Google Inc. was finally awarded a patent for speech-recognition technology that allows users to search the Internet by uttering a few words. But whether that’s good news or bad news for a handful of wireless software developers is far from clear.
Voice-activated technology-which for years has been “the next big thing”-has gained attention in wireless recently as the industry looks to simplify ways for consumers to discover and purchase content. A consumer looking for ringtones by the Black Eyed Peas, for instance, could simply say the band’s name into the phone instead of clicking through menus on a carrier’s deck or triple-tapping to enter the band’s name.
While its patent was awarded just three weeks ago, the Internet behemoth is no newcomer to speech-recognition technology. The company applied for the patent more than five years ago, and has offered the service on Google Labs, a pre-beta-test site. Alexander Franz and Brian Milch, who helped author the patent application, partnered on an academic paper in 2002 titled, “Searching the Web by Voice.”
“Spoken queries are a natural medium for searching the Web in settings where typing on the keyboard is not practical,” Franz and Milch wrote.
Speech-recognition technology allows searches to occur “above the deck so that everything is one click, one word away,” according to Brady Bruce, vice president of mobile business development for Promptu Corp., a California-based developer. Promptu recently changed its name from Agile TV in an effort to expand its speech-recognition software from cable and on-demand television offerings into wireless.
The Google patent covers a system that receives voice commands and delivers results based on a weighted Boolean query in an effort to increase the accuracy of searches.
“Through the use of a language model, phonetic dictionary, and acoustic models, a server generates an n-best hypothesis list or word graph,” according to the patent. “As a result, the server is capable of returning relevant search results for even queries containing a few words.”
Although the patent doesn’t specifically mention wireless-and Google has been tight-lipped about plans to deploy a voice-powered search offering-analysts say mobile phones are a natural fit for the technology.
“Ultimately, voice search means hands-free search, and that means the target market will be the mobile crowd,” said Dan Miller, senior analyst at Opus Research Inc. “There’s extreme relevance for mobile search.”
That could spell trouble for the handful of developers looking to entice users to talk to, rather than through, mobile phones. Promptu and V-Enable Inc., a San Diego-based developer, each issued blanket statements predicting Google’s patent likely won’t affect their respective offerings and welcoming the Internet giant’s efforts in the space.
V-Enable called Google’s technology complementary to its software, adding “Google’s endorsement indicates increased activity in the voice search industry and an advancement in technology.”
A third player, BeVocal Inc. of Mountain View, Calif., declined to discuss potential implications of Google’s patent.
Others on the playing field include VoiceSignal Technologies Inc., which earlier this year announced the shipment of its 50 millionth handset worldwide, and Nuance Communications Inc., formerly ScanSoft. Nuance, a 14-year veteran in the space, has deals with T-Mobile USA Inc., Verizon Wireless and Virgin Mobile USA L.L.C., among others.
The space is once again attracting investors, as well. V-Enable closed a $6 million round of funding earlier this year, and Promptu raised $22 million in financing last May. BeVocal raised $10 million last year.
While speech-recognition patents are nothing new-IBM Corp. has secured hundreds of them, Opus’ Miller said-the Google patent may threaten developers if the company opts to aggressively protect its proprietary technology. But for now the new patent may be a win for the speech-recognition space in general.
“When and if Google productizes this,” said Miller, “I think it has the potential to elevate the perception of speech recognition across a lot of truly useful applications. The linkage of search with the spoken word is the Holy Grail.”
