Slashdot Mirror
Talk ... Without Speaking
mjm7 writes "Finally, we might be able to get rid of all those annoying people yelling over the static on their cell phones! CNN has an article about a new technology that senses muscle movements in your face and then translates them into sound. This way all you have to do is mouth words into the phone...not actually speak!" Somehow I suspect that we'd lose a lot of the
subtleties of communication, but it sure would be nice every time hemos calls me from the discotheque.
"Slashdot is about legos and staplers." -Cmdr. Taco
The Anderson partner called his secretary on his cell phone and said:
Ship the Enron documents to the Feds
But she heard:
Rip the Enron documents to shreds
It turns out that this was all just a case of bad cellular...
Life is the leading cause of death in America.
We'll finally be able to understand what the hell mimes are doing! Rejoice!
"I only speak the truth"
Karma: null(Mostly affected by an unassigned variable)
Not good news for those that like to mutter curses to the morons on the other end of the phone.
Imagine what a world of difference this would make to the mute or to people who had lost the use of their voice due to throat cancer. It seems weird they didn't mention the applications this would have for people who have lost or have never had the use of their voice.
yes i run a goth/punk/emo porn site.
Words like this may cause some minor misunderstandings.
Lord, bless my users that they may stop being such fucking idiots!!
I'd also have to say this should be made mandatory for all people that would otherwise force me to listen to their loud cell phone conversations.
With keyboards we successfully took away peoples needs to physically write something... with this we won't need people to verbally speak... next it'll be visual impulses shot right into your head so you really don't need your eyes anymore... sheesh...
This might help voice recognition catch on as a means of PC input, too. I'd feel slightly less stupid sitting in my office mouthing words at my computer than I would actually talking to it.
Please donate your spare CPU cycles to help fight cancer and other diseases
Yeah, this sounds like just the thing for people who want voice dictation, but work in a "noisy" environment. :-)
Alternatively, you could even have a microphone attached so that when you actually did speak, it would automatically disable the recognition - no more accidentally transcribing your half of a phone conversation for example. Wait a minute, I have to patent that idea!
You either believe in rational thought or you don't
What kind of a sound would it make if I held my middle finger up to it?
I mean really, if the static is so bad that you can't get a good enough signal to hear the person, how is the "face recognition" signal going to get transmitted?
"History doesn't repeat itself, but it does rhyme." Mark Twain
Think about it, don't most people move the muscles in their mouths slightly different when they are mouthing words, as opposed to actually speaking them? I would venture that the technology wouldn't be able to discern the subleties in the way we speak.
Other than that, it sounds like an interesting technology.
Attention all planets of the Solar Federation! We have assumed control! - Neil Peart
It also seems to me that sounds are not necessarily made due to the movement of the jaw. I'd imagine that non-vowel sounds emanate from the vocal cords and tongue. And, what could this end up looking like? Think Nintendo Power Glove...
sarchasm: The gulf between the author of sarcastic wit and the person who doesn't get it.
This would be/will be great. Now we just need cellphones that come in vibrate-only mode and I can finally have a peaceful meal in a restaurant without some moron ten tables away disturbing the whole restaurant with an incoming call (and the subsequent conversation).
Question: if this can eventually recognize what sounds the person is meaning to make with 100% accuracy, does that mean that voice recognition has arrived? Instead of spitting out an audio signal, it could output text instead. THAT would be AWESOME.
I'm so excited. :)
Mr. Ska
"and scream without raising your voice."
-jc
Talking on my phone
I twitch, about to sneeze hard.
Phone thinks I said "F*CK."
I can hear it now...
"Domo Arigato, Mister Roboto"
Neat idea...(I didn't anyway) it looks like all they can detect right now are vowels.
I wonder how they will work out the consonant issues. The way an S is produced is pretty similar to a Z. At least they are pretty similar in my mouth anyway.
