Compressing a Movie to 1MB and Reconstructing It With Wan 2.2
Willsolo · reddit · 2026-07-19
The author compressed a movie into less than 1MB of text and then reconstructed it into a video with sound using Wan 2.2.
Process
- Sliced the video into about 2,000 shots using PySceneDetect
- Used Gemini Flash-Lite to generate a structured description of around 100 words based on 8 frames extracted from each shot
- The text was then compressed to roughly 320KB using xz
- Shots were sequentially fed into a self-hosted Wan 2.2 TI2V-5B on a RunPod A6000 to generate visuals
- Audio was generated using MMAudio (sound effects) + MusicGen (soundtrack), with dialogue restored at original timestamps using ElevenLabs TTS
Main Challenges
- Character consistency was the hardest part: since each shot was generated independently, the author clustered character descriptions and injected them into the prompts
- For shots exceeding 5 seconds, a "previous frame to next frame" chained generation was used
Cost & Resources
- Estimated cost is about $30 per movie
- Provided more detailed examples, cost comparisons, write-up links, and a GitHub repository
The author is also asking for feedback.
More from coding & agent
- BELL: Open-Source Stack Combines Browserbase, Eve, iMessage and Stripe Link for Personal Agents — jeff_weinstein · 2026-10-27
- Legora runs 2,000 parallel agent evals on Modal Sandboxes for legal AI — blelbach · 2026-10-27
- Two engineers rebuilt a FedEx supplier's billing core in 3.5 months vs an estimated 7-8 — alex_verem · 2026-10-27
- GPT-6.1 Terra scores 48 on AA Coding Agent Index at one-third the cost of GPT-5.6 Terra — ArtificialAnlys · 2026-10-27
- Shift-1 router cuts coding agent spend up to 55% while keeping SWE-bench accuracy — ypatil125 · 2026-10-27
- Non-coder vibe-codes a Godot JRPG with Claude Code in two weeks, music and art synced — ADHDavid · 2026-10-27