h a l f b a k e r yRight twice a day.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Why stop at voice recognition for speed dialling... why not set
up a phone that can do voice recognition for all tasks...
especially sending SMS txt messages...
Voice Signal's ELVIS
http://www.voicesignal.com/prod2.html Embedded speech recognition for mobile systems. Runs on the ARM7 used in many mobile phones, but needs 3Mbytes data ROM, not found in many mobile phones. [pottedstu, Feb 19 2002, last modified Oct 04 2004]
IBM embedded ViaVoice
http://www-3.ibm.co...prise/ms_evvee.html Embedded as in "32Mb RAM Pocket PC", which is slightly above the computational power of most mobile phones. [pottedstu, Feb 19 2002, last modified Oct 04 2004]
[link]
|
|
Hmm, yeah would be nice, except for the usual issues with voice recognition. I suppose if you combined it with a more expansive, phrase based rather than word based perdictive text it might sort out the possible errors. Or it might just make for an amazingly garbled message.
Hold on, talking into a phone and having that message conveyed to another phone...wasn't this baked in 1876 by Alexander Graham Bell? |
|
|
More garbled than current SMS messages? |
|
|
I work for a big corporation involved in providing silicon and firmware for mobile phones. We have considered this, although we're first working on software to read out SMS messages, being as that's easier. Although SMS messages are composed of a garbled nonsense code, they have the advantage that being short, people are more prepared to deal with clunky and inefficient interfaces, which means they need a lower level of technology than speech recognition or text-to-speech for long documents. The big problem with speech recognition is the huge data tables you need, meaning vast amounts of memory, many times that needed by a GSM encoder/decoder (say 2Mbytes vs. 200 kbytes), so this can't be easily provided as a little add-on. So, although this isn't baked to my knowledge, give it time. |
|
|
ah .. well... forgive me Rods... i have no idea how it would
be done - speech recognition software ? its pretty half
baked... ! |
|
|
Text messaging has more advantages than just being able to communicate when you can't hear. It is usually much more pithy and can be saved in your phone (which is useful for those who have trouble with their memory). Also, it is a damn sight cheaper than calling another mobile. |
|
|
A good texter (to use someone else's parlance) will make use of predictive text, write in sentences and be able to amuse all in one message.
A great texter will use predictive text, sentences, punctuation, a smattering of a foreign language and their communications will be close to poetry. |
|
|
So, on the basis that this idea would turn texts into endless rambling answerphone message transcripts, I must respectfully fishbone. |
|
|
// The problem is that say if you are in a crowded area like a pub ,with music blaring etc you want to receive your message in some form that u can understand it ie SMS.
True George, however, I suspect voice activated SMS may suffer in this environment too - you may have stumbled on a real limitiation. |
|
|
Why not just have alphanumerioc characters committed to the phone memory, or an intelligent system to recognise phenomes and insert them into the message. Phenomes can't be harder to understand than the text message language used these days. |
|
|
SMS Dictation for Embedded Devices - this is the subject of my Master Thesis! This forum still gives me some ideas. So if you tell me your expectations - or ideas, I can implement it!
Thank you! |
|
|
Why not have somebody SPELL their entire SMS out, letter for letter. Example: T H A N K S -SPACE- F O R -SPACE- G E T T I N G -SPACE- B A C K -SPACE- T O -SPACE- M E. etc...
(each letter pronounced individually, as in a grade 1 phonics/English class). |
|
|
That way, the memory required by the phone to store each letter would be far less than the memory required to store an entire vocabulary. Also, there would be less CPU requirements as all that would need to be processed would be the numbers 0-9 and letters A-Z. That way, while driving, walking, whatever, you can spell out your SMS and have it sent in half the time. Speak fast, the system will buffer it and process in time. Phones are already capable of analysing and processing numbers for voice activated dialling, so why not extend that and allow the phone to process individual alphanumeric characters. Hands Free Kits will start flying off the shelf... |
|
| |