| 摘要 |
PURPOSE:To varies the pitch and timbre of a synthesized musical sound variously with a small amount of stored information while leaving the features of the source sound. CONSTITUTION:This method divides the audio signal Ai of a representative sound into plural frequency bands, converts the signals of the respective frequency bands to specific low frequencies and stores them in a memory 5, and controls the amplitudes of the signals of the respective bands with timbre information, performs frequency conversion from the original frequency bands to shifted bands with pitch information, and then puts the signals of the respective bands together into a musical sound having a necessary timbre and pitch at the time of reproduction. Here, the frame length when the audio signal of the representative sound is divided into the frequency bands is varied by a sampling frequency conversion part 4 to match the period of the waveform of a fundamental frequency component included in the audio signal with the frame length. Then a waveform of one or several cycles of the waveform of the fundamental frequency component included in the audio signal Ai is stored in the memory 5. |
