Improving Tokenization Expressiveness With Pitch Intervals - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Poster Communications Year : 2022

Improving Tokenization Expressiveness With Pitch Intervals

Abstract

Training sequence models such as transformers with symbolic music requires a representation of music as sequences of atomic elements called tokens. State-of-the-art music tokenizations encode pitch values explicitly, which complicates the ability of a machine learning model to generalize musical knowledge at different keys. We propose tracks for a tokenization encoding pitch intervals rather than pitch values, resulting in transposition invariant representations. The musical expressiveness of this new tokenization is evaluated through two MIR classification tasks: composer classification and end of phrase detection. We release publicly the code produced in this research.
Fichier principal
Vignette du fichier
Kermarec-late-breaking-2022.pdf (414.98 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-03877642 , version 1 (05-12-2022)

Licence

Attribution

Identifiers

  • HAL Id : hal-03877642 , version 1

Cite

Mathieu Kermarec, Louis Bigo, Mikaela Keller. Improving Tokenization Expressiveness With Pitch Intervals. 23rd International Society for Music Information Retrieval Conference (ISMIR 2022), Late-Breaking Demo Session, Dec 2022, Bangaluru, India. . ⟨hal-03877642⟩
105 View
116 Download

Share

Gmail Facebook X LinkedIn More