Add text-to-video production guide for creative team#170
Add text-to-video production guide for creative team#170alexwelcing wants to merge 1 commit intomainfrom
Conversation
Practical, step-by-step guide bridging the gap between script and final video. Covers the 6-stage pipeline (script → shot list → prompt → reference image → video clip → post-production), with worked examples from THE REACHING scenes, model selection decision tree, prompt engineering principles, era-specific modifiers, common mistakes, daily workflow, and a glossary for non-technical team members. https://claude.ai/code/session_01Gzq579HEKh2fQRbEPh1Es6
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 43fcfecfaf
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| weakness but from the effort of reaching, her hand extends toward a white wall, | ||
| she holds the position, completely still except for a slight tremor in the | ||
| fingers, warm golden winter light, silence, the camera does not move, thirty | ||
| seconds of absolute stillness after the hand reaches its full extension |
There was a problem hiding this comment.
This walkthrough instructs a single generated clip with thirty seconds of absolute stillness, which conflicts with the guide’s own earlier constraint that video models are best for 4–10 second clips and should be broken into short shots. In practice, asking for 30 seconds in one pass is likely to produce temporal drift/jitter and wasted retries for the team, so this step should either split the beat into multiple short generations or explicitly say the 30-second hold is assembled in post.
Useful? React with 👍 / 👎.
2 participants
Practical, step-by-step guide bridging the gap between script and final video. Covers the 6-stage pipeline (script → shot list → prompt → reference image → video clip → post-production), with worked examples from THE REACHING scenes, model selection decision tree, prompt engineering principles, era-specific modifiers, common mistakes, daily workflow, and a glossary for non-technical team members.
https://claude.ai/code/session_01Gzq579HEKh2fQRbEPh1Es6