Tim Bunnell, Ph.D.

Research Interests

  • Acoustic Phonetics
  • Speech Perception and Production
  • Speech Processing, Recognition, and Synthesis
  • Pattern Recognition
  • Biological Signal Processing

Current Projects

  • Phenotypes and familiality in speech disorders - Over the past two decades, significant advances have been made in speech analysis and speech pattern recognition techniques, however, the penetration of these advances into the speech disorders research arena has lagged, and penetration into the clinic is virtually non-existent. We propose to address this lag by adapting and extending speech recognition technology based on Hidden Markov Modeling (HMM) to an analysis of speech from children with speech delay of unknown origin. We will demonstrate the efficacy of this approach by (a) establishing the effectiveness of the proposed analysis techniques in identifying acoustically defined categories of segmental distortions (endophenotypes), (b) examining these acoustically-defined endophenotypes for evidence of heritability and (c) searching for evidence of genetic linkage of the obtained endophenotypes to a region of chromosome 3 that has been implicated in speech disorders. [Work supported by NIDCD]
  • Phonetic acquisition patterns in children with phonological disorders - The goal of this research is to improve diagnosis and treatment of children with phonological disorders. Children with phonological disorders often substitute segments that are in their phonetic inventory for segments they have not acquired, for instance, substituting /w/ for /r/. Several investigators have suggested that these substitute segments are acoustically distinct from their homophonous intended variants. For example, an error /r/ that sounds like /w/ (call this /w[r]/) may be acoustically distinct from an intended /w/. Additionally, it is argued that this distinctive marking is observed only in cases where children are prepared to acquire age-typical articulation of the error segment. If this is the case, such acoustic markers could be used to guide speech language pathologists in planning therapy for children by allowing them to concentrate on acquisition of segments they are most prepared to master. [Work supported by Nemours Biomedical Research]
  • Personalized Synthetic Speech Using ModelTalker: Development and Evaluation - This project encompasses development and testing of a concatenative text-to-speech (TTS) synthesis system known as ModelTalker and software called InvTool which guides individuals in creating personal synthetic voices for use with ModelTalker. The overall system is intended to be of particular interest to Augmentative and Alternative Communication (AAC) device users who depend upon speech synthesis for communication. Concatenative synthesis technology offers speech that is both more natural sounding and more intelligible than the rule-based formant synthesis systems commonly used in AAC devices (Granstrom & Carlson, 1997; Venkatagiri, 2002 See also Section C.). In addition to improved naturalness and intelligibility, the ModelTalker and InvTool software uniquely offer the capability of rapid development of personal synthetic voices. With InvTool, individuals such as those with Amyotrophic Lateral Sclerosis (ALS) who are at risk to lose the ability to speak can record their own speech for conversion to a personal synthetic voice for the ModelTalker TTS system. This voice banking capability has already been used successfully by a number of ALS patients. [Work supported by NIDCD]

Research Group

  • Suzanne M. McCahan, Ph.D. - Assistant Research Scientist and Bioinformatics Staff Specialist (Ph.D., Princeton University). Gene expression in pediatric inflammatory bowel disease.
  • Daniel Deng, M.S. - Bioinformatics Research Associate (M.S., George Mason University).
  • James Polikoff, M.S. - Senior Research Associate (M.S., Villanova University). Laboratory manager, Editorial assistant for Language and Speech.
  • James Mantell, B.A. - Research Assistant (B.A., Millersville University). Acoustic phenotyping of speech delayed children, Acoustic markers in speech disorders.

Selected Publications

  • Bunnell HT, Pennington C, Yarrington D, Gray J. Automatic personal synthetic voice construction. In: Proceedings of the Eurospeech 2005. Lisbon, Portugal; 2005.
  • Gray J, Pennington C, Yarrington D, Bunnell HT. A system for creating personalized synthetic voices. In: Proceedings of ASSETS 2005. Baltimore, MD; 2005.
  • Bunnell HT, Gray J, Pennington C, Yarrington DM. Automatic Construction of Concatenative Speech Synthesis Databases for AAC. In: 2004 ASHA Convention. Philadelphia, PA; 2004.
  • Bunnell HT, Polikoff JB, McNicholas JE. Spectral Moment vs. Bark Cepstral Analysis of Children's Word-initial Voiceless Stops. In: Proceedings of the Eighth International Conference on Spoken Language Processing. Jeju, Korea; 2004.
  • Polikoff JB, Hammond J, McNicholas JE, Bunnell HT. Spectral moments versus Bark cepstrum classification of children's voiceless stops. J Acoust Soc Am. 2004;115:2628.
  • McNicholas J, Polikoff JB, Lilley J, Bunnell HT. Personalized Synthetic Speech for AAC Devices: The ModelTalker Project. In: American Speech Language Hearing Association Convention. Atlanta, GA; 2002.
  • Bunnell HT, Yarrington D, Polikoff JB. STAR: Articulation Training for Young Children. In: Proceedings of the Sixth International Conference on Spoken Language Processing. Vol. 4. Beijing, China; 2000:85–88.
  • Bunnell HT, Yarrington DM, Polikoff JB. Using Markov models to assess articulation errors in young children. J Acoust Soc Am. 2000;107(5 pt 2):2903.
  • Yarrington DM, Hoskins SR, Polikoff JB, Bunnell HT. Personalized Synthetic Voices for AAC. In: Proceedings of ISAAC 2000. Arlington, VA; 2000.
  • Polikoff JB, Bunnell HT. The Nemours database of dysarthric speech: A perceptual analysis. In: Proceedings of the XIVth International Congress of Phonetic Sciences. Vol. 1. San Francisco, CA; 1999:783–786.
  • Bunnell HT, Hoskins S, Yarrington D. A biphone constrained concatenation method for diphone synthesis. In: Proceedings of the Third International Workshop on Speech Synthesis. Jenolan Caves Mountain House, Australia; 1998:171–176.
  • Bunnell HT, Hoskins S, Yarrington D. Prosodic vs. segmental contributions to naturalness in a diphone synthesizer. In: Proceedings of the Fifth International Conference on Spoken Language Processing, ICSLP-98. Vol. 5. Sydney, Australia; 1998:1723–1726.
  • Yeni-Komshian GH, Bunnell HT. Perceptual evaluations of spectral and temporal modifications of deaf speech. J Acoust Soc Am. 1998;104(2 Pt 1):637–647.

Head, Speech Research Laboratory, and Director, Bioinformatics Core Facility, A.I. duPont Hospital for Children

Phone: (302) 651-6835

Fax: (302) 651-6895

Email: bunnell@udel.edu

Office: A.I. duPont Hospital for Children

Address:
Alfred I. duPont Hospital for Children
1600 Rockland Rd.
Wilmington, DE 19803

Education

  • B.S. - University of Maryland
  • M.S., Ph.D. - Pennsylvania State University