Skip navigation

Interactive Avatars as on-screen characters, which are virtually undetectable from the real thing?

That's the claim from Televirtual MediaLab the UK technology company which combines award-winning TV and Computing activities.

The technique they're calling Videogrammetry merges conventional TV recorded behaviour patterns, speeches and gestures, with interactive versions of the same human subject that may be controlled automatically by a computer. Unlike simpler representational avatars that appear in games and Internet communities like SecondLife, the result is an on-screen automaton which accurately reflects the original. Studio acquisition is at HDTV fidelity.

An early example - Karena - may already be found, delivering regional news headlines and weather forecasts on the Internet. See www.televirtual.com/newsfast

Also See Karena image as attached.

'I wouldn't call them all-singing, all-dancing, at this stage,' said Televirtual's Creative Director, Tim Child, 'but they're certainly all-talking, all-emoting and progressively gesturing.'

Televirtual believe one emergent market for their work will be the creation of Internet Surrogates - web-based versions of major film, TV and public personalities whose value as 'brands' are such that hi-tech stand-ins could meet fan base demand for thousands of personal contacts.

To perform as surrogates, the video-based doppelgangers will be coupled to an advanced linguistic-based artificial-intelligence system, from Icogno Ltd. 'Machine learning, plus scripted purposefulness,' says Founder, Rollo Carpenter, 'allow our AI Creator package to bring artificial personalities to life.'

In such fashion the combined result of Televirtual and Icogno technologies is an on-screen/online character, which looks like the real thing, responds like the real thing and will conduct thousands of unique, concurrent conversations with its admirers.

'The weakest remaining link is speech,' says Child. 'Currently, the performance of synthetic speech, or text-to-speech engines, lags behind the reality values that can be brought to appearance and behaviour.'

But that's changing also. Custom speech engines - meticulously sampled from a real personality - are proving ever more convincing; a recent example being the Doctor Who (Tom Baker) voice commissioned from speech technologists, Nuance.

Televirtual and Nuance have been working closely to address the issues, and Child believes high fidelity cost-effective solutions are moving ever closer.
But in any case, the ambition is not to deceive, but to produce an interactive experience with unique informational and entertainment values.

Email: tim@televirtual.com Web: www.televirtual.com

Email: Icogno: rollo@icogno.com
Web: www.icogno.com or www.jabberwacky.com

This press release was distributed by ResponseSource Press Release Wire on behalf of Intermedialab Ltd in the following categories: Entertainment & Arts, Media & Marketing, Retail & Fashion, for more information visit http://pressreleasewire.responsesource.com/about.