Leonard & Edgar – Making of

© Celia Kaspar

3D Generalist Julien Kaspar shows us how he brought his sister's adorable
sketch to life

This is a breakdown of my workflow of making Leonard & Edgar in Blender and Photoshop. I'll show you the process from the initial concept art to the final image and some tips and advice along the way.

Step 01: Concept art

I usually like to model my characters on existing concept art or by doing a few sketches myself. I like having a rough or clear reference for what I am going to make and not go in blindly. There are also often some parts in the concept that wouldn't make sense in a 3D space, some off proportions and oddities in the perspective. Not everything from the concept art can or should be modeled exactly as it is. If done so the model can feel weird and/or flat when shown from different perspectives.

Original Concept Art of Leonard & Edgar by Celia Kaspar

© Celia Kaspar

Step 02: Base mesh

In Blender I started by sketching out the simple shapes as a base mesh and roughly translating the pose and silhouette of the concept into 3D. I used the Skin Modifier and Spheres and sculpted them with the Snake Hook, Draw and Inflate brushes. I prefer the Snake Hook over the usual Grab brush because of its falloff while moving the mesh. When I am happy with the result I move on to sculpting with Dynamic Topology.

Basemesh with rough pose and shape

Step 03: Dynamic Topology sculpt

I usually keep my Dyntopo settings at 3-5 pixels detail size and subdivide collapse. From there I sculpt the rough shapes with the Clay Strips and Crease brushes. With hair strands and similar parts I use the Snake Hook brush with the strength curve set sharp and adjusting the thickness with Inflate and Smooth. After it gets more detailed I switch the Dyntopo to Subdivide Edges so no sculpted detail accidentally gets destroyed. I also keep a lot of separate objects to sculpt them without affecting other nearby surfaces.

Dynamic Topology sculpt

Step 04: Retopology

I first merged some of the objects using the Boolean Modifier or using the Sculpt UI Addon. For the retopology, add-ons like F2 and LoopTools are very handy. I try to keep it very simple, subdivide the topology and Shrinkwrap it to the surface of the sculpt. Sometimes when I have little time I also use external software like ZBrush for quickly retopologizing the objects as a foundation and adjusting the topology further in Blender. Because the model doesn't have to be animated the topology doesn't have to be super clean.

Finished retopology with ZRemesher and adjustment in Blender

Step 05: High resolution sculpt

After the retopology is completed I subdivided the mesh one or two additional times and Shrinkwraped it to the original sculpted mesh to pick up most of the shapes and details. After that I smoothed out some stretched areas and sculpted further detail with the Multiresolution modifier. Because the models are mostly smooth and don't have much surface detail I didn't have to overkill the polycount.

Sculpted high resolution mesh

Step 06: Vertex painting

I like to paint with vertex colors at a relatively high subdivided mesh by applying the Multiresolution modifiers. That way I don't have to worry about texture size and UVs and can directly paint but it's important to keep an eye on the polycount and isolating the objects you paint because the viewport tends to get really slow at this point.

High Resolution Mesh with Vertex Colors

Step 07: Rendering

I added some area lights in a 3-point-lighting setup and gave the models a soft plastic like shader. Then I rendered multiple passes of the lights, Diffuse, Specular, Shadows and Ambient Occlusion and recreated the original rendered image in Photoshop. This gives me more freedom by tweaking and masking out some areas by adjusting the individual passes and thus slightly changing the look of the render.

Used passes for the composition (not including the AO)

Step 08: Composition

As soon as the passes were combined in Photoshop I changed the background color to something that fit the mood and color scheme, adjusted the brightness of certain areas like Edgar and Leonard's socks, gave the rim light more contrast and painted in some extra shadows. Then I made some more color corrections and painted over the test tubes to give them some glow, sparks and smoke. I try to keep painting over the models to a minimum so it doesn't end up looking too illustrated.

simple breakdown of the steps in compositing

Step 09: Final Image

After I'm satisfied with the result I export the final image and call it done. Thanks for reading and I hope you got something out of it. Have a nice day!

Final Image after Composition

Related links

Check out more of Julien's work here
Looking to learn something new? Check out our eBooks for lots of helpful tutorials
Follow 3dtotal on Facebook