With the advent of, apparently, a workable pet translation device (see link), it is time for the inventors of our age to work on a problem that has plague humanity since the dawn of slang - the inability to understand someone who, supposedly, speak the very same language you do.
Similar ideas have been posted about before, and yet none have gone to the crux of the matter and attempted to devise a device that could logically take the destruction of a native tongue and convert it into a pure form.
So, with the current developmental rate of voice recognition technology (such as the new generation Nokia phones. Quite good - I've heard a female deepen her voice to use a man's voice activated recording: and it worked!) and with wireless communications developing at an equally sufficient rate, it should technically be possible to build one now, let alone in a few years once the technology has had a little more time to mature.
The main stumbling block is finding people able to create a large database of slang and other such words and compile them into one place. Once this is done, the software for the vocal translation would simply need to run the input sample (the speaker) at a few different pitches (to allow for people with deeper or higher voices) and then take the match and transmit the translation to the listener.
The system could use the mobile (cell) phone network technology in place, similar to the on demand WAP access. For example:
Person 1: Yo, sup my homie!
Person 2: What? (Points translation device - a unidirectional microphone attached to a cellphone - at the speaker)
Person 1: I said, Yo, sup my homie! (Speaking sends words via cellnet to database servers. Phrase is translated and sent to the holder of the device via a standard earpiece much like the one currently seen on mobile phones. Person 2 hears "Hi! How are you my friend?" instead)
The beauty of this is, however, that with a sufficient database and server hardware, you could use this for language translations (though you would get the 'babelfish' effect, no doubt) as well as customisable translation effects.
Examples:
* Removing Politically correct speech * Removing swearing * Having everyone you hear speaking in archaic english
Those are just a few basic ideas which you could do.
Once the technology became wide spread and advanced enough, a set of wireless headphones could be developed (much like the hearing impaired aid device) that intercepts spoken words, translates them and THEN passes it on, much like the HHGTTG babelfish did, though admittedly less reliably.
While not possible to date, the technologies are advancing at a rapid enough rate that I don't see this being halfbaked in 10 or 15 years from now.-- Freelancer, Mar 23 2003 Dog Translator http://www.reuters....ews&storyID=2424064So for Dogs, why not people? [Freelancer, Oct 05 2004, last modified Oct 17 2004] Bork Bork! http://www.google.com/intl/xx-bork/ [madradish, Oct 05 2004, last modified Oct 17 2004] Kind of the other way around http://www.super-fr....uk/ask/index.shtmlSee the Emineminator [Bert6322, Oct 05 2004, last modified Oct 17 2004] Hippo's Valley Filter http://www.80s.com/.../Slang_20TranslatorLike, Totally not the same, y'know? [yamahito, Oct 05 2004, last modified Oct 17 2004] Urban Dictionary http://www.urbandictionary.comDictionary of slang (often vulgar) [Acme, Oct 17 2004] //Having everyone you hear speaking in archaic english//
Yes! I want one!
It would also be funny to make my boyfriend sound like the sweedish chef from the muppets.-- madradish, Mar 24 2003 This would be a great invention to help me understand Snoop Doggey dogs language........You get the "Schnizzle dizzle???"-- theThinker, Mar 24 2003 <re:mandrish's link>Can't stop laughing. Those crazy googleers. How many hours did they waste on that?</>-- Worldgineer, Mar 24 2003 Definitely a Wibni. Not least because slang, like technical jargon, is specifically invented to keep outsiders from understanding it.-- DrCurry, Mar 24 2003 [DrCurry] Sure, but isn't almost everything posted here a WIBNI? The way I've tried to explain it the idea should be viable, though expensive and clunky.
Cut down explanation:
* Someone speaks. * Speaking is picked up by microphone and sent by on-demand WAP (which exists, though is slow atm) to a remote server location * Remote server runs sample through voice recognition software (which exists, and is rapidly getting better) (varying the pitch, frequency, tempo etc of the sample) until it gets a match for each word against the neutral voice sample. * Server grabs the translation and uses the same pitch etc modifications to make the neutral sample as close to the submitted sample as possible * Server sends sample via the mobile/cell network to the person who submitted the sample with their microphone. * Person hears a translation of what the person said.
Again, like I said, the hardest bit is setting it up - both in money, the program, and getting a sufficient database.
I never liked jargon any way...-- Freelancer, Mar 24 2003 "rapidly getting better" *cough*-- bristolz, Mar 24 2003 <admin> - changed category.
I admit to uneasiness on this one - since slang is pretty much just another language/subdialect, would this machine not just be the universal translator (tm) - making this WICTTISITMWhatever... I have to admit, I would have thought it was redundant, also, but despite having much common ground with above ideas, nothing seems to cover it in the same way - I'll leave the rest to better brains than mine...-- yamahito, Mar 24 2003 *Grins* It is, bristolz.
Just from experience alone, the voice recognition on my nokia 5510 is far inferior to my collegue's nokia 6310i, where his wife was able to lower her voice an octave and could then access the voice dials which he recorded in his own voice - that isn't possible with the recog I have on mine, where it barely recognises MY voice, let alone someone else deepening their own.
