Inspiration from Nature for Voice Synthesis and Spectral Analysis

Electronic techniques for voice spectral synthesis and analysis have seen huge advances during the 20th century, most notably during its last quarter. The ubiquity of PC machines in the home and office that are more than capable of running spectral analysis algorithms has made it possible for many to make use of these techniques; indeed, there is much software available as freeware. This paper explores current voice synthesis and spectral analysis techniques that are in common use, and suggests what the next steps might be by taking account of what we know about natural systems and the world around us. It suggests that the application of knowledge relating to the peripheral human hearing system, or psychoacoustics, will play a significant part in the design of new algorithms, particularly for voice spectral analysis. For voice synthesis, the world of physical modeling is explored, whence significant advances in terms of output sound naturalness have been gained in music synthesis, evoking terms such as organic amongst users and listeners alike. Both gain their advantage by taking account of what we observe about the world in which we live