CST 325 - Graphics Programming (Week 7 and 8)

 Shadow Mapping & Final

Concepts

• Path tracing: realistic shadows appear naturally as a consequence of the algorithm. Hard shadows have path tracing straightforward. 

• Two main approaches: volume – based, image – based. 

• Volume based: used by many from 2000 to 2010. It extrudes polygons away from light source. Since its polygon based, it has problem with geometries that have lots of mesh. 

• Image based: also known as shadow mapping since 2006, and it is pixel-based. The algorithm involves comparing the distance from the light perspective to create a shadow map and store in the texture. 

• Shadow mapping: Computing the shadow mapping ranges from -1 to 1 and redistribute it to 0 to 1 range for color. Use the light as a camera to create a projection matrix and get clip space coordinated. Divide by W to get coordinates and distance Z to do the comparison.  

• Problems: aliasing (blocky shadows) and surface acne that looks like the shadow goes in and out repeatedly. Another problem is byproduct of reprojection, which can be solved by adding bias offset or increase shadow map’s resolution. However, adding too much bias can result in peter panning (the shadow drifts away).

Application

The final project was surely overwhelming but also fun, allowing room for me to be creative. At the beginning, I was struggled with given code from gltf-viewer since it is different from the json geometry I used to work with. I remained calm and worked one by one, starting with the skybox. In order to use the json files and the geometries correctly, I had to play around and observe the objects’ behaviors. Of course, the problem started when involving the gltf-viewer, causing many errors. Even though Professor updated the file and solved the problem, I felt proud for finding a temporary resolution to load the models with gltf-viewer.

The problem piled up one after another when I kept working on the project. In the end, I managed to solve them, and there are two most notable moments. One is when I had to read through gltf-viewer, the models’ scene file in order to load the texture and emissive correctly. To my surprise, I managed to add or change some lines of code to achieve my desire (with explanation from ChatGPT as well). Previously, when facing with overwhelm files with lots of code like these, I tend to leave them be and use what I need (some functions). But now I gain more confidence with reading, understanding and manipulating the files to do what I desire.

Another moment is when I was working with creating the algorithm for randomizing the missile and fighters. Complicated matrices, variables for forward direction didn’t work as expect, utilizing the lookAt function, and many other obstacles. I could just remove those features completely since they are extra and not required for the project, but I decided not to. After many trials, many approaches and countless debug, I managed to find the recipe to randomize the objects, their directions and make them move. If I have extra time, I could turn the whole fleet into a battle (only if the website can run it), because the projectiles can share the algorithm of lookAt.

Last but not least was the UI elements. Again, I could call my final project done and move on. But I chose not to and decided to spice up the UI, since this is the chance for creativity that I don’t want to miss. Previously, I learned from Software Design course how to add UI through html file. But the question was how can I make it more interesting with different shapes, colors and features. So, I searched around, found many documents and UI website. However, there were much new knowledge in those codes. I could simply just copy the codes but doing so limit my creation as well. Therefore, I spent time understanding what those codes do and applied them to my own UI elements. The results were not just satisfied, but I felt proud to overcome the obstacles and made them my own.

After all the resources I poured into the final project, it’s hard to tell how proud I am now. It was a great time working on the project, I can tell because I was willing to spend more time and effort. Moreover, I wanted to add more features if there was more time to work with. I also shared my project’s video with my friends on Discord and Facebook many times as well. I think they are enough to prove how much this final project means to me and how proud I feel about it.

Visual Documentation