IBM Corp and Nokia Oyj have signed an agreement to collaborate on developing speech recognition technologies for the mobile device and handset market. The two companies want to push speech recognition as a common interface for mobile and other devices – from handsets and PDAs to consumer products such as video recorders and set-top boxes.
The companies claim that a speech interface – based around IBM’s ViaVoice voice recognition software – will become an important, integrated part of future personal mobile terminals and communications devices. The companies say that they are working to drive open industry standards; so that third parties can use any specifications that emerge from the work in their own mobile devices. The research work will concentrate on speech recognition, speaker recognition, text-to-speech and conversational systems. Nokia and IBM will also put funds into long term research on VoiceXML – a subset of the language for enabling phone-to-web access – and applications for it.
Putting a speech interface onto mobile devices has long been seen as a solution to the problems of inputting and viewing data on small-screen or even no-screen devices. Nokia and IBM would seem to be ideally placed to develop a speech user interface, indeed the pair have already been working together on IBM’s Pervasive Computing project. Nokia will bring a hardware platform and mobile phone marketing muscle to the speech project, while IBM has the software and microchip know-how to drive the platform.
IBM Microelectronics has already talked about developing an on- chip version of ViaVoice for command and control applications. Running the speech engine on silicon would remove a lot of the memory concerns that surround voice software and mobile devices. However, problems with increased power consumption and developing a good microphone for mobile applications remain. This is why IBM and Nokia have not given any definite dates of when we will see products from the project and describe it as a long term project.
