Forgot your password?
typodupeerror

Japanese Lab Creates 'Da Vinci' Voices 183

Posted by CowboyNeal
from the speaking-with-the-dead dept.
Mikki writes "Using methods employed in criminal investigations, the Japan Acoustic Lab has analyzed the skeletal structures of Leonardo Da Vinci and Mona Lisa's faces to replicate how their voices would have sounded." While Da Vinci is cool, I can think of a slew of other deceased notables worth talking with as well.
This discussion has been archived. No new comments can be posted.

Japanese Lab Creates 'Da Vinci' Voices

Comments Filter:
  • by Solra Bizna (716281) on Friday May 19, 2006 @04:48AM (#15364024) Homepage Journal

    *is brutally killed before finishing the meme*

    -:sigma.SB

  • Ergh - yuk. (Score:3, Insightful)

    by Whiney Mac Fanboy (963289) * <whineymacfanboy@gmail.com> on Friday May 19, 2006 @04:50AM (#15364028) Homepage Journal
    1) Promotion of lame movie.

    2) IE 6 Only.

    Please don't post this sort of crap (that's so hard to watch) again.
    • Re:Ergh - yuk. (Score:3, Informative)

      by jginspace (678908)
      2)IE 6 Only.

      Huh? TFA opens fine in Opera and IE...

      The "(Leonardon) Da Vinci referece = promotion of lame movie" stance I shall ignore.

      • Re:Ergh - yuk. (Score:3, Interesting)

        Huh? TFA opens fine in Opera and IE...

        And the article contains a link to the MSN IE 6.0 only site, where you can actually listen to the clips the article discusses (they appear to be wmp only audio files too)

        Utterly typical of MS to attempt to force their crap software on the world (but thank god its only a link to their crap content).
        • Utterly typical of MS to attempt to force their crap software on the world (but thank god its only a link to their crap content).

          God forbid they use the software they developed on the website they created to create the content they're providing to you at no cost.

          It's no different than other proprietary formats like Real Audio. It sucks that they exist, but don't make it into more than it is.
          • God forbid they use the software they developed on the website they created to create the content they're providing to you at no cost. It's no different than other proprietary formats like Real Audio. It sucks that they exist, but don't make it into more than it is.

            The ability to do a thing is not sufficient justification to do a thing.

            You can get better results (quality per bitrate wise) by using fully open codecs than by using Microsoft's. They chose to use them, and fully deserve any beratement

        • It's on MSN, what did you expect? Ads for the new Mac laptops?
      • TFA opens fine in Opera and IE...

        Sorry ... meant to say "Opera and Firefox".

        (And sorry about the atrocious typing.)

      • The "(Leonardon) Da Vinci referece = promotion of lame movie" stance I shall ignore.

        People who cared about art cared about Da Vinci before. Now that the book/movie has hit American (and subsequently lots of other parts of the world unfortunately) pop culture, everyone all of a sudden has an interest in this particular great master. It's sad but predictably true. This would never have gotten the press, or possibly would have even happened at all, had it not been for the book/film. *deep sigh*

      • Huh? TFA opens fine in Opera and IE...

        Not TFA referenced that has the actual content.

        The "(Leonardon) Da Vinci referece = promotion of lame movie" stance I shall ignore.

        From TFA "The voices are part of the intense promotion of the Hollywood film on Microsoft's Japanese site."

        And the whole story is bullshit anyway. The methodology:

        For Leonardo, Suzuki made his voice around the time when he was 60 years old to match his bearded face in the famous sketched portrait. "Because the beard covers his jaws i

    • http://promotion.msn.co.jp/davinci/voice.htm [msn.co.jp]

      Cracking "The Da Vinci Code" could have been easier -- well, maybe -- if the characters had enlisted the Japanese lab which has "recreated" the voices of Leonardo and Mona Lisa.

      Using methods employed in criminal investigations, the Japan Acoustic Lab says it has analyzed the skeletal structures of the historical figures' faces to replicate how their voices would have sounded.

      The voices are part of the intense promotion of the Hollywood film on Microsoft's Japanese
    • 1) Promotion of lame movie.

      2) IE 6 Only.

      Must... destroy.... MS...

      *falls over*
    • 2) IE 6 Only.

