Annotation of vocalizations of wild chimpanzees and additional sound types in recordings of the Xenocanto collection

Bioacoustics research often involves identifying and labeling events of acoustic activity within audio recordings, a process known as annotation. The automation of this time-consuming process is a key objective in the field, but its implementation requires manually annotated data to train machine-learning models. To create a model effective at annotating audio files of various nature, annotations from different species and sound sources are necessary. Here we provide a set of fully annotated audio files where all acoustic events originating from biotic and abiotic sources were annotated. The audio files were obtained from the free online Xenocanto audio file repository. The recordings presented here originate from data collected on wild chimpanzees. Chimpanzee vocalizations were annotated with additional detail, including call type and individual variation.