🔇

Noise Remover

Reduce background noise from audio recordings

📚
Learn more — how it works, FAQ & guide
Click to expand

Free Online Noise Remover — Clean Up Audio Recordings

The Noise Remover is a free, browser-based audio cleaning tool that reduces background noise from your recordings. Using the Web Audio API, it applies a combination of high-pass filtering, low-pass filtering, and dynamics compression to suppress unwanted sounds like hum, hiss, fan noise, and low-frequency rumble. All processing happens locally in your browser, ensuring complete privacy for your audio data.

Understanding Background Noise in Audio

Background noise in audio recordings generally falls into predictable frequency ranges. Low-frequency rumble from traffic, air conditioning, and building vibrations typically occupies the 20-200 Hz range. Electrical hum from power lines sits at 50 Hz (Europe) or 60 Hz (North America) and their harmonics. High-frequency hiss from preamps, analog circuits, and poor-quality microphones concentrates above 4000 Hz. By strategically filtering these frequency ranges, you can significantly improve the clarity of voice recordings, podcasts, and interviews.

How the Noise Reduction Chain Works

This tool processes your audio through three Web Audio API nodes in sequence:

  • High-Pass Filter (BiquadFilterNode): Removes all frequencies below the cutoff point. Default is 80 Hz, which eliminates most rumble without affecting speech. Raising it to 150-200 Hz is useful for removing persistent hum but may thin out deeper voices.
  • Low-Pass Filter (BiquadFilterNode): Removes all frequencies above the cutoff point. Default is 8000 Hz, which preserves most speech clarity while cutting high-frequency hiss. Lowering it to 4000-5000 Hz provides more aggressive noise removal but reduces sibilance.
  • Dynamics Compressor (DynamicsCompressorNode): Acts as a noise gate by compressing signals below the threshold. The threshold determines which signal levels are considered "noise" and suppressed. A setting of -40 dB is a good starting point for typical room noise.

Adjusting Settings for Your Audio

The effectiveness of noise reduction depends heavily on the characteristics of your specific noise and the content you want to preserve. Here are guidelines for common scenarios:

  • Voice recordings with fan noise: High-pass 100 Hz, low-pass 7000 Hz, threshold -35 dB, strength 60%.
  • Podcast with mild hiss: High-pass 80 Hz, low-pass 9000 Hz, threshold -45 dB, strength 40%.
  • Phone call recording: High-pass 200 Hz, low-pass 3500 Hz, threshold -30 dB, strength 70%.
  • Music with subtle noise floor: High-pass 30 Hz, low-pass 10000 Hz, threshold -50 dB, strength 25%.

Always use the Before/After toggle to compare results. If the cleaned version sounds muffled or artificial, reduce the strength slider and widen the filter range.

Waveform Visualization

The canvas-based waveform display shows the amplitude envelope of your audio file. Tall peaks indicate loud sections (speech, music), while flat areas indicate silence or low-level noise. This visual representation helps you identify problematic sections and evaluate how your settings affect the overall signal. After processing, you can compare the original and cleaned waveforms to see where noise was reduced.

Limitations of Browser-Based Noise Reduction

This tool uses frequency-domain filtering, which is effective for noise that occupies specific frequency bands consistently. It cannot perform spectral subtraction, which requires analyzing a "noise profile" from a silent section and subtracting it from the entire recording. For professional noise reduction that adapts to changing noise patterns, dedicated software like Audacity (free), iZotope RX, or Adobe Audition is recommended. However, for quick cleanup of common noise issues, this browser-based approach delivers substantial improvement without any software installation.

Tips for Cleaner Recordings

  • Record in the quietest room available, away from windows and HVAC systems.
  • Use a directional microphone (cardioid pattern) to reject off-axis noise.
  • Position your microphone 6-12 inches from your mouth for optimal signal-to-noise ratio.
  • Turn off fans, air conditioning, and unnecessary electronics before recording.
  • Record a few seconds of "room tone" (silence) at the start for reference.

How to Use the Noise Remover

  1. 1

    Upload your audio

    Drag and drop an audio file or click to browse. Supported formats include MP3, WAV, OGG, FLAC, and AAC.

  2. 2

    Adjust noise reduction settings

    Use the sliders to configure noise gate threshold, high-pass filter (cuts low rumble), low-pass filter (cuts high hiss), and overall strength.

  3. 3

    Apply and preview

    Click Apply to process the audio. Use the Before/After toggle to compare the original and cleaned versions side by side.

  4. 4

    Download the result

    Click Download to save the cleaned audio as a WAV file to your device.

Frequently Asked Questions

What kind of noise does this remove?
This tool is effective against consistent background noise like hum, hiss, fan noise, air conditioning, and low-frequency rumble. It uses frequency filtering (high-pass and low-pass) plus dynamics compression. It is not spectral subtraction and will not remove intermittent noises like coughs or clicks.
Will it affect voice quality?
Some voice quality impact is unavoidable with filtering. The high-pass filter removes low rumble but can thin out deep voices if set too high. The low-pass filter removes hiss but can dull sibilance. Start with moderate settings and adjust based on preview.
What is the noise gate threshold?
The noise gate uses a DynamicsCompressorNode to suppress audio below the threshold level. Signals quieter than the threshold are compressed aggressively, effectively silencing low-level background noise while preserving louder speech or music.
How accurate is the waveform display?
The waveform is drawn by sampling the raw PCM data from the AudioBuffer onto a Canvas element. It shows the amplitude envelope of your audio, not a frequency spectrum. Louder sections appear taller, silent sections appear flat.
Can I process long recordings?
Yes, but processing time and memory usage scale with file length. A 10-minute WAV file requires approximately 100 MB of RAM for the uncompressed AudioBuffer. Most modern devices handle this well. For files over 30 minutes, processing may take several seconds.
Is my audio uploaded to a server?
No. All processing happens entirely in your browser using the Web Audio API. Your audio data never leaves your device. No server, no cloud storage, no third-party access.

You might also like

🔒
100% Privacy. This tool runs entirely in your browser. Your data is never uploaded to any server.