A TECHNIQUE FOR REAL-TIME DIGITAL SYNTHESIS OF COMPLEX AUDIO SPECTRA
Abstract
A class of harmonically rich signals for use in subtractive music synthesis is introduced. These signals are attractive for digital implementations due to their inherent ability to prevent audible aliasing. This basic signal generation method is then extended to allow more precise harmonic control and spectral shaping at a low computational cost. Fundamental frequency dependent architectures are developed which provide many of the advantages of additive synthesis methods but at a greatly reduced computational burden. Also presented are several algorithms for computer estimation of synthesis parameters directly from digitized samples of natural or existing voices.
Degree
Ph.D.
Subject Area
Electrical engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.