Cantoría Dataset

Cantoría dataset is a multi-track dataset of 11 songs performed by the professional vocal quartet Cantoría, specialized in the performance of vocal polyphony from the Iberian Golden Age repertoire. Cantoría dataset is released as part of the following Ph.D. dissertation: Helena Cuesta. Data-driven Pitch Content Description of Choral Singing Recordings. PhD thesis, Universitat Pompeu Fabra. 2022 (to appear). The dataset includes the multi-track recordings and automatically extracted F0 trajectories. Particularly, we provide F0 trajectories extracted with pYIN [1] and CREPE [2]. Cantoría dataset contains the following songs: Sus sus sus, written by Bartomeu Cáceres. Riu riu chiu, an anonymous villancico. El Jubilate, written by Mateo Flecha “el viejo”. Virgen Bendita sin par, written by Pedro de Escobar. Hoy comamos y bebamos, written by Juan del Encina. La Negrina, written by Mateo Flecha “el viejo”. Teresica hermana, written by Mateo Flecha “el viejo”. Corten espadas afiladas, an anonymous secular villancico. La Justa, written by Mateo Flecha “el viejo”. La Bomba, written by Mateo Flecha “el viejo”. Yo me soy la morenica, an anonymous secular villancico. The recorded pieces are accompanied by an organ, recorded at the beginning as a reference track with electronic organ. After the accompaniment recording, each singer was recorded separately, singing all songs. Cantoría dataset comprises one full run of each song, performed by the SATB quartet and the organ. It includes each individual audio track and the SATB mixture of the four singers, with and without the organ. The dataset is compressed into a zip file, which also includes a README file with specific information about the folder structure and filenames. [1] Mauch, Matthias, and Simon Dixon. "pYIN: A fundamental frequency estimator using probabilistic threshold distributions.". In Proceedings of the IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP). 2014. [2] Kim, J. W., Salamon, J., Li, P., & Bello, J. P. "CREPE: A Convolutional Representation for Pitch Estimation. In Proceedings of the International Conference on Acoustics, Speech and Signal Processing (ICASSP). 2018