CrePal Introduces AI Storyboard Workflow for Visual Continuity

CrePal has introduced a new workflow that utilizes the Qwen-Image-2.1 AI model to create storyboards and improve visual continuity. The aim is to identify and correct potential visual inconsistencies before the video production process begins, thereby reducing costly errors and rework.

The new approach focuses on providing clear visual instructions to the AI. According to CrePal, a common problem with storyboards is that individual frames may look good, but the continuity between them fails. This can manifest as changes in character, product, or logo as the image sequence progresses. Qwen-Image-2.1 allows the use of up to ten reference images, and CrePal advises carefully selecting the cleanest and clearest images for each element, such as character, product, and environment.

The workflow also includes the removal of conflicting or unlicensed materials. Outdated logos, obsolete packaging designs, or style images borrowed from different campaigns should be eliminated to prevent the AI from generating erroneous elements. Each reference image used should be assigned a clear task; for example, a character image should define the character's appearance, and a product image should detail the product.

CrePal emphasizes that while Qwen-Image-2.1 supports the use of multiple reference images, it is crucial to give each reference a single, clear visual job. This facilitates corrections when a need for change arises. Finally, visual continuity is checked across the frames, paying attention to elements like character faces, product shape, and text. Approved storyboard frames are handed over to video production with clear identifiers and notes.

The model does not eliminate all issues; moving images can reveal even more subtle inconsistencies. Additionally, CrePal reminds users of the limitations of the Qwen Research License, which permits the use of supplied materials for non-commercial purposes only, unless a separate commercial license is obtained.