Clear podcast audio comes from a repeatable process, not one magic preset. This guide gives you a practical workflow for reducing noise, controlling mouth clicks, shaping speech with EQ and compression, setting a sensible loudness target, and checking the final export on the devices your audience uses.
Overview
Before editing, make a copy of the original recording and listen all the way through. Note the problems you hear: steady room noise, hum, plosives, harsh consonants, inconsistent volume, echo, or distracting pauses. Fix the most noticeable problems first. Heavy processing can make a recording technically cleaner while making the voice sound unnatural.
A reliable podcast cleanup chain usually follows this order:
- Repair: remove unwanted sections, obvious clicks, severe pops, and isolated noises.
- Noise reduction: lower constant background noise without creating metallic or watery artifacts.
- High-pass filtering and EQ: reduce rumble and improve clarity.
- Compression: narrow large volume differences between quiet and loud speech.
- De-essing or gentle dynamic control: manage harsh sibilance when necessary.
- Limiting and loudness measurement: prevent peaks and prepare the episode for delivery.
- Quality control: listen on more than one playback system and inspect the exported file.
The exact controls differ between a digital audio workstation, a podcast editor, and automated audio note taking tools. The principles remain the same: make small changes, bypass each processor to compare it, and stop when the voice sounds natural.
Checklist by scenario
1. Cleaning steady room noise
Find a short section where nobody is speaking. Use it as a noise profile if your editor supports that feature, then apply noise reduction conservatively. Start with the smallest reduction that makes the noise less distracting. Large reductions can create a swirling, underwater texture between words.
- Remove the noise before compression, because compression can make a quiet background more obvious.
- Use a noise gate only when it sounds natural. A gate that closes too quickly can cut off word endings and breaths.
- For electrical hum, identify the frequency and use a narrow notch filter rather than removing a wide range of the voice.
- If the room has strong echo, do not expect ordinary noise reduction to solve it. Reduce the effect with room treatment, closer microphone placement, or a tool designed for reverberation.
2. Reducing mouth clicks and handling noise
Use a click-removal tool when available, but review its decisions. Automatic repair can mistake consonants for clicks, especially in fast speech. For a small number of problems, zoom into the waveform and reduce or cut the individual sound. Short fades at edit points help prevent new clicks.
Mouth noise is often easier to manage at the recording stage. Keep water nearby, avoid touching the microphone stand, and leave enough space between the mouth and microphone to reduce plosives. A pop filter and an off-axis microphone position can also help with strong p, b, and t sounds.
3. Applying podcast EQ settings
EQ should support intelligibility rather than impose a particular sound. Begin with a high-pass filter to remove low-frequency rumble that does not contribute to speech. Set the cutoff by ear: raise it until rumble decreases, then back off if the voice starts to sound thin.
For muddiness, try a gentle, wide reduction in the low-mid range. If the voice lacks presence, a small boost in the upper-mid range may help, but listen carefully for nasal or aggressive tones. Sibilance usually sits higher in the frequency range; use a de-esser or a narrow cut only if the problem is consistent. Different voices and microphones require different settings, so treat suggested ranges as starting points, not rules.
- Compare EQ with the processor bypassed at matched volume.
- Do not boost high frequencies to compensate for a noisy recording.
- Check pauses and breaths as well as spoken words; an EQ change that sounds good during speech may make background noise obvious between phrases.
4. Using podcast compression settings
Compression makes speech easier to follow by reducing the difference between quiet and loud passages. Start with a moderate ratio, a threshold that catches louder phrases, and an attack and release that preserve the natural rhythm of speech. If the compressor is working constantly, the threshold may be too low or the ratio too high.
Listen for pumping, flattened emphasis, and breaths that become unnaturally loud. A useful workflow is to adjust the threshold until the loudest ordinary phrases receive noticeable but controlled gain reduction, then match the processed and unprocessed levels before judging the tone. Apply separate treatment to each speaker when their microphones or recording environments differ.
5. Preparing dialogue recorded in different places
When a guest sounds much louder, darker, or more reverberant than the host, volume matching alone will not solve the mismatch. Edit obvious gaps, reduce distracting noise, and use EQ and compression separately on each voice. Keep some natural contrast, but make sure listeners do not need to change the volume between speakers.
For remote interviews, ask for the original local recording when possible rather than relying only on a compressed call recording. Organize the files and feedback in a shared workflow; the creator collaboration workflow guide covers practical file, permission, and delivery habits.
