/videoSto...RelatedVideo-3

 

Very cool.

 

I'm waiting for the web-app version to be available where you can just upload JPEG pictures. Maybe in 15 years?

 

Hmm, I wonder how Charlie Brown's voice would sound if you applied the same methodology you would to an anatomically correct non-cartoon human child.

[Tom Mattson]

90% accuracy? Err...As compared to what?

[ecoli]

I'm guessing that they've done this with painting of alive people, and get it to around 90% accuracy. So, it should be safe to assume that they would do just as good a job with the Mona Lisa.