I suspect everyone produces consonants in a slightly different manner. I mean, when you are learning to speak, you don't stick your hand in someone else's mouth to figure out what their tongue is doing... You just maneuver your own until you make a similar sound.
So there are probably several different tongue configurations that work to produce a sound. Not to mention the shape of one's mouth may require a specific and unique tongue configuration to produce a particular sound as compared to someone else.
Sounds (hehe) like they have their work cut out for them in this area.
--Scott
it sure would be nice every time hemos calls me from the discotheque
:)
Still, this is just a one-way solution. You will be able to hear the person talking in the crowd, but how will the person on the other end be able to hear anything? Will the phone be able to display the message in the form of text or something similar? Or will it just make funny faces at you?
does anyone remember the "my teacher is an alien!! series? plot synopsis: 4th grader finds out teacher is an alien (suprise, suprise), teacher/alien sees him seeing him, and keeping glactic security safe, takes him up into the New Jersey (mega-big spaceship), and they cruise about, saving the universe.
anywho, i read (and probably own) the whole series in probably 4th grade, i'm 18 1/2 now. on one of their missions, they had special devices like this; except it attached to your throat muscles, which is probably a whole lot easier and less conspicious. the funny part was that they had to whisper, otherwise they'd "yell" right into the other people's earsets. good to know this stuff is comming to fruit
my teacher is an alien on amazon.com
the interesting thing about the series, is that it explains in amazingly simple terminology, using a large noodle, how hyperspace works. i'd explain more, but i don't want to get modded offtopic TOO much. and i have to go to work.
moox. for a new generation.
Speaker for the Dead and Xenocide, had Ender and later Miro subvocalizing to Jane, the sentient entity that "lived" in the network of ansibles. It might continue past that, I have only read up to Xenocide.
"Rotate the pod please, Hal..."
... I could see your lips moving ...
Dave
-Ev
I stole this sig from someone cleverer than me.
The French word for 'library' is 'bibliothèque'
Discotheque means the same thing in both French and English - a place where records are played and people dance.
-Vercingetorix
"Necessitas non habet legem." -St. Augustine
Is that this would be great for people who for one reason or another no longer have voiceboxes.
I had a great-aunt who lost a decent portion of her lungs to cancer and cigarettes, and up until her death a few years ago she had to use one of those darth-vader vibration-amplifier things like the "Ned" character does on south park. I was terrified of her when i was six.. (Give me a break, i was six years old and stupid.)
Anyway, i can imagine that technology like this would be just about perfect for people disabled in a similar manner through tobacco, cigarettes or who knows what. No? At least it would keep such people from having to deal with their idiot six-year-old-nephews reactions to the harsh sounds of the vibration amplifier box..
and really, even beyond that, tech like this would be just about the only option for people who are going through whatever that intensive vocal-node-therapy thing is where you're banned from speaking for six months. and i know a number of theatrical singers who would be intensely happy to have one of these so that they could rest their voices between performances without cutting themselves off from the world...
I hope that once this complete, they'll sell a unit where the voice-synth thing outputs into speakers rather than a phone.. I'm sure they would have looked into this possibility by now, right?
(P.S.: While we're on the subject, sort of.. just in case anyone reading knows: This came up as an argument the other night when we were watching the Oscars and examining how much pain Enye appeared to be in from having to exert her voice. What's the difference between a vocal node and a vocal nodule?
Irritable, left-wing and possibly humorous bumper stickers and t-shirts
to those with Tourette Syndrome.
...in a 70s liesure suit... a la "Saturday Night Fever"...
It's rather a scary vision... sorry for sharing it with you.
BlackNova Traders
Was anybody else immediately reminded of the old Simon and Garfunkel tune, sounds of silence in particular the line about "people talking without speaking" (the link is a poor transcription).
For all intensive purposes, "whom" is no longer a word. That begs the question, "who cares"?
Will pointing the phone in someone else's direction enable you to eavesdrop on their conversation?
