New talking heads on the block

She was an absolute hit on the web in 2000. On April 19, 2000 she spoke the famous first words: "Hello world. Here is the news. And this time it's personal." It was Ananova, the synthetic figurehead of the British Press Association, performing as a woman newscaster, who reads the personal selection of news items to a web user 24/7. Her designers Digital Animation described her as 28 years old, 1.75 meters tall, pleasant and intelligent.

Her appearance sparked excitement. She was a technological milestone; computer-simulated animation, syntetic voice. She was an embodied agent. The speech system was the speech synthesis using rVoice from Rhetorical Systems, now Nuance Communications. This generated quite some interest. It looked like the text-to-speech industry was about to break through. Around the time of Ananova’s launch I was in Cambridge (UK) at the BT laboratory, where they demonstrated text-to-speech applications for mobiles. Funny enough the applications have never become real commercial services.

For the press world she was the symbol of innovation: personal news delivered by a personality. Besides services like personal news services on internet, the mobile world showed interest. A combination of personal news and mobile would deliver a new service. Just imagine: every morning the personal news selection is downloaded to the mobile phone; you start driving, a pretty person appears on the screen of the mobile and starts reading your personal selection of news items.

It only lasted two months before the mobile company Orange picked Ananova up for 126 million euro. In 2002 they started a news services for their mobile subscribers, but this was just a text service. They kept Ananova alive till 2004, but never did implement the technology behind it for a mobile news service. In 2004 the company discontinued the animated service on the web with the announcement that “Ananova video is currently under development. Come back soon to check the latest.” Three years later nothing has happened and the announcement is still on the page. By now Ananova is out of sync with the web and never transformed the delivery of news.

Seven years later there is another try in the UK. If Ananova had been live on the web, one could have introduced her to Brian first and more recently to Karena. Brian and later on Karena are the talking heads of EDP24 in Norwich. First there was Brian, a rather dull newscaster. He was replaced by Karena, a pretty lady. The technology differs from Ananova. For Brian and Karena, real people have been filmed. Using a combination of state-of-art TV technology and advanced graphic computing skills, the award-winning team at Norwich's Televirtual MediaLAb, have transformed Carina into the high-tech Karena.

The technique, videogrammetry, involves splicing and blending taped TV sequences of Carina, with reconstructed still images of a 35-year-old housewife. Reconstructed, because, because individual features such as eyes, mouth, nose and eyebrows, have all been captured piecemeal to allow interactive computer control of the assembled image. Karena replaces Brian who came online on March 14, 2007, as part of a unique project pioneered by three Norwich companies and a national sports information service.

Compared to Ananova, Karena is less slick and changes in the movements of the face can be noticed. The speech technology is from the same company as was Ananova’s. The Karena system will need some more fine-tuning graphically. Despite the fact that Karena looks artificial mostly due to the changes in the face, the technology might go someway.

Blog Posting Number: 910

Tags: newscast, talking head, text-to-speech