First, identify the noise

Background noise is not one problem. A steady fan hum, an air conditioner hiss, keyboard clicks, traffic, and room echo all need different tools. Listen on headphones to a quiet part of the voice memo. Constant noise responds well to noise reduction and a high-pass filter. Intermittent noise is better handled with a gate or manual editing. Echo and reverb are not noise, and heavy noise reduction will make them worse. If the noise is as loud as the voice or the recording is clipping, re-record when you can. Cleaning works best when the voice is clearly louder than the background.

Try your voice memo app first

iPhone Voice Memos

  1. Open the memo and tap Edit.
  2. Tap Enhance Recording, usually a magic wand or waveform icon.
  3. Play it back. If the voice sounds thin or metallic, turn Enhance off.
  4. Trim silence at the start and end, then tap Done.

Enhance Recording reduces steady noise and boosts speech. It can also boost room reflections, so treat it as a quick first pass.

Android recorders

Google Recorder on Pixel has Clean or Noise reduction in some versions. Samsung Voice Recorder has a Noise reduction toggle on some models. Open the recording, look under Edit, and find Noise reduction, Clean, or Enhance. Stronger settings remove more noise but can make the voice sound processed. Keep the original file before saving.

Zoom and cameras

For new Zoom recordings, go to Settings, Audio, Background noise suppression, and choose Low, Medium, or High. Low is safest for speech. High can cut soft consonants. Existing Zoom recordings cannot be changed after the fact, so you must clean the audio file. On cameras, a wind cut or low-cut filter helps rumble but not hiss or room tone.

Clean it in Audacity with real settings

Audacity is free and gives you more control than phone apps. These steps are for a typical voice memo with steady background noise.

  1. Make a copy. Never edit the only version. Keep the original M4A or WAV untouched.
  2. Import the file. Open Audacity, then File, Import, Audio. Install FFmpeg if Audacity asks for M4A support.
  3. Select a noise-only section. Choose one to two seconds where nobody is speaking but the fan, hum, or hiss is audible.
  4. Get the noise profile. Go to Effect, Noise Removal and Repair, Noise Reduction, then click Get Noise Profile.
  5. Apply noise reduction. Select the whole track. Open Noise Reduction again. Start with Reduction at 6 to 10 dB, Sensitivity at 5 to 6, and Frequency smoothing at 3. Preview. If the voice sounds watery, undo and lower the reduction to 4 to 6 dB. Apply.
  6. Remove rumble. Use Effect, EQ and Filters, High-Pass Filter. Set 80 Hz for a deep male voice or 100 Hz for most other voices with a 12 dB per octave rolloff. This cuts desk bumps, traffic rumble, and fan thumps.
  7. Notch out hum. If you hear a steady buzz, use Effect, Notch Filter at 50 Hz in Europe and much of Asia, or 60 Hz in North America, plus harmonics like 100/120 Hz. Set Q between 10 and 20. Only if the hum is obvious.
  8. Gate the gaps. Use Effect, Noise Removal and Repair, Noise Gate. Set threshold between -40 and -30 dB, attack around 10 ms, and decay around 100 to 200 ms. A gate can chop breaths and word tails, so use it lightly.
  9. Optional EQ and compression. Cut below 80 Hz and add a mild boost at 2 to 5 kHz for clarity. For compression, try threshold -18 dB, ratio 2:1, attack 10 ms, release 100 ms. Fix noise first.
  10. Export. Choose WAV for best quality or MP3 at 192 kbps or higher. Do not export at 64 kbps; it adds new artifacts.

Use a free online cleaner for the fastest pass

If you do not want to install Audacity or learn noise profiles, a free tool like ClearVoiceHQ (clearvoicehq.com) is built for this exact job: voice cleaning for mobile recordings, cameras, and voice memos. Upload the file, run the cleanup, and download the result. It works well as a first pass because you can hear whether the noise is removable before spending time on manual editing. If the result sounds over-processed, try a lighter setting or go back to Audacity for finer control. ClearVoice is useful when you need to clean a phone recording on a laptop or Chromebook without installing anything.

If the voice memo is attached to a video

You can clean the audio and put it back without touching the video. This works for camera clips, gaming captures, and screen recordings.

  1. Extract the audio: ffmpeg -i input.mp4 -vn -acodec pcm_s16le audio.wav
  2. Clean audio.wav in Audacity or an online cleaner.
  3. Replace the audio: ffmpeg -i input.mp4 -i cleaned.wav -c:v copy -map 0:v:0 -map 1:a:0 -shortest output.mp4

The video stream is copied, so quality stays the same. Check sync after export. If the cleaned audio is delayed, trim the start of cleaned.wav or use an offset. A hand clap at the start of camera clips makes sync easy.

Settings and tradeoffs that matter

Quick recipes by situation

When cleaning is not enough

If background noise is as loud as the speaker, if the audio is clipped or distorted, or if several people talk over each other, cleaning will fail or sound unnatural. Re-record instead. Put the phone in airplane mode, get within a hand span of your mouth, record in a soft room, turn off fans and air conditioners, and record five seconds of silence first for a noise profile.

Final checklist

  1. Duplicate the original file.
  2. Try the voice memo app's Enhance or Noise reduction setting.
  3. If needed, use Audacity: noise profile, 6 to 10 dB reduction, high-pass at 80 to 100 Hz, light gate.
  4. For a fast online pass, try a free cleaner like ClearVoice.
  5. Export as WAV or 192 kbps MP3.
  6. Listen on phone, earbuds, and laptop.
  7. Compare with the original. If the cleaned version sounds worse, use less processing.

Start with the least aggressive fix, listen after every step, and stop when the voice sounds natural. Background noise removal is not about making a file silent. It is about making the voice easy to listen to.