guide · static export
Audio

Mix a Podcast in Audacity Without Harsh Sibilance

Two voices at different levels, one room, no budget. A fixed chain of four effects gets a clean result that sounds deliberate rather than loud.

by@guidox updated:10/31/26 time:45 min level:Intermediate tested on:Windows 11 24H2, Audacity 3.7 works on:Windows 10, Windows 11 tools:Audacity
TL;DR

Normalise to **-3 dB peak**, apply **noise reduction** on a sampled profile, gentle **compression** and a **de-esser** style filter. Export at **-16 LUFS** for platforms that accept it.

Before you start

Requirements

  • Audacity
  • A recorded two-track session
  • Headphones
Step by step

Instructions

01

Clean the tracks before anything else

Select a second of room tone and build a noise profile, then apply noise reduction with modest settings. Heavy reduction makes voices sound underwater and is hard to undo later.

Effect > Noise Removal and Repair > Noise Reduction
⚠ warningSix dB of reduction is plenty for a normal room. Twenty dB removes breaths and consonants along with the hum.
02

Level the two voices first

Set each track so both speakers peak at the same point before you compress anything. Mixing two quiet tracks up at the end amplifies the room along with the voice.

Effect > Volume and Compression > Normalize
03

Compress gently

A ratio around 3:1 with a slow attack and a moderate release evens out the loud passages without flattening the dynamics. Listen for pumping between phrases, not for loudness.

Threshold -18 dB, Ratio 3:1, Attack 20 ms, Release 200 ms
04

Tame the sibilance

Sharp S sounds live between 5 and 8 kHz. A narrow cut of a few dB on the worst syllables is enough; a broad cut dulls the whole voice.

Filter Curve EQ: -4 dB at 6-7 kHz
tipSelect one harsh syllable, loop it and adjust until it stops clicking in your ears. Then apply the same curve to the track.
05

High-pass and finish

Cut everything below 80 Hz on the voice tracks to remove rumble, then normalise the final mix to a consistent peak and check the loudness against a reference podcast.

High-pass filter: 80 Hz, 12 dB/octave
06

Export with the right settings

Export as WAV for an archive and MP3 or AAC at 192 kbit/s or better for publishing. Mono is acceptable for a single voice and halves the file size.

File > Export > WAV / MP3 192 kbit/s
Troubleshooting

Common issues

The voices sound hollow after noise reduction

Too much reduction. Undo it and redo at six dB, or record again with the microphone closer.

One speaker is much quieter

Fix the gain at the source next time. For this file, normalise the tracks separately before mixing.

The mix clips after compression

Compression raises the average. Normalise down before export and leave headroom for the encoder.

There is an echo on one track

A reflective room. A short reverb reduction or a closer microphone position on the next recording helps more than EQ.

FAQ

Frequently asked

Do I need a loudness meter?

For a platform with a target, yes. Most accept -16 LUFS mono and -14 LUFS stereo.

Should I record in mono or stereo?

One microphone per person, recorded as separate mono tracks. Mix down at the end.

How to undo this

Keep the untouched original recording. Audacity undo history disappears on close, so save a copy before the first effect.

Changelog · 1 entry
  1. 2026-10-31
    First version
Interactive version: checklist, feedback, related guides →

Free · no account needed · tracks your progress