If I were Apple I'd launch the iPhone with Nuance speech recognition

Stuart Corner
Sunday, 03 December 2006 10:34

Your IT - Mobility

After multiple 'sightings' in recent weeks the iPhone has taken on the mythic status of a Loch Ness Monster or a Yeti: no one is quite sure if it exists or what it looks like, or how it works. Neither am I, but I have some suggestions.

Everyone is expecting Apple to pull something dramatic out of the hat. And let's face it, Apple will have to do just that. The likes of Nokia and Motorola have been making cellphones for two decades or more and have got it down to a fine art. With the iconic status of the iPod Apple can't just come up with a 'me too'. Worse still it can't come up with something that looks like a forced marriage between a cellphone and iPod: It will need to be cool, funky and different. So, for what it's worth here is my suggestion.

A month ago, Nuance, the people who brought you the Dragon Naturally Speaking speech recognition software for PCs, launched the Nuance Mobile Speech Platform, billed as a "groundbreaking architecture of tools and components that provides mobile application developers with the ability to enhance their offerings with a seamless speech experience – from advanced speech recognition to natural text-to-speech – that dramatically enhances usability."

That mouthful really does not do it justice. In a nutshell what it promises is to render redundant any need to input commands to a cellphone using buttons or by navigating through screen menus: You simply talk to the phone. Your speech is digitised and sent via GPRS, or other packet data channel back to a central server which does the voice recognition. Whether you want to use 'phone' functions such as dialling or sending SMS, or you want to search for and download content such as ringtones song tracks and movies, Nuance can make the process hands-free.

Speech recognition has come a long way of late, on two fronts: the ability to recognise spoken words without prior training (Nuance claims 98 percent out-of-the-box accuracy for the latest version of Dragon Naturally Speaking) and the ability to understand free from speech.