How to Prepare Audio for Publishing

Workflow · 9 min read · Delivery

Publishing is where separate tools become one pipeline. Whether the destination is a podcast host, a course platform, or a social clip, the same questions repeat: is it loud enough, is it clipped, is it the right length, and is it the right file type?

Checklist (print or keep open)

Step 1 — Quality Check first

Open Audio Quality Check and treat the report as your punch list. If clipping is flagged, do not “fix it” with more boost. If LUFS is very low, plan a normalize pass.

Step 2 — Structure the file

Trimmer for edges. Silence Finder if you need to see dead air inside long lectures before cutting.

Step 3 — Level

Volume Booster for under-level material. Re-run Quality Check. Compare against the targets you use for speech or music (see the LUFS guide).

Step 4 — Format

Convert only at the end with the Audio Converter or a specific route like WAV to MP3. Avoid encode → decode → encode loops.

Common upload rejections and what they mean

• Wrong format or sample rate: convert to the format the host lists.
• File too large: lower the bitrate or switch to mono for speech.
• Too quiet or too loud: adjust level and check peaks.
• Clipping detected: return to an unclipped version instead of boosting.
• Long silence at the start or end: trim the edges.

Keep a lossless master

Save an uncompressed or lossless version of the finished audio before you export lossy files. If a platform changes its requirements or you need another format later, you can re-export from the master instead of encoding an already compressed file again.

Platforms differ — your order should not

Hosts disagree on exact LUFS numbers, but almost none want clipped files or thirty seconds of silence. Keep the order stable: measure → edit length → level → convert → measure once more.

Build a repeatable publish ritual

Teams waste time when every episode invents a new order of operations. Write your checklist once: trim → measure → level → convert → measure. The tools can change; the order should not. New collaborators can follow the same recipe without guessing which site tab to open first.

Store a short note with each published file: date, integrated LUFS, peak, and format. When a host rejects an upload, you know whether the problem is loudness, length, or container without re-listening for an hour.

Speech, music, and mixed shows

A pure talk episode and a show with loud stingers need different expectations. Music beds push integrated LUFS up even when speech feels normal. If the host complains that “ads are fine but the talk is quiet,” you may need more even dialogue levels, not a hotter master that smashes the stingers harder.

Last mile: naming and backup

Name exports with episode numbers and status (ep12-master.mp3, not final-final-3). Keep the WAV or high-quality intermediate until the episode is live. If the MP3 sounds dull, you can re-encode from the intermediate instead of from a lossy copy.

Example pass on a real episode

Suppose a thirty-minute talk show lands at −22 LUFS with peaks at −6 dBFS and three seconds of silence before the cold open. You trim the silence, normalize toward −16 LUFS, re-check that peaks stay under 0 dBFS, export MP3 at a host-safe bitrate, and measure once more. Total time is usually under fifteen minutes once the habit is familiar — far less than fixing a rejected upload the next day.

If the second measurement still shows clipping, undo the aggressive normalize and use a smaller gain. Publish-ready means stable, not maximally loud.

Frequently asked questions

← All workflows & recipes