The C-ORAL-ROM book and DVD provide a unique set of comparable corpora of spontaneous speech for the main Romance languages, French, Italian, Portuguese and Spanish. The corpora are accompanied by comparative linguistic studies, models and standard linguistic measures of spoken language variability. Each corpus is built to the same design using identical sampling techniques, and each corpus is presented in multimedia format, allowing simultaneous access to aligned acoustic and textual information. Texts are headed with information about provenance, participants, etc. and the transcriptions show changes of speaker. Speech acts are tagged according to the evidence of prosodic criteria. Each corpus totals 300, 000 words and presents formal and informal speech in a variety of contexts of use, dialogue structure and text genres, semantic domains and speech act typologies. The corpora have great statistical relevance for spoken language structures and can address key issues in human language technology such as speech recognition in unrestricted discourse, the suitability of speech synthesis in natural prosody, and multilingual applications of the spoken language interface. The work provides new data and innovative theoretical perspectives that are relevant for corpus linguistics, romance linguistics, syntactic theory, speech and prosody research, and second language acquisition.pus, where these elements were not considered, we achieved a success rate of 91.5%), if we excluded these errors from the error rate ... Since a large part of these elements was mistagged, a manual revision of this subcorpus was mandatory.
|Author||:||Emanuela Cresti, Massimo Moneglia|
|Publisher||:||John Benjamins Publishing - 2005-05-09|