Signal Processing Approach for Music Synthesis Using Bird's Sounds

Abstract

Music is generated by a sequence of musical notes which are arranged in a particular manner. Indian classical music contains 12 notes in an octave. Similarly bird's sound also contains such notes which are responsible for the sensible voice of birds. So we can generate musical tone using bird's sounds. For generating music using bird's sounds we have to extract these 12 notes. For extracting these notes from bird's sounds and arranging these notes in a particular Raga give us a musical tone. These notes would be sound of different birds. Selection of notes from bird's sounds is done using different signal processing approaches. Using Ant Colony Optimization we can select best 12 sounds as notes whose dominant normalized frequencies give minimum mean square error with already existing normalized frequencies (reference normalized frequency) of Indian classical music. In this paper these notes in form of bird's sounds have been determined using different signal processing approaches and tried to implement with some Raga. Benefit of this selection is to generate musical tones with actual bird's voice to feel different taste in music composition with already available musical instruments and bird sounds. The experiments are performed using the original sound of different perching birds.