Dear Teddy, Automatic Speech Recognition (ASR) it's a really amazing field in CS, and also a very broad one. Many things can be achieved with a OS5 device (they definitively have the required processing power for them) but some don't. I guess the first thing you must decide is what category of ASR do you need: - Speaker-dependant (requires training) or s-independent recognition (no training required, large databases) - Discrete (single words, small phrases) or continuous (detect words within phrases) recognition - Dictation or grammar-based recognition - In-the-Palm recognition or server-based (i.e. in a PC) recognition
Some combinations are definitively possible: speaker dependant, discrete, in-the-device recognition can be achieved with (relative) low processing power (like cell phones, that have the name dial feature), Microsoft recently announced Voice Command for PPC in equipments having comparable CPU power like T3. We have also played with PC-based ASR and on-the-fly recognition, streaming audio from the Palm. Contact me privately if you need more details on this. I guess the higher feature levels you may need, the more R&D involved. Some levels are reachable only for a few "selected" ;-) players: to have a good speaker-independent ASR engine took 10+ years to Microsoft, more or less the same as to some other key players like IBM, Nuance, ScanSoft ASR (formerly Dragon ASR), Speechworks, Philips, etc. IMHO, Fonix (http://www.fonix.com/) is the company with more experience in embedded ASR, if I'm not mistaken they have an ARM-based ASR engine. Sensory Inc. also have some of this. And for more information on the Computer Science state-of-the-art research status, Carnegie Mellon University is the best place to start (and they have hundreds of links to other research labs). HTH, Cheers, Sebastian Massacane [EMAIL PROTECTED] |-----Original Message----- |From: [EMAIL PROTECTED] |[mailto:[EMAIL PROTECTED] Behalf Of Teddy |Law |Sent: Monday, January 05, 2004 9:20 AM |To: Palm Developer Forum |Subject: Seeking for help | | |Dear all Experts/Professional, | |I'm just a normal user on Palm, but my boss had assigned me to evaluate the |feasibility of developing speech recognition application on Palm |device such |as Tungsten T3. | |I bought several Palm programming reference books, and surfing on the |Internet, but it's still too little information on this topic. | |After the research, I only know that Sound Manager API are included in Palm |OS Library, but it still so virtual for me to answer my boss the |feasibility. | |I have many queries, such as... |1. Does Code Warrior 8.0 can handle this task ? |2. Which version of Palm OS start to support Sound capturing? |3. Do all models of Palm have build-in microphone can fulfil my need ? |4. Does it really feasible ? |... | |Does any expert or professional can give me a helping hand ? | |Thks & Rgds. |Teddy Law | | | |-- |For information on using the Palm Developer Forums, or to |unsubscribe, please see http://www.palmos.com/dev/support/forums/ | -- For information on using the Palm Developer Forums, or to unsubscribe, please see http://www.palmos.com/dev/support/forums/
