You recorded a phone call and played it back only to hear hiss, hum, or a low rumble underneath the conversation. It happens. The good news is that Audacity — a free, open-source audio editor — has a solid noise reduction tool that can clean up most recordings in under a minute. This guide walks you through the entire process, from opening your file to exporting a clean version.
What You Need
- Audacity — Free download from audacityteam.org. Available for Windows, Mac, and Linux.
- Your audio file — WAV, MP3, M4A, FLAC, or any other common format. Audacity opens them all. If you recorded a phone call with RECAP S2 or similar hardware, you likely have a WAV or MP3 file ready to go.
That’s it. No plugins or paid add-ons required.
Step 1: Open Your Recording
Launch Audacity and go to File > Open. Navigate to your audio file and select it. Audacity will import the file and display the waveform on screen.
If you recorded in a format like M4A, Audacity may need the FFmpeg library to import it. The program will prompt you if that’s the case — follow the on-screen instructions to install it once, and you won’t need to think about it again.
Step 2: Select a Noise Sample
Before Audacity can remove noise, it needs to know what the noise sounds like. Find a section of your recording where nobody is talking — just the background noise by itself. This is your noise sample.
Click and drag in the waveform to highlight 1 to 2 seconds of that silent-but-noisy section. The longer the sample, the more accurate the noise profile, but even half a second works in a pinch.
Look for gaps between sentences, the beginning of the recording before the call connects, or any pause in the conversation. You want pure noise with no speech mixed in.
Step 3: Get the Noise Profile
With your noise sample still selected, go to Effect > Noise Reduction. In the dialog that opens, click Get Noise Profile.
The dialog will close immediately. It may look like nothing happened, but Audacity has now analyzed your selection and built a profile of the noise frequencies. This profile is what it will use to separate noise from speech in the next step.
Step 4: Apply Noise Reduction
Now select the entire track by pressing Ctrl+A (or Cmd+A on Mac). You want the noise reduction applied to the whole recording, not just the sample.
Go to Effect > Noise Reduction again. This time, you’ll adjust the settings and apply them.
Recommended starting settings:
| Setting | Value | What It Does |
|---|---|---|
| Noise reduction (dB) | 12 | How much quieter the noise will become. Higher values remove more noise but risk distorting speech. |
| Sensitivity | 6 | How aggressively Audacity identifies sounds as noise. Higher values catch more noise but may also catch parts of speech. |
| Frequency smoothing (bands) | 3 | Smooths out the reduction across neighboring frequencies. Helps avoid a “musical” or watery artifact in the result. |
Click Preview to hear a few seconds of the result before committing. If the voice sounds natural and the noise is reduced, click OK to apply.
A word of caution: Don’t crank the noise reduction to 24 dB on the first pass hoping to eliminate every trace of noise. Over-reduction makes voices sound robotic or like they’re underwater. Start at 12 dB, listen, and increase by a few dB at a time if you need more. You can always run the effect again — applying two passes of moderate reduction often sounds better than one pass of aggressive reduction.
Step 5: Normalize the Audio (Optional)
After noise reduction, your audio may be quieter than you’d like. Normalizing brings the volume up to a consistent level without clipping.
Go to Effect > Normalize and set the peak amplitude to -1.0 dB. Leave the “Remove DC offset” box checked. Click OK.
This ensures the loudest point in your recording sits just below the maximum level, giving you a full, even volume throughout the file.
Step 6: Export Your Clean Recording
Go to File > Export Audio. Choose your format:
- WAV — Lossless quality. Best if you plan to edit further or archive the recording.
- MP3 — Smaller file size. Good for sharing or uploading. Use 128 kbps or higher for voice recordings.
Name your file, choose a destination folder, and click Save.
Bonus: Quick EQ for Phone Call Audio
Phone calls carry a narrow band of audio frequencies — roughly 300 Hz to 3,400 Hz. Anything outside that range is not voice; it’s noise. You can use Audacity’s EQ to cut those unwanted frequencies and get a noticeably cleaner result.
Go to Effect > Filter Curve EQ (or Graphic EQ, depending on your version). Then:
- Roll off everything below 200 Hz. This removes low-frequency rumble, electrical hum, and HVAC noise.
- Roll off everything above 4,000 Hz. This removes high-frequency hiss and static.
Apply this before or after noise reduction — either order works, though applying EQ first can make the noise reduction step more effective since there’s less noise for it to deal with.
Tips for Better Recordings
If you’re recording phone calls with RECAP S2, the audio signal itself is typically clean. RECAP taps the headset audio directly at line level, so there’s no ambient room noise in the signal. When noise does show up, it usually comes from the recording device — a computer’s built-in mic input picking up electrical interference, gain set too high in the recording software, or a ground loop between devices.
A few things that help:
- Use proper input levels. Set your recording software’s input gain so peaks hit around -12 dB to -6 dB during normal speech. This leaves headroom and keeps the noise floor low.
- Use a USB audio interface if your computer’s built-in audio input introduces noise. Even an inexpensive one will be quieter.
- Check your cables. A loose or damaged 3.5mm connection can introduce crackling and static.
For a full walkthrough on connecting RECAP to your computer, see our guide on how to record cell phone calls on a computer. For help choosing recording software, see best software for recording phone calls on PC.
FAQ
Can I remove background noise without Audacity?
Yes. If you use OBS Studio for recording, it has built-in noise suppression filters you can apply in real time during the recording — no post-processing needed. Other options include paid tools like iZotope RX and Adobe Podcast’s online enhancer. Our software guide covers several options.
Will noise reduction affect voice quality?
It can, if you overdo it. Aggressive settings strip out frequencies that overlap with speech, making voices sound thin, hollow, or robotic. Start with a noise reduction value of 12 dB, preview the result, and increase only if needed. Two moderate passes sound better than one heavy pass.
What if I can’t find a section of pure noise in my recording?
Look at the very beginning or end of the file — there’s often a second or two before the call connects or after it ends. If your entire recording has voice throughout, find the quietest gap between sentences. Even a fraction of a second can work, though a longer sample gives Audacity more data to build an accurate noise profile.