PolSeT: Polish Semantics of Timbre Dataset

PolSeT (Polish Semantics of Timbre) is a dataset containing semantic descriptions, acoustic features, and psychoacoustic ratings of musical timbre by native Polish speakers. It was created to facilitate cross-cultural research in music perception and to provide a "ground truth" for the semantics of timbre in the Polish language. It provides data enabling the creation of connections between acoustic, perceptual and semantic spaces. The dataset consists of two experiments: Experiment 1 - Lexicon Generation Free verbalization task where 60 participants described 11 sound stimuli with as many words as they wished. They listened to the sounds via headphones and were required to provide at least one word to describe each sound. All participants were native Polish speakers. Experiment 2 - Semantic Ratings: Semantic differential task where 105 participants rated 18 sound stimuli on 8 scales between opposing verbal descriptors. Participants listened to the sounds via standardized headphone setups and could relisten to the sounds multiple times. After rating all sounds, 5 random sounds were repeated to enable the assessment of listener reliability. All participants were native Polish speakers. Polish scales used in experiment Scales translated into English Ciemny – Jasny Dark – Bright Zabrudzony – Czysty Dirty – Clean Ciepły – Zimny Warm – Cold Dźwięczny – Głuchy Resonant – Dull Ostry – Łagodny Sharp – Gentle Miękki – Twardy Soft – Hard Wybrzmiewający – Stłumiony Ringing – Muted Syntetyczny – Naturalny Synthetic – Natural Associated paper: A data report for this dataset, detailing its creation and the provided data, is available as a preprint: Jasiński, J. (2025). PolSeT: Polish Semantics of Timbre Dataset. arXiv:2606.19987. https://arxiv.org/abs/2606.19987 Please cite this paper when using the dataset.Contact: [e-mail]