sound.speech.voices

index · sound.speech

Overview

Voice presets for the formant TTS (sound.speech) — named bundles of the synthesis knobs. A caller can pass a preset name, a preset name plus knob overrides, or a bare knob table (see sound.speech).

knobdescription
f0base pitch, Hz
f0_rangeintonation depth: 0 = dead monotone, 1 = natural, >1 theatrical
speedspeech rate multiplier (1 = normal)
formantformant frequency scale = vocal-tract length (≈1.15 reads female, <1 bigger/deeper, >1.2 small creature)
breath0..1, trades voicing for aspiration (1 = full whisper)
flutterslow random f0 wobble (0 = dead steady, robotic)
growl0..1, alternate-pitch-period modulation (vocal fry / creak) — the gravelly roughness; 0.2–0.5 is the useful range
quantizehold synthesis parameters in blocks of N 5 ms frames — the stepped, quantized "machine voice" effect (0/1 = off)
gainoutput gain

These presets are deliberately generic (an application maps its characters onto knobs); tuning them is a listening exercise — see tests/sound/manual/speak.lua.