Improvements in Speech Synthesis

Beschreibung

Naturalness in synthetic speech is one of the most intractableproblems in information technology today. Although speech synthesissystems have improved considerably over the last 20 years, theyrarely sound entirely like human speakers. Why is this so, and what can be done about it? * Prosodic processing must be rendered more varied and moreappropriate to the speech situation * Timing, melodic control and the relationships between the variousprosodic parameters need increased attention * Signal processing systems must be developed and perfected thatare capable of generating more than just one voice from adatabase * A better understanding must be achieved of what distinguishes onevoice from another, and of how speech styles differ between simplyreading aloud numbers and sentences and their use in interactivespeech * New evaluation methodologies should be developed to provideobjective and subjective measurements of the intelligibility of thesynthetic speech and the cognitive load imposed upon the listenerby impoverished stimuli * Adequate text markup systems must be proposed and tested withmultiple languages in real-world situations * Further research is required to integrate speech synthesissystems into larger natural-language processing systems Improvements in Speech Synthesis presents the latest research inthe above areas. Contributors include speech synthesis specialistsfrom 16 countries, with experience in the development of systemsfor 12 European languages. This volume emerges from a four-yearEuropean COST project focussed on "The Naturalness of SyntheticSpeech", and will be a valuable text for everyone involved inspeech synthesis.

Autor*in

E. Keller

Themen in »Improvements in Speech Synthesis«

Audio & Speech Processing & Broadcasting Audio-, Sprachverarbeitung u. Übertragung Communication Technology Electrical & Electronics Engineering Elektrotechnik u. Elektronik Kommunikationstechnik

Stimmen zu »Improvements in Speech Synthesis«

Details

Link teilen