Shiva

Making 3D feel painted

My work rests on one idea: digital work can feel physical and textured. That's why I started making my work look painterly.

Sep 27, 2026 · @Shiva Viswanathan

Read the original PDF

Every piece starts as a Blender render, built in 3D with standard textures and lighting. Then I post-process it with stroke-based rendering so it looks like it came off a canvas.

Painterly sunflowers at sunset

Five years of making something every day

I've made a piece of art every day for almost five years, and in that time I've tried every style I could think of.

Artwork and process illustration from the Paintify writeup, page 2, image 1
Artwork and process illustration from the Paintify writeup, page 2, image 2

I started with drawing. Drawing every day taught me the basics: design, lighting, shading, and recreating what I see with a pen. That still carries into everything I make. It also taught me to recreate things in a non-photoreal way. NPR, or non-photoreal rendering, is any rendering that isn't trying to look like a photo. I'll use that term a lot.

After about a year and a half of daily drawing, my friend Andri got me into 3D, mostly by accident. I started with Womp, which builds models out of goopy materials. It was an easy way to learn how 3D works, but it was still traditional 3D with traditional materials.

Artwork and process illustration from the Paintify writeup, page 3, image 1

Then I moved to Spline, which is built for 3D on the web. Spline lets you push post-processing filters hard: pixelation, chromatic aberration, outlines and other NPR effects. That's where I really started experimenting, and it got me closer to the painterly look I'm known for.

Then came Blender. I'd known about it for about 10 years and was too scared to learn it. When I finally did, I fell for its NPR tools right away: Grease Pencil, toon shading, dithering, halftones, anything that stops a model looking photoreal. I'm now past 1,000 hours in it.

Artwork and process illustration from the Paintify writeup, page 4, image 1
Artwork and process illustration from the Paintify writeup, page 4, image 2

Why I don't chase photoreal

I avoid photoreal because it's very hard. It takes a lot of detail, textures have to look weathered and worn, and even then it often doesn't work.

When beginners try photoreal 3D, you can tell right away that something is off. 3D has an uncanny valley, like the human face. Masahiro Mori coined the term for robots that look almost human, and CGI humans fall into the same trap.

Artwork and process illustration from the Paintify writeup, page 5, image 1
Artwork and process illustration from the Paintify writeup, page 5, image 2

Photoreal comes from stacking imperfections. Erosion, wear and scratches, everything beyond an object's base color, are what make it read as real. That takes time, and I make something every day. I can't make something photoreal every day.

A simpler style saves time, so I settled on painterly, simple and non-photoreal. The next question was how to get there.

Artwork and process illustration from the Paintify writeup, page 6, image 1

Ways to make 3D look painted

There are a few ways to make 3D look painted, and I tried to learn from all of them.

The first is to paint the object itself: take the model into another program and paint on it directly. That's how Arcane and Puss in Boots got their look, with artists painting the textures by hand.

The second is to change the normals. A normal map fakes surface detail like bumps, dents and scratches without adding geometry. Simplify that map from smooth gradients into flat splotches of color, light the object with it, and it reads as painted.

Artwork and process illustration from the Paintify writeup, page 7, image 1
Artwork and process illustration from the Paintify writeup, page 8, image 1
Both take time, so I looked at algorithms. Kuwahara is an old image filter that flattens color into splotches.
Artwork and process illustration from the Paintify writeup, page 9, image 1

There are also tools built for this. The Tradigitals Live Paint Filter add-on for Blender and Fotosketcher both lay physical-looking strokes over an image, which was exactly what I wanted.

Artwork and process illustration from the Paintify writeup, page 10, image 1
Artwork and process illustration from the Paintify writeup, page 10, image 2
So I had my references. I could have just used these tools, but each one had a problem:

The Tradigitals Live Paint Filter is paid, runs on dense Blender node trees, and I couldn't follow its math. Fotosketcher is closed source, slow, and wasn't built to output video. Kuwahara flattens color but has no texture, so it never looks like strokes. Painting the object or the normals by hand is far too slow for something I make every day.

None of them fit, so I built my own.

Building on Hertzmann

I set four rules for my own method:

  • Open source
  • Good for video
  • Easily accessible
  • Automated batch export
Artwork and process illustration from the Paintify writeup, page 11, image 1
Artwork and process illustration from the Paintify writeup, page 11, image 2

I started from Aaron Hertzmann's stroke-based rendering papers from 1997 and 1998. Hertzmann did his PhD in computer science at NYU under Ken Perlin, pioneered stroke-based rendering, and is now a researcher at Adobe. He's worked in computer graphics and vision for about 30 years, and his papers are public.

Reading his work, I saw how close it was to the Impressionists. They layered brush strokes until the viewer got an impression of the subject. Keep layering and you end up with a full painting.

His research became the backbone of the project. It's all code, with no neural networks or AI tools, and nobody charges me to paint brush strokes. I started with simple flat strokes layered on top of each other and adjusted the opacity until the look felt right.

Paintify

The first version was a web app called Painterly Image Creator Web. It's now Paintify, the first in my series of painting tools.

The workflow is simple. I build and animate a scene in Blender and render it as an image sequence.

Then I load the batch into Paintify and run it. The controls are brush radii from coarse to fine, stroke length, curvature, error threshold, grid factor and opacity.

The painted frames go back into Blender's video editor with music, and that's the finished piece.

The goal of Paintify is that the input can be anything and the painterly look sits on top. You control what goes in, and the tool handles the style.

What sets it apart from other painterly software:

  • It's built on Hertzmann's research, which spans 30 years.
  • It uses salience: it finds the objects in the frame and the boundaries between colors.
  • It knows the output is a video. Each frame builds on the last instead of being repainted from scratch, which stops most of the flicker.
  • It has almost no dependencies, with C++ for the GPU version and JavaScript for the web.

And I made it for me, which is the part I care about most.

Paintify / across tools

The Paintify family

These studies come from Paintify-GPU. Its companion projects bring the painterly process to Blender, TouchDesigner and OBS.

01 / Paintify-GPU motion study
Painterly yellow flowers made with Paintify-GPU
02 / Painterly detail
A blue and violet abstract form processed with Paintify-GPU
03 / Another input, same process

Why this matters to me

My reasons are close to what happened in the 1800s. Painters were the people you went to when you wanted the real world recorded. There were no photographs, just people who were very good at making things look real.

Then the camera arrived, and painters no longer had to record reality. So they went abstract and looked for meaning, color and vision beyond what was in front of them. I think that's beautiful, and it's close to what I'm after with this work.

What's next

Paintify now has versions for the GPU, Blender, TouchDesigner and OBS. I'm still working on the desktop version and seeing how far it can go. You can try the web app at paintify.shivav.space or look through the repos below.

Heavily inspired by Maxime Heckel.

blog.maximeheckel.com/posts/on-crafting-painterly-shaders

Follow along on Instagram at @daysofshiva for the daily art and @shivaviswan to talk, or on Twitter at @shivaviswan.