      I started up IE to view the site instead of Firefox, and it seemed to have some editorializing of its own to do. I didn't have the Japanese language pack installed, so all the non-English characters were replaced by squares. On my task manager, it looked like it was yelling "MS Nooooooooooooooooo," Anakin Skywalker-style.
  • Fine, but... (Score:5, Insightful)

    by jolyonr (560227) on Friday May 19, 2006 @04:53AM (#15364035) Homepage
    how will bone structure determine regional accent?

    If you make assumptions about where someone was brought up and who by, this kind of thing could work - but let's see a blind test. Let someone do a recording of their voice, get these guys to analyse their facial structure (in silence) and see if their prediction matches reality. It's easy to say what dead people noone alive has heard sounds like.

    Jolyon
    • Re:Fine, but... (Score:3, Insightful)

      by datafr0g (831498) *
      I was thinking the same thing... to add to this - How would one be able to predict that vocal cords are even intact from looking at a skeletal structure??
    • Re:Fine, but... (Score:3, Insightful)

      by joe 155 (937621)
      that's a good point, but even if you did know the region, and even what accent was deffinately in that region at that time (because it's changed so much over the last 400 years) I think an even bigger point is the shape of the tounge; even the slightest change in size would change how your vioce sounded far more than any factor like head size/shape.
      • Re:Fine, but... (Score:3, Interesting)

        by Eivind (15695)
        It's worse than that. The skeletal anatomy is only going to give you a very vague idea, if that. There's just too many variables that are unaccounted for, a short list of examples:
        • Where did the person grow up.
        • What dialect did the people he spoke most to speak (parents, friends, teachers, relatives, neighbours etc)
        • What where his vocal-chords, tongue, lips and mouth like ? (for example, skeletal analysis will not tell you if someone has been smoking for 30 years or not)
        • Was he allergic to anything ? Nose o
        • If you'd RTFA, you'd realize they take as many of these factors into account. They're not just using the skeleton (in fact, they *aren't* using the actual skeleton at all. This is doubly true for the Mona Lisa).

          It's very clear that they are using assumptions, and assumptions can be wrong to varying degrees. That doesn't change the fact that it's possible to get a good approximation using assumptions. Whether this works here I don't know, but I do know that people's appearance seems to correlate to how their
    • by umbrellasd (876984) on Friday May 19, 2006 @05:17AM (#15364108)
      Probably doesn't work too well for eunuchs either, :).
      • Probably doesn't work too well for Unix either, :).

        Well, as other posters have pointed out, the site is IE6 only...

        But apart from that, read this quote, and draw your conclusions:

        A former police engineer who specializes in audio analysis, Suzuki says he assumed the woman in the legendary famed Leonardo painting was 168 centimeters (5 foot, 6 inches) tall, giving her a relatively low tone for a woman.
        • A former police engineer who specializes in audio analysis, Suzuki says he assumed the woman in the legendary famed Leonardo painting was 168 centimeters (5 foot, 6 inches) tall, giving her a relatively low tone for a woman.

          How does being 5'6 make you have a low tone? My best friend is 5'6 and she has a high tone. In fact, our voices are so similar (even though I'm 5'2) that people (even my mom) can't tell the difference between us on the phone. I think this whole concept of height = vocal tone is bu
          • I think this whole concept of height = vocal tone is bullshit.

            IMHO, the whole concept (of skeletal features imply voice) is bullshit.

            As other people have pointed out, there are so many other elements that affect voice (regional access, non-skeletal physical features such as tongue and vocal chord shape, how you "move" your vocal chords and tongue, ...).

            But that Mona-Lisa-had-a-manly-voice snippet was just too funny to pass, hehe.

      • A true advance in science

        I've always wanted to know what Snoopy's [snoopy.com] voice sounded like....you never hear him talk in the cartoons. Now thanks to this revolutionary skeletal analysis technique, hearing his voice is within our reach.

        And after that, I'd like them to map out Morn [garrisonent.com] from Deep Space Nine. He never spoke either.

        Great mysteries are about to be solved.

        However, when this work is complete, these guys can devote their spare cycles to folding protiens [stanford.edu]..another worthy cause.
    • Shouldn't DaVinci and Mona Lisa sound the same?
    • Re:Fine, but... (Score:3, Informative)

      by JWSmythe (446288) *
      I agree.

      Accents change a lot by the way the local accent is spoken.

      Even my own voice, I know depending on where I am, my voice changes. There are a few places that I've spent a good bit of time, so I easily slip into the local accents. There are a few bad fake accents I do too.

