The Million Song Dataset
Open accessAbstract
We introduce the Million Song Dataset, a freely-available collection of audio features and metadata for a million contemporary popular music tracks. We describe its creation process, its content, and its possible uses. Attractive features of the Million Song Database include the range of existing resources to which it is linked, and the fact that it is the largest current research dataset in our field. As an illustration, we present year prediction as an example application, a task that has, until now, been difficult to study owing to the absence of a large set of suitable data. We show positive results on year prediction, and discuss more generally the future development of the dataset.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
LP-MusicCaps-MSD is derived from the ECALS subset of the Million Song Dataset and consists of 520k 30-second clips with a vocabulary of 1054 labels, with each music clip associated with an average of 10.2 labels.
“This dataset is derived from the ECALS subset (Doh et al., 2023c) of the Million Song Dataset (Bertin-Mahieux et al., 2011) and consists of 520k 30second clips with a vocabulary of 1054 labels encompassing various categories such as genre, style, instrument, vocal, mood, theme, and culture.Each music clip is associated with an average of 10.2 labels”
From MusiLingo: Bridging Music and Text with Pre-trained Language Models for Music Captioning and Query Response (2024) · cites this paper as Bertin-Mahieux et al., 2011
The models are trained with the MagnaTagATune dataset (19k training songs) and the Million Song Dataset (200k training songs).
“The MTT models are trained with the MagnaTagATune dataset, and the MSD models are trained with the Million Song Dataset.”
From musicnn: Pre-trained convolutional neural networks for music audio tagging (2019) · cites this paper as [1]
Cited by
Sorted by citations
Kell, A. J. E., Yamins, D. L. K., Shook, E. N., Norman-Haignere, S. V., & McDermott, J. H. (2018) · Neuron · doi:10.1016/j.neuron.2018.03.044 · Cited by 544 · see citing works
musicnn: Pre-trained convolutional neural networks for music audio tagging
Pons, J., & Serra, X. (2019) · arXiv · doi:10.48550/arxiv.1909.06654 · Cited by 67 · see citing works
“These models are trained with two different datasets: the MagnaTagATune dataset (the MTT of 19k training songs [3])1 and the Million Song Dataset (the MSD of 200k training songs [1])2.”
MethodSlokom, M., Hanjalic, A., & Larson, M. (2021) · Information Processing & Management · doi:10.1016/j.ipm.2021.102722 · Cited by 31 · see citing works
COALA: Co-Aligned Autoencoders for Learning Semantically Enriched Audio Representations
Favory, X., Drossos, K., Virtanen, T., & Serra, X. (2020) · arXiv · doi:10.48550/arxiv.2006.08386 · Cited by 20 · see citing works
Deng, Z., Ma, Y., Liu, Y., Guo, R., Zhang, G., Chen, W., Huang, W., & Benetos, E. (2024) · Findings of the Association for Computational Linguistics: NAACL 2024 · doi:10.18653/v1/2024.findings-naacl.231 · Cited by 15 · see citing works
LP-MusicCaps: LLM-Based Pseudo Music Captioning
Doh, S., Choi, K., Lee, J., & Nam, J. (2023) · arXiv · doi:10.48550/arxiv.2307.16372 · Cited by 16 · see citing works
“The data sources are MusicCaps [12], Mag-natagtune [31], and Million Song Dataset [32] ECALS subset [13].”
MethodBootstrapping Language-Audio Pre-training for Music Captioning
Lanzendörfer, L. A., Pinkl, C., Perraudin, N., & Wattenhofer, R. (2025) · ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp49660.2025.10887618 · Cited by 3 · see citing works
SonicVerse: Multi-Task Learning for Music Feature-Informed Captioning
Chopra, A., Roy, A., & Herremans, D. (2025) · arXiv · doi:10.48550/arxiv.2506.15154
“This dataset comprises subsets such as LP-MusicCaps-MC, built from 10-second clips in the MusicCaps dataset (Agostinelli et al., 2023); LP-MusicCaps-MTT, using 30-second clips from Magna-Tag-A-Tune (Law et al., 2009); and LP-MusicCaps-MSD, drawing on the Million Song Dataset (Dataset, 2011).”
- Authors
- Bertin-Mahieux, T., Ellis, D. P. W., Whitman, B., & Lamere, P.
- Affiliations
- Columbia University (US)
- Oracle (United States) (US)
- Venue
- Columbia University
- Publisher
- Columbia University
- Published
- 2011
- Pages
- 591-596
- OpenAlex
- W1556219185
- Cited by
- 944
- References
- 0
- Citations by year
- 2026: 2 · 2025: 27 · 2024: 42 · 2023: 33 · 2022: 32 · 2021: 99 · 2020: 97 · 2019: 100 · 2018: 91 · 2017: 91 · 2016: 76 · 2015: 71 · 2014: 75 · 2013: 58 · 2012: 36
- FWCI
- 83.99
- Type
- journal article
- Language
- en
- Open access
- green
- doi.org
- Locations
- Columbia Academic Commons (Columbia University)
- International Symposium/Conference on Music Information Retrieval
- Topics
- Physical Sciences › Computer Science › Signal Processing › Music and Audio Processing
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Music Technology and Sound Studies
- Social Sciences › Arts and Humanities › Music › Music History and Culture
- Keywords
- Million Song Dataset, audio features, music metadata, popular music, music information retrieval, audio classification
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
