Synthesis of speaking styles with corpus- and HMM-based approaches

Abstract

In this paper we compare two state-of-the-art speech synthesis techniques (corpus- and HMM-based) in terms of expressive speech synthesis. Two corpora were composed with different speaking styles (broadcast news and literature reading) from the same female speaker. Our aim was to determine to what extent the different technologies reproduce these styles. The corpora and the synthetic expressive speech samples were evaluated based on objective measures, and a carefully designed perceptual test was carried out in order to evaluate naturalness, quality and style identification rates of the generated samples. In our objective assessment we focused on prosodic features that principally influence the speaking style: F0 contour, average values and articulatory speed. Our evaluation of the perceptual test shows that both techniques were able to capture the main features of expressive speech and although listeners preferred the HMM-based voice, the speaking style was recognizable in case of both methods.