"Dead City" is one of the fourteen locations which I'm preparing for my music clip. I don't want to say too much about the story, because some things can change (there's still a lot of work to do) and - what is more important - I might kill the punchline. Anyway, In this scene I wanted to show abandoned city - dirty, dusty, sun-baked - one can say: without any signs of life, but simultaneously carring hope that life (or life forms) can exist in most adverse conditions, most unexpected place and it resurrect very quickly.

Most often the workflow looks like this: sketch, modeling, texturing, lightning, rendering, post-production. But I found out long time ago that I'm not a painter and I'll never be able to reach level of realism that nature presents. I could make 100 layers in Photoshop with dirt, rust, cracks,finally I added a few for almost every texture but nature will be better in that area. So, that's why I decided to make real photos first and than - based on prepared textures - start to modeling. I borrow digital camera (canon 400d), took my own, old, dusty analog camera (Revueflex SD1) and I started to looking for interesting places to take photos. With help Of my sister I took about 1200 digital photos and about 360 analog photos. Rusty cars, ruins, broken roofs and walls, old trains - I photographed everything that can be usefull in making the clip.

There were 2 problems with photos. The first one: When people saw a man who is running across the streets and taking photos of every building they started to think, that I'm a thief or some kind of cryminalist - so, not everytime I could take photo at right angle and position - because of that there was a problem with correct perspective on photos. The second: There was a lot of unwanted elements on pictures - like antennas, cars, flowers, cables, etc. I run photoshop and started to retouching. Below you can find example of work with one of the buildings.

When textures were ready I started to model buildings. Huge advantage of this style of work is that you have all dimensions, distances, etc. right on the picture. You don't need to worry if "that window" is 160 or 180 cm height, or is "this tier" have 250, 293, or 310 cm... All models (except plants) were made in XSI.

Streetcar - like the one in my scene - has move across my born town (they weren't so dirty and rusty of course). I thought that placing this object in my scene will increse mystery mood. Left on a crossroads, with open doors, overgrown by grass and shrubs... Strange. People have to abandoned this place really quickly... When I was creating this streetcar I was using traditional workflow: first I founded reference photos, than I modeled it, textured and placed it in the scene. Textures for streetcar were made without any "photo base". I mixed and painted several layers of dirt and rust and I added ambient occlusion layer (baked before).

Making vegetation was the hardest part of the work. Not because I did't know how to do it, but because I want to keep rendering times as low as possible and - simultaneously - freedom of camera movement - that's why I didn't want to use billboards. I'm just not sure what the final camera movement will be. Finally I used X-FROG for trees and shrubs, "Ivy generator" for Ivys and ingenious plug-in "DPIT Nature Spirit" for grass and dry leaves. DPIT is a plug-in which gives possibility to "paint" (clone) plants and other objects on surfaces (by mouse or pen - which I used for dry leaves) or use bitmaps for this task (which I used for grass). Textures for plants were taken from XFROG.

When I started to make renderings with GI (some time ago), I was trying to find everything about it in the internet. And I found several articles about linear workflow. If you are interested in details, please look at the internet - there's a lot of materials on this topic. In the brief, the whole thing is about gamma. Monitor displays pictures with gamma 2,2 (1,8 for MAC), but the render engine works in linear scale (gamma=1,0). All you have to do to work in LWF is that you should give your renderer pictures and colors in linear scale. You can do that by adding gamma correction with parameter 0,455 (this means 1/2,2 - for PC). There's a lot of advantges when using LWF: light goes deeper in the scene, rendering is natural and physically correct, you don't have to worry abour color mapping. Of course you don't have to use LWF (you can use color mapping, desaturation, contrast, etc.), but for me it is the best and the easiest way to make correct renderings.

Below - quick comparison:
In every rendering I used the same materials and render settings. No lights (one of the sphere is illuminating).


on the left - linear workflow (Color mapping - linear multiply, Dark mult.: 1, Bright mult.: 1, Gamma: 2,2)
in the middle - no linear workflow (Color mapping - lHSV Exponential, Dark mult.: 3, Bright mult.: 3, Gamma: 1) (too dark)
on the right - no linear workflow (Color mapping - llinear multiply, Dark mult.: 1, Bright mult.: 1, Gamma: 2,2) (light but colors are washed out)

I really don't know how this shot (it will last about 10 or maybe 20 seconds) will look between previous and next one, that's why I decided to make very natural, well-lighted rendering - without any mood. The reason is that I want to have a lot of posibilities when editing this final image. I can brighten it up or darken it down, I can make this picture to look greenish or leave it as it is - in brown shade. In that scene I used only one light - physical sun and physical sky. (For rendering I used Cinema4D and V-Ray).

As I said before I want this scene to look dirty, dusty, sun-baked. I had to use color correction. I give to this picture "caffe-cocoa" shade, I made the light and dark glow also, I "changed" the sky and... that's all.

Thank you for reading this making-of. I hope you like it. Good luck in CG world.