How about all those times you get a phone call and you realize you don't want to talk them and as they drone and drone and drone you mouth to anyone around you "SHUT THE F-CK UP!!!" Now they will hear that.
RonB
It is human nature to take shortcuts in thinking.
This technology, assuming it works, might initially fail to gain popularity if it's not priced right. I doubt many people would pay, say, $100 extra for a phone with this feature. And that's because many people simply don't care if they're irritating the people around them.
But I'd love to see such places as restaurants and bus lines require their customers, who insist on using cellphones on their premises, to use this product. I bet the bulk of customers would support such a rule, and everyone would benefit.
I'm generally "Interesting," "Insightful," and even "Funny" here. What the hell happens to me at parties?
No, the "whore hose" is attached to me. The "whore house" is the place to use it.
This next song is very sad. Please clap along. -- Robin Zander
Just like now there is a vibrate feature so the phone can ring without anoying every one in a room. Now there is an other feature to try to stop those jirks who talk answer their telephone and talk into in in an unaproprate situation to talk. When this technoligy is realease I think we should have legal rights to smack anyone whos Cell phone goes off in an unaproprate situation (because they should have a vibrate) and then Kick the person who Starts talking to the cell phone where we can hear them. I dont know about anyone else but What is more annoying then having a cell phone go off in an unaproprate place, is when they start talking loudly without leaving the room.
This technology may not stop this from hapening but it would give us a reason to force them to stop. Where the answer oh this is an important call, will be become complete BS.
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
As long as the cell phone makes real noise, rather than inserting a probe into your ear canal and manually manipulates your eardrum so that you hear the conversation without sound...
(It's just a JOKE! I know I'm not the first to think of it.)
NetInfo connection failed for server 127.0.0.1/local
Usually when I mouth a word into my phone it usually means I DON'T want the other person to hear it. I'm not sure what the learning curve would be on a device like this but chances are that until person hits it they are going to have a lot of explaining to do!!
I stole this Sig
LLNL has been researching micropower impulse radar to 'image' the vocal chords, mainly for speech recogonition. The main site seems down, but you can get to it with google cache. Also check out ucdavis
HIV Crosses Species Barrier... into Muppets
Problem is most speech synthesis still sounds robotic. Imagine if she sounded like WOPR from war games.
MCC would you like to play a game?
Now that would scare me as an adult even!
Couldn't someone use the movements in addition to the sound to filter out the actual speaker's voice from the background noise? This seems almost like a nonlinear Kalman filter application (though I am by no means an expert on such things), if you had a (presumably nonlinear) model for speech as a function of the movements of the mouth. The article didn't give too much detail. Oh well, it sounds interesting in the very least.
The last thing we need is technology that allows our wives to be able to figure out what we are muttering at them under our breath...
Oh, I forgot. Most geeks aren't married...in fact, most probably have no clue how to perform even the most basic intereactions with one (except of course when their mothers call down the basement stairs to them that dinner is ready)!
You're using her as bait, Master!
A system like this, would either need to incorporate some amount of voice recognition, or use a vibration sensing mechanism.
You know what?
but modern cell phones aren't that bad for you anymore.
Um...
Cell phones always have been subject to a limitation of 600 mW, and more recently subject to specific absorption rate limitations.
If anything, the newer PCS phones that operate at higher frequencies are worse, since you are getting closer and closer to the resonant frequency of your eyeballs.
The only risk that has any real evidence for it is heating risk, so if you feel your brain starting to cook, then you might want to turn the phone off, otherwise, it's harmless.
(Hint, 600mW isn't enough to heat much of anything.)
Anyway, nothing had changed in cell phones to justify your conclusion that they are any less dangerous than they were (they weren't dangerous at all to start with).
I've had enough abrasive sigs. Kittens are cute and fuzzy.
Am I the only one who says nice things aloud into the phone while muttering "fscking azzhool" under my breath? How refreshing honest our communications will become!
