Google Is Improving Speech Recognition On Android

One feature of Android that sets it apart from other smartphone operating systems is its speech recognition and the ability to use speech to text with any text input field. From the moment that Google first added speech recognition about two years ago, it did an incredible job of correctly translating what I say.

When Google first implemented speech recognition it used a standard set of speech models that represent a broad group of users. Google has now enhanced speech recognition by adding the ability for it to build a speech model that is specific to how you talk. The result should be even more accurate recognition of your speech over time. To get the new personalized speech feature you need to install the Voice Search update.

The speech model is built on Google’s servers, which means that samplings of your speech will be stored on their servers. Because there might be privacy concerns, Google has made the personalized speech recognition opt-in. If you want to take advantage of personalized recognition on your phone, you need to tap Menu, Settings, Voice Input & Output, Voice Recognizer Settings, and select the Personalized Recognition option as shown in the screen shot. If you turn on the setting and later decide you don’t want Google collection samplings of your voice, you can turn the service off and delete the data using the Google Dashboard.