@enjoys/voice-changer

Plug-and-play real-time browser voice changer. A Rust DSP core compiled to WebAssembly runs inside an AudioWorklet, so the microphone is transformed on the audio render thread (128-sample quanta) — low latency, no per-frame round-trips to JS.

Built-in characters span age & gender — kids (5/10), teens (18), adults (25/30/35), old lady / old man, plus female/male variants and the robot · alien · villain effects, with live control over pitch, drive, ring modulation, vibrato and tone.

Microphone ─▶ WASM voice changer ─┬─ pitch shift (+ vibrato)
                                  ├─ ring modulator   (robot / alien)
                                  ├─ waveshaper drive (villain grit)
                                  └─ tone (HP / LP) ─▶ <audio> / WebRTC / recorder

The presets pitch-shift the whole spectrum (pitch and timbre move together — the classic "deep / chipmunk" character). This is great for fun/disguise effects. True formant-preserving male↔female conversion or sounding like a specific person (neural voice cloning) is a separate, much larger problem and is intentionally out of scope here.

Install / serve the assets

The package ships the compiled voice_changer.wasm and the worklet. They must be served as static files by your app.

1. Build the wasm (needs the Rust wasm target — rustup target add wasm32-unknown-unknown):

   npm run build:wasm    # from the voice-changer/ folder

2. Copy the two assets into your app's public directory:

   dist/voice_changer.wasm            ->  <public>/voice_changer.wasm
   worklet/voice-changer-processor.js ->  <public>/worklet/voice-changer-processor.js

   (Inside this repo, npm run build:vc already copies them into public/.)

   If you serve them from other paths, pass wasmUrl / workletUrl in options.

React

import { useVoiceChanger } from "@enjoys/voice-changer/react";

export function VoiceChangerDemo() {
  const vc = useVoiceChanger({ preset: "robot" });

  return (
    <div>
      {!vc.running ? (
        <button onClick={() => vc.start()}>Start mic</button>
      ) : (
        <button onClick={() => vc.stop()}>Stop</button>
      )}

      <div>
        {Object.entries(vc.presets).map(([key, p]) => (
          <button
            key={key}
            data-active={vc.preset === key}
            onClick={() => vc.setPreset(key as keyof typeof vc.presets)}
          >
            {p.label}
          </button>
        ))}
      </div>

      <label>
        Pitch {vc.config.pitch}
        <input
          type="range" min={-12} max={12} step={1}
          value={vc.config.pitch}
          onChange={(e) => vc.setConfig({ pitch: Number(e.target.value) })}
        />
      </label>

      {/* play the changed voice back (use headphones to avoid feedback) */}
      {vc.stream && (
        <audio
          autoPlay
          ref={(el) => { if (el) el.srcObject = vc.stream; }}
        />
      )}
    </div>
  );
}

The hook returns: start, stop, ready, running, error, config, setConfig, setPreset, preset, presets, stream, level, engine.

Framework-agnostic (vanilla / Vue / Svelte / SIP.js)

import { VoiceChanger } from "@enjoys/voice-changer";

const vc = new VoiceChanger({ preset: "villain" });

const mic = await navigator.mediaDevices.getUserMedia({
  audio: { echoCancellation: false, noiseSuppression: false, autoGainControl: false },
});
const processed = await vc.start(mic);   // a MediaStream

// Play it, or send it to a call:
audioEl.srcObject = processed;
// await vc.replaceSenderTrack(peerConnection);  // swap the outgoing SIP/WebRTC track

vc.setPreset("alien");
vc.setConfig({ pitch: -3, drive: 0.4 });  // live, while talking
vc.setEnabled(false);                     // bypass (raw mic)
await vc.destroy();

Configuration

| Field | Range | Effect | | --- | --- | --- | | enabled | bool | Master on/off (off = raw mic passthrough) | | pitch | -24..24 st | Pitch shift (− deeper, + higher) | | drive | 0..1 | Waveshaper grit / distortion | | ring.hz / ring.mix | Hz / 0..1 | Ring modulator (robot/alien timbre) | | vibrato.hz / vibrato.depth | Hz / semitones | Periodic pitch wobble | | tone.highpass / tone.lowpass | Hz (0 = off) | One-pole tone shaping | | mix | 0..1 | Dry/wet blend | | gain | 0..4 | Output gain |

How it works

| Layer | File | | --- | --- | | DSP core (WASM) | src/lib.rs | | Pitch shifter | src/pitch.rs (two-tap crossfading delay line) | | Ring mod / drive / tone | src/fx.rs | | Worklet host | worklet/voice-changer-processor.js | | Engine | ts/engine.ts | | Presets | ts/presets.ts | | React hook | ts/useVoiceChanger.ts |