Final Cut: Voiceover, Subtitles, Editing, Export
💡 Goal: Follow this stop's roughly 38 minutes of lessons to turn the batch of shots you have into a publishable AI short drama final cut—voiceover, subtitles, editing, and export done in sequence, producing a video file that meets your target platform's specifications.
🎯 Stage Goals
- Assemble scattered shots into a complete short film: with voiceover, with subtitles, and with a workable pace.
- Produce a video file that meets the target platform's export specifications.
- Complete this course's final-cut checklist: voiceover → subtitles → editing → export → acceptance. This is exactly the last step of Runway's official long-video workflow (story conception → storyboard → character reference → generating multiple short clips → editing and assembling); the checklist itself is a course design, not a built-in feature of any software.
⏳ Estimated Time
- The five lessons total about 38 minutes (8 + 7 + 9 + 7 + 7); with a ±30% buffer, that's about 27–49 minutes
- This stage is mostly local machine work, with no rendering waits like the previous stage; but export transcoding time depends on your machine, and the first export will most likely need rework—don't schedule your time too tightly
- Short-on-time route: 5.1–5.4 is required—only after finishing it will you have a publishable final cut; 5.5 graduation exercise can wait until after the export is done, during the review
🚀 Before You Start
- Finished the previous stage shot generation (including 4.5 shot retry and replacement strategy): you have a batch of usable shot clips in hand
- Dialogue finalized: 2.3 dialogue trimming has been applied—the final lines after trimming, which the voiceover reads from
- Storyboard in hand: 2.4 storyboard is finalized—editing follows it to arrange shot order
- Assets archived per 3.5 asset naming and version management: versions won't overwrite each other during rework
- Decide on the publishing platform: export specs follow the platform (this course uses YouTube's official recommended values; for other platforms, refer to their respective creator documentation)
- Remember the boundaries: the practice film is still single-scene, few characters, 30–60 seconds; in the final-cut stage, don't add new shots or new characters—only integrate
- A cup of water ☕—the first export is rarely perfect, don't get frustrated
First time following this course? The overall route is in Quick Start: From Zero to Your First AI Short Drama.
⚠️ Three licensing reminders—read them now
- edge-tts is free to use, but it calls Microsoft Edge's online service, and the commercial terms are not stated in the project documentation—verify them yourself before official publishing.
- A model license is not the same as asset authorization: Wan2.2's Apache 2.0 only covers claims about the model itself and does not imply that voiceover voices, music, or assets can be used commercially.
- This course does not recommend specific music asset libraries; using real people's likenesses requires their authorization—all practice uses AI-generated characters.
📚 Lessons in This Stage
Complete the five lessons in order, and your scattered shots become an AI short drama final cut. Why this order:
- Voiceover first: the length of each line's voiceover determines how long each shot stays on screen—it's the time baseline for every later step
- Subtitles depend on the voiceover: only with an audio timeline can subtitles be aligned
- Editing comes after all assets are ready: assemble only after shots, voiceover, and subtitles are all in place, to reduce rework
- Export last: after all modifications are done, export once according to the target platform's specifications
| Lesson | Content | Duration |
|---|---|---|
| 5.1 Voiceover and Licensing | Generate line voiceovers with edge-tts (which can also produce SRT subtitles at the same time), or record them yourself; also confirm commercial licensing | 8 minutes |
| 5.2 Subtitle Production and Proofreading | Align lines to the voiceover timeline, proofread the SRT, and burn it into the picture | 7 minutes |
| 5.3 Editing and Volume Check | Use FFmpeg to assemble shots in storyboard order, unify durations and frame sizes, and check volume | 9 minutes |
| 5.4 Export and Pre-Publish Check | Export using YouTube's officially recommended encoding settings; for other platforms, refer to their respective creator documentation | 7 minutes |
| 5.5 AI labeling and micro-drama regulations | add an AI disclosure label to each episode, decide whether filing is needed, and add a one-line label with FFmpeg | 8 minutes |
| 5.6 Graduation exercise | review this film with the acceptance checklist and plan the next one | 7 minutes |
This course demonstrates editing and subtitle burning with FFmpeg; you can also switch to graphical tools like Jianying (CapCut), but their operations have not been verified in this section.
🎒 What You'll Gain
- A complete AI short drama final cut: 30–60 seconds, with voiceover, with subtitles, meeting the target platform's export specifications, ready to publish directly
- Practical usage of edge-tts: one command yields both the voiceover audio and an SRT subtitle file (free tool; verify commercial licensing yourself before publishing)
- A reusable set of FFmpeg commands: assemble shots, check volume, burn in subtitles, and export to platform specifications
- A final-cut acceptance checklist you've walked through once: voiceover → subtitles → editing → export → acceptance—use it to review your next film