MUSIB: musical score inpainting benchmark

Simple item page
dc.article.number19
dc.catalogadorcrc
dc.contributor.authorAraneda Hernandez, Mauricio
dc.contributor.authorBravo Marquez, Felipe
dc.contributor.authorParra Santander, Denis
dc.contributor.authorCádiz Cádiz, Rodrigo Fernando
dc.date.accessioned2023-05-12T14:12:59Z
dc.date.available2023-05-12T14:12:59Z
dc.date.issued2023
dc.date.updated2023-05-07T00:02:44Z
dc.description.abstractAbstract 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.
dc.fechaingreso.objetodigital2023-05-07
dc.fuente.origenAutoarchivo
dc.identifier.citationAraneda-Hernandez, M., Bravo-Marquez, F., Parra, D. et al. MUSIB: musical score inpainting benchmark. J AUDIO SPEECH MUSIC PROC. 2023, 19 (2023). https://doi.org/10.1186/s13636-023-00279-6
dc.identifier.doi10.1186/s13636-023-00279-6
dc.identifier.eissn1687-4722
dc.identifier.urihttps://doi.org/10.1186/s13636-023-00279-6
dc.identifier.urihttps://repositorio.uc.cl/handle/11534/66957
dc.information.autorucEscuela de ingeniería; Parra Santander, Denis ; 0000-0001-9878-8761 ; 1011554
dc.information.autorucInstituto de música; Cádiz Cádiz, Rodrigo Fernando ; 0000-0001-5902-1170 ; 4402
dc.language.isoen
dc.nota.accesoContenido completo
dc.pagina.final15
dc.pagina.inicio1
dc.revistaEURASIP Journal on Audio, Speech, and Music Processing
dc.rightsacceso abierto
dc.rights.holderThe Author(s)
dc.rights.licenseAttribution 4.0 International (CC BY 4.0)
dc.rights.urihttps://creativecommons.org/licenses/by/4.0/
dc.subjectMusic generation
dc.subjectMusic inpainting
dc.subjectMusic infilling
dc.subjectBenchmark
dc.subjectEvaluation
dc.subjectReproducibility
dc.subject.ddc780
dc.subject.deweyArtees_ES
dc.titleMUSIB: musical score inpainting benchmark
dc.typeartículo
dc.volumen2023
sipa.codpersvinculados1011554
sipa.codpersvinculados4402
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
13636_2023_Article_279.pdf
Size:
3.16 MB
Format:
Adobe Portable Document Format
Description:
Download
License bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
license.txt
Size:
0 B
Format:
Item-specific license agreed upon to submission
Description:
Download
Collections
Artículos de revistas