Para a compilação do c-oral-angola: um corpus de fala espontânea informal do português angolano

Abstract

The paper introduces the architecture and compilation criteria for an Angolan Portuguese spontaneous speech corpus. After a brief introduction about the linguistic scenario in Angola, we present an in-depth description of the recording modalities and treatment related to the multiple sociolinguistic variations documented, with special attention to diaphasic variation. The first twenty-seven recorded texts are then detailed. These will make up a minicorpus, portraying at least 30,000 words. The minicorpus will be prosodically segmented and will display text-to-speech alignment. The last part of the article is dedicated to the methodological steps taken for the corpus compilation: acoustic quality definition, transcription criteria, prosodic segmentation procedures, revision, alignment and statistic validation.