You recorded the meeting, the interview or the webinar, played it back, and it sounds worse than it did live. Voices are thin and slightly robotic, there is a constant hiss, someone's keyboard is clattering, and one speaker is twice as loud as the other.
That isn't bad luck. Video-call apps are built for live conversation, not for recordings, and a few of their design choices show up clearly on playback. Here is why, and what you can do about it.
Why call recordings sound worse than the call
The audio is heavily compressed. Calls have to work on weak connections, so the app squeezes audio into a small amount of data. Live, your brain fills the gaps. On a recording you listen to closely, the thin, slightly metallic quality becomes obvious.
Each participant's microphone is different. One person is on a good headset, another on laptop speakers in a kitchen. The recording mixes them all together, so the worst microphone sets the tone.
The app's own noise suppression has trade-offs. Live suppression has to decide in a few milliseconds what is voice and what is noise. When it guesses wrong it clips the start of words or lets noise burst through between sentences, and that pumping is very noticeable on playback.
Levels jump around. Automatic gain control raises quiet speakers and lowers loud ones on the fly. On a recording that can mean the background swelling every time someone stops talking.
Echo and feedback. Someone without headphones lets the call's audio leak back into their microphone. The app tries to cancel it, but traces often remain.
What you can fix after recording
Some of this can be repaired; some cannot. Being clear about which saves time.
- Background noise: yes. Fans, air conditioning, traffic, keyboard clicks and household noise around a voice are exactly what speech-focused noise removal is built for. Our speech AI keeps the voices and removes the sound around them, which also reduces the pumping left by the app's own suppression.
- Uneven loudness: mostly. Setting a loudness target brings the whole recording to a consistent level, so viewers aren't turning it up and down. It won't perfectly balance two speakers on one mixed track, but it helps a lot.
- Compression artefacts: only partly. Detail that the call app threw away is gone. Cleaning makes the result easier to listen to, but it can't restore studio quality.
- Echo and reverb: no. Room sound is part of the voice. Prevention is the fix (see below).
How to clean a call recording
- Find the original recording file (MP4 for video, M4A for audio-only). Avoid a copy that has been re-shared through a messaging app, as each re-share compresses it again.
- Drop it into the tool and preview a section with the most noise.
- Choose a loudness preset: YouTube if it is going online, Podcast for an interview you'll publish as audio.
- Download. For video, the picture is copied as-is and only the audio track is replaced, so the slides and faces look exactly as they did.
Start with the noise remover for audio or go straight to the tool.
How to make the next recording better
- Ask everyone to wear headphones. It is the single biggest fix for echo and feedback.
- Use a headset or a USB microphone rather than the laptop's built-in one, and keep it close to the mouth.
- Pick a quiet room and close the window. Soft furnishings help with echo.
- Mute when not speaking, especially if you type during the call.
- Record locally if the app offers it. Some apps can save each participant's audio on their own machine or as separate tracks, which avoids much of the call compression.
- Do a ten-second test recording before an important session and play it back.
A call recording will rarely sound like a studio, but with the noise removed and the loudness evened out, it can easily sound clear and professional enough to publish.