Podcast chapters jump to the wrong moment? Check the final cut

A chapter called “Choosing the microphone” is less useful if tapping it lands in the introduction. Before rebuilding the chapter list, check which audio revision those times describe. A list made before an intro was shortened can remain readable while pointing to the wrong places.

Keep the final listening file and its chapter list together. Check a few transitions against that exact file, then test the chapter source your intended player actually uses. A valid metadata file and a useful listening destination are separate checks.

Start with one named audio revision

Write down the episode, export filename, duration and revision. Keep the earlier export and chapter list until the replacement has been checked. If two exports share a filename, add a revision to your working copy so the listening result has an identifiable reference.

Use the blank final-audio chapter check ledger to record the chapter source, intended topic and what you actually hear. Its rows begin as NOT CHECKED. It does not create chapters or certify a feed.

Do not adjust all timestamps from one observation. In this invented example, removing ten seconds before every later chapter would move those later destinations by ten seconds. A second edit in the middle would affect only later destinations again. That is why a single global offset can repair one chapter while leaving another wrong.

Check the format before changing the numbers

The Podcast namespace JSON Chapters specification uses numeric seconds for startTime. Its document requires version and chapters; each chapter requires startTime. A title is optional in that format, and the list is assumed to be ordered by increasing start time.

For an invented boundary at three minutes and seven seconds, the JSON number is 187. Writing 3.07 would describe just over three seconds. This arithmetic example is not a tested episode or a complete chapters file.

Podlove Simple Chapters is a separate XML format. It requires a chapter title and a start time expressed in Normal Play Time, which accepts strings such as 3:07. Use the requirements for the format you are exporting. Renaming a file does not convert its structure or units.

Listen before and after each destination

Open the named final audio export. Choose a chapter near the beginning, another after a substantial edit and one near the end. For each, listen from shortly before the proposed start through the transition into the topic. Record whether the title describes what follows and whether the jump cuts away a useful introduction.

These samples are a first check. Before publishing, review every chapter destination. Give each title enough context to make sense on its own, and avoid promising a topic that starts well after the jump.

If the final file sounds right but a player lands elsewhere, record the player and version, the episode revision and the chapter source being tested. Compare against a fresh import or the host's documented refresh procedure. Do not assume that editing show-note timestamps changes an embedded chapter list or a separate JSON file. Keep a result marked uncertain until you can identify which source was used.

Generated titles still need a final-file check

Azin Ghazimatin and colleagues' PODTILE paper studies generating chapter boundaries and titles from podcast transcripts. It evaluates title quality and segmentation separately. The useful lesson for an editor is modest: a plausible title does not establish that its destination fits your final audio. The paper does not validate this ledger or guarantee accurate timestamps for your episode.

For a practical creator's perspective, John Voorhees's MacStories review of Chapters describes placing markers while navigating an episode. It is an older review, so treat it as workflow background rather than current compatibility or pricing guidance.

Keep the handoff small and traceable

Send the final audio revision, the corresponding chapter source and the completed listening notes together. If the audio changes again, mark the previous chapter check as superseded and review the affected destinations against the new export.

If the question is whether the recipient received the same file, use the audio checksum and listening guide. A matching checksum establishes a different fact from a chapter landing at the right topic. Neither replaces listening.

This guide is based on documentation and published research reviewed on 3 November 2026. No podcast hosting account, feed update, player import or real episode transfer was tested for it.