How to Clean Up Podcast Interview Audio for Professional Sound
The quality gap between hobbyist and professional podcasts often comes down not to the host's content or chemistry, but to the audio production. When you record remote interviews — over Zoom, Riverside, Zencastr, or phone — guests record in their own environments with their own equipment. The result is often a mix of different microphone qualities, room acoustics, background noises, and recording levels that need careful cleanup.
Learning how to clean up podcast interview audio is one of the most valuable skills a serious podcaster can develop. This guide covers the complete workflow, from managing different guest recording qualities to meeting Apple Podcasts and Spotify loudness standards.
The Remote Interview Audio Challenge
Remote podcast interviews present a unique set of challenges:
Inconsistent microphone quality: Your guest might be recording on a headset, their laptop microphone, or an excellent condenser mic. You have no control over this.
Uncontrolled acoustic environments: Guests might be in their home office (reflective walls, keyboard and mouse noise), a café (background chatter, espresso machine), or any other space.
Level mismatches: Different recording gains mean one speaker might be 10-15 dB louder than another.
Echo and room reverb: Untreated rooms produce reverb that makes speech sound muddy and distant.
Connection artifacts: Occasional digital glitches, compression artifacts from video call encoding, or internet drops that create brief audio problems.
Your goal is to process all of this into a cohesive episode where both voices sound similarly clean, clear, and consistent.
Workflow: Separate Tracks or Mixed?
Before starting cleanup, know what you're working with:
Separate tracks (best case): Recording platforms like Riverside.fm, Zencastr, and Squadcast record each participant's audio locally and deliver separate audio files. Clean up each track independently before mixing.
Mixed track: Zoom recordings deliver a single mixed file with all speakers combined. This is harder to work with — you can't adjust individual speaker levels or apply different processing to different participants.
For serious podcast production, always use a platform that delivers separate tracks.
Step 1: Level Matching
Before any frequency processing, address volume inconsistencies between tracks:
- Listen to both tracks, note the average level difference
- Apply gain adjustment to the quieter track to roughly match the louder one
- Aim for both tracks averaging around -18 to -12 dBFS
This step makes subsequent processing much more consistent — you're not fighting huge level differences while trying to address noise.
Step 2: Noise Removal Per Track
Process each guest's track separately based on what you actually hear in their recording:
Laptop mic in quiet room: Light broadband noise reduction (10-12 dB) to address microphone self-noise. High-pass filter at 100 Hz.
Headset with background hum: Notch filter for electrical hum (60 Hz and harmonics) plus light noise reduction.
Kitchen or café background: iZotope RX Dialogue Isolate (or Audacity noise reduction with more aggressive settings). Noise gate to silence pauses. Accepts more processing artifacts as trade-off.
Good microphone in treated room: Minimal processing — perhaps just light noise reduction and a gentle high-pass filter.
Match the processing intensity to the problem. Over-processing a clean track introduces artifacts unnecessarily.
Step 3: EQ for Clarity and Consistency
Equalizing different microphone characters to sound more similar is part of professional podcast production:
Thin, harsh laptop microphone:
- Low shelf boost +3 dB at 200 Hz (adds body)
- Mid cut -2 dB at 1-2 kHz (reduces harshness)
- High cut at 10-12 kHz (removes digital harshness)
Boxy, midrange-heavy headset:
- High-pass filter at 120 Hz
- Broad dip -3 dB at 400-600 Hz (reduces boxiness)
- Presence boost +2 dB at 3-5 kHz (improves intelligibility)
Good condenser mic (too bright):
- High shelf cut -2 dB at 8 kHz (gentle de-brightening)
- No other changes needed typically
The goal isn't perfectly identical frequency response, but a sense that both voices come from the same "space" rather than obviously different recording environments.
Step 4: De-reverb for Room Acoustics
Remote guests often record in rooms with reflective walls, hardwood floors, or low ceilings. The room reverb creates a "swimming" or "bathtub" quality that makes the voice feel distant.
iZotope RX De-reverb:
This is the best available tool for this problem. The Reverb Reduction amount determines how aggressively reverb is removed — start at 50% and listen carefully for processing artifacts (a "metallic" quality to consonants) that indicate you've gone too far.
Manual workaround without RX:
Some reverb issues can be partially addressed with high-pass filtering (reverb energy concentrates in the low-midrange) and a gentle reduction in the 300-600 Hz range. Not as effective as dedicated de-reverb, but an improvement.
Step 5: Compression for Consistent Dynamics
After noise and EQ, apply compression to even out level variations within each track:
Voice compression settings:
- Threshold: -24 to -18 dBFS (adjust based on typical speaking level)
- Ratio: 3:1 to 4:1
- Attack: 10-20 ms (allows natural transient punch through)
- Release: 100-300 ms
- Makeup gain: +4 to +6 dB (to compensate for gain reduction)
After compression, both tracks should have much more consistent, even levels throughout.
Step 6: Final Mixing and Loudness Normalization
Mix the processed tracks:
- Set relative levels so host and guest are balanced (guest slightly lower is often natural and correct)
- Check for level conflicts where both speak simultaneously (rare but needs addressing)
- Apply a final limiter at -1 dBTP to prevent any clipping
Loudness normalization for podcast delivery:
- Apple Podcasts: -16 LUFS integrated, -1 dBTP true peak
- Spotify: -14 LUFS integrated
- General: -16 LUFS hits a good balance for most platforms
Use Audacity's Loudness Normalization (Effect > Loudness Normalization) or iZotope RX Loudness Control for accurate LUFS-based normalization.
Advanced Techniques: AI-Powered Podcast Cleanup
Dedicated AI tools have emerged specifically for podcast audio cleanup:
Adobe Podcast Enhance: Online tool (free) that applies AI noise reduction and enhancment to voice recordings. Often produces impressive results for simple background noise.
Descript Studio Sound: If you use Descript for editing, Studio Sound applies AI enhancement with one click. Results are good for most common podcast recording environments.
NVIDIA RTX Voice: Excellent real-time noise suppression if you use RTX GPU hardware.
These tools are convenient but less controllable than manual processing. For podcasts where guest quality varies widely, manual control over each track's processing usually produces better results.
When to Use a Professional Service
For podcast production where quality directly affects audience growth and retention, professional audio cleanup makes sense:
- You have a compelling interview with a famous guest but their recording quality is poor
- You're producing a podcast at scale (multiple episodes per week) and don't have time for manual cleanup
- You need broadcast or premium streaming quality (BBC, Spotify Exclusive, NPR)
- You've tried DIY cleanup but guest tracks sound obviously processed or still distracting
WefixSound offers podcast interview cleanup with a free 60-second sample — upload your most challenging guest recording and see the improvement before committing. We understand podcast production timelines and offer efficient turnaround for regular podcast clients.
For podcast producers who publish regularly, contact us about ongoing production pricing — consistent, professional audio quality across every episode builds listener loyalty.
Checklist: Podcast Interview Cleanup
Before publishing, verify:
- Both voices at approximately equal loudness
- Background noise removed or reduced to acceptable level
- No processing artifacts (watery voice quality, metallic consonants)
- Room reverb reduced if excessive
- Integrated loudness between -16 and -14 LUFS
- True peak below -1 dBTP
- No clicks, pops, or digital artifacts
- Transitions between speakers sound natural
Related Articles
Professional podcast interview audio doesn't happen automatically — it requires systematic cleanup that addresses each recording's specific problems. With the right workflow and tools, even difficult remote recordings can sound polished and professional. For high-value episodes or ongoing production, WefixSound delivers consistent professional quality.