> ML_DATASET // AUDIOSET-GOOGLE_v1.0
AudioSet: A Large-Scale Dataset of Manually Annotated Audio Events
Google Research (Gemmeke et al. 2017) · General Audio & Sound Event Detection · 2,084,320 10-second YouTube sound video clips spanning 527 sound classes organized in an ontology
General Audio & Sound Event DetectionCC-BY-4.0 (Annotations) / YouTube Terms of Service (Media)2,084,320 10-second YouTube sound video clips spanning 527 sound classes organized in an ontologyopen
Dataset Profile & Characteristics
Label Type:Multi-label audio event classifications organized in hierarchical ontology
Languages:en
License Tier:permissive-open-source
Modalities:audio
Intended Use
- Acoustic event detection, audio tagging, and sound recognition pretraining
Prohibited / Discouraged Use
- Eavesdropping audio surveillance systems
Bias, Leakage & Privacy Risk Analysis
Privacy / Sensitive Data Risks:
YouTube consumer video clips; ambient human voices, household sounds.
Known Bias:
YouTube video upload bias; music and speech co-occurring with background sounds.
Known Benchmark Leakage:
Standard train, balanced-train, and evaluation partitions.