Google Personalizes Voice Search

Each human being has a different voice, a unique way of talking, a slight variation in pitch or speed or accent. Today, Google announced personalized voice search, a program that will begin to learn and recognize how you–as an individual–speak.

If you opt in to Google personalized recognition, the application will record words that you ask it to record and file them into a memory bank that will “build a speech model particularly for you.”

According to Google, “This speech model enables us to deliver greater recognition accuracy. Although subtle, accuracy improvements begin fairly quickly and will build over time.”

Personalized voice association, as it is also called, will be available to Android 2.2 users in the United States starting today, and will eventually spread to other countries and languages.

According to a Google response in Read Write Web, the enhanced recognition will come from a vast statistical model that covers all of the sounds of a language (phonemes). It will take that model and compare it to what you said to come up with its best guess at what you said. Over time, the guesses become more honed, more adept, until finally Google has a complete model of your voice.

Google’s response in Read Write Web read, “Knowing what you said in the past allows us to build specialized models that are designed to match your voice and your words.”

The update is free in the Market as I’m sure Google can’t wait for you to get in there and start talking to it. It’s learning all of our locations, preferences, and now our voice patterns by Osmosis. Given Eric Schmidt’s plans for the company, I wouldn’t be surprised if Google turns into the real-life version V.I.K.I. from iRobot.

Sorry, comments are closed for this post.