How to get publish-ready podcast audio in 3 steps — no upload
- Step 1Bring your finished edit — Drop the completed episode edit (audio or video). This is the version you're about to upload — intro/outro assembled, content edits done.
- Step 2Choose the platform target — Match the preset to where you publish: Apple Podcasts (-16), Spotify / YouTube (-14), Amazon Music (-14 / -2 dBTP), or EBU R128 (-23) for broadcast. Multi-platform? Master once to -16 LUFS.
- Step 3Choose the delivery format — WAV or FLAC if your host re-encodes (give it a clean source); MP3 or M4A if you self-host or your host wants a finished file. Lossy output uses a 192 kbps working bitrate.
- Step 4Run the publish pass — Click Master. The chain denoises, strips silence, then normalises to your spec with the -1 dBTP ceiling — producing a mono, distribution-ready file.
- Step 5Verify against the spec — The report shows the achieved integrated LUFS and true peak. Confirm they match the platform target before uploading. If it says 'dynamic', the level may drift — see the cookbook.
- Step 6Tag, then upload — Add episode title and artwork with the id3-editor (the master chain doesn't write tags), then upload to your host. Most hosts add chapters at this stage.
Platform targets and what they mean
How each preset maps to a real platform's playback normalisation. LUFS/TP values are read from the engine; platform playback behaviour is documented industry practice.
| Publish to | Pick this preset | Why | Playback note |
|---|---|---|---|
| Apple Podcasts | Apple Podcasts (-16 LUFS) | Apple normalises playback to -16 LUFS | Master at -16 and Apple passes it through unchanged |
| Spotify (podcasts) | Apple Podcasts (-16) or Spotify (-14) | Spotify normalises to ~-14 LUFS | -16 master is raised slightly; -14 master plays as-is |
| YouTube | Spotify / YouTube (-14 LUFS) | YouTube normalises loud uploads down toward ~-14 LUFS | Master at -14 to avoid being turned down |
| Amazon Music | Amazon Music (-14 LUFS, -2 dBTP) | Stricter -2 dBTP peak ceiling | Use the dedicated preset for the tighter ceiling |
| Radio / broadcast | EBU R128 broadcast (-23 LUFS) | Broadcast reference is -23 LUFS, tighter 7 LU range | Deliver WAV; keep a -16 LUFS master for the podcast feed |
What 'publish-ready' includes (and excludes)
The chain produces a technically correct master. These items it does, and these it does not.
| Aspect | Done by the chain? | Detail / next tool |
|---|---|---|
| Noise floor cleaned | Yes | RNNoise removes steady hum/hiss |
| Dead air removed | Yes | Silence strip at -40 dB / 0.5 s |
| Loudness on spec | Yes | 2-pass EBU R128 to your target |
| True-peak safe | Yes | -1 dBTP ceiling (-2 on Amazon) |
| Stereo image | No | Output is mono |
| Episode tags / artwork | No | Use id3-editor |
| Chapter markers | No | Added by your host on upload |
| File size for a host cap | Indirectly | Use audio-compressor to hit a size |
Output containers for delivery
Containers offered for the publish master, with the right pick per host type.
| Container | Codec | Choose when |
|---|---|---|
| WAV | 16-bit PCM | Host re-encodes for you — give it the cleanest source |
| FLAC | Lossless | Lossless delivery at smaller size than WAV |
| MP3 | MP3 (LAME) | Self-hosting or host wants a finished MP3 |
| M4A | AAC | Apple-friendly, efficient quality-per-byte |
Cookbook
Pick-the-right-target recipes for each major destination, plus the publishing gotchas that come up after the numbers look right.
Publish to Apple + Spotify from one master
The most common case: you publish through one host feed that distributes to both Apple and Spotify. One -16 LUFS master serves both.
Target : Apple Podcasts (-16 LUFS) Format : MP3 (or WAV if the host re-encodes) Result : -16.0 LUFS / -1.1 dBTP, mono Apple : plays through unchanged Spotify: nudged up ~2 LU on playback (inaudible loss) One file -> both platforms, consistent with other shows.
YouTube video-podcast upload
You post the episode to YouTube too. YouTube normalises loud uploads down, so master to -14 LUFS to avoid being turned down.
Target : Spotify / YouTube (-14 LUFS) Format : M4A (AAC) Note : output is audio-only; pair it with your video in the editor. Result : -14.0 LUFS / -1 dBTP. YouTube leaves -14 uploads close to as-is.
Amazon Music's stricter peak
Amazon Music wants a -2 dBTP ceiling. Use the dedicated preset rather than the generic -14.
Target : Amazon Music (-14 LUFS, -2 dBTP) Why : extra peak headroom for their pipeline Result : -14.0 LUFS / -2.0 dBTP, mono Delivery: WAV or FLAC recommended for the cleanest ingest.
Hitting a host's file-size cap
Your host caps uploads at, say, 200 MB. The master is fine on loudness but the WAV is too big.
Problem: master.wav = 310 MB, host cap = 200 MB
Fix:
1. Master to spec (target -16 LUFS) as MP3 -> much smaller, OR
2. Master to WAV, then audio-compressor to a target size:
/audio-tools/audio-compressor (set targetMb)
Loudness spec is preserved either way.Numbers look right but it sounds quiet next to other shows
The integrated LUFS is on target, yet the episode feels softer. Usually a wide dynamic range: peaks are loud, but the average sits low.
Report : -16.0 LUFS integrated, BUT LRA 14 LU (wide) Why : loud bursts pull peaks up; quiet talk dominates perception Fix: speech-leveler first (compress dynamics) -> then this master Result : tighter range, episode sits forward like polished shows.
