ECHO REMOVER
Remove echo from audio
Recorded in a bare room, a kitchen or a hall? This tool measures how long the room rings, then pulls down the echo tail behind each word so the voice sounds closer. Compare before and after, then download.
Hear a generated voice recorded in a simulated echoey room, then measure the reverb and remove it.
Watery or metallic voice? Lower the strength or the maximum reduction. Still distant? Raise the strength.
Remove echo from a recording in three steps
- Choose a voice recording or a video. The page estimates the reverb time from the pauses between words.
- Keep the measured reverb time, or set it yourself, and choose the strength and the maximum reduction.
- Press Remove echo, compare the before and after waveforms and audio, and download. Video keeps its picture.
How the echo remover works
Room echo decays at a steady rate: in a room with a one-second reverb time, sound fades by 60 dB in one second. The tool first measures that rate from your own recording, by finding the moments after a word ends and fitting a line to how fast the level falls (a T20 fit, the same idea acoustic engineers use with a test signal). It then splits the audio into short overlapping frames of about 23 ms and, in every frequency band, predicts how much of the current sound is just the echo of what was there 50 ms earlier. That predicted echo is subtracted, the direct voice is kept, and the frames are joined back together. The method is late-reverberation spectral subtraction, published by Lebart, Boucher and Denbigh in 2001.
Getting a good result
Start with the measured reverb time and the default strength of 1.5. If the voice still sounds distant, raise the strength towards 2 or the reverb time a little. If consonants sound clipped or the voice turns watery and metallic, lower the strength or set a smaller maximum reduction, such as 10 dB. Small rooms with a short, boxy echo respond best. Large halls with several seconds of reverb improve, but some of the space will always remain, because the early reflections arrive too close to the voice to separate.
When the measurement says nothing
The reverb time can only be measured when the recording has pauses where the echo dies away on its own. Continuous music, constant background noise or a recording that was already gated may give no reading; the page then starts at 0.6 seconds, a typical furnished room, and you can adjust it by ear. A reading under 0.3 seconds means the room is already fairly dry and this tool will change little; noise, not echo, is probably what you hear.
Limits and privacy
No tool can fully undo a reverberant room, and this is signal processing, not an AI model, so it will not rebuild a voice that is buried in echo. It works best on speech; on music it also shortens the natural decay of notes. It does not cancel acoustic echo from speakerphones or video calls, where your voice comes back from the other side. Record close to the microphone with soft furnishings around you to avoid echo in the first place. Everything runs in your browser and the file is never uploaded.
Frequently asked questions
How do I remove echo from a recording?
Choose the file, let the page measure the reverb time, and press Remove echo. Compare before and after, adjust the strength if needed, then download the result.
Can it remove echo from a video?
Yes. Choose the video and the cleaned soundtrack is put back with the original picture, or pick audio only to get an MP3.
Is this the same as noise reduction?
No. Noise is a steady hiss or hum that is always there. Echo follows each word and dies away. Use the noise reducer for hiss and this tool for room echo; you can run both.
Why does my voice sound watery after removing echo?
The strength is too high for the recording. Lower it to about 1, or set the maximum reduction to 10 dB, and run it again.
All processing runs locally in your browser memory via WebAssembly and Web Audio. Files never leave your device. Built-in peak limiting prevents audio clipping.
AudioClover Engineering · Updated October 2026KEEP IN THE LOOP
A better workflow for your next recording.
Get tool updates and audio tips. Batch presets and a podcaster plan are on the roadmap.