How to Remove Background Noise in Audacity (Step by Step)

Audacity has been the go-to free audio editor for years, and one of the main reasons is its noise reduction tool. Unlike some editors that bury this feature inside a bigger toolset, Audacity’s Noise Reduction effect is dedicated, reliable, and free. If you’ve recorded a voiceover, podcast, or interview and picked up a hum, hiss, or fan noise in the background, Audacity can clean most of that up in just a few steps.

In this guide we’ll cover exactly how to remove background noise in Audacity using its built in Noise Reduction effect, the Noise Gate as a backup option, a manual EQ approach for stubborn low-end hum, and a quick AI option for noise that’s too heavy for Audacity to fully handle on its own.

Why Your Audio Picks Up Background Noise

Before fixing anything, it helps to know what you’re actually dealing with. Most noise in Audacity projects falls into a few common types:

• Room noise, like fans, air conditioners, or a fridge humming nearby
• Electrical hum from cables, chargers, or poor grounding, usually a steady low tone
• Wind noise from recording outdoors
• Traffic or street sounds bleeding in through a window
• Hiss from cheap microphones or gain that’s turned up too high
• Echo from recording in a room with hard walls and no soft furniture

Audacity’s Noise Reduction effect works best on the first few, steady, repeating noise sitting underneath the voice. Like most tools, it has more trouble with noise that changes, such as someone talking in the background or a sudden clatter. Knowing which type of noise you’re dealing with helps you know what to actually expect from the tool.

Remove Background Noise in Audacity Using Noise Reduction

This is the main way most people clean up audio in Audacity, and it works well for voiceovers, podcasts, and interviews.

Step 1: Import your audio file
Open Audacity and drag your file into the project, or use File then Import to bring it in.

Step 2: Find a section with only background noise
Look through your waveform for a few seconds where nobody is talking, just the room noise on its own. This is usually easiest at the very start or end of a recording, before someone starts speaking or after they stop.

Step 3: Select that noise-only section
Click and drag across that quiet section with your mouse to highlight it. This selection is what Audacity will use as a reference for what noise actually sounds like in your file.

Step 4: Open Noise Reduction and get the noise profile
Go to Effect, then Noise Reduction. In the window that opens, click Get Noise Profile. This tells Audacity exactly what to filter out later. The window will close automatically after this step.

Step 5: Select your full track
Press Ctrl+A, or Cmd+A on Mac, to select the entire track, or just the section you want cleaned.

Step 6: Open Noise Reduction again and apply it
Go back to Effect, then Noise Reduction. This time you’ll see three sliders: Noise Reduction (dB), Sensitivity, and Frequency Smoothing. Start with the default settings and click OK to apply.

Step 7: Listen and adjust if needed
Play back the result. If it still sounds noisy, undo and try again with the Noise Reduction slider pushed a bit higher. If the voice sounds thin, robotic, or has a warbling quality, lower the Sensitivity slider and try again. This effect often takes two or three attempts to get right, which is normal.

Understanding Audacity’s Three Noise Reduction Sliders

A lot of people apply Noise Reduction once, don’t love the result, and give up. Understanding what each slider actually does makes a big difference.

Noise Reduction (dB). This controls how aggressively the noise is reduced. Higher numbers remove more noise but increase the risk of artifacts. Start around 12, and only push higher if the clip still sounds noisy after listening back.

Sensitivity. This decides how much of the sound gets treated as noise versus voice. Too high and you’ll hear a warbling or bubbling sound in quiet parts of the voice. Lowering it usually fixes that at the cost of leaving a bit more background noise.

Frequency Smoothing. This blends the reduction across neighboring frequencies instead of applying it in sharp, isolated bands. A little smoothing usually sounds more natural, but too much can soften the clarity of speech.

The default settings work fine for light noise. For heavier noise, it’s worth spending a few extra minutes nudging these three sliders individually rather than only adjusting the first one.

Using Noise Gate as a Backup Option

Noise Gate works differently from Noise Reduction. Instead of reducing noise throughout the whole clip, it mutes audio that falls below a certain volume, which is useful for cutting silence-level hiss between sentences.

1. Go to Effect and find Noise Gate
2. Set a threshold level, this is the volume below which audio gets muted
3. Adjust the attack and decay time so the gate doesn’t cut off the start or end of words
4. Preview and adjust until pauses sound clean without chopping off speech

Noise Gate isn’t a replacement for Noise Reduction, since it doesn’t touch noise that happens while someone is actually speaking. It’s more of a finishing touch used alongside Noise Reduction, not instead of it.