6. Setting loudness and exporting
Measure the complete episode, including its intro and outro, rather than judging loudness from one short section. A commonly used starting point for spoken-word delivery is around -16 LUFS integrated for stereo or around -19 LUFS integrated for mono, with a true-peak ceiling near -1 dBTP. These are working targets, not universal guarantees. Check the current requirements or recommendations of the destination platform and keep the show consistent from episode to episode.
Do not raise loudness simply because a meter shows a lower number. If an episode is already heavily compressed or clipped, adding gain makes the distortion more obvious. Leave sufficient headroom during mixing, then use a transparent limiter at the final stage to control occasional peaks. Export in the format required by your host, and listen to the exported file rather than assuming it matches the session.
What to double-check
Use this before-publish audio checklist:
- Edits: No accidental words, clipped breaths, abrupt cuts, or repeated phrases remain.
- Noise: Room tone is consistent, and noise reduction has not introduced metallic artifacts.
- Clicks: Mouth clicks, edit clicks, keyboard taps, and cable handling are addressed where they distract from the conversation.
- Speech: Both sides of an interview are understandable without constant volume adjustment.
- EQ: The voice sounds balanced on headphones and does not become boomy or harsh on speakers.
- Dynamics: Compression controls variation without removing natural emphasis or making breaths jump forward.
- Peaks: The limiter is not being driven so hard that it produces audible distortion.
- Loudness: Integrated loudness and true-peak readings have been checked against your current delivery plan.
- Export: The file opens correctly, plays from start to finish, and has the intended sample rate, channel format, and metadata.
- Real-world test: Check at least headphones, ordinary speakers, and a phone or small speaker. Also listen briefly at a quiet volume, where harshness and missing detail can become easier to notice.
A troubleshooting table can help you choose the next action:
| Problem | Likely cause | First adjustment |
|---|---|---|
| Voice sounds distant | Microphone placement, room reflections, or excessive noise reduction | Improve placement or reduce processing before boosting high frequencies |
| Audio is muddy | Low-mid buildup or room resonance | Try a gentle low-mid cut and compare at matched volume |
| S sounds hurt | Sibilance or an overly bright EQ | Reduce the high-frequency boost or add restrained de-essing |
| Volume jumps between speakers | Different recording levels or inconsistent compression | Level speakers separately, then apply moderate dynamics control |
| Background rises between words | Compression is lifting the noise floor | Reduce compression, repair noise earlier, or use a natural-sounding gate |
| Audio distorts | Clipping during recording, overdriven plugins, or excessive limiting | Find the stage causing overload; do not try to hide severe clipping with more gain |
Common mistakes
Starting with a preset and never checking it. Presets can be useful starting points, but they are not matched to every microphone, room, or voice. Treat them as suggestions and adjust by ear.
Trying to remove every breath. Cutting all breaths can make speech sound artificial and can create audible gaps. Reduce only breaths that are unusually loud, distracting, or part of an edit.
Using EQ to solve a recording problem. EQ cannot fully repair a distant microphone, severe echo, or clipped input. Better recording technique is usually more effective than increasingly complex processing.
Confusing loudness with quality. A louder file is not automatically clearer. Preserve intelligibility, dynamics, and a comfortable listening experience before pursuing a louder meter reading.
Checking only inside the editor. Playback systems reveal different problems. A voice that sounds polished on studio headphones may be too bass-heavy on speakers or too sharp on a phone.
After the audio is approved, you can use a podcast show notes workflow to turn the final transcript into an accurate episode page, or follow the guide to repurpose a podcast episode into social content. Create those assets from the approved episode so timestamps and quotations match the final audio.
When to revisit
Return to this checklist before every episode, especially after changing microphones, rooms, editors, recording platforms, or delivery hosts. Recheck the workflow when a guest uses a different recording setup, when a seasonal series introduces new music or ad segments, or when listeners report volume, harshness, or intelligibility problems.
Review your loudness and export notes whenever a destination platform changes its guidance or your publishing format changes from mono to stereo. Keep a short record of the settings that worked: microphone, distance, high-pass approach, EQ moves, compression behavior, limiter ceiling, integrated loudness, and the devices used for quality control.
For the next episode, do not start by adding more plugins. Start with the checklist: listen to the raw recording, identify the largest problem, make one controlled change, compare at equal volume, and verify the exported file. That repeatable habit is the foundation of professional-sounding podcast audio.
