Quantitative Modeling of Segmental Duration

In natural speech, durations of phonetic segments are strongly dependent on contextual factors. Quantitative descriptions of these contextual effects have appfications in text-to-speech synthesis and in automatic speech recognition. In this paper, we describe a speakerdependent system for predicting segmental duration from text, with emphasis on the statistical methods used for its construction. We also report results of a subjective listening experiment evaluating an implementation of this system for text-to-speech synthesis purposes. 1. I N T R O D U C T I O N This paper describes a system for prediction of segmental duration from text. In most text-to-speech synthesizer architectures, a duration prediction system is embedded in a sequence of modules, where it is preceded by modules that compute various linguistic features ~ from text. For example, the word "unit" might be represented as a sequence of five feature vectors: (< At/, word initial, monosyl labic , . . . , >) • " (< / t / ~ r s t , w o r d final, monosyllabic, . . . , >). In automatic speech recognition, a (hypothesized) phone is usually annotated only in terms of the preceding and following phones. If some form of lexical access is performed, more complete contextual feature vectors can be computed. Broadly speaking, construction of duration prediction systems has been approached in two ways. One is to use generalpurpose statistical methods such as CART 2 or neural nets. In CART, for example, a tree is constructed by making binary splits on factors that minimize the variance of the durations in the two subsets defined by the split [2]. These methods are called "general purpose" because they can be used across a variety of substantive domains. There also exists an older tradition exemplified by Klatt [3, 4, 5] and others [6, 7, 8, 9] where duration is computed with duration models, i.e., simple arithmetic models specifically designed for segmental duration. For example, in Klatt's lWe define a factor, FFi, to be a partition of mutually exclusive and exhaustive possibilities such as {1-stressed, 2-stressed, unstressed}. A feature is a "level" on a factor such as 1-stressed. The feature space F is the product space of all factors: Fl × -. × Fn. Because of phonotactic and other constraints, only a small fraction of this space can actually occur in a language; we call this the linguistic space. 2Classification and Regression Trees [1 ]. model the duration for feature vector f E F is given by