how to remove background noise from audio

How to remove background noise from audio recordings

Quick answer: To remove background noise from audio, first listen to a silent section to identify the noise, then use gentle noise reduction on steady sounds such as hum and hiss, cut or mute isolated sounds such as coughs and clicks, and use equalisation and a noise gate where they help. Work in small steps and compare with the original often. Heavy processing damages the voice, so prevention — a quieter room and better microphone placement — matters as much as repair.

Background noise is the faint hum, hiss, rumble, or echo that sits behind a voice in a recording. It is often invisible while you are recording, because your brain filters it out in the room, but it becomes obvious later, when a listener plays the recording through headphones. Learning how to remove background noise from audio is therefore a useful skill for podcasters, video creators, and anyone who records speech at home. Gentle, targeted fixes almost always sound better than one aggressive effect applied to the whole track.

What counts as background noise, and what types are there?

Background noise is any sound in a recording that is not the intended voice or instrument. Identifying the type is the first step, because each type needs a different fix. Steady noise stays roughly constant — the hiss of a cheap microphone preamp, the hum of a computer fan or air conditioner, and the low buzz of mains electricity. It is the easiest type to reduce, because an editor can learn its sound and remove a similar amount from the whole track.

Intermittent noise comes and goes — traffic, a door closing, a phone vibrating, a dog barking, or a chair creaking. These are harder to remove with a single effect, because they change over time. The best approach is usually to cut or mute the short section where the noise happens, if nobody is speaking. Room echo is a third type: sound bouncing off hard walls and floors, which makes a voice sound hollow. Echo is difficult to remove after recording, and it is usually better prevented than repaired.

What causes background noise in a recording?

Most background noise comes from the room, the equipment, or the way the microphone is used. Room noise includes fans, heating systems, refrigerators, traffic, and other people in the building. A microphone placed far from the speaker picks up more of this room sound, because the voice is quieter relative to the background. Moving the microphone closer increases the voice level and reduces the relative level of the room, which is often the single biggest improvement you can make.

Equipment noise includes hiss from a microphone preamp when the gain is set too high, interference from a poorly shielded cable, and a laptop fan close to the microphone. Handling noise comes from touching the microphone or desk, and wind noise from air moving across the microphone. Understanding the cause helps you choose the right fix. If you are new to these ideas, our sound engineering fundamentals material explains gain, frequency, and room sound in plain language.

How do you find and identify the noise before you fix it?

Do not start by applying effects. Start by listening. Find a section where nobody is speaking — the few seconds before the speaker starts, or a pause between sentences — and listen to it on closed-back headphones at a moderate volume. That silent section is your noise sample. Is it a low hum, a high hiss, a rumble, or a series of clicks? Is it constant, or does it rise and fall? Write down what you hear, because that description guides every later decision.

Next, look at the waveform if your editor provides one. A steady hum often shows as a regular pattern, while a click shows as a sudden spike. Check whether the noise is present throughout the recording or only in certain places. If it changes — for example, a fan switched on halfway through — treat the sections separately.

How do you remove steady noise with noise reduction?

Noise reduction is the main tool for steady hum and hiss, and it works in two stages. First, you select a short sample of noise alone, usually half a second to two seconds from a silent section, and the editor learns the noise profile — the frequencies that make up that sound. Second, you apply the effect to the whole track, and the editor reduces those frequencies. Most editors, including widely known free tools such as Audacity, include a noise-reduction effect that works this way.

The key is to use gentle settings. Start with a modest reduction amount and listen: the voice should sound the same, only cleaner. If it starts to sound thin, metallic, or underwater, the setting is too strong. It is better to leave a small amount of noise than to damage the voice. Apply noise reduction before compression, because compression raises quiet sounds — including any remaining noise. Our overview of essential audio processing tools explains the other tools you will use most often.

How do you fix hum, hiss, rumble, and echo?

Hum, hiss, rumble, and echo each respond to a different approach. Mains hum is a low, steady tone, often around 50 or 60 cycles per second, with harmonics above it. A narrow notch filter, which cuts a very small frequency range, can target it precisely, but cutting too much can thin the voice. Hiss is a broad, high-frequency sound, often from a preamp or a low-quality microphone. Noise reduction usually helps, and moving the microphone closer next time will prevent it.

Rumble is very low-frequency sound from handling noise, wind, or a vibrating desk. A high-pass filter, which reduces frequencies below a chosen point, is the usual fix, and for speech a gentle one can remove rumble without affecting the voice much. Echo is the hardest of the four to fix: a small amount can be reduced with a de-reverb effect, but heavy echo usually cannot be removed cleanly. If echo is a regular problem, add soft furnishings to break up reflections, and record closer to the microphone.

How do you remove clicks, coughs, and other short noises?

Short, isolated noises are best handled by editing, not by effects. If a cough, click, or door slam happens during a pause, select that short section and delete it, or mute it. If it happens while someone is speaking, a very short click can sometimes be repaired with a de-click tool. A longer noise under speech, such as a cough, may need to be accepted, or the sentence re-recorded.

