Here's a preview images from the last few days since PK stopped posting them here.
The attempt to replace rendering programs with AI to improve graphical quality has caused the work to regress significantly. Renderers like Daz3D, Blender, or Maya are designed to provide precise control over every aspect of a scene, from lighting and textures to poses and facial expressions. These programs offer robust tools that allow for detailed adjustments, producing highly realistic or stylized results depending on the artist’s vision.
On the other hand, using AI to replace renderers is still an emerging field, and in most cases, these tools lack the same level of precision. AI models focused on image generation, such as Stable Diffusion or MidJourney, are excellent for creating images from textual descriptions or enhancing low-resolution images. However, they lack the interactivity and detailed control necessary for projects like comics. Current AI solutions still struggle with maintaining character consistency across different poses and angles, which is critical for visual storytelling.
Replacing a traditional renderer with poorly integrated AI results in work that feels more like a collection of disconnected scenes than a cohesive production. AI can be a powerful complementary tool, for instance, to enhance textures or add details, but it is not yet ready to replace an entire rendering pipeline. In practice, this leads to a decline in the quality of the final work, especially for projects that rely on visual and narrative continuity.
Rather than adopting rushed solutions, the ideal approach would be to integrate both technologies. Using the renderer to create the foundation of the work and AI to refine elements, such as lighting or fine details, could result in an efficient process with visually impressive results. The issue with the current approach lies in attempting to replace the creative process with a solution that lacks the necessary technical maturity, ultimately compromising both the quality and the artistic identity of the work. Furthermore, when I make edits using AI on works based on PigKing's material, I always incorporate references from renowned artists to enhance the quality. This results in a more refined and artistically aligned outcome. The problem is that PigKing doesn’t understand anything about art, making it hard to expect anything beyond the bare minimum. In fact, even the works I create turn out better than what they’re currently presenting.
