Performance of the genetic annealing algorithm in DFM synthesis of dynamic musical sound samples

Abstract

A qualitative analysis of the performance of a genetic annealing algorithm (GAA) is presented. The GAA is applied in the double frequency modulation DFM parameter optimization problem. It is shown that the GAA can optimize the solution to the global minimum by plotting the solution spaces of one DFM equation for all samples. Dynamic musical sounds were synthesized using the DFM synthesis technique. A recycling process is also proposed to improve the efficiency of the GAA process. The resulting dynamic sounds synthesized by the DFM are presented and compared with those of real samples.