| 摘要 |
<p>PURPOSE:To provide the encoder which reduces quantizing error in an encoded output, simplifies system configuration and easily changes a quantizing error correcting range. CONSTITUTION:A predictive error value ei (n) is calculated by subtracting a preceding value decoded value xip outputted from a predicter 16 from an input sampling value xi by a subtracter 11, a predictive error value ei (n-1) delayed for one frame period according to this predictive error value ei (n) and a delay circuit 31 is inputted to a fuzzy inference part 32, and quantizing error DELTAepsilon(n+1) to be corrected at an (n+1)th frame is inferred according to this information. The inferred quantizing error is subtracted from the predictive error at the (n+1)th frame by a subtracter 33, and an encoded output yi is obtained by a quantizer 12.</p> |
