We recently completed our work on Windows Vista, which has eight languages of speech recognition (SR) built into the OS. This integration provides an end-to-end speech experience: in addition to dictating documents and emails, users can navigate and control their computer and applications just by speaking. Currently, we're developing SR in many languages to enable unified communications such as controlling your email and calendar via mobile phone. Topics we could discuss at ETech include:

• Audio indexing and the rise of new speech-related services that enable web crawlers to look through podcasts and other large audio files to make them searchable
• Voice-based search from mobile phones to let you easily find information relevant to people on the go
• Living room and media center remote control that lets you speak the title of the song or movie you want to watch, making a much easier and natural UI than scrolling through a long list of options on a small input device
• New speech interfaces that can make SPOT watches and other very tiny devices more usable

We believe speech can be a natural and effective modality for interacting with devices. This technology is just now becoming part of the mainstream. Speech recognition in Vista will expose the possibilities to a large number of computer users. In addition, new VoIP applications are driving higher quality (16Khz vs 8khz) audio; and faster, multicore CPUs are increasing the number of new applications (like SR) that can run on a single device.