Edge cases and what actually happens
On-spec LUFS but episode sounds quiet vs peers
ExpectedIntegrated LUFS is an average; a wide loudness range (LRA) means loud peaks coexist with lots of quiet talk, so perceived loudness sits low even at -16 LUFS. Compress the dynamics first with speech-leveler, then re-master — the same LUFS target will feel noticeably more present.
Loudnorm reports 'dynamic' — final LUFS drifts
ExpectedWhen the source loudness range exceeds the preset's LRA target, FFmpeg switches loudnorm to dynamic mode and the achieved LUFS may not exactly match the spec. The report flags it. For a guaranteed on-spec linear result, level dynamics first with speech-leveler.
Master is mono — fine for Apple, but you wanted stereo music
By designThe chain outputs mono (denoise resamples to 48 kHz mono). Apple/Spotify accept mono podcasts happily and it halves file size. If your show needs stereo (music, immersive), use loudness-normalizer standalone to hit the spec while preserving stereo, doing denoise separately or not at all.
True peak edges above the ceiling after MP3 encode
SupportedThe loudnorm pass targets -1 dBTP, but MP3/AAC encoding can re-introduce inter-sample peaks slightly above that. For a hard guarantee, master to WAV/FLAC (no re-encode) or use the Amazon preset's -2 dBTP headroom. The report's measured true peak lets you verify.
Host says the file has no chapters or artwork
Out of scopeThe master is audio-only and writes no metadata. Add title/artwork with id3-editor before upload; chapter markers are typically added in the host's dashboard at publish time.
Uploaded a video and got audio back
By designDropping a video extracts and masters the audio track, returning an audio file (the video is discarded). For a video-podcast deliverable, take the mastered audio and swap it into your video in an editor, then upload the video.
Episode exceeds the tier's duration cap
Tier limitFree preview: 30 min / 10 MB. Pro: 120 min / 200 MB. Pro+Media and Developer: unlimited duration / 100 GB. A long-form interview needs Pro+Media. Size and duration are checked independently.
Two platforms with different specs
SupportedIf you deliver to both a podcast feed and a radio station, run the chain twice — once at -16 LUFS (MP3 for the feed) and once at -23 LUFS EBU (WAV for broadcast). The source is processed locally each time; no re-upload.
Music-bed episode degraded by denoise
Limited effectRNNoise can damage music because it's a speech model. For music-forward episodes, skip the all-in-one chain and hit the spec with loudness-normalizer plus true-peak-limiter, preserving the music and stereo.
Frequently asked questions
What loudness should a published podcast be?
-16 LUFS integrated with a -1 dBTP true peak is the safe universal target. Apple Podcasts normalises playback to -16 LUFS (so it passes through unchanged) and Spotify to about -14 LUFS (so a -16 master is only nudged up). Mastering to -16 lets one file publish cleanly to both.
Which preset for YouTube?
Use Spotify / YouTube (-14 LUFS). YouTube turns loud uploads down toward roughly -14 LUFS, so mastering at -14 avoids having your audio attenuated on playback. The output is audio-only; combine it with your video in an editor.
Why does Amazon Music have a different ceiling?
The Amazon Music preset uses a -2 dBTP true-peak ceiling instead of -1, giving extra headroom for their delivery pipeline. Pick that preset specifically when you publish to Amazon; otherwise -1 dBTP is standard.
Do I need to master separately for each platform?
Usually no — a single -16 LUFS master is safe for Apple and Spotify. Master separately only when a destination has a distinct spec, like Amazon Music's -2 dBTP ceiling or broadcast's -23 LUFS. In those cases run the chain again with the matching preset.
What format should I deliver to my host?
If your host re-encodes the audio for distribution, give it WAV or FLAC so its encoder starts from a clean source. If you self-host or the host ingests finished files, MP3 or M4A is fine. All four hit the same loudness spec.
Why is my published file mono?
The chain outputs mono because the denoise stage resamples to 48 kHz mono. Podcast platforms accept mono and it halves file size. If you need stereo, hit the spec with loudness-normalizer standalone, which preserves channels.
The LUFS is right but it sounds quiet — why?
Integrated LUFS is an average. A wide dynamic range means loud peaks raise the peak meter while lots of quiet speech keeps perceived loudness low. Compress the dynamics with speech-leveler before mastering for a more 'forward' sound at the same LUFS.
Will the file clip on the platform?
No — the loudnorm pass enforces a -1 dBTP ceiling (-2 on Amazon), keeping it below full scale. Lossy encoding can add tiny inter-sample peaks; master to WAV/FLAC or use the Amazon headroom if you need an absolute guarantee.
Does it add my episode title and artwork?
No. The master is audio-only and writes no ID3 tags. Add title and artwork afterward with the id3-editor, then upload. Chapter markers are normally added by your host.
My host has a file-size limit — how do I meet it?
Master to MP3 (much smaller than WAV) at your loudness spec, or master to WAV then run audio-compressor targeting the host's size cap. The loudness target is preserved through either path.
How long an episode can I make publish-ready?
Free preview up to 30 minutes (10 MB), Pro up to 120 minutes (200 MB), Pro+Media and Developer unlimited duration (100 GB streamed). Long-form interviews need Pro+Media. Duration and size limits are enforced separately.
Is the audio uploaded to check the spec?
No. Measurement and normalisation both happen locally via FFmpeg 8.1 in WebAssembly. The integrated LUFS and true-peak readout in the report are computed in your browser, so nothing about your episode leaves the device.
Privacy first
Every JAD Audio tool runs entirely in your browser via FFmpeg (WebAssembly) and RNNoise. Your audio files never leave your device — verified by zero outbound network requests during processing.