Google has updated its Voice Search app for Android 2.2 or higher, adding an interesting new feature: personalized recognition. Once enabled, the feature will tailor its speech algorithm to a linked Google account by recording the words that you ask it to recognize and use them to automatically building a custom speech model. The speech model will then in turn deliver greater recognition accuracy over time.
The personalized recognition functionality is currently only available in English (United States) though Google is planning to support other countries and languages in the near future. The new app also includes some improvements to name recognition and speed, especially over 3G and EDGE connections.
The first time you launch the Voice Search app, you'll be asked to turn on personalized recognition, which you can disable at any time in the Google voice recognition settings. You can also disassociate your voice recordings from your Google Account via the Google Dashboard in the Speech section.
Google Voice Search launched more than two years ago, but only now is its accuracy being improved. The app still has speech models that are broad enough to accommodate factors such as gender, age, accents, variations in pitch, pace, and so on. With personalized recognition, though, the app will try to learn your voice and to provide more accurate results.