A noise gate can help with low-level noise between words. A gate turns the track down when the signal falls below a threshold — during pauses — and lets it through when the voice is present. Used gently, it cleans up the gaps between sentences; used aggressively, it cuts off quiet words. Set the threshold just above the noise level, with a short attack and a gentle release.

How do you clean up a recording step by step?

A repeatable workflow keeps the work in the right order. Work on a copy of the original file. Listen first and identify the noise. Then make structural edits — remove false starts and long pauses — before you apply noise reduction, so you do not waste time cleaning audio that will be cut. Then apply noise reduction gently, followed by equalisation for rumble, then a noise gate if the gaps between words still sound noisy.

After cleanup, level the voices, and only then apply compression, which raises quiet sounds and can bring back noise if applied too early. Listen to the whole recording at normal speed before you export. If a section sounds worse after processing, undo that effect and try a gentler setting. The table below summarises the main methods.

MethodBest forWatch out for
Noise reductionSteady hum and hissToo much makes voices thin or metallic
High-pass filterRumble and handling noiseToo high a cutoff thins the voice
Notch filterNarrow mains humWrong frequency can damage speech
Noise gateNoise between wordsToo aggressive cuts off quiet words
De-click / repairShort clicks and popsCan soften consonants if overused
Cut or muteCoughs, doors, isolated soundsDo not cut half a word by mistake
Re-record the lineNoise under important speechMatch the tone and distance of the original

How can AI tools help remove background noise?

AI-assisted noise removal has become a practical option, especially for beginners. Some tools analyse the recording, separate the voice from the background, and rebuild a cleaner version of the speech. These tools can be effective on steady noise and on some types of echo, and they can save time on a long file. Our guide to AI audio editing workflows explains how these tools fit into a normal production process.

AI tools have limits. They can misjudge a quiet word as noise and remove part of it, or leave a faint, unnatural quality in the voice. Treat an AI pass as a first draft: listen to the result at normal speed, compare it with the original, and be ready to use a gentler setting. If you record in a browser-based tool, our guide to testing recording permissions and latency is also worth reading, because browser recording can introduce its own noise.

How do you prevent background noise in your next recording?

Prevention is more reliable than repair. Choose the quietest room you can, and turn off fans, air conditioning, and appliances that hum while you record. Close windows and doors, and put your phone on silent and away from the desk. Record closer to the microphone, and use a pop filter or windscreen to reduce breath and wind noise.

Check your gain before you start: speak at your normal volume and set the level so the loudest moments stay well below the maximum. Too little gain forces you to raise the volume later, which raises the noise too; too much gain causes clipping. Do a short test recording, listen to it on headphones, and fix any noise before the real recording begins. For podcasters, our guide to starting a podcast covers recording setup alongside planning and publishing. Once the voice is clean, keep the mix simple and let the voice sit clearly above any music, as our tips on giving your mix the Nugen Stereoizer treatment illustrate.

Frequently asked questions

Can I remove background noise completely from a recording?

Usually not completely, and you should not try. The aim is to reduce the noise until it no longer distracts the listener, while keeping the voice natural. A small amount of clean room tone sounds better than an over-processed voice.

What is the best free tool for removing background noise?

Audacity is a widely used free option, with noise reduction, a high-pass filter, and tools for cutting short noises. GarageBand offers similar cleanup tools for Mac users. Gentle settings in a free editor will beat aggressive settings in an expensive one.

How do I remove background noise from a voice recording on my phone?

Start with the recording itself: move to a quiet room and hold the phone closer. Then transfer the file to a computer and clean it in an editor, using noise reduction for steady hiss and cutting for isolated sounds. Phone apps that promise one-tap cleanup can help, but listen to the result before you use it.

Why does my voice sound metallic after noise reduction?

That usually means the noise-reduction setting was too strong, or the noise profile was taken from a section that included speech. Undo the effect, take a clean noise sample from a silent section only, and apply a gentler reduction.

How do I remove echo from an audio recording?

A small amount of echo can sometimes be reduced with a de-reverb effect, but heavy echo is very hard to remove cleanly. Prevention works better: record in a smaller, softer room, add curtains, rugs, or cushions, and move closer to the microphone.

Should I use a noise gate or noise reduction first?

Use noise reduction first, then a noise gate if the gaps between words still sound noisy. A gate only helps during pauses; it does not remove noise under speech. After both, listen to the transitions between speech and silence, because that is where an aggressive gate will cut off quiet words.

Conclusion

Learning how to remove background noise from audio comes down to a simple habit: listen first, identify the noise, fix it gently, and check your work. Steady hum and hiss respond to careful noise reduction, rumble to a high-pass filter, and isolated sounds are usually best cut or muted. Combine careful cleanup with better recording habits — a quieter room, a closer microphone, and a short test recording.

Similar Posts

Leave a Reply

Your email address will not be published. Required fields are marked *