How to Remove Background Noise in Audacity: Noise Reduction vs Noise Gate, Best Order, and When to Clean the Exported File Online

Quick Answer

To remove background noise in Audacity for spoken voice, use Effect > Noise Removal and Repair > Noise Reduction like this:

1. Select a section that contains only the unwanted noise.
2. Go to Effect > Noise Removal and Repair > Noise Reduction.
3. Click Get Noise Profile.
4. Select the full voice track.
5. Open Noise Reduction again.
6. Start with moderate settings, preview, and apply only if the voice still sounds natural.

For speech, Audacity works best on steady background noise like hiss, fan noise, constant buzz, or hum-like sounds, as explained in the official Audacity Noise Reduction manual and Audacity support guide. It works poorly on changing noise like traffic, barking, wind gusts, crowd noise, clipped audio, or music under speech. If you do not have a true noise-only sample, or the result turns metallic, watery, or dull, it is often faster and safer to clean podcast audio with SimpleClean after export instead of pushing Audacity too hard.

Official Audacity docs also make an important distinction: Noise Reduction removes constant noise across the recording, while Noise Gate only turns down quieter sections. A gate can help in pauses, but it is not the same thing as true noise removal. See Audacity's Noise Gate manual and Alternative Noise Reduction Techniques.

What Audacity can actually fix well

Audacity's own manual says Noise Reduction is intended for constant background noise. In real spoken-word editing, that usually means:

• steady mic hiss
• computer fan or AC noise
• constant electrical buzz
• some hum-like or whine-like background tones

That is why Audacity can work well for podcasts, voiceovers, webcam recordings, course lessons, and screen recordings made in a room with a predictable background sound.

Where creators get into trouble is expecting the same tool to fix everything. Audacity is not at its best when the noise keeps changing shape or volume, which the official manual also warns about.

Usually good candidates for Audacity cleanup

• solo podcast with steady room noise
• YouTube voiceover with light hiss
• screen recording with constant laptop fan
• interview track with mild HVAC rumble and a stable noise floor

Usually bad candidates for basic Audacity Noise Reduction

• street traffic rising and falling
• wind bursts on phone or camera audio
• barking dogs or crowd noise between words
• music under speech
• clipped or distorted voice

If that sounds like your recording, skip ahead to the section on when to export and use SimpleClean.

Before you start: get a real noise sample

The most common mistake is sampling the wrong thing. Audacity needs a true noise-only region to build the profile. The official manual is clear that a noise profile should come from audio that contains only the unwanted sound, and a longer sample is better than a tiny one.

That means:

• Find a section where nobody is speaking.
• Use a real continuous section of noise if possible.
• Do not duplicate a tiny snippet over and over and expect a better profile.
• If you cannot find noise alone, results may be poor.

Also note: the noise profile is not permanent. Audacity resets it when the program closes, so if you reopen later, you need to capture it again, according to the Noise Reduction documentation.

Best order for spoken-word cleanup in Audacity

For voice, the safest workflow is usually:

1. High-Pass Filter for low rumble, if needed
2. Notch Filter for mains hum or narrow whine, if needed
3. Noise Reduction for steady background noise
4. Noise Gate only for pauses and gaps, if needed
5. Light EQ or compression after cleanup

This order matters because you do not want Noise Reduction wasting effort on deep rumble or a narrow hum that could be reduced more cleanly with targeted filtering first.

1) High-Pass first for low rumble

If your waveform has obvious low-end thumps, HVAC rumble, stand vibration, or handling noise, start with Audacity's High-Pass Filter. This can reduce low frequencies before denoising so the later step has less unwanted energy to fight. See the Audacity High-Pass Filter manual.

2) Notch hum or whine before denoise

If the problem is a narrow hum or whistle, a notch-style approach can be cleaner than aggressive full-band noise reduction. This is especially useful when the rest of the voice is fairly clean.

3) Use Noise Reduction for the steady noise bed

Once rumble or hum is tamed, apply moderate Noise Reduction to the full spoken track.

4) Use Noise Gate only for pauses

A Noise Gate does not remove background noise from speech itself. It simply reduces quieter passages below a threshold. Audacity's documentation warns that gates can cause pumping or unnatural transitions if pushed too far. See Noise Gate and Alternative Noise Reduction Techniques.

So for voice, think of a gate as a pause cleaner, not a universal fix.

5) Tone-shape after cleanup

If the voice sounds slightly dull after denoise, do any gentle EQ or compression after the cleanup, not before.

Exact Audacity steps: Noise Reduction for voice

1. Open your recording in Audacity.
2. Listen through once and identify the noise type: hiss, fan, hum, rumble, or changing outdoor noise.
3. If needed, remove low rumble first with Effect > EQ and Filters > High-Pass Filter.
4. Select a clean noise-only section.
5. Go to Effect > Noise Removal and Repair > Noise Reduction.
6. Click Get Noise Profile.
7. Select the whole track.
8. Open Noise Reduction again.
9. Start with conservative settings and use Preview.
10. Apply the effect only when the speech still sounds natural.

These steps match the official Audacity support instructions. If you are cleaning audio from a video project, this is also the stage where it may be smarter to export the audio and use a dedicated cleanup workflow first, then bring the cleaned file back into your editor. If you also need subtitles later, Best AI Captions can fit naturally after audio cleanup.

