Klatt formant cascade · no samples, no network · every waveform computed here

TTS Studio

Sample rate
—
Engine
—
Timing
Mora

A bench for the FormantTTS synthesiser. Type text, shape the voice, then read back exactly what the engine will say — the kana it derived and the phonemes it will articulate — before you hear a sound.

Console

input

0 characters limit 2,000

Language

Detecting…

118 Hz
1.00×
Voice presets

Ctrl/⌘ + Enter speaks

Signal

output
Press Speak to plot the signal
Duration — Samples — Rate — Peak —

Phonetic analysis

—

Reading

—

Phoneme tape

consonant vowel pause