Plan for Arland "Barry" Woodham

My Senior Project Car Based Deathmatch Now With Cell Shading

I should call this plan a successful failure and if you read through you will find out why. I will explain the failures and my processes first. So if you don't want to read it skip down to the success section.


Level Design Woes

Well this has been a big miserable failure so far. I was hoping to do an update no later than two weeks after the last .plan I posted. However, such is the way of life. My original idea for my level involved curving tunnels that lead from one level of the arena to the next. In the time since the last .plan I have tried so many methods to get this to work that I have lost count. This includes many QuArK attempts involving brush subtractions and path extrusions. Also I tried using DeleD without much success. The main problem there was the exporter does not like curved surfaces and tries to break them up in a manner that leaves lots of gaps in the level. Something like this I could have built in about a day and a half in UnrealEd but, instead I am no farther now on the level several weeks later than when I started. However in all this time I have seen where QuArK's strengths are and how it can be a good editor under its intimidating interface.


So Close and Still So Far

My best attempt came when I started examining the path duplicator however It does not like the nodes being moved in more than two dimensions at any one time. For an explanation of what I mean by this you can examine this post I made. www.garagegames.com/mg/forums/result.thread.php?qt=22410Realizing that the duplicator wouldn't quite work right I attempted applying woodworking theory to my level design. What I mean by this is I would take a section of the tunnel (1) and overlap it with another one following a guide (2) then I would take those pieces and cut them to fit. (3) In reality creating a digital form of mitering. I continued that around the guide creating my loop on a gradient (4). At this point it became obvious that there were problems with where the sections would join together. So, I created column and overhang pieces to hide some of these problems (5) and then finished both sides with a path between them (6). Basically this got me close to what I want with little line up problems all over the place (7+8). Unfortunately tweaking in QuArK does not work properly and if you move a vertex it likes to move the whole edge even if the other vertex is frozen. This left me with no level and no real hope to finish this one (9). I decided to start a new futuristic arena with lots of angles, something I know QuArK can handle, but, it became obvious that if I want to finish this project for the interactive show I would have to have something I could work on much faster.


Images explained above used small images because of number of them used in this posting so details may be hard to make out.
1 2


3 4


5 6


7 8


9


Yay Success ...... Well Sort Of

In class I talked with my professor and she suggested that I go with a more stylized design rather than the realism I was striving for before. The way I interpreted this is something along the cartoon design look. I knew I was not going to get any work done in class because I was going to change the whole look of this project so I decided to help one of my classmates with his coding project. I was able to help him without much difficulty and asked him how his project for this class was going. He has been trying to create a Tim Burton / German Expressionistic Unreal Level. What that means for the most part is distorted shapes with outlines. He had been using a environment cube mapping trick to try to achieve this cell shading which worked fine on some objects but, others were far from perfect. I had read an article somewhere about using inverted normals to create an outline effect before and thought we could give it a shot. Long story short we made it work and the effect was better than I imagined it might be.

So right now you are probably asking what does this mean for my project. In wanting to create a more stylized look I had already thought about cell shading but decided it would be too much work to implement. I was wrong though after seeing how easy it was in Unreal I figured I could do it the same in Torque. After some muddling around with a combination of Wings3D for modeling and then Milkshape3D for texturing and exporting I succeeded. This is what made me even want to post anything in my .plan. The fact that something is working right for me. With a cell shading look in the game I can go for much more simplified shapes and textures meaning I can get the project done in less time. In turn I should be ready for the interactive show so here's hoping. Anyways I will end this rather long .plan with my victory shots of working cell shading. Also maybe after this quarter is over I could put a tutorial together on how I achieved it. That's it for now time to get back to work.


Wireframe of actual object


Cell shading show off shots