Real Time Visual Displays in Speech and Singing .

Abstract

Many branches of the speech, hearing and singing research community use analyses of aspects of the voice source and overall spectral changes. By these means, patterned variations have been noted in both the voice source and the acoustic output from the vocal tract during speech and singing, the speech/singing of professional voice users and pathological speech. The widespread availability of fast computers and dedicated signal processing devices has enabled many of these techniques to be made available as real-time displays to a wide range of users. This paper describes a number of real-time visual displays of voice source, the acoustic output and articulatory gestures during speech and singing. The development of analysis techniques based on a model derived from the results of contemporary psychoacoustic experiments into the nature of the human peripheral hearing mechanism is discussed. This model depends on parallelism which is ideally suited to implementation on transpires which are specifically designed to operate in parallel. Developments such as these may further our understanding of the relative importance of voice source and acoustic cues in speech and singing, whether normal, professional or pathological, and enable a new generation of real-time visual displays to be implemented.