A corpus consisting of transcribed speech or conversations, used for phonetic or sociolinguistic analysis.