Docs: multipitch.evaluation macro scores

[Laubeee]

I'm confused by the docs:

mir_eval/mir_eval/multipitch.py

Line 348 in 4d52264

"""Compute multipitch metrics. All metrics are computed at the 'macro' level

in what sense is this the "macro" level? I thought this is what is called the 'micro-average', i.e. average across every single instance, whereas "macro" means per-class scores that are then averaged. Are the docs wrong here?

[craffel]

@rabitt