To mix a podcast, you clean up each voice track, balance levels between speakers, shape tone with EQ and compression, then set the whole episode to a consistent loudness — usually around -16 LUFS for stereo podcasts. The goal is simple: every speaker is clear, easy to listen to at any volume, and roughly the same loudness as other shows. Here is how to mix a podcast from raw recordings to a finished export.
Get organised before you mix a podcast
Start with each speaker on its own track. If you recorded everyone into one file, the mix options are limited, so record separately whenever you can. Listen through once and note the problems: background hiss, a quiet guest, plosives, mouth clicks, a humming fridge. Knowing what is wrong tells you what processing each track actually needs — most tracks need less than you think.
Step 1: Clean up the recordings
Fix issues at the source before adding tone-shaping:
- Noise reduction: sample a quiet moment and reduce broadband hiss or hum. Apply gently — heavy denoising makes voices sound underwater.
- High-pass filter: roll off everything below about 80 Hz to remove rumble, handling noise and air conditioning.
- Edit out distractions: long pauses, “ums,” coughs and false starts. Leave natural breaths in so it sounds human.
Step 2: EQ each voice
EQ makes voices clear and distinct. A common starting point: a gentle cut around 200–400 Hz if a voice sounds boxy, a small boost around 3–5 kHz for presence and intelligibility, and a touch of air above 10 kHz if the mic sounds dull. Make small moves and compare with the EQ bypassed. If EQ is new to you, our EQ and compression fundamentals guide explains the why behind each move.
Step 3: Compress for consistency
Speech naturally jumps between loud and quiet. Compression evens that out so listeners do not reach for the volume. Try a ratio around 3:1 to 4:1, a medium attack and fast-to-medium release, and pull the threshold down until you see about 3–6 dB of gain reduction on the loudest words. Then add make-up gain. The aim is control, not obvious pumping. If the controls feel unfamiliar, our guide on how to use a compressor breaks down each knob.
Step 4: De-ess and tame harshness
Bright mics and close talking exaggerate “s” and “t” sounds. A de-esser targeting roughly 5–8 kHz softens sibilance without dulling the whole voice. Use the lightest amount that does the job; over-de-essing gives speakers a lisp.
Step 5: Balance levels between speakers
Set the host and each guest to the same perceived loudness so nobody disappears. Ride or automate volume where one person leans in or pulls back. If you mixed multiple voices to the same intelligibility, the conversation feels effortless. Keep music beds and stings 10–15 dB below speech, and duck them under voices so talking always wins.
Step 6: Set the final loudness
Podcasts are normalised by loudness, not peaks. The widely used target is -16 LUFS for stereo (-19 LUFS for mono) with true peaks no higher than -1 dB. Put a loudness meter on the master, mix to taste, then nudge the overall level to hit the target. A limiter on the master catches stray peaks and adds a little glue. Our LUFS explainer walks through measuring and hitting these numbers, and the mixing and mastering hub has more on master-bus processing.
Why a good recording beats a good mix
Mixing can rescue a flawed recording, but it can never fully replace a clean one. The two biggest variables happen before you touch a fader: the room and the mic technique. A small, soft-furnished room with a duvet or some panels behind the speaker kills the boxy reverb that no plug-in removes cleanly. Consistent distance — roughly a hand-span from the mic, slightly off-axis to avoid plosives — keeps the tone and level steady so your compressor and EQ have less to fight. When a guest records remotely, ask them to use headphones, sit in the quietest room they have, and record locally rather than relying on the call audio. Five minutes of setup saves an hour of repair later, and the finished episode simply sounds more professional.
How to set the right amount of processing
The hardest skill in podcast mixing is restraint. Each tool should solve a specific, audible problem, not be added by reflex. A useful workflow is to bypass everything, listen to the raw voice, and add one process at a time, asking after each whether the voice is genuinely clearer. If it is not, take the process off. Reference a podcast you admire at a matched loudness — your ear adapts fast, and a known-good comparison stops you from over-EQing or over-compressing. When in doubt, do less: a slightly raw but natural voice is far easier to listen to for an hour than a heavily processed one that sounds brittle or pumped.
Common podcast mixing mistakes
- Over-processing: stacking aggressive noise reduction, EQ and compression until the voice sounds thin, robotic or “underwater.” Gentle and natural wins.
- Mismatched speakers: a loud host and a buried guest force the listener to keep adjusting volume. Balance perceived loudness, not just the meters.
- Music too loud or not ducked: intros and beds that bury the talking, or that never drop under the voice. Speech should always win.
- Ignoring loudness standards: exporting at peak-normalised levels instead of -16 LUFS, so your show is much quieter or louder than everything else in the feed.
- Editing out every breath: removing all pauses and breaths makes speech sound unnatural and tiring. Leave the human moments in.
DAW notes: Audacity and GarageBand
In Audacity, use Effect > Noise Reduction, the built-in EQ and Compressor effects, then Loudness Normalization to set -16 LUFS before exporting. In GarageBand, use the Channel EQ, Compressor and a De-esser plug-in per track, set levels with track volume automation, and check loudness with a metering plug-in since GarageBand has no built-in LUFS meter. Whatever you use, export as a 128–192 kbps MP3 or AAC. For capture tips that make mixing easier, see our home podcast recording guide.
Frequently asked questions
What loudness should a podcast be mixed to?
Target -16 LUFS integrated for stereo podcasts (or -19 LUFS for mono), with true peaks at or below -1 dB. This matches most platforms and keeps your episode’s volume in line with other shows.
In what order should I apply effects when mixing a podcast?
Noise reduction and editing first, then high-pass filter, EQ, compression, and de-essing on each voice. Balance the speakers, add and duck any music, then finish with loudness normalisation and a limiter on the master.
Do I need separate tracks for each speaker?
It helps enormously. Separate tracks let you EQ, compress and balance each voice independently and fix one person’s problems without affecting the others. Record each speaker to their own track whenever possible.
How long should mixing a podcast take?
Once you have a saved chain of EQ, compression, de-essing and loudness settings, a clean recording can take well under an hour to mix. Save your settings as a preset or template so each new episode starts from a known-good point and you only adjust for that day’s recording.
Can I fix a noisy or echoey recording in the mix?
Up to a point. Gentle noise reduction tackles steady hiss and hum, and careful EQ can reduce some boxiness, but heavy room reverb and severe background noise are very hard to remove without making the voice sound artificial. Improving the recording space and mic technique will always give a better result than relying on repair plug-ins.