"with their freedom lost all virtue lose" - Milton
...but it sure would be nice every time hemos calls me from the discotheque.
I don't know what's more scary -- a new cellular electrode attachment or Hemos heating up (literally) the floor under a giant mirrored ball.
"I'll just chip in a bit for RedHat: I actually have that installed on my university machine." - Linus, '95
You may not be aware of this or have thought of it this way, but a microwave oven is basically just a big, unmodulated radio station broadcasting in the microwave band instead of the radio band.
Are you a real physics genius, or do you just play one in front of your liberal arts friends ;)
Seriously, though, the promised "killer application" for over a decade now has been voice recognition, and we're STILL at a point where the inaccuracy rate leads to it being generally useless in anything other than "ooh, isn't that neat" kinds of demos (for instance it was a laugh to see voice recognition as a hyped feature of Office XP : Now tell me how many people on the planet are actually using it? While I applaud them for adding it for the handicapped, of the general public it seems neat, but when you have to babysit every word it dictates you relegate it to the unused feature list).
So we've barely gotten voice recognition down, despite being "just a wee bit more" type of promise for so long now, and someone is claiming that they'll read your lips? Fat chance in hell, is all I can say. Unless we concatenate our language to about 4 words, there isn't a chance.
cell phones have caught up to deaf people.
go get it
Wouldn't everybody start sounding like the vietnam vet on south park??
So, I wonder how the system works with inflection and stressed syllables. Would be a disaster for those domestic husband/wife disputes (not to mention Japanese which is almost *entirely* inflective):
*I* put the dishes away.
I *put* the dishes away.
I put the *dishes* away.
I put the dishes *away*.
Looks like we will still that Sprint guy hovering around for a while....
"Just because you're a genius doesn't make you a smart guy!" -- Narrator, Powerpuff Girls
All I told her was, "I wanna fig newton!"
"And like that
Not new, old technology. They even have a guy who uses a handheld one on South Park!
"Live Free or Die." Don't like it? Then keep out of the USA
And X-Ray machines are just big unmodulated radio stations broadcasting at higher frequencies, we should put those in every room of our homes too.
Color flashing, thunder crashing, dynamite machines.
Anybody ever pay attention to the sounds that the handlink makes on Quantum Leap? For example, it kind of goes 'waaaaahhhh' when he smacks it. That's the most obvious one, but if you listen a little more carefully, the sounds that little device makes start to emote. You can get an idea what he's reading on the screen before he actually states it.
Tom and Jerry is similar, to a degree. I ran across a cartoon of Tom and Jerry on the web a few days ago and watched it. I noticed something very interesting. The music in the cartoon responded to every little movement that the characters made. You listen to the music, for example, and tell if Jerry was tiptoe'ing or running. That was a very interesting dimension to Tom and Jerry. That is the type of element that would allow you to watch a slideshow of the show with the sound track and still keep track of what's going on.
This article was very interesting because I think it may be the start of making computer interfaces take advantage of audio responses that don't even require words. I've spent a great deal of time assigning different sounds in Windows to different events. For example, I have a very distinctive sound that ICQ makes when I recieve a message. I even went as far as to provide different people with different sounds. I noticed something very interesting, when I went to use ICQ on another machine, I ached to hear the sounds again. It was so strange not hearing them!
I hope one day Windows (or whatever OS I use in the future...) spends more effort into providing a sound-enhanced interface. That would truely provide better a better multi-tasking experience. It'd be cool if, for example, the window on the screen causing the sound was played through the right or left speaker based on where the window is on the screen. Maybe muffle it if a window is under it.
Anybody know of any products for Windows that do this today?
"Derp de derp."
I have a minor speech impairment (not very clear) so it would probably be useful to me. :)
Ant(Dude) @ Quality Foraged Links (AQFL.net) & The Ant Farm (antfarm.ma.cx / antfarm.home.dhs.org).
