IBM, Microsoft and partners are pushing speech technologies into the mainstream, with new systems to operate Linux, web browsers and other software.
At the Avios SpeechTek 2004 conference in San Francisco last week, IBM said it would support Linux alongside Windows and AIX in an imminent release of its WebSphere Voice Server that hosts speech recognition and text-to-speech software. The release will also add support for a proposed standard called Media Resource Control Protocol (MRCP), allowing multiple vendors' speech-recognition and text-to-speech engines to be integrated. IBM also upgraded its speech development toolkit.
Among IBM's supporters is Opera Software, whose browser is popular with technical users. A future version will incorporate IBM's ViaVoice technology to enable users to ask questions and then receive audio responses from their computers.
Also at the US show, Microsoft showed off its new Speech Server 2004, offering speech recognition, text-to-speech and developer tools that tie into Visual Studio .Net. Microsoft is relying on partners to build applications, including ScanSoft and Pronexus, which announced VeoBill, a bill-payment application. Pronexus is also launching a tool for Speech Server called VBSalt to help Visual Studio developers use the Salt speech specification that Microsoft supports.
Some observers fear that divided support between Salt and rival VoiceXML, which IBM backs, could fork development but others are more optimistic.
"Two years ago there were no standards so although there are two now, they are helping to drive applications and taking cost out of the process," said Mark Erwich, ScanSoft international marketing manager.
See also:
Speech recognition is still a niche application, arguably for good reason. But will perceptions change? 08 Sep 2003All Developer