Podcast Audio Quality Checklist — 15 Steps Before Publishing

Publishing a podcast episode with audio problems is permanent. Unlike a blog post, you cannot quietly edit it after listeners have downloaded it. Once an episode is in someone’s podcast app, the original audio is cached locally. A re-upload creates a new entry, not an update.

This checklist covers every audio quality check you should run before hitting publish. Each step has a specific target value or pass/fail criterion — no vague “make it sound good” advice.

Pre-Edit Checks

1. Verify Source File Integrity

Before editing, confirm your raw recording is intact.

Check: Play the entire raw recording from start to finish. Listen for:

• Digital corruption (clicks, pops, sudden silence, zipper noise)
• Incomplete recordings (file ends abruptly)
• Wrong microphone source (recorded from laptop mic instead of external mic)

Target: Complete, uncorrupted recording from the intended microphone.

Common trap: Recording software sometimes silently switches to the default input device after a Bluetooth disconnection. If your guest’s track suddenly sounds like a laptop mic, it probably is.

2. Check for Clipping in Raw Audio

Clipping occurs when the signal exceeds 0 dBFS (the maximum digital level). Clipped audio has flat-topped waveforms and sounds harsh and distorted. Clipping cannot be repaired — once the waveform is flat-topped, the original shape is lost.

Check: In your DAW or editor, look at the waveform for any segments that are perfectly flat at the top or bottom. Most editors highlight clipped samples in red. Alternatively, use a peak meter and scan for samples that hit exactly 0 dBFS.

Target: Zero clipped samples in the raw recording.

Fix if clipped: If only a few words are clipped, you can sometimes reduce the gain on those segments by 3-6 dB and apply a soft clipper to round off the harsh edges. If the clipping is extensive, you need to re-record.

3. Confirm Sample Rate and Bit Depth

Target: 44.1 kHz or 48 kHz, 24-bit or 32-bit float.

Why it matters: If you recorded at 48 kHz but your editor project is set to 44.1 kHz (or vice versa), the software will resample. This is usually fine, but some editors do not resample transparently — they just play back at the wrong speed, making the audio sound slightly pitched up or down.

Check that all tracks in your project have the same sample rate and that it matches your recording settings.

Edit Phase Checks

4. Remove Dead Air and Long Pauses

Dead air — silence longer than about 1.5 seconds — makes podcasts feel slow and unprofessional. Short pauses (0.3-1 second) are natural and should be kept. Long gaps (2+ seconds) should be trimmed to about 0.5-0.8 seconds.

Check: Scan the waveform for extended flat sections. Most editors have a “silence detection” or “truncate silence” feature.

Target: No silence longer than 1.5 seconds between speech. Natural breathing pauses preserved.

Tools like Hearably Studio’s silence remover automate this with configurable thresholds for minimum silence duration and target gap length.

5. Remove Filler Words (If Appropriate)

Filler words (“um,” “uh,” “like,” “you know,” “basically,” “actually”) are a stylistic decision. Some podcasters remove them all for a polished, edited feel. Others leave them in for authenticity.

Guideline: Remove fillers that interrupt the flow of a sentence or occur more than once every 15-20 seconds. Leave fillers that act as natural transition markers or reflect the speaker’s personality.

Check: Listen through at 1.5x speed — fillers become more noticeable at higher playback speeds, which is how many listeners consume podcasts. AI-powered filler word removal can flag them automatically.

6. Check Edit Points for Clicks and Artifacts

Every cut in your edit is a potential click or pop. If you cut in the middle of a waveform (not at a zero-crossing), the discontinuity creates an audible click.

Check: Scrub through every edit point slowly. Listen for clicks, pops, or unnatural transitions.

Fix: Apply a 5-10ms crossfade at every edit point. Most DAWs can apply crossfades automatically. The crossfade smooths the transition between the two segments.

Target: Zero audible clicks or pops at edit points.

Processing Phase Checks

7. Noise Floor Level

The noise floor is the ambient sound level when nobody is speaking. It includes room tone, mic self-noise, preamp hiss, and environmental sounds.

Check: Solo each track and listen to a silent section. Measure the RMS level.

Target: Noise floor below -60 dBFS for a professional sound. Below -50 dBFS is acceptable. Above -40 dBFS is noticeable and distracting.

Fix if too high: Apply noise reduction. Spectral subtraction (Audacity’s noise reduction, iZotope RX) works well for steady-state noise. For a quick browser-based option, Hearably Studio’s noise reduction targets common podcast noise types. Never apply more than 12-15 dB of noise reduction — beyond that, artifacts become audible.

8. EQ — Remove Problems, Enhance Clarity

EQ should fix problems first and enhance second.

Essential cuts:

• High-pass filter at 80 Hz — removes low-frequency rumble from handling noise, HVAC systems, and traffic. Voices have very little useful content below 80 Hz.
• Notch at the room’s resonant frequency — if you hear a consistent hum, it is likely at 50 Hz (Europe) or 60 Hz (North America) electrical hum, or at a room mode frequency. A narrow 6-10 dB cut at the offending frequency removes it without affecting voice quality.
• Cut at 200-300 Hz if “boxy” — cheap microphones and small rooms emphasize this range, making voices sound like they are inside a cardboard box. A 2-3 dB cut opens up the sound.

