VOICE SYNTHESIZER

摘要

PURPOSE: To make it possible to more accurately draw a human voice by linearly executing the inter-frame interpolation of data by a variable interpolating circuit in a linear predicting filter using an encoded reflection coefficient for preparing a digital signal expressing a human voice. CONSTITUTION: A synthesizer 10 is a linear predicting filter type synthesizer and prepares a synthetic voice from non-coded parameters. These parameters are transferred from a code RAM 32 to a ROM/RAM loading parallel/series converter 38 to be read out to a linear predicting filter in series. The parameters outputted from the converter 38 are connected to an array multiplier/interpolator 50. A shift register 66 is used for connecting excitation data or magnification data from a ROM 78 to the multiplier/interpolator 50 through a multiplexer 58 and recording logic 52 and the multiplier/interpolator 50 executes interpolation based upon the magnification data.