Pitch distributions in a very large corpus of spontaneous Finnish speech

Speakers differ in the pitch range they use in their speech. In order to analyze the functional aspects of pitch, the typical pitch range of each individual is needed as reference. However, systematically collected pitch data from a sufficiently large corpus have not been previously available. We analyze the pitch distributions of individual speakers in a subset of the Donate Speech Corpus, collected from speakers of Finnish in 2020-2021. We report pitch analysis results based on samples from 8197 speakers and 1475 hours of speech. We compare the results obtained from male and female speakers in different age groups.