| 摘要 |
PURPOSE: To improve reproducing tone quality of a voice signal encoded at a low bit rate. CONSTITUTION: A prescribed linear predictive coefficient is generated by a molecule coefficient calculation circuit 25 based on a spectrum parameter outputted from an encoder. A prescribed correction coefficient is generated from the linear predictive coefficient and the spectrum parameter by a correction filter coefficient calculation circuit 35. A spectrum post filter 20 generates a prescribed transfer characteristic based on the spectrum parameter and the linear predictive coefficient and filters an inputted regenerative signal. A correction filter 30 forms the prescribed transfer characteristic based on the correction coefficient to correct the output of the spectrum post filter. The level of the output of the correction filter is adjusted by a gain adjustment circuit 40, and the corrected regenerative signal is outputted to the outside. |