      I will say my nice clear broadcaster voice with a midwestern accent (i.e., plain) is a whole lot different than say my southern drawl. And like when I do my totally bogus 80's valley w
    • I'm not an anatomist, but it might be the fact that facial structure might give enough information to estimate the tone and pitch. Obviously, you can add in random accents, but accents don't exactly change your tone / pitch that much. It also appears that body height is factor in determining tone.

      I'll also be curious on how they fare when comparing with live demos. They say they experimented on Osama bin Laden's facial structure. It'd be interesting to see how that came out.
    • An oil painting can record sound. The canvas vibrates with the sound in the studio, creating ridges in the paint left by the artist's pallette knife or brush. I remember some research in the 1970's (?) that claimed to have reproduced some noises, though no intelligible words, from light-slice microscopy of some Dutch masters.

      The Mona Lisa was painted on wood. Not much chance of "enough of the knock-knock jokes, you silly cow, this is supposed to be a serious portrait" spoken with a Tuscan accent.

    • "how will bone structure determine regional accent?"

      What, you don't think Leonardo spoke Modern English with a thick Japanese accent?
  • with them? (Score:5, Funny)

    by iogan (943605) on Friday May 19, 2006 @04:56AM (#15364039) Homepage
    While Da Vinci is cool, I can think of a slew of other deceased notables worth talking with as well.

    Yeah, um.. you won't actually get to talk with them though, you'll just get to figure out what their voices might have sounded like. Sorry if that ruins it for you.
    • you'll just get to figure out what their voices might have sounded like

      In Japanese.

      Maybe.

      The reasoning sounds pretty shaky, though - to the extent that I am curious as to who was rash enough to fund it.

    • he had one of those high pitch girl voices a man shouldn't have!
    • you won't actually get to talk with them though


      Actually, he can talk with them as much as he like, and he doesn't even need a special device or anything. He just won't have any answers, true, but he can talk.
    • Apparantly, these voices will be barely audible through white noise, and will only be available in japanese video format. I hear that bad things happen when you become too engrossed in trying to figure it all out though.
  • Mike Tyson ;)
  • I'd rather know what they'd have to than how they'd sound saying it.
  • Da Vinci (Score:4, Funny)

    by dr_d_19 (206418) on Friday May 19, 2006 @05:12AM (#15364090)
    No wireless. Less paint than Monet. Lame.
  • by Burb (620144) on Friday May 19, 2006 @05:27AM (#15364126)
    ... we all know that we way we talk is completely determined by skeletal structure. Your native language, culture, education, temperament, mood, and state of health are completely irrelevant.

    Mind you, it would be funny if he sounded like Tom Hanks.

    • ...likewise, skeletal structure alone doesn't determine shape of the soft tissue involved in speech (such as the larynx). In criminal investigation, there are huge problems involved in determining, for instance, the exact shape of the nose. These can only be guesstimated. As assumptions need to be made, this whole thing is going to be wildly inaccurate. Also, although modeling the entire physics of the head is definitely a Very Advanced form of speech synthesis, you'd need to model the exact movements of a
  • Could we please please pretty please have a ban on the use of the name "da Vinci" for at least a year?

    I'm totally overmarketed.

    The tragic thing is that I was a big fan of the man himself until that trashy novel came out.

  • Analyze the facial features of what is a Da Vinci painting, make the skeletal structure of a painting, and analyze it to see how would a Da Vinci painting sound if it could say something?

    That sound kinda farfetched to anyone?

    Plus, why take on the easy job? Let'em try and analyze what a Picasso painting would sound like...
  • The Japanese have been weird about the Mona Lisa for a long time. I remember seeing something on TV 10 or more years ago about a device to simulate Mona Lisa' voice. People were getting plastic surgery to look like Mona Lisa (don't think it was just woman either).
    • don't think it was just woman either

      hihihi!

    • The Japanese have been weird [...] for a long time.

      That sentence is universally true. That's what makes them so interesting.
      • That sentence is universally true

        It is true from your perspective, but not to everyone in the universe hence it's not universal.
        Behaviour of Americans is ultimately weird to me, doesn't mean my perception is universal.

        • Okay, you're right. I should have said "That sentence is always true". "Always" implies that there's always someone who finds your behaviour weird; "universally" implies that you do so, too.

