Soundata: reproducible use of audio datasets

Soundata is an open-source Python library for working with audio datasets in a programmatic and standardized way. It removes the need for writing custom loaders and improves reproducibility by providing tools to validate data against a canonical version. It speeds up research pipelines by allowing u...

Descripción completa

Detalles Bibliográficos
Autores: Fuentes, Magdalena, Plaja-Roglans, Genís, Cortès Sebastià, Guillem, Khandelwal, Tanmay, Miron, Marius, Serra, Xavier, Bello, Juan Pablo, Salomon, Justin
Tipo de recurso: artículo
Estado:Versión publicada
Fecha de publicación:2024
País:España
Institución:Universitat Pompeu Fabra
Repositorio:Repositorio Digital de la UPF
OAI Identifier:oai:repositori.upf.edu:10230/70528
Acceso en línea:http://hdl.handle.net/10230/70528
http://dx.doi.org/10.21105/joss.06634
Access Level:acceso abierto
Palabra clave:Audio
Environmental-sound
Bioacoustics
Dataset
Urban-sound
Descripción
Sumario:Soundata is an open-source Python library for working with audio datasets in a programmatic and standardized way. It removes the need for writing custom loaders and improves reproducibility by providing tools to validate data against a canonical version. It speeds up research pipelines by allowing users to quickly download a dataset, validate that the dataset is complete and correct, and load it into memory in a standardized and reproducible way. It is designed to work with bioacoustics, environmental, urban, and spatial sound datasets; to be easy to use and easy to contribute to; and to increase reproducibility and standardize the usage of sound datasets in a flexible way.