Click here to register.

General Discussion

Flat
New iPhone voice dialing apps
User: kmaclean
Date: 9/29/2008 9:56 pm
Views: 3458
Rating: 9

Here are a couple of new apps for the iPhone that use Open Source speech recognition engines.  Although the apps themselves are closed source, it is interesting to see what can be done with PocketSphinx and Julius.

Cactus Voice Dialer (no cost - based on PocketSphinx)

from their website:

Cactus is a free voice dialer for the iPhone. It allows you to 
  call any contact in your address book by saying commands like:
  • "Call Kate Bell"
  •  "Call Anna mobile"
  •  "Call David Jones at home"
Because Cactus runs completely on your device, it does not send any private 
  information such as contact list to any server and works fast even on an EDGE 
  data connection.
  Customizable settings allow you to:
 
  • tune the system's sensitivity by displaying one or several results for each command
  • set the system to automatically launch calls or not
  • use Main numbers by default for contacts that have several numbers
Cactus was written by Antoine Raux. It is based on PocketSphinx, 
  an open source speech recognition engine for embedded devices developped at 
  Carnegie Mellon University (big thanks to David Huggins-Daines, the main author 
  of PocketSphinx).

Vocalia ($3.99USD - uses the Julius engine)

from their website:

Speech recognition running directly on your iPhone.
Do you have many contacts in your iPhone's address book?
Finding quickly the one you need is quite difficult and boring.
You should add a powerful speech recognition tool to your iPhone!
Vocalia helps you find quickly and easily the person or company you want to call or send an email to.
No network connection required.
No private information sent to some unknown and far away servers.
No additional work required from you to record your contacts' names. 
Key features
Vocalia...
• Uses Julius engine for multi-locutor, speech recognition (English only at this time)
• Runs entirely on iPhone, works without network connection
• No "Tap to speak"
• Provides default working phonemic transcriptions for all your contacts
• Shows up to 10 recognized names

• Displays visual effects responding to sound and gravity (a.k.a "Vocalia's Eye").

With Vocalia, you can... 
• Define additional vocal shortcuts ("My best friend") in addition to the name ("John Doe")
• Split first and last name (saying "John" will display all corresponding entries of your address book)
• Shake the iPhone to try again.

--- (Edited on 9/29/2008 10:56 pm [GMT-0400] by kmaclean) ---

PreviousNext