Optional boosts:

• +2 dB at 3 kHz — presence boost, makes voices “pop” and sound more forward
• +1 dB at 8-10 kHz — “air” boost, adds brightness and openness

9. Compression — Even Out Dynamics

Compression reduces the difference between the loudest and quietest parts of the recording.

Recommended settings for podcast voice:

• Threshold: -18 to -24 dBFS (start compressing when the signal crosses this level)
• Ratio: 2:1 to 4:1 (for every 2-4 dB the signal exceeds the threshold, output increases by 1 dB)
• Attack: 10-20 ms (fast enough to catch transients, slow enough to preserve consonant punch)
• Release: 100-200 ms (matches natural speech cadence)
• Makeup gain: adjust to bring the compressed signal back to the same perceived loudness

Target: 6-10 dB of gain reduction on the loudest passages. The goal is to narrow the dynamic range to about 10-12 LU (loudness units).

Check: A/B the compressed audio against the uncompressed original. The compressed version should sound more consistent in level without sounding “squashed” or lifeless.

10. De-essing (If Needed)

Sibilance — harsh “s” and “sh” sounds — is common with condenser microphones, especially when the speaker has prominent sibilance naturally.

Check: Listen for harsh, piercing “s” sounds, especially in words like “system,” “session,” “specifically.”

Target: Sibilance should be present but not painful. If you wince at any “s” sound, it needs de-essing.

Settings: A de-esser is a frequency-specific compressor. Set it to target 5-8 kHz with 4-8 dB of reduction and a fast attack (1-2 ms).

11. Level Match Between Speakers

If your podcast has multiple speakers, they should be at the same perceived loudness.

Check: Measure the LUFS (Loudness Units Full Scale) of each speaker’s track individually, on representative 30-second segments.

Target: All speakers within 1 LU of each other. A 3 LU difference is noticeable. A 6 LU difference is distracting.

Fix: Adjust individual track gain before the master bus processing. If one speaker is 4 LU quieter, add 4 dB of gain to their track.

Master Phase Checks

12. Integrated Loudness (LUFS)

This is the most important number in podcast mastering. Every podcast platform normalizes to a target loudness:

Platform	Target LUFS
Spotify	-14 LUFS
Apple Podcasts	-16 LUFS
YouTube	-14 LUFS
Amazon Music	-14 LUFS
Overcast	-16 LUFS

Target: Master to -16 LUFS integrated. This is the safe middle ground — platforms that target -14 LUFS will boost your episode by 2 dB (harmless), while platforms targeting -16 LUFS will play it as-is.

Check: Use a loudness meter (every major DAW has one, or use the free Youlean Loudness Meter plugin) and measure the integrated LUFS of the entire episode.

13. True Peak Level

True peak is the actual maximum sample value, accounting for inter-sample peaks that can occur during digital-to-analog conversion.

Target: True peak below -1.5 dBFS. This prevents clipping during encoding to lossy formats (MP3, AAC, Opus) where the codec’s reconstruction filter can create peaks that exceed the original peak level.

Check: Use a true peak meter (not just a sample peak meter). Your DAW’s loudness meter usually includes true peak measurement.

Fix: Apply a limiter with the ceiling set to -1.5 dBFS. A loudness maximizer can handle both the LUFS target and true peak ceiling in one step.

14. Export Format and Encoding

Recommended format: MP3 at 128 kbps CBR (constant bit rate) for speech-only podcasts. 192 kbps for podcasts with music segments. CBR is preferred over VBR because some podcast players have seeking issues with VBR files.

Alternative: AAC at 96-128 kbps if your host supports it. AAC sounds better than MP3 at the same bitrate due to more efficient encoding.

Check: After exporting, play the MP3/AAC file back and listen for encoding artifacts — usually audible as a “swishy” or “bubbly” quality on sibilants and high-frequency content. If you hear artifacts at 128 kbps, increase to 192 kbps.

Metadata: Embed episode title, show name, episode number, and artwork in the file’s ID3 tags (MP3) or M4A metadata (AAC). Many players display this information.

15. Final Listen-Through on Multiple Devices

The most important check is the simplest: listen to the exported file on the devices your audience uses.

Check on:

1. Headphones — your reference. You know how these sound.
2. Laptop speakers — many listeners use these. Check that dialogue is clear and the low end is not overwhelming.
3. Phone speaker — surprisingly common for podcast playback. Voices should be intelligible even through a tiny speaker.
4. Car audio (if possible) — road noise competes with podcast audio. Check that levels are adequate.

Listen for: Any problems the previous 14 steps missed. Trust your ears over meters. If something sounds wrong, it is wrong — even if the meters say it is fine.

The Quick Version

If you only have time for five checks:

1. No clipping (0 clipped samples)
2. Noise floor below -50 dBFS
3. Speakers level-matched within 1 LU
4. Integrated loudness at -16 LUFS
5. True peak below -1.5 dBFS

Hit those five targets and your episode will sound professional on every platform. The other ten steps refine the quality further, but these five are non-negotiable.

For automated processing that handles loudness normalization, noise reduction, and compression in a single pass, Hearably Studio processes everything in your browser without uploading your audio to any server.