Neural Network-Based Modeling of Phonetic Durations
This work addresses duration modeling for TTS and ASR applications, but it is incremental as it applies existing DNN methods to phonetic data without major methodological breakthroughs.
The paper developed a DNN model to predict phoneme durations in US English, identifying pre-pausal lengthening, lexical stress, and speaking rate as key factors, and found that duration prediction is poorer for ASR training due to noisy, casual speech and children's variability.
A deep neural network (DNN)-based model has been developed to predict non-parametric distributions of durations of phonemes in specified phonetic contexts and used to explore which factors influence durations most. Major factors in US English are pre-pausal lengthening, lexical stress, and speaking rate. The model can be used to check that text-to-speech (TTS) training speech follows the script and words are pronounced as expected. Duration prediction is poorer with training speech for automatic speech recognition (ASR) because the training corpus typically consists of single utterances from many speakers and is often noisy or casually spoken. Low probability durations in ASR training material nevertheless mostly correspond to non-standard speech, with some having disfluencies. Children's speech is disproportionately present in these utterances, since children show much more variation in timing.