Skip to main content

Crate elara_voice

Crate elara_voice 

Source
Expand description

ELARA Voice State - Parametric voice encoding

This is NOT audio PCM or Opus codec. This is the STATE OF SPEECH for reality synchronization.

§Philosophy

Traditional voice: Encode audio samples → transmit → decode ELARA voice: Extract speech parameters → transmit state → synthesize

Key concepts:

  • Voice as parametric state (pitch, energy, spectral envelope)
  • Prediction under packet loss (voice continues)
  • Graceful degradation (L0: full voice → L5: presence only)

§Degradation Ladder (Voice)

  • L0: Full parametric voice (pitch, formants, prosody, emotion)
  • L1: Reduced formants, simplified prosody
  • L2: Pitch + energy only, robotic quality
  • L3: Symbolic (speaking/silent indicator)
  • L4: Presence pulse only
  • L5: Identity heartbeat

Re-exports§

pub use encoding::*;
pub use frame::*;
pub use prediction::*;
pub use state::*;
pub use synthesis::*;

Modules§

encoding
Voice Encoding - Wire format for voice state
frame
Voice Frame - Time-sliced voice state
prediction
Voice Prediction - Continuity under packet loss
state
Voice State - The state of speech
synthesis
Voice Synthesis - Reconstruct voice from parametric state