p@int_sm3ar: Viscous Oil Paint Motion LoRA

This LoRA is designed to animate the physics of thick oil paint, heavy impasto textures, and fluid pigment mixing. It was trained on high-quality macro footage of oil and acrylic mediums to capture the way light reflects off wet, grooved surfaces and the specific "drag" of a brushstroke.

🎨 Model Details

  • Trigger Keyword: p@int_sm3ar
  • Effect: Generates realistic, high-viscosity paint smears, palette knife textures, and flowing artistic pigments.
  • Base Model: Wan-2.2 (Image-to-Video)
  • Training Focus: Motion dynamics, light reflectivity on wet surfaces, and physical texture.

πŸš€ Usage Instructions

To get the best results, combine the trigger word with descriptive and detailed prompts including artistic terms like impasto, high-gloss, pigment, and brushstroke. Samples are generating using an LLM to prompt (recommended). Images used for sample videos are generated using Z-image.

Prompt Example:

p@int_sm3ar, a macro video of vibrant sunset-colored oil paints merging on a dark canvas, heavy impasto ridges, high-shine wet texture, static camera, 4k.

p@int_sm3ar a top-down shot featuring a dark, rectangular object with gold lettering on a black sand-textured surface. Deep teal, glossy fluid streams emerge from behind and beside the object, moving in parallel vertical lines. The liquid is thick and opaque, with a high-sheen finish that contrasts with the matte, grainy texture of the surrounding environment.

p@int_sm3ar, The video opens on a richly textured impasto oil painting depicting a woman's profile, eyes gently closed, her right hand raised near her face. Her features are sculpted with thick, visible strokes of brown and subtle pinks, while a vibrant red headscarf, a solid mass of crimson paint, frames her face. Behind her, a radiant halo of thick yellow and orange paint radiates outwards. Abruptly, the entire canvas comes alive: the dense impasto strokes of the woman's form, the red headscarf, and the glowing background begin to subtly shimmer, then actively swirl and deconstruct. The thick paint globs detach from their original forms, blending and dragging across the canvas. The vibrant reds, warm yellows, and deep browns from the initial image churn into a dynamic vortex of color. From this swirling mass, new, sharp-edged strokes of vibrant yellow and orange paint rapidly coalesce and push forward, meticulously forming the illuminated top and front surfaces of a large, three-dimensional block letter 'B'. Simultaneously, deep purple paint strokes swiftly animate into existence, defining the shadowed side planes of the 'B', giving it depth and volume. The background darkens and solidifies with broad, textured strokes of deep black paint, consuming the remnants of the previous scene. Below the evolving 'B', smaller, distinct strokes of orange and yellow paint rapidly form, spelling out "BANOdoCO" in a blocky, textured style. The animation concludes as the canvas settles into a static, highly textured oil painting of the bold geometric 'B' and accompanying text, rendered in striking impasto against a dark, textured background.

🎞️ Sample Gallery

πŸ“½οΈ Dataset Breakdown

The training data consists of 10+ high-bitrate macro videos featuring:

  • Surface Interactions: Primed canvas, matte black backgrounds, and stone textures.
  • Camera Motion: A mix of static macro shots, zooms, and tracking shots that follow the flow of the paint.
  • Lighting: Studio cool lighting, dappled natural light, and dramatic high-contrast side-lighting.

Created as a specialized tool for digital artists and motion designers looking to bridge the gap between static oil paintings and fluid animation.

P.S.> Readme generated using Gemini. Get in touch on discord for any support - (https://discord.com/users/1138907661602136245)

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support

Model tree for Ashmotv/paint_smear

Adapter
(177)
this model