eduzhai > Applied Sciences > Engineering >

Automatic Composition of Guitar Tabs by Transformers and Groove Modeling

  • king
  • (0) Download
  • 20210507
  • Save

... pages left unread,continue reading

Document pages: 8 pages

Abstract: Deep learning algorithms are increasingly developed for learning to composemusic in the form of MIDI files. However, whether such algorithms work well forcomposing guitar tabs, which are quite different from MIDIs, remain relativelyunexplored. To address this, we build a model for composing fingerstyle guitartabs with Transformer-XL, a neural sequence model architecture. With thismodel, we investigate the following research questions. First, whether theneural net generates note sequences with meaningful note-string combinations,which is important for the guitar but not other instruments such as the piano.Second, whether it generates compositions with coherent rhythmic groove,crucial for fingerstyle guitar music. And, finally, how pleasant the composedmusic is in comparison to real, human-made compositions. Our work providespreliminary empirical evidence of the promise of deep learning for tabcomposition, and suggests areas for future study.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×