Terminal analog synthesis of continuous speech using the diphone method of segment assembly

Abstract

Diphone segmentation consists of the stylization and simulation of the perceptually meaningful dynamic portions of the acoustic continuum of speech. A primary requirement of the stylization is that any segment must be usable in a variety of morphemic environments. Diphone segment assembly is a technique for synthesizing a potentially unlimited variety of continuous utterances under computer control. A major requirement of this method of segment assembly is that there be no perceptually distracting discontinuities between contiguous segments. The presentation consists of two parts. The first is a brief description of the terminal analog synthesizer and the system for generating, storing, and assembling the control signals for diphone segments. The second part is a discussion, by example, of the analysis, synthesis, and assembly rationale for simulating continuous utterance. Parameters used for the simulation of prosodic features are discussed, and the evaluation methods are described.