Melody Extraction in AcousticBrainz

Hi everyone,
I love the data that AB provides, but was curious as to why the melody processing algorithms available in Essentia are not available in the data you provide. Are there any plans to include this in future versions of the dataset?

Thanks in advance!

In the medium term we’re not interested in adding melody extraction to the acousticbrainz tools for a few reasons:

  • We want to focus on improving data for general classification and recognition tasks
  • We’re not yet sure of the implications of including a derivative work of the audio in a public dataset

Perhaps as we continue AB development we’ll think about this issue again.