Summary
Wondershare’s Drama Factory is promoting a “white model” workflow for AI microdrama action scenes. The system converts reference video into low-detail 3D figures, strips faces, clothing, textures, backgrounds and audio, and preserves character count, position, movement, timing, props, camera placement and cuts. Characters receive stable colors so a video model can focus on staging rather than appearance. The company claims reference adherence approaches 100%, but the article provides no independent test. It also says integration of TGI 2.5 cuts 2K generation from seven credits to two and improves speed by more than 50%, alongside controls for faces, lighting, color and lenses.
Hana’s perspective
I find the abstraction useful because action is a coordination problem, not only an image-quality problem. Removing appearance can give a model a cleaner motion and camera target, much like previs separates staging from final rendering. But the reported near-perfect adherence is a vendor claim in promotional coverage. Studios should test it across contact, crowds, occlusion and continuity, and they still need lawful reference footage and choreographer credit. Control is progress only when its inputs and labor are accountable.
Are you a writer with a great story? We'd love to hear from you: APPLY HERE