How to Clean Up Podcast Audio With Auphonic
Auphonic processes audio files server-side: it analyzes loudness, applies adaptive leveling, runs noise and hum reduction, and exports to a loudness target you pick. Typical output is -16 LUFS for podcasts or -14 LUFS for Spotify. Free accounts get 2 hours of processing per month; paid credits start around $11 for 9 hours.
It is not an editor. You cannot cut. It is the step after the edit, replacing roughly 20 minutes of manual compressor work per episode.
What does Auphonic do that a DAW does not do faster?
Nothing, if you already know your DAW. Everything, if you do not.
The value is that it makes the judgment calls. Adaptive leveling rides gain continuously to hold speech at a target, where a manual compressor needs a threshold, a ratio and a makeup gain you have to hear before you trust. For a two-person interview with mismatched mic positions, that is the difference between a 25-minute and a 4-minute fix.
How do you process a podcast episode, step by step?
- Export a clean WAV from your editor. Not a compressed MP3 — processing artifacts compound. 44.1 kHz or 48 kHz, at least 16-bit.
- Upload to Auphonic and pick a production type. “Podcast” preset is the safest starting point; “Interview / Multi-track” if you have separate stems.
- Set loudness target and true peak. -16 LUFS / -1.5 dBTP for podcast RSS. Spotify normalizes to -14 LUFS, so exporting at -16 means it gets turned up, not down.
- Enable Adaptive Leveler at around 60-70% strength. Higher values flatten dynamics and make laughter sound robotic.
- Add noise reduction only if you hear hiss. Applied to clean audio it thins out consonants. Check the before/after on a sibilant word.
- Leave the output format on MP3 VBR V0 or AAC 192 kbps. Auphonic’s MP3 encoder is fine; re-encoding it again elsewhere is not.
- Download, spot-check three timestamps, then upload to your host. Do not trust the render blindly — intros with music beds get leveled oddly.
How much does Auphonic cost?
| Tier | Price | Processing hours | Notes |
|---|---|---|---|
| Free | $0 | 2 hrs/month | Full feature set, non-commercial terms |
| Credits | ~$11 | 9 hrs | Valid 2 years |
| Business | ~$99/mo | 100 hrs | API, team accounts |
Credits expire after two years, which is generous. The 2-hour free tier is about four half-hour episodes monthly — enough to test, not enough to run a show.
One caveat: free-tier output is for non-commercial use under the current terms. A podcast with sponsors is commercial.
Multi-track interviews: what to actually do
Feed separate tracks per speaker and let Auphonic cut on silence. It detects speech segments and produces a single mixed file with automatic ducking.
Caveats worth knowing before you rely on it:
- Crosstalk gets muddled. Simultaneous speech on two tracks can produce phase issues.
- Silence cutting changes pacing. A 500 ms pause becomes 250 ms. That reads as rushed to some listeners.
- Room reverb does not get fixed. Leveling and noise reduction cannot undo a bad room.
Who should not use this
Anyone mastering music. Adaptive leveling destroys dynamic range by design. Also anyone on a DAW with a preset chain they already trust — the extra round trip adds a file transfer for marginal gain.
If you record in a treated room with consistent levels, you probably do not need it. If you record in hotels and kitchens, you do.
Key Takeaways
- Auphonic normalizes loudness and reduces noise server-side; it does not edit.
- Free accounts get 2 hours of processing per month under non-commercial terms.
- Standard podcast target is -16 LUFS with -1.5 dBTP true peak.
- Adaptive Leveler above ~70% flattens dynamics; keep it at 60-70%.
- Multi-track mode handles interviews but crosstalk and silence-cutting need review.
