The Workshop included an extended group of presentations by selected US government agencies and an invited guest from the European Community. These presentations, amplified in the following pages, indicated the importance given to speech and natural language research and development, and the approach of each agency to adapt their philosophy and activities to those aspects which they considered ...

We present a study on the relation between fundamental frequency (F0) and its perceptual effect in the context of text-tospeech (TTS) synthesis. Features that essentially capture the intonational (macro-prosodic) properties of spoken speech are introduced and analysed with regard to the following questions: (i) How does the prosodic variation of TTS signals differ from natural speech? (ii) Is t...

This study investigates masking effects occurring during speech comprehension in the presence of concurrent speech signals. We examined the differential effects of 4to 8-talker babble (natural speech) or babble-like noise (reversed speech) on word identification. We measured phoneme identification rates. Results showed that different types of linguistic information can interfere with speech rec...

We have used data on patents and publications, and from an Internet-based survey, to analyse corporate technological activities in Automatic Speech Recognition (ASR) and Natural Language Processing (NLP) technologies. Two distinct clusters of firms exist: large firms mainly in telecommunications, desktop computing, and consumer electronics; and small firms specialising in speech technologies. T...

When speech comprehension becomes effortful, e.g. in adverse conditions, it is necessary to understand which processes are the sources of the effort. By means of pupillometry, this paper investigates whether the automatic process of lexical competition increases effort, and whether mismatching cues, degradation of the signal, and listener’ experience with degraded signals additionally affect th...

We present an evaluation of the perception of foreign-accented natural and synthetic speech in comparison to accent-reduced synthetic speech. Our method for foreign accent conversion is based on mapping of Hidden Semi-Markov Model states between accented and non-accented voice models and does not need an average voice model of accented speech. We employ the method on recorded data of speakers w...

Forced alignment both for words and phones is a challenging and interesting task for automatic speech processing systems because the difficulties introduced by natural speech are many and hard to deal with. Furthermore, forced alignment approaches have been tested on Italian just in a few studies. In this task, the main goal was to evaluate the performance offered by the proposed systems on Ita...

This paper describes Hidden Markov Model and its application in natural language process, first introduces the basic concept of Hidden Markov Model, then introduces the three basic issues and the basic algorithm to solve the problems, finally gives the demonstration of the application of Chinese part-of-speech tagging and speech recognition via Hidden Markov Model.

Infants and adults are able to match auditory and visual speech but the cues on which they rely may differ. Here we provide an initial assessment of the relative contribution of temporaland phonetic cues available in the AV signal. Adults (N=52) and infants (N=18) matched 2 trisyllabic speech sounds, either natural speech or SWS, with visual speech information. Adults saw two articulating faces...