EMOPIA

dataset·music emotion audio midi·active

Pop-piano audio and MIDI clips labeled for perceived emotion, supporting recognition and emotion-conditioned generation.

Verified facts

Official site	https://annahung31.github.io/EMOPIA/ ↗
Geography	Taiwan
arxiv	2108.01374
size	clips: 1087 · songs: 387 · annotators: 4
scope	Perceived emotion in pop piano for audio and symbolic tasks.
genres	pop piano; anime; K-pop; Western pop; soundtracks
license	CC BY-NC-SA 4.0
steward	EMOPIA authors
creators	Hsiao-Tzu Hung; Joann Ching; Seungheon Doh; Nabin Kim; Juhan Nam; Yi-Hsuan Yang
languages	not_applicable
modalities	audio; MIDI; clip-level emotion labels
geographies	East Asia; Global
access method	Project page, GitHub and Zenodo.
intended uses	music emotion recognition; emotion-conditioned generation
related papers	EMOPIA: A Multi-Modal Pop Piano Dataset
annotation method	Four dedicated annotators assigned valence-arousal quadrant labels.
collection method	Pop-piano performances curated and segmented into musically coherent clips.
current availability	Project links available.
commercial use limits	Commercial use prohibited.
us market scope basis	Publicly available or materially used in US-facing music/audio-AI research.
limitations biases disputes	Small, culturally concentrated pop-piano corpus.; Perceived-emotion labels are subjective.
redistribution restrictions	Attribution, noncommercial and share-alike conditions apply.
related models tools benchmarks	MARBLE