A Brit taking on Apple’s Siri with ‘Evi’ (Wired UK)

Evi vs. Siri speech recognition comparison 

This article was taken from the June 2012 issue of Wired magazine. Be the first to read Wired’s articles in print before they’re posted online, and get your hands on loads of additional content by subscribing online.

The quest to beat the Turing Test, which tests a machine’s ability to engage in intelligent conversation with a human, took a step forward in January with the launch of Evi. Developed by Cambridge-based company True Knowledge, Evi has gone head-to-head with Apple’s Siri and Google’s voice-search app. But Evi — available for iPhone (all models) and Android - claims a better understanding of real English and returns results in the same conversational style. Ask it a question (by speaking or typing) and it aims to give a real answer, rather than a list of links (Google) or the suggestion that you use the web (Siri).

The app uses the same Nuance voice-recognition technology as Siri, but relies on a far wider knowledge base. “The platform draws on a vast store of ideas, lexical uses and common sense that lets it understand a plain-English question much more accurately,” says William Tunstall-Pedoe, 43, True Knowledge’s CEO (above). Using the phone’s location capability, you can ask Evi “Where is the nearest petrol station?” and get an actual address.

Tunstall-Pedoe is something of an artificial-intelligence prodigy; he has been making and selling software since he was 13, such as Crossword Maestro.

Evi was almost victim of its own success — more than 250,000 downloads in its first week meant early users found its servers often unresponsive, before more server space eased the congestion. Extra features are on the way, including Siri-like command functions for Android. Tunstall-Pedoe is positioning Evi as a multipurpose platform. “A personal-assistant product is a framework in which you can do almost anything,” he says. “Evi is the final app.”

via http://www.speechtechnologygroup.com/speech-blog - Evi vs. Siri speech recognition comparison  This article was taken from the June 2012 issue of Wired magazine. Be the first to read Wired’s articles in print before they’re posted online, and get your hands on loads of additional content by subscribing online . The quest to beat the Turing Test , wh ...

Radio World: FCC Course Change on TTS Takes Hold

For now, emergency alert system manufacturers don't need to provide text to speech functionality.

FCC Course Change on TTS Takes Hold

on 05.08.2012

     

The FCC has changed its Part 11 rules governing the Emergency Alert System so that broadcasters may, but are not required to, use text-to-speech functions for the Common Alerting Protocol delivery of alerting messages.

Publication in the Federal Register triggered the change, which went into effect May 7. RW reported in April that this was the FCC’s intention.

The commission previously had banned TTS, saying it didn’t have a lot of information on the record that convinced the agency “it could go another way,” the FCC’s Tom Beers told attendees at the recent NAB Show.

Broadcasters, EAS equipment manufacturers and FEMA lobbied the agency to change its stance, saying without the use of TTS, some equipment might just pass along alert tones, but no audio.

The commission did state in April that it was deferring action, not making a final decision; so it’s possible that it might revisit this question at some point. 

via http://www.speechtechnologygroup.com/speech-blog - For now, emergency alert system manufacturers don't need to provide text to speech functionality. FCC Course Change on TTS Takes Hold on 05.08.2012       The FCC has changed its Part 11 rules governing the Emergency Alert System so that broadcasters may, but are not required to, use text-to-speech f ...