AI AUDIO LAB · PRIVATE STUDIO DSP

Noise Removal

Filter rumble and background hiss. Fast, browser-based, lossless, and free forever.

✦
⚡
DIRECT ANSWER / SUMMARY:

The Audio369 Noise Removal tool applies two filters: a high-pass at 80 Hz that removes rumble, handling noise and mains hum, and a low-pass at 9 kHz that removes hiss above the speech range. It is a filter pair rather than a neural denoiser, so it removes noise that sits outside the voice and cannot remove noise that sits on top of it.

↑
Drop your audio file here, or browseFast in-browser processing · Zero uploads · Unlimited free exports
MP3, WAV, M4A, FLAC, AAC, OGG, WEBM
Don't have a file ready?

How to Remove Background Noise, Microphone Hiss and Low Rumble

This page is in the AI section of the site and it is not an AI tool, so it is worth saying that first. It is two filters in series, and being clear about what they do is more useful than borrowing the label. The high-pass at 80 Hz removes everything below that frequency. That band contains almost nothing of a human voice and a great deal of what makes recordings sound bad: traffic through a wall, air conditioning, a fan, footsteps through a desk, the thump of a microphone being adjusted, and mains hum at 50 or 60 Hz. Removing it is close to free, since there is nothing down there worth keeping in a speech recording, and it frequently makes the largest single improvement available to an amateur recording. The low-pass at 9 kHz removes everything above. Tape hiss and some electrical noise live up there, and speech carries very little above 8 kHz, so the cost is modest for voice. It is a real cost though: cutting at 9 kHz makes a recording slightly duller, and on music, where cymbals and air live above that line, it is usually too much. Now the limitation that matters most, because it is the reason people are disappointed by filters like this. Broadband noise, the constant hiss of a room or a cheap preamp, occupies the same frequencies as the voice. Filtering can only remove frequency ranges, so removing noise that overlaps speech means removing the speech with it. This tool cannot fix that. Genuine noise reduction uses spectral subtraction, which learns the profile of the noise from a quiet section and subtracts it, or a trained model that has learned what voice looks like. Neither is what this does. So the honest advice: run this on any recording with rumble or hiss outside the voice band, where it will help a lot for no effort. If the problem is a constant hiss underneath the speech itself, this is not the tool, and the fix is at the microphone.

How to Use the Audio369 Online Noise Removal (Step-by-Step)

1. Load the recording

Speech benefits most. On music the low-pass usually takes too much, so use the equaliser's bottom band instead if you only need rumble gone.

2. Export and compare

The two filters are applied across the whole file at once. Listen to the before and after on headphones, where low-frequency changes are audible.

3. Check what you lost

The low-pass makes things slightly duller. If that matters more than the hiss did, the recording did not need this tool.

4. Then set the level

Noise removal before volume is the right order, because boosting first raises the noise you were about to filter out.

Technical Architecture & Audio Engine Specifications

High-pass80 Hz, removing rumble and mains hum
Low-pass9 kHz, removing hiss above the speech band
TypeTwo biquad filters in series
Model involvedNone, despite the category name
NotSpectral subtraction or neural denoising
Cannot removeNoise sharing frequencies with the voice
CostA slightly duller top end

What this removes and what it cannot

NoiseRemoved?Why
Traffic and air conditioning rumble✓ YesIt sits below 80 Hz, clear of the voice
50 or 60 Hz mains hum✓ YesBelow the high-pass, along with its lowest harmonic
Tape and preamp hiss above 9 kHz✓ YesAbove the low-pass
Constant room hiss under the voice✓ NoIt occupies the same frequencies as speech
A dog, a door, a cough✓ NoFilters work on frequency, not on events

Who Uses Audio369 Noise Removal? (Practical Creative Workflows)

Recordings made near a road or a fan

Low-frequency rumble is the most common fault in home recordings and the easiest to remove, because nothing you want is down there.

Cleaning up before transcription

Speech recognition works better on filtered audio, and removing rumble costs the recogniser nothing it was using.

Mains hum from a bad cable

The 50 or 60 Hz buzz from an unbalanced cable or a laptop charger sits below the high-pass and disappears with it.

Freeing headroom before boosting

Inaudible low-frequency energy eats level. Removing it first means a louder voice at the same peak.

Older recordings with tape hiss

Hiss above 9 kHz is straightforwardly removable, and on speech the loss of air above that point is barely noticeable.

Frequently Asked Questions (FAQ)

Why is there still hiss after I run this? +

Because that hiss lives in the same frequency range as the voice, and a filter can only remove a range. Taking it out would take the speech with it. This is a genuine limit of filtering rather than a setting you have missed.

Is this AI noise removal? +

No. It is a high-pass at 80 Hz and a low-pass at 9 kHz. Neural denoisers and spectral subtraction can separate noise from speech in the same band; this cannot, and says so.

Will this remove a cough or a door slam? +

No. Those are events across the whole spectrum, not a frequency range. Cut them out with the Audio Cutter instead.

Should I use this on music? +

Usually not. The 9 kHz low-pass removes cymbals and air that music genuinely needs. For rumble alone, cut the bottom band in the 10-Band Equalizer.

Before or after other processing? +

First. Filtering before boosting or compressing means those stages are working with the noise already gone rather than amplifying it.

Why does cutting frequencies below 80 Hz eliminate air conditioner rumble? +

HVAC systems, traffic rumble, and microphone stand vibrations resonate between 20 Hz and 60 Hz. A steep 80 Hz high-pass filter eliminates low rumble without affecting vocal clarity.

How does filtering high frequencies above 9 kHz reduce microphone hiss? +

Cheaper audio interfaces and preamp thermal noise create consistent broadband hiss in high frequencies. Gently rolling off content above 9 kHz attenuates hiss while keeping speech consonants clear.

Can this tool remove room reverb or echo from a voice recording? +

No. Bandpass filters remove continuous low and high-frequency noise bands. To reduce acoustic echo or reverberant reflections, recording in a room with soft furnishings is the best solution.