Just wrapped up a trial using DALL-E 3 for generating consistent storyboard frames for a short film project. I had high hopes, especially with its "improved" coherence.
The results? Honestly, pretty disappointing. Each prompt for a character or scene, even with detailed descriptions, produced wildly different styles and interpretations. The character never looked the same twice. Felt like starting from zero every time.
For quick, standalone concept art, it’s fun. But for any kind of sequential visual storytelling where consistency is key, it’s currently unusable. The outputs are just too generic and unpredictable.
Has anyone else tried this for sequential art? Found any clever prompt engineering tricks to lock in a style or character across multiple images? I’m curious if it’s just me.
Another tool to try!