          BTW, yes, Americans are weird. As are Germans (guess where I'm from). As is pretty much everyone.
    • don't think it was just woman either

      Which is not as bad as you may think: I recall reading (perhaps a decade ago) that analysis shows that a fair fraction of the the Mona Lisa's face is really based off of da Vinci's. He probably didn't have the subject sit for him for very long, and then finished the painting without her.

      • This looks like as good a place as any to insert my thought... please forgive me if it seems too off base relative to your comment. :-)

        ... and Mona Lisa's faces ...

        When I read this part of the subject post, I read it as Mona Lisa's feces! Eek.

        Okay, I feel better now for having shared. Thank you, sir, for putting up with me.

  • 15 When they had finished breakfast, Jesus said to Simon Peter, "Simon, son of John, do you love me more than these?" He said to him, "Yes, Lord, you know that I like you." He said to him, "Feed my lambs."

    16 He then said to him a second time, "Simon, son of John, do you love me?" He said to him, "Yes, Lord, you know that I like you." He said to him, "Tend my sheep."

    17 He said to him the third time, "Simon, son of John, do you like me?" Peter was distressed that he had said to him a third time, "Do you

    • by Jesus_666 (702802)
      22 When they had finished breakfast, Jesus said to Simon Peter, "Simon, son of John, do you love me more than these?" He said to him, "Yes, Lord, you know that I like you." He said to him, "Feed my lambs."

      23 He then said to him a second time, "Simon, son of John, do you love me?" He said to him, "Yes, Lord, you know that I like you." He said to him, "Tend my sheep." He said to him, "But I just fed them." He said to him, "I don't care, feed them again."

      24 He said to him the third time, "Simon, son of Joh
    • Pardon my ignorance, but I don't get it. Was Jesus forgetful? Is that what this is teaching us?

    • "And btw, in that infamous "last supper" picture, there is more than one character that looks like a woman..."

      Now, if only one of them actually had the facial structure and skin tone of an ancient Israelite, rather than a cracka ass Anglo-Saxon, then you might have a case.
    • And btw, in that infamous "last supper" picture, there is more than one character that looks like a woman...

      Not to take this in a serious direction, but my favorite part about the amateur symbologists that graduated from Dan Brown University is how somehow a direct correlation is drawn between a painting and the truth.

      Is it more likely that ol' Leo was just having a bit of fun with an androgynous figure in his painting, or that he was a member of a secret organization that knew about the cover-up of Jesu
  • Paint and Sound (Score:3, Interesting)

    by MichaelSmith (789609) on Friday May 19, 2006 @05:59AM (#15364214) Homepage Journal

    I remember something from about 10 years ago about people running an LP pickup through the grooves made in paint by a painters brush. The idea is that sound makes the brush vibrate and records the sounds in the paint.

    Apparently they were able to get the sound of the word "blue" out of a patch of blue paint so this painter must have been talking to himself (or somebody else) while he worked.

    • Re:Paint and Sound (Score:2, Insightful)

      by Threni (635302)
      > I remember something from about 10 years ago about people running an LP pickup
      > through the grooves made in paint by a painters brush. The idea is that sound
      > makes the brush vibrate and records the sounds in the paint.
      >
      > Apparently they were able to get the sound of the word "blue" out of a patch
      > of blue paint so this painter must have been talking to himself (or somebody
      > else) while he worked.

      It's hard to imagine with half a brain anyone believing for more than a second that this
    • Re:Paint and Sound (Score:2, Informative)

      by $sjfsjf (862288)
      Indeed, Archaeoacoustics [chello.se] However this technique has never been used to recover sounds from actual historical pots or paintings. I wonder why not, if their tests have been so sucessful... Maybe it only works if you know what you are supposed to be hearing, 'Here's to my sweet Satan' anyone?
      • Thanks for the link. Its interesting that they really only tried low tech approaches. Lasers have more recently been used to digitise phonograph records, and I imagine that you could attack a surface with an electron microscope and digitally convert the profile to sounds.

        Maybe a laser could even detect the original surface of a painting, under coatings which were added later.

        • Non-visible spectral imaging (also known as infrared and ultraviolet photography) are typically used to detect alterations to a painting. You don't need an expensive laser, just a nice IR-pass filter and some tungsten lights.
  • Or does Leonardo Da Vinci kind of sound like Jabba the Hut?
  • So (Score:5, Funny)

    by nickthisname (630860) <gmoore AT friendlycity DOT net> on Friday May 19, 2006 @06:17AM (#15364262)

    We now know what Da Vinci would have sounded like when he said:
    "Someone please shoot Dan Brown."

