Nov AI Video Generation Tips

Fix the character first before making the video: use multi-angle reference images and a six-grid expression sheet to lock facial shape and expression boundaries, preventing facial features from collapsing when the character makes big expressions.

To avoid the cookie-cutter influencer look, you can reverse-convert an anime character into a realistic person, borrowing the facial proportions of the original artwork to create a distinctive look.

Write storyboard prompts shot by shot: first describe character locking and positioning relationships, then for each shot specify duration, shot size, camera movement, expressions, and dialogue.

For fight shots, you can use a fixed formula: camera position plus camera movement plus action (no more than two), then add effects. Each shot carries less information, so the model is more likely to follow it.

For MV-style works, edit first and then break down the shots. Create first and last frames for each shot and check them carefully, then apply unified color grading. Segments with problems can be reworked on their own.

Generate only 3 to 5 storyboard videos at a time. Do a rough edit first to check whether the mood and conflict land, then go back and revise the prompts. This can save a lot on generation costs.

Short films don't need heaps of effects: repeat the same action with one character design, using consistent costumes and composition as visual anchors, which also sidesteps the weakness of complex interactions in long takes.

Comic dramas and short dramas can follow an assembly line: first produce character and scene reference images, then arrange storyboards, generate clips, and add voiceovers, and finally export a CapCut draft for manual fine-tuning.