Methodology and software for the semi-automatic annotation and analysis of speech: Human Language Technology meets Linguists
Brigitte Bigi, Daniel Hirst and Dafydd Gibbon
Dresden - September, 6th, 2015
Presentation
Summary
Introduction (15')
Selection of annotation software (15')
Corpus development methodology (45')
Momel and INTSINT (30')
SPPAS (30')
Time Group Analysis (45')
Conclusion and references (15')
Presenters
Brigitte Bigi
Researcher at the CNRS, Laboratoire Parole et Langage, Aix-Marseille Université, Aix-en-Provence, France
Brigitte Bigi
Computer Scientist working in the field of corpora and annotations:
formalization/constitution of corpora,
automatic annotation (mainly at the phonetic level, also at the discourse level),
multimodality (annotation, exploration, extraction of annotated data),
multilinguality (methods and algorithms).
Author and developer of SPPAS - Automatic Annotation of Speech
Daniel Hirst
Directeur de Recherches Emeritus at the CNRS, Laboratoire Parole et Langage, Aix-Marseille Université, Aix-en-Provence, France
Daniel Hirst
Linguist and phonetician, who has been working in the field of prosody and phonology for the last forty years
Founder and former President of the ISCA Special Interest Group on Speech Prosody (SproSIG),
He has developed software for the automatic analysis of speech prosody:
Momel - an algorithm for the automatic factoring of fundamental frequency contours
INTSINT - designed as a descriptive tool for linguistic annotation
ProZed – to test prosodic models of rhythm and melody using an analysis by synthesis paradigm
Dafydd Gibbon
Fakultät für Linguistik und Literaturwissenschaft, Universität Bielefeld, Germany
Dafydd Gibbon
Computational linguist and phonetician, specialising in speech prosody, particularly computational models for tone languages.
He has developed numerous software tools for both research use and advanced teaching of linguistics and phonetics:
the Verbmobil Lexical Database for spoken language,
an industry strength implementation of the inheritance lexicon language DATR, with Grigoriy Strokin, then Lomonossov University, Moscow,
more recently the TGA (Time Group Analyzer) software for speech annotation analysis
Tutorial scopes
This tutorial will report on methodology for the manual and/or automatic annotation and analysis of a recorded speech corpus.
We illustrate the steps to take in the perspective of:
obtaining rich and broad-coverage speech annotation
and initial analysis of such a corpus.
The levels of annotation contain information ranging from Phonetics to Prosody.
Corpus annotation "can be defined as the practice of adding interpretative, linguistic information to an electronic corpus of spoken and/or written language data. 'Annotation' can also refer to the end-product of this process"(Leech, 1997).
