Hi everyone,
I’ve just released a new preview dataset from the Human Vocality Primitives series by Harmonic Frontier Audio:
Plosives and Non-Lexical Consonant Bursts (Preview)
This preview provides a small, representative subset of gesture-level vocal events capturing plosive articulations and short-duration non-lexical consonant bursts. The focus is on controlled vocal tract closure and release, transient airflow behavior, and clean onset/decay characteristics — recorded and structured explicitly for AI training and research workflows.
What’s included
-
3 high-fidelity mono WAV files
-
Recorded at 96 kHz / 24-bit
-
Non-lexical, non-semantic vocal material
-
Structured metadata aligned with the Human Vocality Primitives schema
-
Rights-cleared, performer-consented recordings following The Proteus Standard™ for dataset provenance
Intended use
This preview is intended for:
-
evaluating dataset quality and structure
-
articulation-aware speech and voice modeling
-
phonetics and speech production research
-
testing systems that model consonant onsets and transient vocal gestures
This is not a full dataset.
The complete Plosives and Non-Lexical Consonant Bursts dataset includes a broader and more balanced articulatory inventory and is available for commercial licensing.
Dataset link:
Plosives and Non-Lexical Consonant Bursts
More information about Harmonic Frontier Audio and the Proteus Standard:
https://harmonicfrontieraudio.com
Happy to answer questions about structure, metadata, or intended use cases.
— Blake