Search by voice from Google has just gotten a little bit smarter, with a new opt-in feature that allows the cloud to adapt to your unique speaking and pronunciation style for greater recognition accuracy.
When Google introduced Voice Search for iPhone about two years ago, it was a prime example of how the cloud can take search to the next level. In a nutshell, the app lets you search by speaking aloud your query into the handset's microphone. The software then uploads a digital voice file up to the cloud where speech recognition algorithms turn data into a regular search query. A few seconds later, text results appear on your handset.
Voice Search was later released on Android and Google kept perfecting speech models to accommodate a wide variety of people, regardless of gender, age, and accents, or variations in pitch, pace, and other factors.
Today, Google enhanced Voice Search with what they're calling personalized recognition. The updated Android app now sports a new opt-in switch for personalized recognition that associates your dynamic voice profile with your Google Account, the company wrote in a blog post:
If you opt into personalized recognition, we begin to associate the recordings of the words that you ask us to recognize with your Google account. We then automatically use these words to build a speech model specifically for you. This speech model enables us to deliver greater recognition accuracy. Although subtle, accuracy improvements begin fairly quickly and will build over time.
Needles to say, privacy advocates could list dozens of reasons why such a feat might file as a privacy nightmare. Google on its part says you can always disassociate your voice recordings from your Google Account, by clicking the Speech section in the Google Dashboard. If you plan on using personalized recognition, bear in mind that it only works for US English at the moment - and only in the Android app - until Google rolls out the feature to additional markets "in the near future."
Source: The official Google Mobile blog