Acoustic analysis of dysarthric speech
نویسندگان
چکیده
This report describes an acoustic analysis of the speech of persons with dysarthria. Sentence-based vs. word-based syllable rate and examples of deviant articulation patterns are presented. Implications for speech recognition systems are discussed.
منابع مشابه
Maximum Likelihood Linear Regression (MLLR) for ASR Severity Based Adaptation to Help Dysarthric Speakers
Automatic speech recognition (ASR) for dysarthric speakers is one of the most challenging research areas. The lack of corpus for dysarthric speakers makes it even more difficult. The speaker adaptation (SA) is an alternative solution to overcome the lack of dysarthric speech and enhance the performance of ASR. This paper introduces the Severity-based adaptation, using small amount of speech dat...
متن کاملDysarthric Speakers' Intrinsic Vowel Durations
This study uses the Nemours Database of Dysarthric Speech and the Buckeye Corpus of Conversational Speech to look into differences in the way vowel quality correlates with intrinsic duration in typical and non-typical populations. Results based on speech material from ten dysarthric subjects indicate that intrinsic vowel duration may indeed play a role as a parameter for acoustic classification.
متن کاملLexical boundary error analysis in hypokinetic and ataxic dysarthria.
This investigation is the second in a series to examine a potential source of reduced intelligibility in dysarthric speech, namely the mismatch between listeners' perceptual strategies and the acoustic information available in the dysarthric speech signal. Lexical boundary error (LBE) analysis was conducted on listener transcripts from phrases produced by speakers with hypokinetic dysarthria, a...
متن کاملAcoustic analysis of dysarthric speech and some implications for automatic speech recognition
This paper describes one part of an EU project, ENABL, in which speech recognition systems are used to provide access by voice to computer software. An acoustic analysis of the speech of fourteen motorically disabled persons, twelve with dysarthria and two without, is presented. The analysis is based on a standard description of dysarthric speech and parameters such as fundamental frequency, sy...
متن کاملAn evaluation of unsupervised acoustic model training for a dysarthric speech interface
In this paper, we investigate unsupervised acoustic model training approaches for dysarthric-speech recognition. These models are first, frame-based Gaussian posteriorgrams, obtained from Vector Quantization (VQ), second, so-called Acoustic Unit Descriptors (AUDs), which are hidden Markov models of phone-like units, that are trained in an unsupervised fashion, and, third, posteriorgrams compute...
متن کامل