Wednesday, May 22, 2013

Post-Apocalyptic breakdown

  • Crafting the idea
  • Reference
  • Lighting/Color correction
  • Detail shots

Crafting the idea
I had wanted to do a post-apocalyptic environment for a very long time, somehow I felt very overwhelmed and couldn't really choose which direction I should take it. I got very inspired to do this thematic environment because of Unreal Tournament 3, Gears of War series and lately a game I'm hyped about is The Last of Us. I always had a warm heart for “the beauty of destruction” but this time I really wanted to step up to the plate and just do it.
As I am writing this I must admit that I started to notice how much I have learned throughout the last couple of months, not just with texturing but also modeling, organization, shader creation and optimization.
During this breakdown it’s obvious to me that I could have used a more efficient workflow and I figured how I could improve this when I start a new project. Also I’ve become more daring to show off my work in progress, when I started the environment I wasn’t sure at all if it was a good idea or if fellow artists would like it at all.

I must say that all this knowledge doesn’t come out of nowhere, I’ve read whitepapers from other developers, I read books about lighting and composition.

Goals: Large post-apocalyptic environment, create detailed textures, lots of high poly asset sculpting, shader experimentation (environmental shader and screen shaders e.g. vignette effect, lens dirt), use DX11 features like Tessellation and IBR.

Timespan: I have been working on this project on and off for the last 6 months, in total I estimate I would have worked on it for 2-3 months. And it’s still on going with additional asset creation.

To do: Vehicles (Modeling/texturing a Taxi and Bus), re-adjust/improve additional buildings, perhaps a fly through video?

During the process I looked up a lot of reference just to establish the main idea or props that I felt would fit the scene. I’ve heard it isn’t always a good idea too just focus on videogame concept art but when I got inspired by The Last of Us it was hard stay away from this style.
When I started to work on the project I didn’t had in mind in which direction I wanted the environment to push. Later on you’ll see my starting point, where I just had been messing around with temporary meshes and started crafting a broken wall in UDK. From there I started looking for possible ideas and came with this post-apocalyptic city.

With all the reference in mind I really wanted to feel like the place had been abandoned for a long time so dust would start to build upon the objects. My idea was to give the polygonal faces (within the shader) that are pointing upwards the dust-blending effect. At the time I didn’t have that much experience how I could apply this in UDK (and figure it might be too heavy for all the asstes) and decided to skip it for this project. Later on I did add the idea to the wooden floors shader where you can use vertex painting to add dust or moss buildup.

During development I started posting some of my progress online and I contacted Bryan Cavett, he is an amazing 3D/2D artist at Nerve Software. Bryan gave me some helpful tips and did a very nice paint over of my initial concept art sketch. The main thing I had to fix was the scaling; I had not used a model as height reference, which was a big mistake because it was very hard to estimate the height of the ceiling for example, compared to the other assets in the screenshot. After this tip I started using it as much as possible for all personal and professional testing.

I must say I regret using some DX11 like tessellation it seemed interesting and easily to use but it turned out to be very time consuming while the end result isn’t always worth the time investment. Thankfully Image Based Reflection did work perfectly in UDK and didn’t cost me that much extra time.

I am a level designer but I during my first internship I had a small project where I did some level design and I’ve read a few things about it. I did to lead the player’s eye, what using the directional light and bright spots to point the direction.
This isn’t only a composition rule but is also a helpful guide. In this case the player would follow the path towards the large building in the distance, not only because it stands out because of the height, also all edges of the surrounding buildings point towards that direction. Besides that, I did use the concrete chunks around the concrete pile as a guide, I thought the brightness of the chunks would trigger the player to move in that direction as well.

Shaders – color shifting
Re-using texture space was crucial and I wanted to take this up a notch by experimenting with color “hacking” as some people call it. At first I made the bottle red colored but the color values are too intense so I figure I should go towards a more subtle color maybe white or beige. At this point color shifting was now possible, to choose from green to blue, you name it. By experimenting some more with the texture, I noticed that tinting a yellow (and white colored) texture is the best color to work with. And I think that having an extra material is more efficient then loading in a unique texture.

I also used this technique on the plastic crates that are scattered throughout the scene.

The first pass with grass modeling was terrible, normally I made them from photo reference but once I had decided to rework these assets, I choose to make them high poly first and afterwards do some touchups to enhance the high poly baked textures. I really enjoy making these with the fairly new technique for me, I also got way better alpha channels at the same time. I also modeled out the tropical plants which also was pretty amazing to experiment with.

Here is also a small production video where I show off the steps I went through to get this composition/result.

Lighting and color correction
Doing color correction was interesting to mess around with. As you can see every screenshot has a different color tint, so it gives a different interpretation of the atmosphere.

Future goals

As I stated earlier in this document, I learned a ton to improve myself and my workflow. Things I’d like to improve for the next time are:
  • ·         Smart and effective asset structure 
  • ·         Better defined block out level 
  • ·         Explore more artistic ways with compositions 
  • ·         Create (more) paintovers to explore ideas

Detail shots

No comments:

Post a Comment