MUSIB: musical score inpainting benchmark

EURASIP Journal on Audio, Speech, and Music Processing

Empirical Research

Mauricio Araneda-Hernandez¹  Felipe Bravo-Marquez¹  Denis Parra²  Rodrigo F. Cádiz³ 

[1] Department of Computer Science, University of Chile, Santiago, Chile;Millennium Institute for Foundational Research on Data (IMFD), Millennium Institute for Foundational Research on Data (IMFD), Santiago, Chile;National Center for Artificial Intelligence (CENIA), Santiago, Chile;Millennium Institute for Foundational Research on Data (IMFD), Millennium Institute for Foundational Research on Data (IMFD), Santiago, Chile;National Center for Artificial Intelligence (CENIA), Santiago, Chile;Department of Computer Science, Pontificia Universidad Catolica de Chile, Santiago, Chile;National Center for Artificial Intelligence (CENIA), Santiago, Chile;Music Institute and Department of Electrical Engineering, Pontificia Universidad Catolica de Chile, Santiago, Chile;

关键词: Music generation;    Music inpainting;    Music infilling;    Benchmark;    Evaluation;    Reproducibility;   

DOI  :  10.1186/s13636-023-00279-6

 received in 2022-09-13, accepted in 2023-03-14,  发布年份 2023

来源: Springer

【 摘 要 】

Music inpainting is a sub-task of automated music generation that aims to infill incomplete musical pieces to help musicians in their musical composition process. Many methods have been developed for this task. However, we observe a tendency for each method to be evaluated using different datasets and metrics in the papers where they are presented. This lack of standardization hinders an adequate comparison of these approaches. To tackle these problems, we present MUSIB, a new benchmark for musical score inpainting with standardized conditions for evaluation and reproducibility. MUSIB evaluates four models: Variable Length Piano Infilling (VLI), Music InpaintNet, Music SketchNet, and AnticipationRNN, and over two commonly used datasets: JSB Chorales and IrishFolkSong. We also compile, extend, and propose metrics to adequately quantify note attributes such as pitch and rhythm with Note Metrics, but also higher-level musical properties with the introduction of Divergence Metrics, which operate by comparing the distance between distributions of musical features. Our evaluation shows that VLI, a model based on Transformer architecture, is the best performer on a larger dataset, while VAE-based models surpass this Transformer-based model on a relatively small dataset. With MUSIB, we aim at inspiring the community towards better reproducibility in music generation research, setting an example for strongly founded comparisons among SOTA methods.

【 授权许可】

CC BY   
© The Author(s) 2023

【 图 表 】

40517_2023_256_Article_IEq99.gif

Fig. 3

40517_2023_256_Article_IEq98.gif

40854_2023_494_Article_IEq2.gif

Fig. 2

41512_2023_147_Article_IEq89.gif

40517_2023_256_Article_IEq87.gif

40517_2023_256_Article_IEq86.gif

40517_2023_256_Article_IEq85.gif

Fig. 4

40517_2023_256_Article_IEq83.gif

Fig. 13

40517_2023_256_Article_IEq81.gif

Fig. 12

Fig. 2

40517_2023_256_Article_IEq79.gif

Fig. 7

40517_2023_256_Article_IEq78.gif

Fig. 1

40517_2023_256_Article_IEq77.gif

Fig. 11

40517_2023_256_Article_IEq76.gif

Fig. 10

Fig. 1

40517_2023_256_Article_IEq74.gif

Fig. 9

40517_2023_256_Article_IEq73.gif

Fig. 9

Fig. 8

Fig. 5

Fig. 3

Fig. 7

Fig. 3

Fig. 1

Fig. 6

Fig. 2

Fig. 1

Fig. 1

Fig. 5

Fig. 4

40517_2023_256_Article_IEq64.gif

40517_2023_256_Article_IEq62.gif

Fig. 1

Fig. 3

40517_2023_256_Article_IEq60.gif

Fig. 1

40517_2023_256_Article_IEq59.gif

Fig. 3

Fig. 4

40517_2023_256_Article_IEq55.gif

Fig. 3

Fig. 3

Fig. 2

40517_2023_256_Article_IEq53.gif

Fig. 2

40517_2023_256_Article_IEq52.gif

40517_2023_256_Article_IEq51.gif

Fig. 2

40517_2023_258_Article_IEq122.gif

40517_2023_258_Article_IEq121.gif

40517_2023_258_Article_IEq119.gif

Fig. 1

40517_2023_258_Article_IEq115.gif

40517_2023_258_Article_IEq114.gif

40517_2023_258_Article_IEq113.gif

Fig. 1

Fig. 1

40517_2023_256_Article_IEq14.gif

Fig. 8

Fig. 7

40517_2023_256_Article_IEq11.gif

Fig. 5

40517_2023_256_Article_IEq5.gif

40517_2023_256_Article_IEq4.gif

41116_2023_36_Article_IEq117.gif

Fig. 2

41116_2023_36_Article_IEq815.gif

41116_2023_36_Article_IEq811.gif

41116_2023_36_Article_IEq763.gif

41116_2023_36_Article_IEq761.gif

41116_2023_36_Article_IEq759.gif

41116_2023_36_Article_IEq754.gif

41116_2023_36_Article_IEq747.gif

40517_2023_258_Article_IEq51.gif

41116_2023_36_Article_IEq702.gif

41116_2023_36_Article_IEq700.gif

41116_2023_36_Article_IEq698.gif

40517_2023_258_Article_IEq34.gif

41116_2023_36_Article_IEq688.gif

41116_2023_36_Article_IEq655.gif

41116_2023_36_Article_IEq502.gif

41116_2023_36_Article_IEq448.gif