A Statistical Timing Model for French

Abstract

Numerous factors influence speech timing. Statistical analysis can identify an order of importance and mutual influences between such factors. A three-tiered (segment-syllable-phrase) model was created by a modified stepwise statistical procedure. It predicts the temporal structure of French of a single, highly fluent speaker at a fast speech rate. The model's predictions correlated with the 1204 syllables of the original corpus at r = 0.846.