The CORP-ORAL project aims to create a European Portuguese spontaneous speech corpus, completed with orthographic transcription and the prosodic marking of speech breaks/boundaries of the entire corpus, as well as phonetic transcription of a selection of chunks. The main goal is to place this structured information available on-line to the scientific community for the creation, training and further improvement of speech synthesis and recognition programs. CORP-ORAL is being structured in the following general on-going work phases: a) The recording of 60 hours of conversations between two European Portuguese speakers per conversation (at a time); with a Marantz PMD670 professional recorder and two Beyerdynamic Opus 55.18 MK II microphones. b) The selection and transcription of the corpus material. c) The selection and phonetic transcription of orthographically transcribed and prosodically marked chunks of the corpus.
