The Technology Behind Auracle
Auracle is built in the Java programming language with a combination of publicly available products and custom developed components. Core technologies include JSyn for sound synthesis and TransJam for networking, both developed by team member Phil Burk.
We developed our own modules for voice analysis and classification. The voice analysis algorithms extract the fundamental frequency and first two formants from a vocal signal. The classification algorithms break incoming vocal input into gestures and create various groupings of those gestures.
Clients send and receive control data — not streaming audio — to each other via a central server; the voice analysis and sound synthesis is performed by each client machine. With this architecture, Auracle can produce high-quality audio with minimal latency, even over relatively slow Internet connections, and the Auracle server can accommodate a large number of simultaneous users.
We intend to release the Java source code for our voice analysis and classification modules. If you are interested in getting a copy, contact us.