Datasets:
metadata
license: cc-by-nc-sa-4.0
task_categories:
- audio-to-audio
- audio-classification
language:
- en
tags:
- MIDI
- bad-midi
- MIDI quality
- bad
- MIR
- MIDI classification
pretty_name: badmidi
size_categories:
- 10K<n<100K
Bad MIDI
25k+ select bad MIDIs from Discover MIDI Dataset for symbolic music AI and MIR applications
Dataset purpose
- Bad or corrupted MIDIs identification and classification
- Bad MIDI MIR and analysis
Dataset scope
- Comprehensive random sample of all bad MIDI types found in the wild in public/open-source datasets
- Dataset contains corrupted, malformed and otherwise bad MIDIs in all genres and styles
- Dataset features MIDI art MIDIs, black MIDIs, etc.