Leeds Multi-Session Corpus of Standard Southern British English

The Leeds Multi-Session Corpus of Standard Southern British English is a publicly available, 16-speaker dataset of Southern Standard British English speech, created at the University of Leeds between 2022 and 2024. The speakers are all females in the age range 18-35. The corpus was designed to allow for substantive analysis of intra-speaker variation. To this end, participants performed a series of commonly employed speech production and control tasks repeatedly, in 6 to 8 near-identical sessions. Consecutive sessions were separated by at least 2 weeks. Audio recordings were segmented by task to allow for the compilation of data subsets suitable for a range of phonetic analyses. The speech production tasks include a word production task, a passage reading task, and a sentence reading task in which participants produced sentences in multiple tempo and style conditions. The control tasks included a finger tapping task and a nonsense syllable repetition task. The dataset should be of interest to a range of speech researchers including those working in forensic phonetic practice.