A system that allows drivers to listen to the contents of e-mails, text messages and other computer-based documents without taking their eyes off the road has been developed by UK speech technology specialist 20/20 Speech.
Field engineers who need to listen to instructions while carrying out complex operations represent another potential user base.
The company claims its Aurix TTS product is the first text-to-speech synthesiser suitable for use in handheld computers and other small mobile devices.
This is because it uses a fraction of the memory needed for other speech synthesis systems, which rely on a massive stored database of recorded words and phrases actually spoken.
These are ‘cut and pasted’ together to produce intelligible speech, but the huge memory needed to run such systems means they are only suitable for server-based computing networks.
Permanently connecting mobile devices to servers via a GSM phone link can be costly and unreliable, especially if a user strays out of network range.
Instead of a database of pre-recorded material, the Aurix system simulates speech by using models derived from the frequencies of the human vocal tract. It then applies these to text-based data to produce speech.
the result, claims 20/20 Speech, is not only more intelligible over long listening periods, but also has a low enough memory requirement to be installed directly onto mobile hardware, eliminating the need for connection to the server.
Dr Roger Moore, chief technology officer at 20/20 Speech, said the company hoped the new system would find a wide range of applications among people needing to access information in a ‘hands-free, eyes-free’ situation.
‘The most obvious users are drivers who need to listen to the contents of a text-based message without taking their eyes off the road,’ said Moore. ‘But we think there are many areas where people are highly engaged on other tasks with their eyes and hands where this technology could make a real difference.’
Aurix TTS is undergoing testing by mobile hardware manufacturers and systems developers in a range of specific industries, including automotive, aerospace and logistics.
Moore said the new product was one of the most significant yet for Malvern-based 20/20 Speech, set up in 1999 as a joint venture by Qinetiq – the technology group spun out of DERA – and NXT, the London-based audio specialist.