SPEECH SYNTHESIS SYSTEM

摘要

Speech synthesis system implementable in an integrated circuit device and capable of converting frames of digital data into analog signals representative of human speech. The frames of data comprise digital representations of values of pitch energy and filter coefficients which are stored in non-volatile memory. The filter coefficients are utilized to control a linear predictive filter of the speech synthesis system with the filter being excited by voiced and unvoiced excitations stored in non-volatile memory. Components of the speech synthesis system include a variable interpolation circuit within the linear predictive filter to allow a variable number of interpolation steps to be calculated between successive values of reflection coefficients, control circuitry for determining the rate at which new frames of data are utilized, a pulse width modulated digital-to-analog converter operable at low voltage control, circuitry coupling the voiced and unvoiced excitation signals to the filter wherein an operator may select an external excitation signal rather than precalculated stored excitation signals and control circuitry for detecting a control signal indicating whether the speech parameters representative of pitch energy and filter coefficients being input to the speech synthesis system are in coded or uncoded format and including decoding circuitry within the system such that all calculations within the system may be based upon uncoded input parameters. The speech synthesis and control circuitry may be constructed by employing metal-insulator-semiconductor technology so as to provide circuits capable of low voltage operation.