A Manual EQ Approach for Stubborn Hum

If you’re dealing specifically with electrical hum, sometimes a targeted EQ cut works better than the general Noise Reduction effect.

1. Select your track and go to Effect, then Filter Curve EQ or Graphic EQ
2. Hum usually sits at 50 Hz or 60 Hz depending on your country’s electrical system, along with multiples of that frequency
3. Pull down a narrow band right at that frequency, rather than cutting a wide range
4. Listen closely, since cutting too broad a range affects the voice as well as the hum

This method takes more trial and error than Noise Reduction, but it can clean up steady electrical hum without touching the rest of the frequency spectrum, which sometimes sounds more natural than a general noise reduction pass.

When Audacity Isn’t Enough

Audacity’s tools are strong for a free editor, but they still have limits. They tend to struggle with:

• Overlapping conversations, where two people talk at once
• Heavy wind noise from outdoor recordings
• Very echoey rooms with a lot of natural reverb
• Background noise that’s almost as loud as the voice itself

In these situations, running your audio through a dedicated AI noise removal tool before bringing it into Audacity often gives a cleaner result than pushing the Noise Reduction sliders to their limit. The workflow is simple: export the noisy file, clean it separately, then bring the cleaned version back into your Audacity project if you need to do any further editing.

If you want to try this, our free noise remover strips out hum, hiss, wind, and background chatter without needing any editing experience, and you can preview the before and after result before deciding to use it. This tends to work especially well on recordings where Audacity’s Noise Reduction leaves a slight warble or thinness in the voice, even after adjusting the sliders.

If your noise problem is actually inside a video file rather than a standalone audio track, it’s usually easier to handle that with a background noise remover for video first, instead of trying to extract and clean the audio manually inside Audacity.

Preventing Noise Before You Even Open Audacity

Fixing noise after recording is always more work than avoiding it in the first place. A few habits during recording make a real difference:

• Record in a smaller, softer room where you can. Carpets, curtains, and furniture absorb sound that would otherwise bounce around and cause echo.
• Keep the microphone close to your mouth, since distance is one of the biggest reasons background noise ends up louder than it needs to be.
• Turn off fans, air conditioners, and anything else humming nearby before you hit record.
• Record a few seconds of silence at the start of every take. This gives you a clean noise profile to grab later, which makes Noise Reduction far more effective.
• Check your gain levels before recording, not after. Audio recorded too hot picks up more hiss and is harder to clean without hurting the voice.

None of this means you’ll never need Noise Reduction again. It just means the tool won’t have to work as hard, which usually means a cleaner, more natural sounding result.

Common Mistakes When Removing Noise in Audacity

• Skipping the Get Noise Profile step. Applying Noise Reduction without first grabbing a proper noise sample makes the effect far less accurate.
• Pushing the Noise Reduction slider too high. This is the most common mistake in every editor, not just Audacity. Too much reduction creates a thin, robotic, or warbly voice.
• Grabbing a noise profile that includes speech. If your selected sample has even a little bit of talking in it, the noise profile gets confused and the results suffer.
• Only listening on laptop speakers. Small speakers hide artifacts that become obvious once you export and listen with headphones or on a bigger system.
• Applying Noise Reduction to music instead of dialogue. This effect is built with speech in mind, and using it on music tracks often strips out parts of the sound that were never meant to be removed.

A Quick Checklist Before You Export

• Grab a clean noise profile from a section with no speech
• Apply Noise Reduction to the full track, then listen closely
• Adjust the three sliders individually rather than only the first one
• Use Noise Gate as a finishing touch if pauses still have residual hiss
• Toggle the effect on and off, or compare with Ctrl+Z, to check the difference
• If Audacity’s tools aren’t enough, clean the audio externally before reimporting

Frequently Asked Questions

Audacity’s Noise Reduction tool covers a lot of ground for a completely free editor. Grab a clean noise profile, adjust the sliders gradually instead of maxing them out, and build a few good recording habits so there’s less to fix later. When a clip is too noisy for Audacity alone to handle, that’s when it’s worth cleaning the audio separately before bringing it back in. You can find more free tools built around exactly this over at Noise Reducer AI.