| 摘要 |
PURPOSE:To obtain a high-quality musical sound waveform when musical sound waveforms are synthesized based on stored waveforms by storing the musical sound waveforms at a high S/N without increasing the number of elements required for the storage. CONSTITUTION:A 24-bit waveform value outputted from an adder 1 is inputted to the first added input terminal and the higher-rank 16 bits of the 24-bit output value of an adder 5a are inputted to a delay circuit 2 as a quantized waveform value, with the remaining lower-rank 8 bits, namely, a quantization error is inputted to another delay circuit 5b. When such a quantization circuit 5 is used, quantization is preformed on values obtained by adding waveform values inputted at every sampling period to the quantization error accumulated values at every sampling period and accumulation is performed on quantization errors produced at every quantization. Since the output waveform value of the adder 1 is quantized after the value is corrected by means of the accumulated value of quantization errors, occurrence of distortions by quantization errors can be prevented and high-quality attenuated sounds can be obtained. |
