Change of Voice: Interview with voice-to-content developer SpinVox
The choice of channels in which humans have to communicate with and update their friends, family and associates has grown inexorable in the last decade. One growing channel is the instant conversion of voice-to-content. New Media Knowledge met a UK innovator pioneering the technology.
The recent Mobile World Congress (MWC) in Barcelona brought together some developers from all spheres of the mobile sector. One exhibitor was UK voice-to-content firm SpinVox whose technology converts the spoken word into text which is used mainly either to transform voicemails into SMS text messages and emails or to blog and microblog. To date the company claims to have enabled more than 100 million message conversions.
NMK met up with the company’s social media specialist James Whatley at the show to learn more about the potential for voice-to-content technology and where it’s currently being used.
Out of Time
“It’s seven times quicker to ‘read’ voicemails than listen to them,” Whatley argued. “That’s a big saving for the consumer over listening to their voicemails.”
SpinVox has developed its own Voice Message Conversion System (VMCS) which the company says fairly represents voice as text content. VMCS differs from Automatic Speech Recognition (ASR), which the company says works well for desktop dictation but is not apt for the telephone as ASR cannot recognise meaning. VMCS contains artificial intelligence semantic systems that assist in finding a meaningful answer.
“The system understands nuance,” said Whatley. “There are around 600,000 words in the English language and a lot of slang, so the system needs to understand and process this information accurately.”
The system can record voicemails up to three minutes long and has so far saved users more than three million hours combined, the company says.
Voice-to-text conversion is tailor-made for the microblogging era, according to Whatley, and even celebrity microblogger Stephen Fry is a user.
“We’re working with the likes of Ping.fm, Facebook, Twitter and MySpace to enable people to share their moments with friends,” he said. To illustrate, Whatley instantly updated his Twitter status via his mobile phone to read that he was currently in an interview with NMK.
“People want to update on the move and talking is so much easier and quicker than typing it all out,” he added.
The world’s first ever voice-to-text sermon took place in November last year at St. Christopher’s Church in Esher, Surrey. The Reverend John Kronenberg’s sermon was delivered via SpinVox directly into parishioners’ mail inboxes live, which enabled church goers who couldn’t attend to hear the service.
SpinVox’s voice-to-content system is also being used by various organisations to better service youth audiences. Debate Mate is the first national peer-to-peer debating programme in the UK and encourages state schoolchildren to volunteer in discussions to “celebrate the power of oral expression”, SpinVox says.
Music television channel MTV also uses the company’s voice-to-text technology extensively. The channel’s BOOM! campaign encourages young people to use creative media skills via their local youth group. MTV also helps its viewers to share stories anonymously to think more about sexual health.
Lost in Translation
The company announced at MWC that is has launched Portuguese and Italian versions of its voice-to-content technology, adding to its existing English, French, German and Spanish offerings.