One for Chip? Speech synthesis
simonl
Posts: 866
Unfortunately my wife was diagnosed with thyroid cancer recently -- thankfully we're being told it's not life threatening -- and this got me thinking all sorts of horrible thoughts; one of them being that the operation goes wrong and leaves her without a voice. Then I thought, hey, with the PChip's excellent sound capabilities, would it be possible to create a device that translates throat vibrations and mouth muscle signals into life-like voice?
Well, I don't know if it's possible, but if anyone wants to try, please let me know -- I only wish I were clever enough.
Take care everyone, and have a happy festive season.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Cheers,
Simon
www.norfolkhelicopterclub.co.uk
You'll always have as many take-offs as landings, the trick is to be sure you can take-off again ;-)
Well, I don't know if it's possible, but if anyone wants to try, please let me know -- I only wish I were clever enough.
Take care everyone, and have a happy festive season.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Cheers,
Simon
www.norfolkhelicopterclub.co.uk
You'll always have as many take-offs as landings, the trick is to be sure you can take-off again ;-)
Comments
I'll keep my fingers crossed(whenever possible.) and wish her the best of luck with the operation.
Your idea...
The first problem is accurately detecting muscle-positions(I have no idea exactly how important the tongue's exact position is, but I feaqr it's pretty d@rn important). then there's the tightening/slackening of the vocal chords, breathing-pattens, jaw-position and mouth shape.
Reading out the mouth shape and the jaw position should be possible with a camera, at least(but that leaves the system desk-bound. You probably want it to be portable and as discrete as possible, but you must learn to walk before you can run)
It may be possible to read the tiny electical impulses from the nerves to the vocal chords.
(Anyone know of some decent sensors, like those used when measuring brain-activity?)
Couple that with breathing-pattern, (possibly using a flexi-force sensor to measure chest-expansion?) and you may be able to pick up some phonemes, I think.
Possibly stick an ultrasonic sensor in a pendant and point it upwards to read jaw position?
(you'll need to tell her to look straight ahead when speaking, though)
That is, if the sensor is fast enough to read it correctly.
you'll also need a mute button...
(A belch may be fun sometimes, but... )
And no, I'm not clever enough to get it to work.
(I'm good at tossing together ideas, though
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Don't visit my new website...
I am not the one either, but I can tell you the device you speak of is allready invented. I saw a disabled veteran one time that used such a device to communicate with. By all means don't give up hope, because it's out there somewhere.
It's a device they hold up to the throat which sends vibrations through their throat, to modulate the airflow.
gives them a rather strange, buzzing voice.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Don't visit my new website...
I hope the best for you and your wife.
The problem is the lack of vibration. The surgery would render the vocal cords inoperative....therefore there would be no vibration. That is the problem. Vocal cords do some remarkable things. They are adjustable guitar strings for lack of better definition....They adjust very fast. What would be cool is an electric larynx that could automatically adjust according to speech patterns.
The mouth, throat, tongue, and lips do most of the other vocal manipulation. The nasal passage does some things....but is more important in singing.
There has been quite a bit of research in this area....at least I believe there has.....I just don't know how you would implement such a thing....unless you could read the nerve impulses.
Without vocal cord variation......mono tone is the result. The voice·lacks meaning/intent without variation in pitch.
Sorry.......I didn't want to burst the bubble.....I'm sure it can be done....just not sure how.
James L
It was a perminant thing, but you may still be right, as it was a buzzing type synthesis I heard. Thanks for the correction.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
OS-X: because making Unix user-friendly was easier than debugging Windows
links:
My band's website
Our album on the iTunes Music Store
anyway, don't knock your own abilities. a lot of usefull things like this are never made because they'd take a LOT of time to get working and nobody has a really pressing reason to build it. Take it slow, do your homework, screw up a lot, and amazing things can happen.
my 2 cents
Marty
Good luck to you and your wife and family.
Chris I.
A person who 'just' lacks the vocal chords would do just fine with a normal 'text-to-speech' translator, and Chip has already done the groundwork for that by giving us phoneme-based sound.
(It just lacks the 'text-to-phoneme' part, and that can be done with a dictionary if needed.)
A pressure-sensitive button on the electronic vooder?
It would take a delicate touch and years of practice to get it to improve the sound instead of distort it, compared with the monotone models used now. (A person must not only learn to 'send' the 'control signals' to his fingers, but also to send them earlier than the signals that should havee gone to the larynx because the signal path is so much longer, and timing is rather critical)
But if you could manage to read out the nerve impulses, and use that to control the box, that would be a great help.
Might even be a marketable product.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Don't visit my new website...