The Synthesis of Soundscape using Genetic Algorithm and Popular Songs
* Presenting author
This paper proposes the concept of soundscape synthesis based on improving existing soundscapes with optimization through Genetic Algorithms. As a goal function for optimization the parameter of some acoustic environment called Total Distraction Coefficient (which depends on loudness, sharpness and loudness growth rate) is chosen by which one can synthesize an improved soundscape. To verify the proposed concept two representative (and familiar to most urban residents) acoustic environments are further analyzed and compared (namely, a children's park and an expressway). Using Genetic Algorithms (with previously known Total Distraction Coefficient) two new soundscapes are synthesized from each acoustic environment - one with lower and one with optimum Total Distraction Coefficient (i.e. the environments supposed to be perceived as "worse" and "better" than the starting one, respectively). The obtained soundscapes are then tested for subjective perception.