Applications of the phase vocoder in the control of real‐time electronic musical instruments
Nicholas J. Bailey, Alan Purvis, Ian Bowler, Peter Manning · Interface · 1993
One of the limitations of real‐time musical instruments is their methods of gestural capture. The normal Man‐Machine Interface is constructed from a series of switches organised as an electronic keyboard, and can typically provide only two or three control dimensions (time struck, velocity of strike and pressure). Conversely, traditional musical instruments are capable of being controlled in a much more subtle and complex way. The Fourier Transformation (FT) is a familiar method of transforming a time‐series of samples to its frequency‐space equivalent. Unfortunately, this process is rather inaccurate when exact control parameters are required for a real‐time electronic musical instrument. The use of the phase vocoder can overcome these inaccuracies by preserving and using the phase information of successive FTs. By using the sound of a traditional instrument to control a synthesiser, it may be possible to take advantage of the former's more subtle control possibilities. The vast quantity of information thus obtained requires look‐up table of high dimensionality, and special techniques will be required if such a table is to be stored in a reasonable amount of memory.