AT&T develops speech recognition service for iPhone and other devices

Stephen Withers
Thursday, 24 July 2008 03:45

Your IT - Mobility

AT&T has demonstrated a speech recognition web service that can be used in conjunction with devices such as the iPhone and other smartphones, and web-connected set top boxes.

Currently a research project, Speech Mashups uses a new version of AT&T's WATSON speech recognition engine, and is implemented using standards including AJAX and JavaScript.

This arrangement means that speech recognition can be used on devices without having to install the recognition engine itself. Instead, speech data is collected by the device and sent to a server, which returns the corresponding text.

"This enables easy and rapid development of new speech and multimodal mobile services as well as new web-based services," AT&T officials said.

A video demonstrating the use of Speech Mashups with the yellowpages.com site along with an architecture diagram can be seen on the AT&T Labs web site.

At this stage there is no indication of how AT&T might commercialise the system. The choice of words in the statement - "a web service to economically bring speech processing technologies to the larger web and mobile developer community" - suggests it will not be completely free. One possibility is that web site operators could pay a fee to voice-enable their forms.

But what about voice dialling? See page 2.