Safe starting settings for spoken voice

There is no single magic preset. Audacity's own documentation stresses previewing and adjusting by ear. For speech, start moderately instead of trying to erase every trace of noise in one pass.

• Use modest reduction, not maximum reduction
• Use preview repeatedly on difficult words like S, F, and quiet sentence endings
• If artifacts appear, back off before applying
• One moderate pass is usually safer than one extreme pass

Good spoken-word goals:

• lower the distraction
• keep consonants clear
• avoid metallic texture
• accept a small amount of residual room noise if the voice remains natural

That preview-first approach is consistent with the Audacity manual.

How to tell when you are over-processing

Audacity provides a Residue option in Noise Reduction, which is useful for troubleshooting. If the residue contains obvious voice components, you are removing part of the speech along with the noise. This behavior is described in the official Noise Reduction manual.

Preview for these warning signs:

• metallic or robotic voice
• underwater or watery texture
• dull, papery, or hollow speech
• chirping or "musical noise" in the background
• word endings that disappear

When that happens, reduce the aggressiveness, improve the noise sample, or stop and use a different workflow. We cover that more deeply in How to Fix Metallic Voice After Noise Reduction.

Noise Reduction vs Noise Gate in Audacity

Best for most voice creators:

• Use Noise Reduction if the noise is present during speech and stays fairly constant.
• Use Noise Gate if the noise is mostly noticeable in gaps between phrases.
• Use both carefully when the track has a steady noise bed plus noisy pauses.

In most cases, Noise Gate should come after Noise Reduction, not before. If you gate first, you may make later cleanup less consistent and create more obvious transitions. Audacity's alternative techniques page supports using a gate as a complementary tool rather than a replacement for denoise.

When Audacity is enough vs when to export for online cleanup

Audacity is usually enough when

• the background noise is steady
• you have a real noise-only sample
• the speech is already intelligible
• you only need mild to moderate cleanup
• you are working on a WAV or simple voice track

Export and clean elsewhere when

• you have no clean noise sample
• the noise is variable or outdoor
• wind, traffic, or crowd noise keeps changing
• the voice gets metallic every time you denoise
• you are cleaning video audio from MP4 or MOV workflows
• you need a faster file-level fix for repeated creator workflows

That is the practical handoff point for remove background noise from exported audio with SimpleClean. It is especially useful when pushing Audacity further only makes the voice worse.

If your recording is part of a larger video workflow, you may also want to clean it before translation or distribution. After cleanup, Translate Dub can help create dubbed multilingual versions, and Mallary.ai can help schedule and distribute the finished clips across social platforms.

Save project vs export audio

Audacity's project file is for editing. Exporting creates the actual audio file you will use elsewhere.

• Save Project when you want to keep editing later in Audacity.
• Export Audio when you want a usable file for upload, video editing, captioning, translation, or publishing.

Audacity's export dialog also mixes down your audible tracks into the output file, so make sure only the tracks you want are active before exporting. See Audacity's Export Audio documentation.

WAV or MP3 before further cleanup?

• WAV is the safer choice before more processing because it avoids another lossy encode step.
• MP3 is fine when you need a smaller delivery file and are done editing.

If you plan to do more cleanup outside Audacity, export WAV when possible.

Do you need FFmpeg for MP4 or M4A?

For some formats, yes. Audacity's own FAQ on opening and saving files notes that FFmpeg is needed for wider import and export support, including formats like MP4 and M4A in certain workflows. If Audacity cannot open or export the media format you need, check your FFmpeg setup first.

If your end goal is cleaned video audio, you may also like our guide on how to remove background noise from video online.

Best-for recommendations

• Best for steady fan or hiss on a podcast: High-Pass if needed, then moderate Noise Reduction.
• Best for AC hum or tonal whine: Notch the hum first, then light Noise Reduction.
• Best for noisy pauses between phrases: Light Noise Reduction, then careful Noise Gate.
• Best for outdoor interview audio: Usually export and use a dedicated cleanup workflow instead of relying on classic Audacity denoise alone.
• Best for MP4 or MOV creator workflows: Export or extract audio, clean it, then return to the video edit.

Practical troubleshooting

Why does noise come back after I fix it?

Usually because the noise is variable, the sample was poor, or the gate only reduced the quiet gaps while the noise remains under speech.

Why does the exported audio sound different?

Check your export format, bitrate choice, and whether all audible tracks were mixed down during export. A WAV export is the safest reference when judging whether the cleanup itself worked.

Can Audacity remove wind noise?

Sometimes lightly, but wind gusts are usually not the ideal case for classic Noise Reduction because they are not constant. Severe wind often needs a different workflow.

Can Audacity clean video audio directly?

Sometimes, depending on import support and FFmpeg configuration. But for many video creators, it is simpler to export or extract the audio, clean it, and then reattach it in the video editor.

Final takeaway

If you want to remove background noise in Audacity without ruining voice quality, the key is not one secret preset. It is choosing the right tool for the right noise:

• steady hiss or fan: Noise Reduction
• low rumble: High-Pass first
• hum or whine: Notch first
• noisy pauses: Noise Gate after cleanup

And if the sample is unusable, the noise keeps changing, or the voice turns metallic, do not keep forcing it. Export the file and clean spoken audio online with SimpleClean instead.