I came across this the other day, a really interesting article about the advances in speech technology for GPS navigation instructions and the impact on driver behaviour [read it here]. Basically, the author summarises the state of play and development for text-to-speech technologies and the use of computer generated voices in satnav. He concludes by illustrating how crude the systems still are in terms of getting emotion into the computer generated GPS voices. Contrast this with pre-recorded GPS voices, like our very own VoiceSkins, which manage to convey a huge range of emotions in each character’s satnav voice: Darth is threatening, seductive, menacing; Homer is goofy, defensive, self-important; Snoop is laconic, critical; Yoda is quizzical, authoritative, encouraging; C-3PO is concerned, obsequious.

Why would you want to listen to a voice that fails to keep your interest and that can’t reproduce anything like the emotional range of a real human voice? I’ve always wondered why you need a text-to-speech voice to tell you to ‘turn left into Acacia Avenue’, as:

a) you rarely have the ability to catch a street sign whilst driving, in which case the additional location specific addition is unnecessary;

b)if you need your GPS to guide you then you probably don’t know / don’t care about the name of the road you are about to turn left into – all you need is a simple instruction like, ‘take the next left’.

Text-to-speech will have to be able to reproduce a far richer emotional range than currently as well as provide the option of real celebrity GPS voices like Homer, Darth Vader, Snoop Dogg, Yoda and others before it will be a real contender IMHO!

Chris.