SOFTWARE HOUSE Google is updating its Chrome web browser to add voice recognition.
The latest build of Chrome beta features support for the HTML5 speech input API, which Google has been working on with the HTML Speech Incubator Group.
The API will allow developers to write web apps for Chrome in HTML5 that allow full speech recognition at the click of a button. The transcription occurs on a speech server after the sound is recorded.
We tested out a demo of the speech recognition option, which works very similar to that within Android. The first thing we noticed is that it's pretty fast, returning text transcriptions within a second of inputing speech.
It recognised "hello" and "how are you". When we said "The INQUIRER" it gave us "the enquirer". We said "for all your technology news" and it returned "phone technology", which suggests it might still be a little rusty around the edges.
Of course, the nature of speech recognition software means that the faster you say something and the more of an accent you have the less likely it is to pick it up accurately. We tried the last phrase again in a slow monotone voice, which it picked up accurately. This might satisfy some, but we still dream of a day when we can talk normally at a reasonable pace while the computer picks it all up completely accurately.
Google is also adding a GPU-accelerated 3D CSS feature, which allows developers to render 3D effects in CSS on web pages. µ