This way all you have to do is mouth words into the phone...not actually speak!
My daughter talks without saying anything. Maybe she could get a job testing these things.
Disconnect your television. Do your own research. Draw your own conclusions. They're probably lying. Don't be a sheep.
Now I can't mouth obscenities about the person I'm talking with without them hearing!!! You can't also hold a "quiet" conversation with the person beside you while "politely" listening to the person on the phone...
Oh well... my boss probably needs to know about what I call him behind his back anyways. q:]
MadCow.
I used to have a sig, but I set it free and it never came back.
Eloi are stupid, throw morlocks at them!
now i can say 'olive juice' to my wife and still earn major points. OTOH, everyone in Italy is gonna think that the entire US population is gay!
Another way to help speechless persons to communicate is the recognition and translation of sign language. If you're interested in that you might want to look here.
Are they going to call it "subvocalizing" like in Enders Game?
Travis
Why? What's the big deal? I can understand if somebody talks too loud, but that's true of anybody, not just a phone user. I was at a McDonald's once grabbing a bite, and I called my dad and talked to him for a bit. The woman in front of me got irritated and muttered that I should get off the phone.
I never did find out what sparked that. If I was talking too loud, for example, she could have just touched her mouth in the 'sssh' sybol and been polite about it. I don't think I was talking that loud. Nobody else even looked up at me. I think she just had a conception that people with cell phones are rude. Well my response to that is 'ITS NONE OF YOUR FUCKING BUSINESS.'
There's 0 difference between me talking on the phone or me talking to somebody in person. If it's okay for me to talk with somebody in person, but not on a phone, then there are some serious social issues that will arise down the road. I bet she'd be tickled to death if any of her kids called her out of the blue just to say hi, but I call my dad (who lives 2000 miles away) and I'm a rude jerk. If it's distracting to her to watch a guy talk to somebody that isn't there, then she can watch Quantum Leap until she gets used to it. I certainly am not turning off my phone for the simple reason that it displeases her.
If anybody is going to discrminate against people with cell phones, make damn sure the reason is unique to the cellphone itself. No phone ringing in a theater: Acceptable. No Cell phones in a Hospital because they interfere with equipment: Acceptable. No talking on a cell phone in a restaraunt: Unaccetpable.
"Derp de derp."
The "100% vowel detection" claim sets off alarm bells for me. Sure, pure vowels tend to show up on the face, but there are lots of characteristics of speech which occur down in the throat, or back of the tongue...how do they plan to distinguish between sh, ch, and j? S and Z? F and V? For now, they don't.
I also just don't see how the claim can be accurate. I can say the "ih" phoneme with my jaw in any position, and I can say "a e i o" without moving my cheeks or jaw at all. What gives?
Human lip readers need *context*, and lots of it. This one I'll believe when I can use the demo myself.
As for losing subtleties of communication, I think the real problem is in synthesis. I work on the opposite side of this problem, generating lip movements from audio (i.e. lip synching). A lot of the subtleties you might think you'd lose are actually there in both signals, the audio and the muscles. For example, you can tell when somebody's smiling over the phone, the change in the shape of the mouth makes the phonemes sound different. Shouting invokes different muscles from normal speech, emphasis might be picked up from the eyebrows, and so on. But even if you detect such things on the face, no voice synthesis engine is capable of rendering the accompanying vocal effects.
To think this will end up as mass market technology is plain wrong. Can anyone really think that people will stop "speaking" into their phones? And for what, to evade cell phone static? Come on. Usage of this technology will only really take off in very niche markets where there's an actual need, like those who's speach is affected frome one form or another. Those are the people who will really benefit from this. The implications there are incredible. Now where's my Crystal Pepsi?
On the other hand, I've sat 30 feet away from people talking to each other across a small table and heard every word, either because they're elderly and deaf or young and just loud -- they're often the same people who yell down the length of the bus to their friends ("YO, my man Duane!" "Yo, yo, yo, Tyrone!").
