Hi everyone —
I’m excited to share the latest addition to the Harmonic Frontier Audio catalog:
Kalimba (Preview)
A high-fidelity, rights-cleared dataset capturing the natural resonance and percussive clarity of the African lamellaphone, recorded at 96 kHz / 24-bit using dual-microphone phase-aligned capture (Oktava MK-012 + RØDE NT1-A).
Dataset page: Harmonic-Frontier-Audio/Kalimba_Preview · Datasets at Hugging Face
What’s Inside
-
12 clean, mono master files (sustain, dyads, arpeggios, chords, rhythmic figures)
-
Consistent metadata with descriptive “content” fields
-
Preserved natural dynamic range — no normalization or compression
-
Noise floor ≈ –80 dB FS, ready for AI training or timbre analysis
All recordings follow The Proteus Standard™ for dataset provenance and ethical audio use.
Use Cases
-
Instrument recognition / classification
-
Timbre modeling & feature extraction
-
Audio synthesis & sample design
-
Cross-cultural sound research
About Harmonic Frontier Audio
HFA is building the first large-scale, rights-cleared catalog of folk and world instrument datasets for transparent, compliant AI music research.
Other recent previews include Highland Bagpipes, Scottish Smallpipes, Irish Tin Whistle, and Subharmonic Phonation / Vocal Fry.
Feedback welcome!
If you use the Kalimba dataset, or have suggestions for additional world instruments you’d like to see represented, I’d love to hear your thoughts.