Statistical wavecycle synthesis

This technique works by synthesising waveforms from a sequence of interpolated breakpoint samples produced by a statistical formula e.g. Gaussian, Cauchy or Poisson. The interpolation may be calculated using a number of functions, such as exponential, logarithmic and polynomial. Due to the statistical nature of the wavecycle production, this technique is also referred to as dynamic stochastic synthesis. 

This technique was developed by the composer Iannis Xenakis at Cemamu (Centre for Research in Mathematics and Music Automation) in France. Xenakis also designed a synthesis system called Gendy (Serra, 1993). Gend) s primary assumption is that no wavecycle of a sound should be equal to any other, but rather that it should be in a constant process of transformation. The synthesis process starts with one cycle of a sound and then 

---

In essence, the technique resembles the statistical wavecycle synthesis technique introduced earlier, with the fundamental difference that here the breakpoints are deterministic, in the sense that the segments are defined explicitly rather than statistically. Despite the deterministic nature of the technique itself, its outcome is often unpredictable, but notwithstanding interesting and unique. Like the binary instruction technique discussed in Chapter 2, synthesis by sequential waveform composition is also commonly referred to as non-standard synthesis. Indeed, in order to work with this technique one needs to be prepared to work with non-standard ways of thinking about sounds, because its s)mthesis parameters by no means bear a direct relation to acoustics. 

---