Answers are grounded in this paper and cite their sources.
This pilot study aims at comparing two speech corpora in order to capture possible features of two different speech formats. The first corpus is a face-to-face speech corpus, the C-ORAL-BRASIL I (Raso and Mello, 2012) and the second corpus is a phone call corpus, part of the C-ORAL-BRASIL II (Raso; Mello; Ferrari, to appear), both in spontaneous informal Brazilian Portuguese. Through a pipeline prepared in Python, we compared: illocutionary complexity, informational density, overlapping, and disfluencies. Our results indicated that the speakers in the telephone calls had to use more vocal elements to convey a message while in face-to-face interactions the message would be divided in speech and gestures. The findings will be of interest to those who deal with speech corpora in different formats and are interested in selecting certain features for the comparison. Furthermore, this study has methodological ramifications because the elements examined here are very likely to occur in spontaneous speech in any format, and need to be compared carefully.
No takes yet. Share an insight, caveat, or question.
Rocha et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: