Every once in a while a totally new way of doing something comes along. VoicePredict by TravelingWave is such an innovation, with a surprisingly simply but powerful method of improving the speed and accuracy of text entry.
Instead of doing straight voice recognition, the system uses voice and text entry to augment each other.
The software, which is currently in beta, features:
- Fast & Easy Mobile Text Input Using Accurate Speech Recognition
- Prediction of words and symbols using 1-2 key presses
- Enhances all types of interfaces: touch, qwerty keyboard, reduced-keypads.
- Automatic switching between voice powered text prediction and plain text prediction
- Seamless voice user interface based on an Always Listeningâ„¢ mode
- No requirements for special speech training
- Recognition robust to noise and variations in speaking-styles and microphone levels
- Voice Prediction speed: 75% reduction in number of key-presses on qwerty-phones, 85% reduction in key-presses on 10-key phones
- Voice Prediction accuracy: 90% for 1-letter input, 95% for 2-letter input, 99% for 3-letter input, and an overall voice recognition task-completion accuracy of near 100%
Of course it is questionably whether anyone would want to speak their facebook entry out loud, but as a way to improve the speed and accuracy of voice dialling I can see the software being pretty effective.
The software is being developed in Windows Mobile, which has always had an open plug-in architecture for data input, and of course would not be viable on the iPhone, where applications always run in a sandbox, largely unable to provide services for other apps.
Read more about the app at TravellingWave here, where one can also apply for a beta download.