Arcadia Labs IO / VDMS-Light

Lightweight observation instead of artificial precision.

VDMS-Light is a compact observational framework describing how synthetic voices behave under tonal, social, emotional and linguistic drift.

Principle

Vectors are observational markers.

Arcadia does not treat voice vectors as objective scientific measurements. The values are lightweight listening markers used to describe perceived tendencies inside a synthetic identity.

The system intentionally avoids excessive mathematical complexity. Its purpose is not biometric certainty, but practical comparison, drift observation and prompt distillation.

Core Fields

A small number of stable dimensions.

Tonal Layer

Warmth & clarity

Describes resonance behavior, tonal softness, intelligibility and perceived vocal density.

Rhythmic Layer

Movement & timing

Observes pacing, pause structure, melodic motion and rhythmic tightness.

Social Layer

Presence & distance

Describes conversational closeness, social pressure and perceived human proximity.

Emotional Layer

Tension & reflection

Tracks subtle emotional pressure without reducing voices to simple emotional labels.

Example Profile

JESPER-DE source profile.

The current source profile is continuously refined through comparative listening and derivative testing.

JESPER_SOURCE:

warmth: 68
clarity: 82
breathiness: 7
rhythm_tightness: 79
melodic_motion: 12
social_proximity: 24
internationality: 91

Prompt Distillation

Voice systems interpret language more reliably than vectors.

Arcadia does not assume that systems such as ElevenLabs Voice Remixing understand vector values as direct commands. Internal vectors are therefore translated into compact natural-language prompts.

Observation → Distillation → Prompt

SOURCE LISTENING ↓ VECTOR OBSERVATION ↓ SEMANTIC INTERPRETATION ↓ EL-COMPATIBLE PROMPT

VECTOR SHIFT:
  melodic_motion: -10
  social_proximity: -8
  breathiness: -4

DISTILLED PROMPT:
  Restrained melodic phrasing,
  composed social distance,
  minimal breath texture,
  while preserving the calm core identity.

Observation Notes

Some values interact unexpectedly.

Melodic Motion

Strong melodic movement can rapidly alter perceived regional identity and social character.

Social Proximity

Increased conversational closeness often destabilizes otherwise stable source identities.

Breathiness

Small breath adjustments can strongly influence intimacy and emotional realism.

Rhythm Tightness

Rhythmic control affects perceived authority, narrative structure and international intelligibility.

Limitations

VDMS-Light is intentionally incomplete.

The framework does not claim exhaustive modeling of human voice identity. Some effects remain unstable, context-dependent or difficult to separate cleanly.

The goal is practical orientation and controlled comparison — not total abstraction of the human voice.

Human Layer

Listening remains more important than numbers.

VDMS-Light only becomes meaningful through repeated listening and contextual interpretation. Human evaluation remains necessary to determine whether a derivative still feels plausible, stable and culturally coherent.

| vdms light