backtop

Print E-mail del.icio.us 22 comment(s) - last by Alpha4.. on Oct 16 at 3:56 PM
We may be able to speak without actually speaking aloud

A new technology that will allow people to speak without actually speaking is a technology of the future currently being researched by the National Aeronautics and Space Administration.  NASA's Human Senses program arranged for the start of the subvocal speech program seven years ago. Dr. Charles Jorgensen's research team was initially supposed to design a system of communication that would be utilized during construction and work in the International Space Station – but they discovered the technology could have a number of practical uses in every day life.

It could be used by people with vocal cord disorders, along with security personnel and a number of potential military uses. The subvocal speech is analyzed sub-auditory speech that is the same as silently reading. Small, button-sized sensors are placed on a test subject's neck to record nerve signals in the throat. All nerve signals collected by the receivers are then sent to a processor before being sent to a computer for analyzation and proper translation into words.

In the first test conducted by scientists, the special software recognized six words and 10 numbers of the English language. With 92 percent accuracy during the first trial the software learned “stop,” “go,” “left,” “right,” “alpha,” “omega” and numbers ranging from zero to nine. The second test will focus on controlling a mechanical device while only using a simple set of commands.

Subvocal research is being conducted at the NASA Ames Research Center in Silicon Valley.

Comments     Threshold


This article is over a month old, voting and posting comments is disabled
Anyone else think of Ender's Game?
By jkresh on 10/14/2006 4:17:31 PM , Rating: 2
Anyone else think of Ender's Game? Or more specifically later books in the series.




RE: Anyone else think of Ender's Game?
By Knish on 10/14/2006 4:48:26 PM , Rating: 2
Absolutely :)


By regnez on 10/14/2006 6:01:50 PM , Rating: 2
Speaker for the Dead, and so on, when he is talking to that computer? Absolutely. Personally, I thought that the second and third books in the series were the best, not the first.


RE: Anyone else think of Ender's Game?
By Zirconium on 10/14/2006 7:20:22 PM , Rating: 1
I was thinking more "Ghost in the Shell."


RE: Anyone else think of Ender's Game?
By AndreasM on 10/14/2006 8:23:52 PM , Rating: 2
And my first thought was "Axis of Time". I guess this is yet another technology that was foreseen by sci-fi. The fact that it has to be trained sounds a bit disappointing, hopefully they will be able to make it more microphone-like so it could be used in daily communication. I don't think they had these in "Ghost in the Shell" though, it was more along the lines of non-vocal wireless communication, but this does come close.


By judasmachine on 10/14/2006 9:59:36 PM , Rating: 2
Hopefully it goes better than MS's demo of their speech recognition software. It seems like it would though since it's actually detecting the movement than the sound. (if I'm understanding it correctly)


By captchaos2 on 10/16/2006 8:07:14 AM , Rating: 2
Oh well, if we don't evolve into "newtypes" like in the Gundam series, then I guess this is the next best thing. Seriouisly though, I'm starting to think that Mamoru Oshii is the new H.G.Wells. He is predicting the future of the combining of man and machine very well in his "Ghost in the Shell" series.


By Ealdric on 10/14/2006 8:21:08 PM , Rating: 2
Also Anne McCaffrey's 'shell people' from The Ship Who Sang and related stories.
http://en.wikipedia.org/wiki/The_Ship_Who_Sang
Maybe.


RE: Anyone else think of Ender's Game?
By lennylim on 10/14/2006 11:26:32 PM , Rating: 2
http://en.wikipedia.org/wiki/Earth_(novel)

Granted, it came out a few years after Ender's Game and Speaker for the Dead. But it was featured more prominently than in OSC's books.


RE: Anyone else think of Ender's Game?
By Heron Kusanagi on 10/14/2006 11:35:07 PM , Rating: 2
And next will be subliminal messages...

Jokes aside, I can see why people will want this technology. But really, is it truly necessary?


RE: Anyone else think of Ender's Game?
By Belegost on 10/15/2006 2:51:19 AM , Rating: 3
This technology is very needed.
Consider places where people need to communicate in noisy environments, e.g. heavy machinery operators. This technology would allow them to talk with other crew members, or give voice commands to equipment, as well as maintain a log of operator speech for safety records. Without needing to deal with separating the speech from the background noise which is currently a difficult task.

Or situations where you don't want to have audible speech: I'm sure the secret service will be quite interested in a system that allows an agent to communicate to the team about a suspicious person without tipping them off by speaking into a microphone.

Also, this technology is going to be used to augment current systems - for instance, hands free radio use, this system can identify when the person is actually speaking, this could be highly useful in places where you need both hands to perform tasks, and you need to communicate but you don't want to leave the mic open. Fighter pilots might find it useful.


By Ringold on 10/15/2006 4:34:51 AM , Rating: 2
I'm sure it could be trained away, but there's definitely thoughts that I might say subvocally in the cockpit that I'd really rather not ATC or other pilots hear. ;)

Actually, wish they could hear it, but if after calling to have a jet pulled off the middle of the taxiway my message wasn't just "Thanks" but "Thanks, idiots" (for parking it there in the first place), well, when I got back in later on I think the full-serv avgas would suddenly become self-serve for me. And I really don't like pumping avgas..


