JESPER-DE
The current Arcadia experiments use the German JESPER voice as the primary source identity. The voice remains comparatively neutral, rhythmically stable and culturally restrained, making it suitable as a baseline for controlled derivative generation.
A source identity is defined by stability under drift.
The source voice is not treated as a fixed or idealized final form. Instead, it functions as a stable reference space against which tonal, social and linguistic derivatives can be compared.
In repeated listening tests, the JESPER source profile shows unusually high stability under tonal and linguistic variation while becoming more fragile when melodic movement and social proximity increase simultaneously.
High articulation stability across multilingual transfer.
Controlled emotional surface with reduced theatrical movement.
Stable rhythmic structure and restrained melodic behavior.
International intelligibility under moderate drift conditions.
Lightweight observational markers.
Arcadia uses compact observational vectors to describe perceived behavior inside a synthetic identity. These vectors are not treated as objective measurements, but as listening-based reference markers.
JESPER_SOURCE: warmth: 68 clarity: 82 breathiness: 7 rhythm_tightness: 79 melodic_motion: 12 social_proximity: 24 internationality: 91
The vector profile is continuously refined through comparative listening, prompt iteration and derivative testing rather than fixed technical analysis alone.
Controlled movement within a recognizable identity.
Arcadia derivatives are not intended to replace the source identity. They explore how the same recognizable core behaves under tonal, emotional, rhythmic and cultural drift.
Public Arcadia derivatives currently focus on carefully selected cultural drift spaces.
Observation remains central.
The source profile is continuously evaluated through repeated listening and comparative drift testing. Some derivatives preserve identity remarkably well under linguistic change, while others destabilize the perceived social character of the voice.
These observations are treated as practical listening outcomes rather than mathematically deterministic behavior.
Identity cannot be reduced to vectors alone.
Human listening remains essential for evaluating whether a derivative still feels connected to its source, when drift becomes replacement and when a technically successful result no longer feels culturally plausible.