Nothing this side of the Cone of Silence will even dent these people. They were loud before telephones were invented, and they're not getting any quieter.
Keep track for a day of all the conversations you hear in public places. How many of them are "cell-yell", and how many are just plain loud people? (For that matter, how many of the ones yelling on the phone hang up and yell at their companions?)
This next song is very sad. Please clap along. -- Robin Zander
Inappropriate capitalization of Apparently randomly selected Words.
This next song is very sad. Please clap along. -- Robin Zander
"Read my lips. No new taxes (today)."
How can you make a phone call if you can't even spea-k.
Maybe speech recognition w/o speaking? Now all you have to work about is a repetitive strain injury of your facial muscles. ;)
The biggest trick the devil pulled was letting lawyers become politicians so they can write the laws.
I don't know Japanese myself, but I'm in the middle of reading The Japanese Language (no link; not carried by Amazon). One of the things that's discussed is how little mouth movement is required in Japanese, in contrast to other languages. So it's somewhat ironic that DoCoMo, a Japanese company, is leading the charge in this field.
Even in non-Japanese languages, guttural sounds like 'g', 'k', and German 'ch' cause very little muscular change--just watch yourself in a mirror some time. The article didn't go into much detail, but it may be infinitely more useful if the sensors paid attention to tongue movements instead of cheek ones.
:wq
I seem to recall a project like this designed for special forces that was funded by CIA and DARPA (Defense Advanced Research Projects Agency). There was also a "silent sound" device that could transmit acoustic information directly into the skull.
Of course there could be many applications for the delivery of this type of thing, but one of the applications that the CIA was interested in subliminal presentation of messages in peoples sleep while the silent transmission of information would obviously be useful to special forces teams that need to communicate without revealing themselves.
Visit Jonesblog and say hello.
How will this differentiate between voiced and unvoiced consonants? "Pat" and "bat" sound different but the two initial consonants are extremely similar outside of vocalization. Yes, the articulation of the "b" is longer than the "p", but it's really miniscule and probably differs from person to person. I wonder if this will take the tack of making the phone "learn" how to discern such, or will it make the person learn how to "speak" in a way that the phone "understands" (kind of like handwriting recognition versus using Graffiti)...
Although the article talks about getting 100% accuracy in discerning vowel sounds, the Japanese language is pretty simple in its vowels -- a, i, u, e, and o, and that's about it. What about vowel sounds like umlauted vowels that occur in European languages? Heck, what about African languages that incorporate clicks and creaky vowels?
This sounds like promising technology, but the article leaves a lot of questions that need to be answered. I guess five more years of research will help, though.
--
http://www.aikiweb.com - AikiWeb Aikido Information
...since the research is being done in Japan.
i s_being_said
Japanese has very few dipthongs.
A word that might be spelled 'Ao' using latin characters,(Â), would be pronounced as 'Ah-ow' (sort of).
Some words do change the vowels, but usually just by extending it. The word Tokyo isn't pronounced 'toe-key-o' as much as it is 'to-u-key-o-u'. The audible differences can be very slight, though. Possibly by sensing the muscle movements, it would be easier to discern the differences.
Another interesting capability would be the ability to discern mood. Consider the following:
'Yes dear, I'd <rolls_eyes>love</rolls_eyes>to have your mother visit this weekend...'
I'm not sure that I'd want my phone telling my girlfriend when I'm being sarcastic. You could have a new groupof 'tags' kind of like those you see on IRC:
roll_eyes
clench_jaw
check_watch
sneer
cringe
shake_head_in_disbelief_at_the_studidity_of_what_
You get the idea...
Cheers,
Jim in Tokyo
-- My Weblog.
Fool's gold. Especially considering this post was copied verbatim from a joke email that was going around not long ago.
