← ResearchEmotion8 Classes2025
Emotion

EmotionalEngine™
8-Class Detection.

How we quantify mood, energy, and emotional expressiveness from raw audio — and why it matters for every downstream processing decision.

Access via API →← All Research

Emotion is encoded in the voice at the acoustic level — in pitch variation, energy distribution, timing patterns, and spectral dynamics. The EmotionalEngine analyzes these correlates to classify vocals into eight primary emotional states.

The 8 Classes

Confident, vulnerable, aggressive, euphoric, melancholic, neutral, tense, and transcendent. Each class has distinct acoustic signatures and corresponding processing implications — a vulnerable vocal needs space and air, an aggressive vocal needs presence and bite.

A vulnerable vocal needs space and air. An aggressive vocal needs presence and bite. These are not style preferences — they are acoustic requirements.

Contents
Overview
The 8 Classes
Acoustic Correlates
Processing Implications
Related ResearchVocalDNA™ EngineNeural Vocal Orchestrator
API
POST /v1/vocaldna/extract
API Docs →