This system is designed to take real-time user input through a microphone and process it in novel ways. The user needs only to open the program and begin speaking. The spoken voice is recorded and analyzed using an FFT, and the spectral data is converted into a 3D matrix using a technique by my former teacher, Luke Dubois. This software is based heavily on his model. The 3D matrix is then scanned to resynthesize the original spoken sample. The novelty of resynthesizing from an image is that visual effects can be applied which change its sonic properties. Over time, the original sample will change substantially--there are no other sound sources or effects introduced in this system: everything heard is derived from the original sample. In this piece, the user says "God" one time. It is then processed according to deterministic processing score that the software follows; nothing is random.
by V.J. Manzo
http://www.vjmanzo.com/cv
All Comments