Rodd, J., Bosker, H. R., Ernestus, M., Ten Bosch, L., & Meyer, A. S.
(2019). PiNCeR: a corpus of cued-rate multiple picture naming in Dutch. PsyArXiv. doi:10.31234/osf.io/wyc6h.
PiNCeR is a corpus of speech recordings from Dutch speakers who named pictures at different speaking rates. Participants named pre-familiarised ˈ(C)CV.CVC words (e.g., snavel [ˈsnaː.vəl] “beak”) from line drawings displayed in groups of 8 arranged on a ‘clock face’. A cursor moved clockwise from picture to picture to indicate at which of three trained rates (fast, medium and slow) participants were required to name the pictures. Annotation was performed using the POnSS tool (Rodd, Decuyper, & ten Bosch, 2019), where manual and automatic segmentation is combined to yield accurate word onsets and offsets. To detect the onset and offset times of syllables within words, we identified excursions of above-average acoustic instability between the vowel of the initial syllable and the first consonant of the second syllable (Rodd, Bosker, ten Bosch, & Ernestus, 2019). This approach was licensed by careful control of segmental content in the target words to maximise correspondence between acoustics and articulation. The PiNCeR corpus was intended for use in modelling control of speaking rate (Rodd, Bosker, Ernestus, et al., 2019), but may be of interest for other purposes. Trial-level recordings from two related experiments are made available for 25 participants (12 for Experiment 1, 13 for Experiment 2), along with the onset and offset times of the words and the syllables.