EMOPIA: A Multi-Modal Pop Piano Dataset For Emotion Recognition and Emotion-based Music Generation
Bibliographic record
Abstract
EMOPIA (pronounced ‘yee-mò-pi-uh’) dataset is a shared multi-modal (audio and MIDI) database focusing on perceived emotion in pop piano music, to facilitate research on various tasks related to music emotion. The dataset contains 1,087 music clips from 387 songs and clip-level emotion labels annotated by four dedicated annotators. For more detailed information about the dataset, please refer to our paper: EMOPIA: A Multi-Modal Pop Piano Dataset For Emotion Recognition and Emotion-based Music Generation. File Description midis/: midi clips transcribed using GiantMIDI. Filename `Q1_xxxxxxx_2.mp3`: Q1 means this clip belongs to Q1 on the V-A space; xxxxxxx is the song ID on YouTube, and the `2` means this clip is the 2nd clip taken from the full song. metadata/: metadata from YouTube. (Got when crawling) songs_lists/: YouTube URLs of songs. tagging_lists/: raw tagging result for each sample. label.csv: metadata that records filename, 4Q label, and annotator. metadata_by_song.csv: list all the clips by the song. Can be used to create the train/val/test splits to avoid the same song appear in both train and test. scripts/prepare_split.ipynb: the script to create train/val/test splits and save them to csv files. ------ 2.2 Update Add tagging files in tagging_lists/ that are missing in the previous version. Add timestamps.json for easier usage. It records all the timestamps in dict format. You can see scripts/load_timestamp.ipynb for the format example. Add scripts/timestamp2clip.py: After the raw audio are crawled and put in audios/raw, you can use this script to get audio clips. The script will read timestamps.json and use the timestamp to extract clips. The clips will be saved to audios/seg folder. remove 7 midi files that were added by mistake, and also corrected the number in metadata_by_song.csv. 2.1 Update Add one file and one folder: key_mode_tempo.csv: key, mode, and tempo information extracted from files. CP_events/: CP events used in our paper. Extracted using this script, and add the emotion event to the front. Modify one folder: The REMI_events/ files in version 2.0 contain some information that is not related to the paper, so remove it. 2.0 Update Add two new folders: corpus/: processed data that following the preprocessing flow. (Please notice that although we have 1078 clips in our dataset, we lost some clips during steps 1~4 of the flow, so the final number of clips in this corpus is 1052, and that's the number we used for training the generative model.) REMI_events/: REMI event for each midi file. They are generated using this script. -------- Cite this dataset @inproceedings{{EMOPIA}, author = {Hung, Hsiao-Tzu and Ching, Joann and Doh, Seungheon and Kim, Nabin and Nam, Juhan and Yang, Yi-Hsuan}, title = {{MOPIA}: A Multi-Modal Pop Piano Dataset For Emotion Recognition and Emotion-based Music Generation}, booktitle = {Proc. Int. Society for Music Information Retrieval Conf.}, year = {2021} }
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.002 |
| Meta-epidemiology (narrow) | 0.003 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.018 | 0.024 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".