Control of prosodic focus in corpus-based generation of fundamental frequency contours of Japanese based on the generation process model

Abstract

A total corpus-based process of generating prosodic features from text is developed. The process first predicts pauses and phone durations, and then generates F <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</sub> contours. Since F <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</sub> contour generation is based on the generation process model, it is rather easy to manipulate the generated F <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</sub> contours in command level. A method was developed for generating sentence F <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</sub> contours, when a focus is placed in one of the ldquobunsetsurdquo of an utterance. The method is to predict differences in the F <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</sub> model commands between with and without focus utterances, and apply them to the F <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</sub> model commands predicted beforehand by the baseline method. The validity of the method was proved by the experiment on F <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</sub> contour generation and speech synthesis.