ah package

This python package builds the Aligned Hierarchies for music-based data streams. For details on the aligned hierarchies, see Aligned Hierarchies: A Multi-scale structure-based representation for music-based data streams by Kinnaird (ISMIR 2016).

Elements of the package

• Aligned Hierarchies - This is the fundamental output of the package, of which derivatives can be built. The aligned hierarchies for a given music-based data stream is the collection of all possible hierarchical structure decompositions, aligned on a common time axis. To this end, we offer all possible structure decompositions in one cohesive object.
  • Forthcoming Distance metric between two aligned hierarchies
• Forthcoming Aligned sub-Hierarchies - (AsH) - These are derivatives of the aligned hierarchies and are described in Aligned sub-Hierarchies: a structure-based approach to the cover song task
  • Forthcoming Distance metric between two AsH representations
• Forthcoming Start-End and S_NL diagrams
• Forthcoming SuPP and MaPP representations

Contributors

The contributors for this package are:

• Aligned Hierarchies
  • Jordan Moody
  • Lizette Carpenter
  • Eleanor Donaher
  • Denise Nava

Matlab code

Original MATLAB code by Kinnaird can be found here.

Funding sources

This code was developed as part of Smith College's Summer Undergraduate Research Fellowship (SURF) in 2019 and was partially funded by the college's CFCD funding mechanism.