[edit] Ah, thank you [yamahito]. I hadn't seen this catagory! I had noticed similar ideas about - and to my knowledge the UT from the Trek series' operates by trying to identify the syntax of the language, wave a magic wand, translation occurs. This is a much more basic thing that requires people to actually create a database first and even then I admit you'll get the babelfish effect. I thought that you could choose what dialect or jargon you wanted to *hear* was fun, too [/edit]-- Freelancer, Mar 24 2003 Well, from my non-techie standpoint, I see that the main problem with this is that the database would very rapidly become limited in usefulness without constant in-field slang gatherers* involved. Where I stay, slang appears to be in a state of flux, with descriptive and confusing terms being minted and slipping into the Patter every day. Equally, a large number of phrases will be lost for any one of a number of reasons (need for term disappears, reference in slang loses popularity etc).
Just thought of another potential problem - accents. Similar terms can be used globally (where globally = over more than one locale) but pronounced differently. This'd mean that each term may have to have a fair number of entries, increasing the size of the database.
[* originally typed as "slag gatherers"]-- my face your, Mar 24 2003 I think that'd be a fun job to have: Slang Gatherer.
Oh I admit it would be a *huge* task, but it's not utterly impossible like the UT from Trek seems to be.
To pay the gatherers and software and hardware sooner or later you'd get advertisements...now *that* is insidious.
Person 1 says: Hi friend! How about I buy you that drink I owe you.
Person 2 hears: Hi friend! How about I buy you a nice, cool refreshing Heinekin, brewed from clear mountain spring water and the finest hops, I can't go past a Heinekin when I choose my drink.
...or...
Mother says: What do you kids want for dinner?
Children say: KFC! Pizza Hut! Burger King!
Mother hears: McDonalds is the only place we want to eat at, Mother dear!-- Freelancer, Mar 24 2003 The mechanics of voice recognition may indeed be getting better but the translation part, the semantic recognition and all that needed to do what you propose . . . loooong way away.-- bristolz, Mar 24 2003 mm-hmm. You'd need some form of syntax recognition to address some of mfy's points - which is what Bristolz is referring to, I think. And that would make it a bit close to the UT, wouldn't it?-- yamahito, Mar 24 2003 Not really. You can use voice pattern recognition combined with a much larger database to cover the most common accents - so a swede speaking swahili is unlikely to work, but a german speaking english or an englishman speaking german are viable alternatives.
There is voice recognition that caters for different accents after all - the Dragon Naturally Speaking software, for example, comes in an American and Australian version (due to the different vowel sounds).-- Freelancer, Mar 24 2003 In australia alone, I can think of three or four different accents. In America, there are many more. Here in the UK, the accent changes from village to village, never mind county to county.
Naturally Speaking only has to cater for one specified voice pattern (per user, anyway): your idea has to be ready for them all - you see my point ;op
BTW - hope I'm not coming across like I'm just shooting down this one idea...-- yamahito, Mar 24 2003 Not at all - and I realise that the database would have to be *huge* (hence why the idea is halfbaked, not baked :p).
By having the software manipulate the samples to cater for some of the more common accents it may reduce your sample requirement size but would increase the server time needed to analyse a single sample (stretching the sound, changing the pitch, frequency etc etc)-- Freelancer, Mar 24 2003 Eye av fort get buzz morn g' thospitul, az getten reet bad yeadawarch.
Bin men bin mam?
'Tint in tin.
Am tekkin th'astra n'gooin croppers. Eye av fort gerruz a pow.
etc.etc.
Exactly whoooo do you imagine is going to program all the accents and dialects and slangs of the known world into a crappy mobile phone thingy??
Speech recognition is one problem, machine translation is quite another. The systems that are available today are a v.v.v.v.v. long way from being able to do anything even remotely approaching this.
Its the Babelfish, hence magic, hence -.-- squeak, Mar 25 2003 Spin hek, ma babba flupe dey. Huh? Your translator can't translate that? Pfft... That's some jemp fuh!-- snarfyguy, Mar 25 2003 Having spent several months in Australia where they are supposed to be speaking English, I would welcome a device such as this. Could have kept me out of several fights, I'm sure. Even if it could be a typed database... +-- ato_de, Mar 25 2003 did some 240lb aussie shove you in the river by any chance?-- po, Mar 25 2003 [squeak] Not at all, read some on the anno's I made - I realise it would be unwieldy and you would get a babelfish effect for other languages at times, but where slang is different is that it tends to have the same grammatical construction as the normal language, simply different words (hence less of the babelfish - systran - effect)-- Freelancer, Mar 25 2003 As a side note, the local communications company (Telstra) in Australia has a voice activated phone menu system. It works at times moderately well regardless of the accent you have (I tested it by speaking with a bad german accent)-- Freelancer, Mar 25 2003 snarf, that's McCartney talking isn't it?-- waugsqueke, Mar 25 2003 //Having spent several months in Australia where they are supposed to be speaking English//
Nah mate, we speak Aussie.-- madradish, Mar 26 2003 Step, [cuban_chi].-- k_sra, Feb 26 2004 Yo homeboys there ain't nothing wrong wit da way i spit me words out and if you don't vibe wit dat den you is just a square-- dan23, Feb 26 2004 A girl showing her undies-- killo, Aug 16 2004 random, halfbakery