WaivOps WRLD-LP Dataset For Music ML

Patchbanks announces the release of WaivOps WRLD-LP Dataset, an open audio resource dataset capturing symbolic sounds of world percussion music. This dataset is designed to support research in machine learning and AI music generation projects.

The WRLD-LP Dataset is the third release under the WaivOps project, containing 3,162 audio recordings recorded in uncompressed stereo WAV format. These recordings were created using an internal sample dataset, complemented by note-dense MIDI files from a code-based music generation system. The dataset also includes output generations from a MIDI transformer model trained on over 30,000 MIDI files.

Potential applications of this dataset include tempo detection and classification, drum rhythm analysis, audio-to-MIDI conversion, source separation, automated mixing, music information retrieval, AI music generation, and signal processing.