  • I think this is a neat idea (although an obvious plant by some marketing parasites) but I have to ask: has anyone tested this?

    Specifically: has anyone recorded a voice, recorded an MRI, and generated a voice? Did they match? Were they close?

    • You don't need to test it to know it's a lot of bunk.

      It's a given that a female will usually have a "female voice" (not just pitch but formants, inflection patterns etc also), and a male a male one. For physiological reasons larger people tend to have deeper voices, and smaller people higher ones, but not always so. A larger nasal cavity (big nose) will make nasalized phonemes more pronounced.

      Beyond that, a skelton will tell you nothing of the quality of someone's voice, their accent, whether it indeed is a
  • Always using their technologies the right way... wow!!!
  • I'm no doctor, but I know enough that our larynx and vocal cords are made of fleshy muscle; not bone.

    In order to accurately figure out what somebody's voice might have sounded like -- minus unknowable unique accent quirks -- would require a DNA sample and technology we don't yet have(1).

    (1) Namely, vastly faster computers that can take some source DNA and quickly "grow" (protein unfolding, etc) an adult human being in simulation, then send the right signals to the nerves to make virtual speech. Complica

  • by Viol8 (599362) on Friday May 19, 2006 @07:11AM (#15364376)
    "Leonardo , I *really* need to go to the toilet! Now!"
  • How many real criminals got away with real crimes while the boys in the lab were abusing the facilities for something so pointless?

    Never mind, if the Japanese are anything like the rest of the world, they probably just hassled a few extra motorists.
  • and Get back to work on the robotic woman.
  • ... and structure of someone who is still alive and see how close they get.
  • by iBod (534920) on Friday May 19, 2006 @08:11AM (#15364548)
    I think he'd have been very interested indeed in the maiden flight of the Airbus A380 yesterday, which received NO coverage whatsoever on Slashdot (stuff that matters!) and would be pissed off by this lame article about some fools trying to cash-in on his name (stuff that matters not).
    • I think he'd have been very interested indeed in the maiden flight of the Airbus A380 yesterday

      I think Leonardo would indeed have applauded the maiden flight of the A380 - on April 27, 2005. (The maiden flight of the aircraft was over a year ago.)

      which received NO coverage whatsoever on Slashdot (stuff that matters!) and would be pissed off by this lame article about some fools trying to cash-in on his name (stuff that matters not).

      It's amusing that you criticize Slashdot for not having coverage of an

  • Oh goody! (Score:2, Funny)

    by CamDawg (970808)
    Now we can look forward to authentic Da Vinci-voiced endorsements for vacuum cleaners. Ah, what a glorious age in which we live.
  • Some people have been listening to Da Vinci for centuries.

    How's that? Da Vinci was a brilliant man who left us not only some wonderful paintings but also a wealth of writings. To listen to the real Da Vinci, all you need to do is look at his art or read his writings -- carefully, with understanding, of course.

    Hearing some lab's claimed reproduction of his physical voice really doesn't help us to understand the man or his thoughts.

  • Please call him Leonardo, only Dan Brown calls him Da Vinci. I don't think Leonardo would have answered to it.
  • by jimwatters (110653) on Friday May 19, 2006 @09:08AM (#15364808) Homepage
    I hear dead people!
  • BRaAAaAaINS!!!! BRAaAAAaAINZZZ!!
  • I seriously doubt this is even close to what Da Vinci sounds like. The true test of this would be to take someone on this project who has never heard of Mike Tyson, give them a photograph of him, and see if they come close to what he sounds like. I suspect they wouldn't be even be in the same ballpark. I know Leonardo did a lot of anatomical drawings, but unless he managed to do a really, really detailed drawing of his own vocal chords, I suspect they're making a very rough guess.
  • Embarrassingly, he sounded like Chico Marx...

    "Ha ha ha ha ha ha... you can't-a fool me. There ain't-a no Sanity Clause."

  • The article will teach you many varriations of "we made it up."
  • Well they should try it on dead people who have had voice recordings of themselves to see how well these scientists can recreate a voice.

    So THIS is now Tupac keeps coming out with albums posthumously...

Overdrawn? But I still have checks left!

Working...