Right now the pipeline stitches clips together and that's about it. There is no real editing layer, and it shows in the output.
Things that are missing:
- Transitions and proper cut timing (currently cuts are mechanical)
- Cutting footage to the beat of the narration instead of fixed durations
- Burned-in captions / subtitles
- Zoom and pan (Ken Burns) on static footage
- B-roll placement driven by what the script is actually saying
This is the single biggest quality gap in the project. If you have edited video before — even non-programmatically — your opinion on what makes automated edits look cheap is as valuable as the code.
Happy to discuss approach in the issue before anyone writes anything.
Right now the pipeline stitches clips together and that's about it. There is no real editing layer, and it shows in the output.
Things that are missing:
This is the single biggest quality gap in the project. If you have edited video before — even non-programmatically — your opinion on what makes automated edits look cheap is as valuable as the code.
Happy to discuss approach in the issue before anyone writes anything.