Making Of 'All Roads Lead to Rome'
Previous to the creation of this image I did a number of renders for a gallery exhibition (Adaption Series). For this series I produced thirteen inkjet prints in total and introduced a low-poly, realistic, astronaut character with no textures whatsoever. This was a starting point for making a completely different image. I obviously wanted to do something new with a character/role I liked, so I reused two elements: a helmet and a backpack and started to model around them (Fig.01 - 02).
This new image was supposed to be the complete opposite of the series I've mentioned; a cartoon image relying on a comical idea with minimal scenography, and something I could produce in two-three days time. The helmet was slightly remodeled by further reducing the polycount. After that I modeled the body with no reference drawings. I knew what I wanted - short legs, long arms and a stronger upper body - so I had no need for a detailed sketch. Also I knew that the astronaut's face wouldn't be visible behind the glass of the helmet so I skipped modeling any unnecessary details (Fig.03 - 04).
Idea Behind the Image
I took a well known phrase - "All roads lead to Rome" - and placed it into a context where it was unlikely to work. For this I modeled one comet with a road across its surface, roadside signs pointing towards Rome and an astronaut strolling across the galaxy.
It was a simple idea and so I wanted to make a fairly simple scene that I could texture, rig and render quickly. Therefore the focus was never on high-end realistic production possibilities, but on a stylized image that could deliver the punch-line in one fast look-over. Inspiration behind the style can be found when browsing through vast quantities of gaming images on sites like 3DTotal, Game Artisans, CG Society etc. I won't be mentioning individual influences because there are so many exceptional artists; one just has to continue clicking.
Modeling (Comet Example)
After the astronaut I started to model the comet and the "Rome" sign. All the objects are low poly quad models that were detailed with textures and additional sculpting in ZBrush. Before sculpting, the comet was previously modeled in 3ds Max (Fig.05), textured in UV Layout and imported as an .obj file into ZBrush.
Mostly Clay and Mallet Fast brushes were used when sculpting (Fig.06). Exported maps included Tangent Normal map, Cavity map and Displacement map. All textures were finished in Photoshop and later placed upon the Oren-Nayar-Blinn shader (Fig.07 - 08).
I decided to add the space background in the composit so my render preparations were pretty simple. First I had to do a basic rig for the character (Fig.09) and pose him. I did the same with the scenery - all the models were linked and I could then position the signs as I liked. Props (such as the backpack for the character) were skin-wrapped to the parent model.
Now I could position the elements, introduce the lights and set up Indirect Illumination. A total of three mr Area Omni lights were used, each with a different light color (light yellow, blue and red) while just one light cast shadows. Final Gather was enabled with mr Photographic Exposure Control defined. The image was scaled to meet the format for most online CG portals.
The final pose is something that took a while to get down. I tried to avoid a classic contra-pose and opened the character to the camera. The second thing that was important was not to over-emphasize the importance of the character in regards to the context. The comet deserved the same attention (Fig.10 - 11). When the pose was done, I set the composition out of balance (Fig.12) and I could now start with the rendering.
The scene was divided into several output channels: Diffuse (with and without lighting), AO, Lighting pass, Reflection, Self Illumination, Specular, Shadow and Z-Depth, with a couple of matte passes that were used to mask specific parts assigned to a specific channel. The background, sun rays and comet mist were painted later and grain was added in the finish. Fig.13 and Fig.14 show the compositing stack adjusted for the final image (Fig.15).
Production of this image taught me a lot. First of all, if you organize your workflow and know how to save time you won't waste additional minutes modeling something you can redefine from earlier projects (for example). Secondly, you need to have your priorities straightened out before you go into production. When you have limited time to produce something don't hang on to a value that can't be executed in the given time frame, such as realistic textures or organic modeling, hang on to your idea and make it visible through the image you display. Have fun with it! To conclude I will quote one of my favorite artists who said one of my favorite sentences: "Banal ideas cannot be rescued by beautiful execution".
For additional images please visit the original WIP thread: