Camera, Mic & Media
Voice Changer - Pitch Shift Your Voice in the Browser
Record your voice and play it back at a different pitch - chipmunk, deep monster, or anything in between. No upload needed, runs entirely in your browser.
How the voice changer works
Your microphone input is captured using the MediaRecorder API and stored locally as
an audio blob. When you play back the recording, the HTML5 audio element's
playbackRate property shifts the pitch - faster playback raises pitch, slower playback
lowers it. This technique is time-domain pitch shifting: it changes frequency
and duration simultaneously (unlike professional phase-vocoder algorithms that decouple the
two).
Audio effects reference
| Effect | How it works | Typical use |
|---|---|---|
| Pitch up (chipmunk) | playbackRate > 1.0 - faster replay = higher pitch | Cartoon voices, comedy |
| Pitch down (deep) | playbackRate < 1.0 - slower replay = lower pitch | Monster voices, radio DJ effect |
| Reverb | Convolution with an impulse response (simulates room reflections) | Concert hall, cathedral |
| Echo / delay | Copies of the signal played back after a set delay time | Slap-back, dub, stadium |
| Distortion / overdrive | Clips or soft-saturates the waveform, adding harmonics | Guitar amp, robot voice |
| Low-pass filter | Removes high frequencies above a cutoff Hz | Telephone or walkie-talkie effect |
| Bit crush | Reduces bit depth, creating quantisation noise | Retro / 8-bit sound |
Privacy
Your voice recording never leaves your device - no server or network request is made.
Phase vocoder vs. playback-rate pitch shifting
The pitch effects in this tool are implemented by changing the audio
playback rate - playing audio faster raises pitch but also shortens duration. Professional-grade
pitch changers use a
phase vocoder, which shifts pitch independently of playback speed by
splitting audio into frequency bands, shifting each band, and reassembling. Phase vocoders can
raise or lower pitch without changing duration at all. Implementing a phase vocoder in the
browser requires an AudioWorklet and significant DSP code; the playback-rate approach
is far simpler and good enough for fun effects.
Reverb types
Reverb simulates how sound reflects off walls and surfaces in a physical space. Two main approaches:
- Algorithmic reverb: uses feedback delay networks to approximate a space. Fast and controllable but can sound artificial.
- Convolution reverb: convolves the dry signal with an impulse response (IR) - a recording of how a real space responds to a brief impulse. Sounds extremely realistic. The Open AIR library provides free IRs of real spaces ranging from small rooms (~100 ms decay) to cathedrals (2–5 second decay).
Use case examples
- Podcast warmth: a gentle pitch shift down 10–15% and a light reverb can add depth to a thin microphone sound.
- Gaming anonymity: voice alteration prevents friends from recognizing your voice in multiplayer games with voice chat.
- Accessibility: pitch shifting can help users with voice conditions communicate more comfortably.
- Teaching demonstrations: exaggerated pitch effects make audio concepts engaging for students.