How to clean a noisy interview recording — ai denoise, no upload
- Step 1Upload your noisy recording — Drop your field or remote interview recording.
- Step 2Run AI noise reduction — AI noise reduction runs locally in your browser.
- Step 3Download cleaned audio — Download the denoised, publication-ready audio.
Frequently asked questions
What types of noise does the AI denoiser remove?+
RNNoise is trained primarily on steady-state and semi-stationary noise: HVAC and air conditioning hum, room tone, fan noise, keyboard and mouse clicks, broadband white noise, and electrical hum (50/60 Hz). It is most effective on speech recordings. Heavy music or non-vocal audio may produce artefacts.
Why does the denoised audio sound metallic or 'watery'?+
This 'underwater' artefact appears when the denoiser aggressively suppresses noise in gaps between speech. It is most common when the original noise floor is very high relative to the voice. Consider recording with a better microphone position first — AI denoising works best when the signal-to-noise ratio is already reasonable (above −15 dB SNR).
Is RNNoise better than Krisp or NVIDIA RTX Voice?+
RNNoise is an older but battle-tested algorithm. Krisp and NVIDIA RTX Voice use larger transformer-based models that handle more complex noise scenarios. For simple background noise on voice recordings, RNNoise is typically sufficient and fast. For severe or variable noise conditions, a commercial denoiser may give better results.
Privacy first
All audio processing runs locally in your browser using FFmpeg compiled to WebAssembly. No file is ever uploaded — only metadata counters are saved for signed-in dashboard stats.