PodPolish

Record it right. Fix less. Know the bill.

How to remove background noise from a podcast, in the right order

Noise removal fails from enthusiasm more than from bad tools: wrong order, full strength, group punishment for one guest’s fridge. This is the bench’s triage — what to fix, what to process, what to re-record, and what each lane costs when free runs out.

from the PodPolish benchpublished July 31, 2026✓ prices checked July 30, 2026
RECORDCLEANEDITPUBLISH

Triage before tools

Not all noise is the same job, and the order of operations is the whole trick. Constant noise — fans, hiss, hum — is the easiest kill: profile-based reduction (Audacity) or one Enhance pass. Intermittent noise — traffic, doors, a barking dog — resists profiles; it’s Enhance’s territory, or an edit if it landed between sentences. Echo is the bully of the three: cleanup tools blur it at best, which is why the recording setup spends so long on the room. And crosstalk noise on interview shows is why separate tracks exist — per-person cleanup beats group punishment.

Clean first, edit second — always. Every splice inherits the room tone around it; give the editor one consistent, already-clean room to cut in.

The order of operations

  1. Listen once, honestly. On headphones, note-taking: what kind of noise, which tracks, how bad. Ninety seconds that picks the right lane.
  2. Fix the free way first. One pass per raw track through Enhance’s free lane — split files at 30 minutes, stay inside the daily hour. For steady-hum-only tracks, Audacity’s noise reduction (learn the profile from your room-tone seconds, apply gently) keeps everything offline.
  3. A/B at 60–80%. Full-strength rebuilds can sound boothy. Mixing the processed track back under the original is the manual strength slider — free, and it preserves your show’s actual room.
  4. Escalate only what earns it. Ums, mouth sounds and dead air too? That’s one Cleanvoice upload instead. Weekly noisy interviews? Stop repairing and start recording clean.
  5. Then edit. Cuts, music, levels — on clean, consistent tracks.

What the lanes cost when free runs out

The cleanup lanes, priced per episode
RouteMonthlyPer episodeHow it covers you
Adobe Enhance, free tier$0.00/mo$0.001 free hour/day covers a weekly show; files split to fit 30-min cap
Audacity, manual pass$0.00/mo$0.00profile-based noise reduction; costs your evening instead
Adobe Enhance Premium$9.99/mo$2.504 h/day, hour-long files, strength slider
Cleanvoice 10 h/month$11.00/mo$2.756 raw hours/month fits the 10 h plan; credits roll over
Assumptions, printed where you can argue with them: 4 episodes a month, 60-minute shows, about 1.5 hours of raw audio each. Krisp and Descript lanes priced on their own pages. List prices verified July 30, 2026.

The full wall — every lane, one table — hangs on the cost worksheet.

Asked at the bench

How do I remove background noise from a podcast for free?
Run each finished track through Adobe Podcast Enhance — free for an hour of audio a day, files up to 30 minutes. For steady hum on a long file, Audacity’s profile-based noise reduction is the offline alternative. Both lanes are walked step-by-step on this bench.
Should I remove noise before or after editing?
Clean first, edit second. Editors make cuts sound natural by matching room tone — if you clean after editing, every splice can exhale a different room. One cleanup pass per raw track, then cut.
Why does my cleaned podcast sound underwater?
Over-processing. Rebuild-style tools at full strength can leave voices synthetic; profile-based reduction pushed past ~15 dB goes watery. Use less, or mix the cleaned track back under the original at 60–80%.

Same shelf