Why Google Meet recordings sound noisy (and what you can actually fix)
Google Meet saves cloud recordings as MP4 files in Google Drive. The audio is already mixed: your mic, guests, system audio. If there is hum, keyboard clicks, room echo, or fan noise, it is baked into that MP4. Google Meet in-call noise cancellation (Settings > Audio > Noise cancellation) can help future calls, but it will not clean a recording you already have. You need to edit the audio track in post.
Below is a workflow that works in Audacity, DaVinci Resolve, Adobe Premiere, or a free online tool. The key is order: diagnose, repair, then sweeten. Do not throw every filter at it.
Step 1: Get the audio out of the Google Meet recording
Open Google Drive, find the Meet recording, download the MP4. Work on a copy. If you are using a video editor, import the MP4 and detach or extract the audio. If you are using Audacity, install FFmpeg so Audacity can open MP4s, or use VLC: Media > Convert/Save > Add the MP4 > Convert > Audio - Vorbis/MP3. For editing, WAV 48 kHz 24-bit is best. If you can only get MP3, use 320 kbps. Avoid re-encoding more than once.
Step 2: Diagnose the noise before you treat it
Listen with headphones at a comfortable level. Then look at a spectrogram. In Audacity, click the track name > Spectrogram. You will see:
- Low rumble/hum: horizontal lines below 120 Hz, often 50/60 Hz and harmonics.
- Hiss: fuzzy high-frequency band across the whole file.
- Keyboard clicks: short vertical spikes, usually 2-8 kHz.
- Room reverb: smeared decay after words, not a separate line.
- Background voices/music: unpredictable, hard to remove without artifacts.
Write down the top two problems. If you have broadband hiss plus keyboard clicks, treat hiss first, then manually mute or repair clicks. If you have hum plus reverb, fix hum first.
Step 3: Remove rumble and hum with a high-pass filter
In Audacity: Select all > Effect > High-Pass Filter. Set frequency to 80 Hz for deep male voices, 100-120 Hz for higher voices. Rolloff 12 dB per octave is gentle; 24 dB is stronger. Preview. If voices sound thin or telephone-like, lower to 70-80 Hz. For steady 60 Hz hum (US) or 50 Hz (EU), add a notch filter at that frequency and its harmonics (120/180 Hz or 100/150 Hz) with Q around 10-20. Do not notch too many harmonics or you will hollow out the voice.
Step 4: Reduce hiss with a noise profile (not a blanket gate)
Find 0.5-2 seconds where only noise is present, between sentences or before someone speaks. Select it. In Audacity: Effect > Noise Reduction > Get Noise Profile. Then select the whole track. Open Noise Reduction again. Start with:
- Reduction: 6-12 dB (not 24 dB)
- Sensitivity: 6
- Frequency smoothing: 3
Preview. If you hear metallic chirps or water-bubbling sounds, undo and reduce Reduction to 4-6 dB. For Google Meet recordings, 6-9 dB is often enough because Meet already applies some compression. If the noise is only a low hum, skip this and use the notch filter instead.
Step 5: Tame reverb and room echo
Reverb is the hardest. Google Meet compression can make a small room sound like a bathroom. In DaVinci Resolve, try Voice Isolation at 30-50% on the Fairlight page. In Adobe Premiere, use DeReverb or Dialogue > DeReverb. In Audacity, there is not a great built-in de-reverb; you can try a noise gate to reduce tail, but gates chop room tone and can sound unnatural. A better fix: use an expander with a slow release, or apply a light gate only between phrases. If the echo is severe, ask a guest for a re-record if possible. For interviews, a lav mic close to the mouth is better than any plugin.
Step 6: EQ, compression, and loudness
After repair, make the voice sound intentional. Use EQ:
- Cut 200-400 Hz by 2-3 dB if it sounds boxy.
- Boost 2-5 kHz by 1-3 dB for presence and intelligibility.
- Low-pass at 12-16 kHz if hiss still creeps in.
Then compress. In Audacity: Effect > Compressor. Start with threshold -18 dB, ratio 3:1, attack 10 ms, release 100 ms. Add makeup gain so peaks hit around -3 dB. For podcasts, aim for -16 LUFS integrated (mono or stereo). For YouTube, -14 LUFS. For internal meetings, -16 to -18 LUFS is fine. If s sounds are harsh, add a de-esser at 5-8 kHz.
Step 7: Use a free online tool when you do not want a full edit
If you need a quick fix for a webinar, interview, or voice memo, you do not have to learn Audacity. A free online option like ClearVoiceHQ (clearvoicehq.com) focuses on noise removal for remote meetings, webinars, and interview recordings. Upload the Google Meet MP4 or extracted audio, run the noise removal, and listen to the before/after. It is useful for steady hiss, fan noise, and light room tone. The tradeoff: automated tools give you less control than manual EQ and compression, and aggressive settings can soften consonants or create a slight underwater texture. Always keep a copy of the original. If the result is close but not perfect, use it as a first pass, then finish with a high-pass filter and compressor in Audacity. ClearVoice is also handy when you are on a Chromebook or cannot install software.
Step 8: Export and put the audio back
Export a WAV at 48 kHz/24-bit if you are editing video. Then replace the audio track in your video editor: import the cleaned WAV, mute the original, align if needed (usually it is already in sync). If you are delivering only audio, export MP3 320 kbps or AAC 256 kbps. Test on three systems: laptop speakers, phone speaker, and headphones. Listen for artifacts, clipped peaks, and whether the voice sounds natural. A/B against the original. If the cleaned version is clearer but harsh, back off the noise reduction.
Step 9: Prevent the next Google Meet recording from needing a rescue
- Before recording, open Google Meet Settings > Audio and turn on Noise cancellation if available. It will not fix an existing file, but it improves the recorded feed on future calls.
- Use a USB mic or headset mic within 6-12 inches of your mouth. Close windows, turn off fans, and put your phone on airplane mode.
- If you control the recording, capture a local backup with OBS or a separate audio recorder. That gives you a clean track to mix with the Meet recording.
- Record 10 seconds of room tone at the start. That noise profile makes post-processing much easier.
The bottom line
Cleaning Google Meet recording audio is mostly about order: extract the audio, diagnose the noise, high-pass the rumble, reduce hiss with a light noise profile, control reverb, then EQ and compress. For a fast pass, a free tool like ClearVoice can handle steady noise; for full control, Audacity or a DAW is still the best. Keep the original, process gently, and check on real speakers before you publish.