Phonetics and Speech Laboratory
Established in 1979, the Phonetics and Speech Lab provides for research and teaching that bridges experimental phonetics and phonology from a linguistics perspective and more engineering based approaches to speech modelling. It participates in a number of international research projects, and has collaborative links with many speech laboratories in Europe and beyond. There is a particular interest in research on Irish, but aspects of Hiberno-English and of many languages have also been investigated.
The Laboratory is described under the following headings:
People
- Prof Ailbhe Ní Chasaide (Associate Professor of Phonetics)
- Dr Christer Gobl (Lecturer in Speech Science)
- John Duggan (Senior Technician/Programming Support)
- Irena Yanushevskaya (Postdoctoral Researcher)
- Harald Berthelsen (Researcher)
- Maria O'Reilly (Researcher)
- Amelia Kelly (Researcher)
- Amelie Dorn (Researcher)
- John Kane (Researcher)
- Neasa Ní Chiaráin (Researcher)
- Christoph Wendler (Researcher)
- John Dalton (Researcher)
Back Row: Irena, Christoph, Amelia, Maria, Raya, Neasa; Front Row: Christer, Harald, Ailbhe, John K, Amelie, Nick
Research strands/Projects
Research for student dissertations and projects may be undertaken in a variety of theoretical and descriptive areas. We list here 4 major research areas and related research projects (projects are also listed with linked descriptions below).
Voice source/voice quality analysis
see projects: CNGL, FastNet, COST 2103
- Development of techniques for source filter decomposition
- Analysis of voice quality
- Linguistic aspects of voice source variation: production studies
- Tone of voice and the communication of emotion, mood, attitude: synthesis based perception studies
- Other relevant past projects: Interaction of voice quality and f0, SPACT SPEECHMAPS, VOX, VERIVOX
Prosody
see projects: Prosody of Irish Dialects
- Analysis of Irish intonation: the four main dialects
- How intonation, voice quality interact in prosody
- Linguistic and paralinguistic functions of prosody
- Other relevant projects: Interaction of voice quality and f0
Speech synthesis and text-to-speech
see projects: CABÓGAÍ II
- Developing text-to-speech for Irish
- Irish corpora for text-to-speech
- Voice quality and the naturalness of synthetic speech
- Other relevant past projects: CABÓIGÍN 1, WISPR, CABÓG, Interaction of voice quality and f0, SPACT, VOX
Articulatory/acoustic/aerodynamic analysis
see past project: ACCOR
- Coarticulation and connected speech processes
- Primary and secondary articulation of the Irish consonants
- Articulatory database: Irish consonants
- Development of software for combined analysis of EPG, EMMA, acoustid and aerodynamic signals
- Other relevant projects: ACCOR II and COST
Projects and collaborations - Current
Current Funded Research Projects/Working Groups. Principal investigators: Ailbhe Ní Chasaide and Christer Gobl.
- Voice parameterisation and analysis of linguistically relevant voice source variation. The research is part of a large national collaboration aiming at personalised multlingual access and customer care.
Web site: www.cngl.ie
- FastNet
Modelling tone of voice and the non-verbal, paralinguistic dimension in spoken communication. Includes voice analysis and tool development.
Funded by: Science Foundation Ireland (SFI) - CABÓGAÍ II
Text-to-Speech (TTS) for Irish
Irish synthesis availabe at: www.abair.ie
Funded by: Foras na Gaeilge - Prosody of Irish Dialects
The use of intonation, rhythm, voice quality for linguistic and paralinguistic signalling
Previously funded by: Irish Research Council for the Humanities and Social Sciences (IRCHSS)
Web site: www.emotion-research.net
Human-Machine Interaction Network on Emotion
Funded by: EU, Sixth Framework Programme
Projects and collaborations - Completed
- Cabóigín (CABÓIGÍN I)
Text-to-Speech (TTS) for Irish
Web site: www.abair.ie
Funded by: Foras na Gaeilge - Interaction of Voice Quality and Pitch in Prosody
Govt. of Ireland Fellowship to Ailbhe Ní Chasaide
Funded by: Irish Research Council for the Humanities and Social Sciences (IRCHSS) - WISPR (2003-2005)
Welsh and Irish Speech Processing Resources
Funded by: EU INTERREG IIIA Community Initiative Programme - FOCAIL (2002-2003)
Foclóir Canúna Il-mheánach: towards a multi-media electronic pronunciation dictionary of Irish Dialects
Funded by: Foras na Gaeilge - COST 258 (1997-2001)
The naturalness of synthetic speech
EU funded COST action - SPACT 1 and 2 (1995-1998)
Application of speech recognition/synthesis technology in the MIR space station.
Funded by the European Space Agency, project no. 11695/95/NL/JE. - VERIVOX (1997-1998)
Voice variability in speaker verification.
EU funded LTR working group under RTD programme in Information Technologies, Project No. 22760. - CABÓG (1993-1996)
A development system for Irish text-to-speech conversion.
Project funded by ACCUTRON, Limerick and the Irish-American Foundation. - SPEECH MAPS (1992-1995)
Sound-to-gesture inversion in speech.
EU funded ESPRIT/BRA project, No. 6975. - VOX (1992-1995)
The analysis and synthesis of speaker characteristics.
EU funded ESPRIT/BRA working group, No. 6298. - ACCOR I and II (1989-1992, 1992-1995)
Articulatory-acoustic correlations in coarticulation
EU funded ESPRIT/BRA project, No. 3279; EU funded ESPRIT/BRA working group, No. 7098.
Teaching
Courses are taught at undergraduate and postgraduate levels.
Undergraduate courses
- Introduction to Speech Science
- Speech Analysis and Synthesis
- Instrumental Phonetics
- Introduction to Phonetics and Phonology
- Introduction to Language Study
- Experimental Techniques in Speech Science and Clinical Applications
Postgraduate courses
- Signals and Systems for Speech Analysis and Modelling
- Introduction to Phonetics and Phonemics
- Speech Synthesis and Signal Processing
- Experimental Phonetics/Phonology
- Speech Production and Perception
- Current Issues in Speech Science
- Experimental design and laboratory techniques
Facilities
- Semi anechoic chamber
- Network of PCs using windows and linux
- Software systems (developed at CLCS) for source-filter decomposition: inverse filtering and voice source model matching
- Multichannel system for simultaneous recording of articulatory, aerodynamic and acoustic signals
- In house software system for EMMA and EPG analysis
- Multispeech and CSL
- SWELL, Matlab, Alladin and various speech analysis packages
- Gaeltec aerodynamic system (flow and pressure measurement)
- Electropalatography (EPG)
- Movetrack
- Photo electroglottograph (PEG)
- Pitchmeter