Skip to content
All case studies

Polymaths · Podcast production · My own show

An hour-long episode, 76 graphics, none of them animated by hand.

Polymaths is the podcast I co-host. Its look calls for a name, a question, a number or a chapter card on screen every minute or so. Building those by hand costs an editor hours per episode, so I built the kit that renders them from a cue file and puts them on the timeline.

Client
Polymaths, the show I co-host
Sector
Podcast production
Built
Graphics engine, timelines and an editor handoff kit
Status
Used on episode one

The problem

Good graphics were eating the edit.

The shows we look up to put something on screen every minute or so: who is talking, the question being answered, the number that carries the sentence. Most of them pay editors to build each one by hand, episode after episode.

Every graphic meant retyping text into a title, matching the fonts and colors, keyframing the same entrance again, and then doing it all over for the vertical clip. Hand that to a hired editor and the look drifts unless every detail is written down.

What I built

Write the cue once. The kit does the rest.

The kit lives in a folder any editor can open. It takes an episode and a list of cues and gives back the graphics and a timeline with every one of them already in place.

  1. Transcribe the episode

    Every word is transcribed and matched to the person who said it, using each speaker's own microphone track.

  2. Plan the cues

    A cue is a name, a question, a stat, a note, a list or a chapter, pinned to a phrase in the transcript rather than a timecode, so it survives a re-edit.

  3. Render with real transparency

    Seven graphic types in the show's own look render as ProRes 4444 files with an alpha channel, in 16:9 for the episode and 9:16 for clips. Text that does not fit is scaled down or stopped with a note saying what to shorten. Nothing is clipped silently.

  4. Build the timeline

    The kit writes timeline files for Final Cut Pro, Premiere Pro and DaVinci Resolve with every graphic on its own track at the right frame.

  5. Hand it to any editor

    Timelines, graphics, the brand fonts and a README in one folder, so an editor can pick it up cold.

What changed

Episode one: 76 graphics, placed and timed.

Nobody retyped a word in the editor, because every word on screen came from the cue file. The same kit renders the vertical versions for clips from the same cues.

The next episode starts from a transcript and a cue list instead of a blank timeline.

graphics in episode one
76
graphic types
7
editing apps
3
formats, 16:9 and 9:16
2

Under the hood

How it’s put together

polymaths.media
  • Seven components: lower third, question, stat, note, kinetic phrase, list, chapter card
  • Frames rendered in headless Chromium, encoded to ProRes 4444 with ffmpeg
  • Timelines as FCPXML for Final Cut and FCP 7 XML for Premiere and Resolve
  • Speaker-matched transcript from each person's own mic track
  • Every render checked before it goes in the kit

More case studies