Device promises ability 'to speak in tongues'

People could soon have conversations in languages they do not know

Reddit Stumble

 

Also In The News

Council to target gas-guzzling cars
Council leaders in a south-west London borough are considering plans to alter the area's residential parking permit system so that owners of high-polluting vehicles are charged more than those with more environmentally friendly cars.

• Global warming linked to Spanish frog decline
• Blair's climate change record condemned
• Carbon markets 'could save' tropical forests

 

Thursday, 26, Oct 2006 02:54

A new device claims to be able to translate words spoken from a person's mouth into a language they have not learned.

Researchers from the Carnegie Mellon University (CMU) in Pittsburgh, Pennsylvania, say that the device makes people appear bilingual by translating their unspoken words into synthetic speech in another language.

Electrodes are attached to the face and neck which are able to detect the patterns of electrical signals sent to the facial muscles and the tongue as the person mouths words.

CMU speech researcher Tanja Schultz told the New Scientist magazine that the effect is like watching a television show that has been dubbed into a foreign language.

The researchers claim that their new device is superior to existing systems as they are based on speech-recognition software that requires the user to speak the phrase out loud, making conversation stilted.

"The ultimate goal is to be in a position where you can just have a conversation," said CMU speech researcher Alan Black.

To use the device to translate from English, someone would have to train the system on the 45 phonemes, or building blocks, used in spoken English.

The device then comes up with a range of sequences and picks the ones with the highest probability.

But the researchers warn that there is still some way to go before the device is able to be used with no mistakes as at present it picks the right sequence only 62 per cent of the time.

Even so, Chuck Jorgensen, who is working on using sub-vocal speech recognition to control robots at Nasa's Ames Research Centre said that the device is a "very significant achievement" as it "is showing that the technology is really within reach".