airtape
A practical guide to cleaning up fan hum, traffic, AC noise, and room reflections — using modern automated tools, plus the recording habits that make cleanup almost unnecessary.
When podcasters talk about background noise, they usually mean one of four very different problems: continuous hum (laptop fan, air conditioning), intermittent sounds (traffic, dogs, kids), room reverb (the boxy or hollow quality of an untreated room), and electrical artefacts (a buzzy USB cable, ground loops). Each one needs a different fix.
The fastest way to remove podcast background noise is to figure out which category your problem falls into, then pick the right tool. A blanket noise reduction filter applied to everything will often make a clean voice sound worse.
Noise removal has changed a lot in the last few years. Where you used to have to fiddle with spectral edits and noise prints, modern ML-based tools handle most cases in one click. The four most useful right now:
Here is a rough mapping from noise type to the tool that handles it best:
Here is the unglamorous truth: no software can fully reverse a bad recording. A great noise removal tool can rescue an OK recording, but a clean source recording will always sound better than a cleaned-up noisy one — and it saves you an hour of post per episode.
The biggest single upgrade most podcasters can make is not better cleanup software. It is recording each participant locally at source quality, so the file going into post is already free of the compression and dropouts that meeting-tool recordings suffer from.
A handful of small changes during recording will cut your post-production time more than any noise removal tool:
A locally-recorded source file is uncompressed and not subject to network artefacts. Tools like Airtape do this by default in the browser.
Closer is louder, and louder means the signal-to-noise ratio is higher. Background hum becomes much smaller relative to the voice.
Turn off the AC, close the window, mute Slack notifications, and ask anyone in the next room. Two minutes of prep saves twenty in post.
Carpet, curtains, sofas, and bookshelves all absorb reflections. A bare-walled room is the hardest thing to fix in software.
Without headphones, your microphone picks up the other person through your speakers. That bleed is one of the noise sources that is nearly impossible to remove after the fact.
Once you have a decent source recording, a sensible cleanup order is:
Most episodes do not need every step. If your source is clean, automated noise removal plus a quick loudness normalize is enough.
Adobe Podcast Enhance is the strongest free option for most podcasters. It handles room reverb, distant-mic recordings, and broadband noise well. For continuous hum on long files, Auphonic's free monthly quota is also a good pick.
Continuous hum (fan, AC, mains buzz) is best removed with a notch filter or a dedicated de-hum module. Auphonic and iZotope RX both handle this automatically, and Airtape applies it in the browser as part of its recording pipeline.
Modern ML tools like Adobe Podcast Enhance can rescue surprisingly rough recordings, especially for room reverb. But they cannot fully recover a dropped section, a clipped signal, or audio recorded through a heavily compressed meeting tool.
Recording cleaner is almost always better. Source quality has a hard ceiling that cleanup software cannot exceed. Spend on a decent mic, a quiet room, and local recording before you spend on cleanup tools.
Because the algorithms remove frequencies that overlap with both noise and voice. Pushing the strength too high removes part of the voice itself, leaving the watery, processed quality. Use the lightest setting that gets the noise to an acceptable level.