retro formant speech / local synth / MP3 / MP4
Type a script, press play, and the browser breaks the words into pronounceable sounds, feeds them to a local formant synthesizer, and renders a rough late-70s computer voice.
instrument / 01
Paste a line, a poem, or a script. The page turns words into phoneme groups, renders them with a local formant engine, and can export either audio or a moving soundmap.
Or drop in a song or any recording and the robot will follow it, note for note, in time.
The script is split into pronounceable sound groups before it reaches the synthesizer.
The formant synth runs locally. Your text does not need a server round trip to become audio.
It is not polished neural speech. It is a brittle, direct, low-quality robot voice for character and texture.
Download the dry robot voice as audio, or render a cobalt soundmap video in the same visual system as this page.
score / 07
Write or paste the source text.
Choose a preset and tune speed, pitch, melody, and volume.
Play, stop, or restart until the rough machine cadence feels right.
Save the rendered MP3 without leaving the page.
score / 08
Plain text in the textarea, then hidden phoneme grouping before synthesis.
Klattsch-style formant synthesis, tuned through local voice presets.
Text and audio rendering happen inside the browser runtime.
Immediate playback plus downloadable MP3 audio or MP4 soundmap video.
archive / 09
No. The charm is the rough robotic formant sound. It is useful when you want texture, not a polished assistant voice.
No. The speech conversion, playback, and MP3 render happen locally in the browser.
The page converts visible words into phoneme groups, applies the selected voice contour, then schedules those targets through the formant engine.
Use MP3 when you only need the voice. Use MP4 when the soundmap itself should become the visual artifact.