Actually, I didn't have the email handy so I had to do it from memory but it should be close. I'm not seeking "gold" as I have been at the karma cap for some time now. I just thought that this was appropriate for this topic. You have to realize that not everyone gets the same email as you do.
Next time, be quicker and post it yourself.
Laugh...
Life is the leading cause of death in America.
The reason people shout into cell phones isn't that the phones don't pick up sound well enough. They do. It's that people don't *THINK* they pick up sound well enough because the phones don't give you any feedback in your own ear. Normal phones do give feedback and people are used to that. When you hear no feedback, you think "hey this phone must not be picking me up very well".
It may be a neat bit of technology they've come up with, but people won't stop shouting into their phones until they get feedback.
Yes, the English phonemes 'g' and 'c' are articulated in the same position, both dorso-velar (dorsum of the tongue contacting the velum, the flap of skin behind the palate). They're both also 'stops' (the passage is momentarily completely blocked). But discerning sounds of identical position is actually somewhat less problematic in English than it might be in certain other languages. You hit upon a really important point when you mentioned 'the air' which accompanies 'c' word-initially in English (called 'aspiration'). Khmer, spoken in Cambodia, distinguishes between aspirated and unaspirated stops (e.g., the first 'k' in 'kook' is an aspirated stop, the second is unaspirated, but English speakers don't distinguish between them). How could this system possibly tell the difference? The only difference between the first 'k' and second 'k' in 'kook', as you point out, is the quick expulsion of air which accompanies the first. Even more confusingly, the first 'k' in 'keel' is not even articulated in the same position at all as the first 'k' in 'kook'. 'k' in 'keel' is palatal (further forward), where 'k' in 'kook' is velar (further back). But, for some reason, in English, we consider them the same phoneme (the subjective perception of what constitutes a unique sound in a given language. 'Keel' and 'kook' start with the same English phoneme, because we can't tell the difference). This is just impressing the point that where a phone is articulated is only a tiny piece of the puzzle. Making a system which understands language on the basis of position alone is ludicrous. That's impossible.
As you point out a workable system would have to detect 'voicing' (the vibration of the vocal cords), as voicing, AFAIK, differentiates at least some phonemes in every language on earth.
What about nasalisation (where the nasal passage is opened in pronouncing a vowel)? The only difference between the French words 'main' (hand) and 'mais' (but) is that the first is pronounced with resonance in the nasal cavity. How is this system to divine that one has opened a tiny passage to one's nasal cavity for the duration of the vowel?
Speaking of point of articulation, how about glottals (articulated in the larynx) and pharyngeals (articulated in the pharynx. We have none in English, but they exist in Semitic languages)? Without a camera rammed down the subject's throat, sensing articulation in there is going to be hard.
If we have some way of determining the position of the tongue, vowels will be comparatively easy to distinguish, as they're distinguished by 'rounding' (i.e., of the lips), position of the tongue and nasalisation alone (a caveat: Japanese has a 'voiceless vowel', but it's a total phonetic red-herring, really). And detecting nasalisation still seems a difficulty.
At any rate, the idea of recognising language mechanically would seem to at least necessitate detection of 1) position and character of vibration in the nasal cavity, pharynx and mouth and 2) exact position of the tongue at all times. At any rate, I'll leave the last word on this 'invention' to others:
Phones that can understand your mouth movements will probably have to translate these movements into some sequence of sounds that correspond to the speaker's langauge words. What I mean is that the phone will have to KNOW what language you are speaking in order to be able to translate your mouth movements into sounds meaningfull for your language. I made a joke a couple of days ago about phones understanding their owners, I was with my girlfriend and told her that it is amazing what my cell phone heard from me so far, for example when I talk to her. Anyway, cell phone companies would love to sell these phones since it will mean more upgrade capabilities - do you want your phone to speak english? russian? german? japaneese? Ha! Another 99.99$!
You can't handle the truth.