← All articles

Podcast Audio Cleanup: What an Engineer Actually Fixes in Post

The Content Factory team · July 14, 2026

People obsess over cameras and microphones, then send us a recording that sounds like it was captured in a moving car. The good news: most audio problems are fixable in post. The better news: a lot of them never happen if you record in a treated room. Here is what audio cleanup actually means, what we can and cannot rescue, and where the line is between a quick fix and a rebuild.

What is podcast audio cleanup?

Audio cleanup is the set of processing steps that turn a raw recording into something people will listen to for 45 minutes without fatigue. It is not one button. It is a sequence, and the order matters.

Here is the rough chain we work through on a typical episode:

• Cleanup and repair first: noise, hum, clicks, mouth sounds.
• Balance next: getting each speaker to sit at a consistent level.
• Tone shaping with EQ so voices sound natural, not boxy or thin.
• Dynamics with compression so quiet moments are audible and loud ones do not spike.
• Loudness last: setting the whole thing to a platform-safe target.

Do these out of order and you fight yourself. Compress before you remove noise and you just make the noise louder.

Noise removal without the underwater sound

Background hum from HVAC, laptop fans, or fluorescent lights is common. Modern noise reduction tools are genuinely good, but they have a cost. Push them too hard and voices develop a swirly, underwater artifact. A careful engineer removes enough to clean the floor and stops before the voice starts to sound processed. The goal is that a listener never notices the noise was there, not that they notice it was removed.

Room echo is the one you cannot fully undo

This is the honest limit. If you record in a hard, empty room, the reflections are baked into the recording. De-reverb tools help a little, but they trade clarity for dryness and often leave a strange metallic edge. Echo is the problem we cannot fully fix in post, which is exactly why recording in a treated space matters more than any plugin. If you want to skip this entire category of problem, that is part of what you get when you book a session in a room built for voice.

Why do my two speakers sound so different?

Two people, two mics, two distances from the mic, two voices. One guest leans in and booms; the other sits back and disappears. Balancing is the unglamorous work that makes a conversation feel like it was recorded together rather than stitched from two phone calls.

We handle it with a mix of level automation and gentle compression per speaker. If someone turned their head constantly or the mic placement drifted, that is manual, clip-by-clip work. It is also the difference between an episode that sounds amateur and one that sounds like a real show.

Plosives, sibilance, and mouth clicks

Three small problems that make a big difference:

• Plosives are the low thumps on hard P and B sounds. A high-pass filter and targeted volume dips handle most of them.
• Sibilance is the harsh S and T. A de-esser tames it without lisping the voice.
• Mouth clicks and lip smacks are those tiny wet sounds between words. Removing them one at a time is tedious, and it is one of the clearest signs a human touched the file.

What loudness should a podcast be?

Streaming platforms normalize loudness, so chasing a super loud master is pointless and often makes your show sound crushed. The widely used target for spoken word podcasts is around minus 16 LUFS for stereo, with true peaks kept below zero to avoid clipping. Video platforms tend to sit a little louder, closer to minus 14.

Why you should care: if your episode is quiet, listeners crank their volume, then get blasted by the next show. If it is over-compressed, it is exhausting on earbuds. Hitting a consistent target across every episode is what makes a subscriber trust the play button.

Can I just use an AI cleanup tool myself?

You can, and for a clean recording you might get most of the way there. One-click tools are great at raising a quiet voice and knocking down steady background noise. Where they struggle:

• Two speakers at wildly different levels in one file.
• Cross-talk, where both mics pick up both people.
• Deciding what is noise versus what is a soft-spoken guest.
• Judgment calls, like leaving a laugh in but cutting the long pause before it.

The tools are a starting point, not a mix engineer. If your recording was captured cleanly on separate tracks, an editor can do far more than any automatic pass. That is a big argument for recording with multitrack in the first place, which our sessions give you by default. If you want a person to take it the rest of the way, our editing services start at fifty dollars an hour and we only touch what needs touching.

What clean audio unlocks later

Cleanup is not just about the full episode. Every clip you pull for Reels, Shorts, or TikTok inherits the audio from the master. If the source is noisy, every one of your short clips is noisy too, and short-form is often watched on phone speakers in loud rooms where noise is brutal. Clean the source once and every derivative benefits.

The same is true for the packaging around your show. Trailers, cold opens, and audiograms all pull from the same clean master. If you would rather hand off the whole pipeline instead of managing it piece by piece, our done-for-you show packages cover the cleanup, the episode, and the assets that come out of it.

The short version

Audio cleanup is a chain, not a filter. Noise and repair first, then balance, tone, dynamics, and loudness. The problems you can fix easily are hum, hiss, plosives, sibilance, and level differences. The one you cannot fully fix is room echo, which is why the room you record in matters more than any plugin.

If you are recording somewhere with hard walls and a lot of ambient noise, the smartest move is to solve it at the source. Our downtown room is built and treated for voice, ten minutes from the airport with parking in the building. Planning a trip in? Our visiting San Diego page has the logistics, and you can book a session whenever you are ready to sound like a real show.