To effectively improve a speech synthesis system, it is important to find and focus on improving the modules whose effect on naturalness degradation in synthesized speech are the largest. In this paper, we describe the design of a perception experiment to measure the effect of each module separately. Synthesized speech stimuli whose intermediate information is modified during a synthesis proces...

This event-related potential study examined the impact of imitating an unfamiliar accent on the processing of spoken words embedded in sentential contexts produced in that accent. The cloze probability effect in two groups of southern French speakers after they had to either listen to or imitate sentences spoken by a Belgian French speaker was tested. Speakers who did not imitate the unfamiliar...

An English utterance was synthesized in four versions using sets of diphones produced under four different prosodic and contextual conditions. The synthesis used either accented diphones only or appropriately located accented and unaccented diphones, with each of these conditions being repeated using neutral-context and differentiated-context diphones. They were presented to two listener groups...

Rules for segmental duration has been studied in the context of a speech database that is under development in our department. The database search procedures include the same kind of context sensitive rules that are used in our speech synthesis project. This gives us the possibility to make a direct comparison to the database durations when developing durational rules for a text-to-speech syste...

Both auditory and audiovisual speech synthesis have been the subject of many research projects throughout the years. Unfortunately, in recent years only very few research focuses on synthesis for the Dutch language. Especially for audiovisual synthesis, hardly any available system or resource can be found. In this paper we describe the creation of a new extensive Dutch speech database, containi...

In this paper we evaluate the performance of 8 different speech rate estimators [1, 2, 3, 4, 5] previously described in the literature by applying them on a multilingual test database [6]. All the estimators show an underestimation at high speech rates and some also suffer from an overestimation at low speech rates. Overall the tested methods obtain high correlation coefficients with the refere...

In this paper I will report on recent extensions to the nonconcatenative speech synthesis architecture employed in the YorkTalk and IPOX systems (Coleman 1992a, 1994, Local 1992, Dirksen and Coleman 1994). These systems have not to date been noted for verbosity: our primary goals have been to achieve the most natural−sounding speech that we can, as well as solving the technical challenges prese...

We are presenting an interactive course on speech synthesis which is designed to support the education in speech communication. In the basic section, the fundamental principles of speech synthesis are explained. To explore a complete text-to-speech (TTS) system, the user is provided with access to the Dresden Speech Synthesizer DreSS. The user may type any text, and he may observe how the syste...

The goal of this paper is to present the work carried out up to now for the development of the Greek Text-To-Speech (GRTTS) system by NTUA. The system under consideration is based on the method of concatenative synthesis and follows the Bell Labs approach to this technique. In order that the input text to the GRTTS is translated into continuous synthetic speech the following modules have alread...

Accurate estimation of segmental durations is crucial for naturalsounding text-to-speech (TTS) synthesis. This paper presents a model of vowel duration used in the Bell Labs Japanese TTS system. We describe the constraints on vowel devoicing, and effects of factors such as phone identity, surrounding phone identities, accentuation, syllabic structure, and phrasal position on the duration of bot...