Acoustic unit coding method for use in e.g. speech synthesis, involves determining interpolation function of spectral envelope model of frame from models of reference frames, and coding unit from modelisation of frames and function

摘要

#CMT# #/CMT# The method involves determining an interpolation function of spectral envelope model of each frame from spectral envelope models of reference frames. An acoustic unit is coded from detailed modelisation of the reference frames and from the interpolation function. Interpolation co-efficients of the spectral envelope model corresponding to the current frame are determined from the spectral envelope models of the reference frames. #CMT# : #/CMT# Independent claims are also included for the following: (1) a program for coding acoustic units comprising instructions for implementing a method for coding acoustic units (2) a device for coding acoustic units, comprising an acoustic unit access unit (3) a method for converting voice. #CMT#USE : #/CMT# Method for coding an acoustic unit in an application such as word synthesis. #CMT#ADVANTAGE : #/CMT# The method utilizes the reference frames and interpolation parameters i.e. format frequency, between the reference frames in order to efficiently code the acoustic unit. The utilization of the interpolation between the spectral envelope models permits the coding of the acoustic unit in an automatic manner. #CMT#DESCRIPTION OF DRAWINGS : #/CMT# The drawing shows a flowchart illustrating a method of coding an acoustic unit.'(Drawing includes non-English language text)'.