Handytool
აუდიოუფასოახორციელდება ადგილობრივად

Voice enhancer

Remove background noise from voice recordings in your browser.

.mp3.wav.ogg.m4a.aac.flac.webm.opus
Loading model…

Runs entirely in your browser.

Voice enhancer შესახებ

Voice enhancer cleans up recordings made on bad microphones, in noisy rooms, or over fans, traffic, and keyboard chatter. Drop in an MP3, WAV, M4A, OGG, or FLAC and Handytool runs RNNoise — a recurrent neural network trained on hundreds of hours of speech — entirely inside your browser to suppress background noise while keeping the voice intact. Nothing is uploaded; your audio stays on your device. Tweak the strength slider to find the right balance between aggressive cleanup and a natural sound, then download the result as a 48 kHz WAV.

Voice enhancer მახასიათებელი

  • 01

    Neural-network noise suppression

    RNNoise is a small recurrent neural network used in production by Discord, Jitsi, and Mumble. It removes hum, hiss, fans, traffic, room noise, and keyboard sounds while preserving voice — far better than a plain noise gate.

  • 02

    Adjustable cleanup strength

    Voices that are already clean don't need full denoising. Slide the strength down to mix the cleaned signal back with the original for a natural result, or push it to 100% on heavily noisy clips.

  • 03

    Runs locally, no upload

    The denoiser is a 125 KB WebAssembly module that loads once and stays cached. Audio is decoded, processed, and downloaded entirely on your machine — no server round-trip, no account, no usage limits.

Voice enhancer ხშირი კითხვები

How do I clean up a noisy voice recording?
Drop your audio file (MP3, WAV, M4A, OGG, FLAC, or WebM) into the tool and click Enhance. The denoiser runs in your browser and gives you a 48 kHz WAV with the background noise removed.
What kinds of noise does it remove?
Steady noise like fans, AC hum, traffic, and street noise is suppressed best. It also helps with keyboard typing, paper rustling, and muffled room reverb. Very loud crashes or another person talking over the speaker are harder cases.
What does the strength slider do?
100% applies the full neural-network output. Lower values mix part of the original signal back in — useful when full denoising sounds slightly robotic or when the noise is light to begin with.
Is the audio uploaded to a server?
No. The model and your file stay in your browser. The denoising network is a small WebAssembly module that runs locally on your CPU, so nothing leaves your computer.
What output format do I get?
A mono 48 kHz WAV file in 16-bit PCM. WAV is uncompressed and supported by every audio editor, podcast host, and video tool. Convert it to MP3 with the Convert audio tool if you need a smaller file.
How long can the recording be?
Files up to 200 MB are accepted. Processing speed is roughly 5–10× real-time on a modern laptop, so a 10-minute interview cleans up in about a minute.

დაკავშირებული ინსტრუმენტები

აუდიო →

სხვა ინსტრუმენტების გამოკვლევა

ყველა ინსტრუმენტი →