In this paper, we proposed a method for asynchronous virtual meeting called AVM. This system provides the easy manner to quote the voice messages and increases the usefulness of the voice-mail system. As the result of our preliminary evaluations, conversations were realized efficiently with the system.
AVM is an application platform for the many kinds of speech processing. We are integrating the spontaneous speech recognition and the dialog tagging functions into our AVM server.
The integration of the AVM and the E-mail or the BBS enables the seamless communication of written words and spoken words. Text-to-speech technology is expected to be used in such a system.
Many streaming media systems on the Internet are developed and used in these days. The HyperAudio system has a radio-like user interface, and the user can listen to the audio contents interactively using the `jump' button. The telephone also can be used as the client system. The HyperAudio with the feature of submitting voices from the listeners will be a system resembles to the AVM. In comparison to such systems, AVM is a metaphor of reading and writing text messages. The data structure of AVM is designed for the bi-directional conversations. It is also extensive for making excerpts of many kind of continuous media, such as video-messages. Some systems, including HyperAudio and RealPlayer (by RealNetworks), use SMIL for describing the structure of the streaming contents. The role of the data which the AVM server generates is similar to SMIL, so an extension to the server enables the SMIL browsers to listen to the conversations on AVM.
Because the AVM server makes the audio contents on demand, many applications can be implemented into the server. Our future works include the realization of an multi-user voice game system using AVM.