Do I understand this?
By saratoga on 10/14/2006 11:00:13 PM , Rating: 2
So basically you talk without really exhauling so that your muscles are moving but you're not actually generating sound, and then something like an electromyogram is used to determine how your muscles are moving. Finally some sort of computer model (HMM, NN or something?) attempts to guess what word you were saying?

Sounds cool, but detecting this sort of speech has got to be even harder then noraml speech recognition, which already is a very hard problem.




RE: Do I understand this?
By Belegost on 10/15/2006 3:06:54 AM , Rating: 2
This actually may be EASIER than recognizing speech in a sound wave.

Some big obstacles to recognizing speech in sound are noise, and individuality.

First up, a person generally isn't speaking in a silent environment, so the computer has to recognize the difference between speech and all other sounds - including background speech. This is usually not too difficult to do, but it certainly increases the complexity of the classifier.

Second, each person sounds different, and a lot of those differences are due to things such as the size and shape of the mouth, the sinuses, the tongue, and to a small extent the lungs. There's a huge space of possible speech producing systems among humans, and the SR system needs to be able to learn to work with any particular one - hence the training that modern speech recognition software requires.

I haven't looked at this particular research, but it very well could be that humans use very similar neural signals to produce the same words, between different individuals. And devices for measuring electrical signals like these are well developed with a high SN ratio. So it is possible for this to actually be an easier method of speech recognition. Of course, it is possible that humans use very different neural signals to create the same words between different individuals, in which case the SR system will still need to be trained to an individual.


RE: Do I understand this?
By alcalde on 10/16/2006 2:31:13 PM , Rating: 2
"So basically you talk without really exhaling so that your muscles are moving but you're not actually generating sound, and then something like an electromyogram is used to determine how your muscles are moving."

I'm immensely excited about this, but for a reason other than the ones being talked about. Stutterers generally don't stutter when whispering. Refinement of this technology could lead to effective fluency. I'm only a mild stutterer and wouldn't need/want something extreme like this, but this could DRAMATICALLY improve the quality of life for the severe stutterers who often can not even speak half a sentence without severe stuttering or blocking. This severely limits career opportunities and social experience. Use of this technology along with human-sounding speech synthesis could enable them to much more effectively communicate and be a life-changing improvement.


Tricky
By Wwhat on 10/15/2006 5:07:22 PM , Rating: 2
It's a bit tricky, if a soldier has it on and he mutters 'moron' about a general's orders the computer will say "MORON!"
If it's too sensitive it means those of us that mutter stuff under their breath have to avoid it or become known as aggresively critical.





RE: Tricky
By Alpha4 on 10/16/2006 3:56:39 PM , Rating: 2
Heh. That sounds like a very likely scenario, but I'm sure the technology can be inversely adapted to allow a muting button as opposed to recording.


ridiculous i know
By judasmachine on 10/14/2006 10:00:48 PM , Rating: 2
My first thought was this is why the gray's mouths are pretty much useless.




The Draka
By qdemn7 on 10/14/2006 11:21:01 PM , Rating: 2
The bio-engineered "New Race" Draka in S.M. Stirling's Domination of the Draka series were especially good at this.




Speaking of things being foretold..
By Ringold on 10/15/2006 4:51:09 AM , Rating: 2
I read 'Eternity's End' not too long ago on a recommendation, and some people chose to have implants in this book which allowed them to have a helpful little AI right in their heads.

This, like the Enders Game references above, I think is a good first step. Nail this down, and we can communicate with computers, they just wouldn't be horribly useful. Improve AI, though, and throw in an ear-bud and when you hear a word or technical term you dont know, you ask it and it tells you, or if you need to get some background on something, it summarizes, say, a Wikipedia entry. Or, ones girlfriend can be talking about her shopping trip while you smile as you listen to your AI read AnandTech's latest article.

It'll stop there with verbal communication because I think we're a long way from knowing had to pull thoughts right out of our minds, or insert thoughts right in to them. The above, though, is only lacking AI at this point.. and I don't see how that would be too hard to take care of..

After that all I'll need is a pair of La Forge's eyes, Picards' heart, and Steve Austin's arm and legs.




Even better than that...
By Turnip on 10/15/2006 7:24:46 AM , Rating: 2
No more "Hello!? Yeah, no, no... I'm ON A TRAIN! What? No? He never? What, really? The whole thing!? Feathers and all? ..."

Perhaps people could talk on the phone in public places WITHOUT making a total arse of themselves ;)




By Shining Arcanine on 10/15/2006 11:12:59 AM , Rating: 2
I find that when people talk on their cell phones in doors to be rude. Now, with this technology, they could talk all they want, without having me hear their conversations. It should work beautifully. :)




"A politician stumbles over himself... Then they pick it out. They edit it. He runs the clip, and then he makes a funny face, and the whole audience has a Pavlovian response." -- Joe Scarborough on John Stewart over Jim Cramer









botimage
Copyright 2010 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki