
Asimov
Reminds me of an Asimov concept where a neckband would act as a microphone replacement by monitoring the movement of the jaw and neck. The point was that, with practice, the user didn't actually need to say anything out loud for it to work.
A US firm has demonstrated a "thought into speech" neckband which converts nerve signals to the vocal cords into a computer-generated voice. Michael Callahan, co-founder of Ambient Corporation, showed off his device, dubbed the "Audeo", at the recent TI Developer Conference 2008, in Dallas: Using the Audeo, which currently …
the first time any Geek learnt that signals are sent to the larynx and tounge and while speaking silently it's an obvious step to monitor and decode them. Not that nearly all ideas don't have a few thousand inventors and one team who actually make it happen, facebook, google, youtube etc.
But this is very old tech, the difficulty has been working out what the person is trying to say and the small increase in word and digit counts compared to Nasa subvocalisation back in 2003 and various attempts prior to this makes this tech a bit pointless.
The only problem remaining still remains, but I really hope they succeed, or computing power provides a brute force solution.
If you replaced the speech output from this (I presume it uses a series of files for this) you could get a sort of universal translator - select output language, think about what you want to say in your native language a d hey presto, no more tedious learning of foreign languages.
That depends on having a translator capable of grammar, not to mention the current problems with homonyms.
Translating tools really butcher sentences, because a sentence structure varies both between languages and between different sentence constructions.
(See the Lost In Translation tool at http://tashian.com/multibabel/ )
> a "thought into speech" neckband which converts nerve signals to the vocal cords into a computer-generated voice.
Can I just point out that most people's brains are not in their necks?
This thing does not "translate thought into speech". It recognizes muscle twitches and plays back canned recordings. It is entirely dependent on the presence of an upstream thought-to-speech translator known as the human brain that does all the actual *difficult* signal processing work.
I've connected a piezo sensor strain-gauge up to my middle finger.
A computer analyzes the signals and, based on training, selects a pre-recorded clip of speech.
The net result is that every time I give Ambient Corporation the finger, a computer voice says out loud "Michael Callahan is full of crap". But I don't call it "thought to speech", even though it *is* what I'm thinking.
So, straight on the sex-offenders register for thinking "Fuc*k me - look at the jugs on that!! - Worth a squirt!!!" as all virgin male geeks do, then find out said target of overheard voice recognition is 15..
Yep. Won't get one o' them gadgets. Wassa frikkin' point? I've a voice, so I need a computer to save the effort of having to open my mouth? Yeah, right...
(Sigh). A cure looking for a desease to fix. (OK, so was the laser...)
(Paris - 'cos I'm thinking loud and hard, and she aint heard me...)
don't think about it sex damn every six sex seconds a bloke thinks about sex sex...
really sex useful.
The possibilities are amazing - hack into it replace the word "and" with somethign a little risque? or replace your wife's nagging voice with Paris or yours with roger moore...
I want one.... christmas toy 09
Simes.
As I understand it, the thing recognises neurological impulses - the electrical currents your brain sends down your nerves to "tell" your muscles what to do, rather than "muscle twitches".
Also, your brain doesn't "translate thought into speech". What happens in your brain IS thought. The sound waves emitted by your vocal cords is speech. What converts the electrical impulses in your brain to the sound waves in the air is the transmission of these electrical impulses, via nerves, to a sound-wave generating aparatus - in most people, the vocal chords.
Basically, in this application, the Audeo is a replacement set of vocal chords. And yes, it can only make a certain range of sounds - but so can your vocal chords. The Audeo's range is way inferior to vocal chords - but it's way better than NO vocal chords...
Anyway - the Audeo translates neurological impulses into sound. If that's not "thought to speech", could you define for me what you understand by the term?
If the thing is around your neck, does it read signals from your spine or nerve endings? I thought if you had motor neurone disease your signals didn't make it to your nerve endings?
Also I think the universal translator should be both in your ear and voice, so your voice translator can go from native to universal, and ear from universal to other native. Then your devices don't have to "understand" every language, just your language and universal? (Encrypted of course, after you exchange temporary keys.)
"But will it also make your mouth move in synch with the words of the alien language it's producing?!"
Does it matter? The actual output ain't gonna be that realistic anyway, it might as well LOOK like a Bruce Lee overdub!
I never said it's be a GOOD universal translator...
/ yes, thats the one, quickly now, before that foreign chappie over there figures out what I really said to him...
I'm a habitual user of foul language, spoken, written and indeed thought 'louder' than normal thinking. I can't imagine being involved in any accident severe enough to FUBAR my vocal chords putting in any better a mood either.
For me, this would be a disaster. In fact, having one strapped to me under any circumstances would be like some terrible kind of punishment, if not for me then for the countless people who'd have to endure a never ending diatribe of Dalek swearies.
"OH NO! NOT THIS FUCKNOZZLE AGAIN! ABORT ABORT!"