Perspektivische Stadtbilder

  For the mission description screens in Pizza Connection 2 we had to create 3D images of each metropolis. Initially we had planned to play a movie sequence that should show a flight over each town, but as we realized what amount of 3d geometry had to be rendered we decided to leave it at a single high resolution image.

Importing layer data

  One of the major goals was to create a nearly perfect duplicate of the town, as it shows up in the game. This included not only the already rendered ground layer, but also all sprites like houses, traffic signs, memorials, parks, trees and so on. For this purpose we needed to import data from the Pizza Connection 2 level editor into Maya. Since all graphic elements that show up in the town already existed as 3d models, we only had to import these objects, duplicate them and calculate their position and rotation. Again Dirk Bialluch programmed a MEL script to deal with this task.

Rendering image layers

As already mentioned the ground layer pass took about 800MB RAM during rendering. Now we had to deal with scene containing some hundred houses, plans and other objects that had to be added to that image. Therefore we needed to separate these objects into different layer. To combine the image layers later on in compositing we created a z and alpha channel. In addition we created a separate shadow pass.

Z depth channel converted to grayscale image. Using z depth made the compositing task a lot easier, because we did not have to rely exclusively on alpha masks.

The render layers introduced with Maya 3.0 came in handy here. We started off creating the house layer. This was the largest scene and had still to be quartered to fit into memory.

		Quarter images of the house layer

Next we created the object layer, including park benches, traffic signs, memorials, street lights and so on.

Object layer

After that we rendered the the Paint Effects pass, generating all plants.

Paint Effects layer

Finally we created a smoke layer using Maya's software particle. Three particle emitter were placed over randomly chosen chimneys.

Smoke layer

Compositing

Compositing was done in Maya Fusion. First we combined the different color layer using z depth information of each image to compose them properly.

Composition of all color layers

The same had to be done for the shadow layer.

Next we combined the color image with the shadow layer.

To get rid of that typical cold CG look, we finally put the image through a rather complex grading and filter process. Based on the z chennel we added fog and depth of field to simulate the atmosphere. Colors were shifted into a yellowish look to get a warm mood.