Segment a broadcast master with AI
Identify probable programme breaks in a master file and build a reviewable segment report with durations, timecodes and representative frames.
Editors can segment a broadcast master with AI before preparing an edit decision list or delivery notes. The program-master skill looks for combined black frames and silence, keeps track of fades and produces local evidence for a human to inspect.
A dark closing shot is not necessarily a programme break. Black frames can begin after a fade has already carried meaningful picture information, and an audio bed may continue across a transition. Looking at the candidate boundaries alongside thumbnails helps an editor retain the intended cut.
What it does
- Find likely separators that meet the configured picture and sound criteria.
- Keep source timecode distinct from time elapsed since the file began.
- Record fade boundaries and flag questionable transitions.
- Choose representative thumbnails and label segments for a report.
- Check the first and last segment against the entire master’s duration.
Install and use
- Install the skill, Python, uv, FFmpeg and ffprobe. Some OCR profiles also require Tesseract. Use an authorised local copy of the master.
- Ask “Prepare a segment report with source timecodes, durations and thumbnails. Mark the breaks I should check before editing.”
- Review boundary frames for critical edits and confirm the final segment reaches the actual end of the programme.
FAQ
Source
Made by timelapsetech. Original skill: program-master, published on GitHub under the MIT licence.
View on GitHub