CAGD 373 - Game Asset Production

Final Game Scene

December 13, 2020

Our scene turned out beautiful.

Overall, this project went very, very well. It was good that I had the chance to work purely on textures without also being the primary modeler. I love modeling, but it was good for me to focus my efforts and work as part of a team. I could definitely see myself being comfortable in this role for an extended period of time.

Our scene really came together in this last week. Paul transitioned our project to Unity's HD render pipeline, and it felt like that motivated us to make it as pretty as possible. On the left is what our scene looked like on Wednesday, and on the right is how it ended up.

Here are all of the models next to their textures. Most of them were modeled by Tyler, while Paul modeled the vases and goblet. I touched up every model to some degree, and re-arranged and occasionally completely remade the UV layout. The chest, statues, and building exterior ended up needing some significant revisions compared to the original models. I modeled the terrain, rocks, and interior of the capital.

Texturing went smoothly for the most part. I have had a lot of practice in the past few months, and I can usually think of solutions to achieve specific goals in Substance Painter fairly quickly. Often times while working on textures, I felt like I was following in the footsteps of Destiny 2's texture artists.

This prop, for example. When I looked at the image from the game, I immediately thought about using a planar projection to overlay a series of circles, then use that as a mask. As I put this together in Substance Painter, it felt like I was doing exactly what the original artist did.

The trickiest part for me was the interior. There were a lot of very large walls that needed generic textures, but not a lot of texture space left. My solution was to texture only some of it, then tile it. I also set up my UVs so the texel density decreases as it gets higher up and further away from the player. I was also very careful with the interior lighting so that it brings attention to the assets that were meant to be highlighted.

My biggest problem has been personal health, probably. I am able to spend way too much time on modeling, texturing, and everything I am doing in other classes. But I promise I am totally not a vampire. I love garlic. I guess I am just used to spending every waking hour working on project. But the end result turned out almost perfectly in my mind, so it must have been worth it! Video evidence below.

 

 

 

 ----------------------------------------------------

Final Game Scene - Week 7

December 9, 2020

Our scene made a lot of progress in a short period of time this past week.We now have solid ground beneath of feet, literally!

 

Rocks! I modeled these islands to hold all of our beautiful models. These were fairly quick to sculpt in blender, and texturing went smoothly as well. I modeled the big one specifically to fit under the capital building, and I plan on going back and adjusting the other islands to fit as well.

 

The scene! Once I had the dreamer capital model in a decent state, and the textures sketched in, I imported it into unity. Then I rearranged our scene a bit to be more interesting. Previously, there was a bridge between the first two islands, and all of the islands were closer together. I staggered the rocks out and raised the first platform to add verticality.

 

The bridge! I gave Tyler some suggestions and requests for adjusting the bridge model. Once he made the changes, it already looked miles ahead of where it was before. The textures transferred over to the new model pretty well, and then I added extra cracks and details.

 

Final Game Scene - Week 6

December 2, 2020

This week, I took the time to combine some of our models into the same texture sheet. Sometimes I like to texture things individually even when they're going to be part of the same texture sheet. Maybe it just creates more work, but maybe I'm just like that

Lastly, I textured this wall. It doesn't stick exactly to the reference material, but that's fine. This texture went a lot more quickly than most of the others. Really, it was just a matter of using masks to make the same kinds of patterns the other props are using.

 

So, I have a lot of work left to do right now. But that's fine.

 

 

Final Game Scene - Week 5

November 17, 2020

This week, I worked on this chest, and what a chest it was! Tyler originally modeled the base for this, and gave it to me to finish.

Also, I textured this lamp model that Paul made. On the left is the reference material, on the right is the new one.

The crystal material I made in Substance Designer was very helpful here! I also rigged this chest so that I can animate it in Blender, before importing into Unity. I spent almost the whole day working on this.

At this point, I have accepted that I am going to be very busy over thanksgiving break. I still need to draw up a plan for the interior of the building, and I will update this blog when I do. Below, you can see some of our most finished assets in the scene.

 

Final Game Scene - Week 4

November 9, 2020

It's time for SUBSTANCE DESIGNER.

Tyler made these new modular bridge models, and I got to texturing them quickly. Pretty soon, I wanted to make some custom materials that I could reuse. So I started up a new project in Designer and started on this crystal material.

I am probably going to do some work on the model too, maybe even make a high poly version. That's all for now!

 

--------------------------------------------------------------------------------

Final Game Scene - Week 3

November 4, 2020 

Whoa, this has been quite a week. In general, that is. Not so much for this project, but we made some important progress on our scene's overall composition.

 

 

I am glad to report that we are now all on the same page. Previously, there was a bit of confusion among our team about the arrangement of our scene and the player's path through it. I was under the impression that the final location for the player was going to be the interior of the capital building.

 

As it turns out, an interior was not in our plan. It is now, and if I have anything to with it you can bet it's going to have purple. We are not necessarily sticking to the interior of that building from the game. We're keeping our options open, and will likely pull pieces from other interiors and smash them together.

 

 

 

For texturing, this was a bit of a slow week on my end. Our modelers made some progress, so I'll have some fun working through them. After I get the textures started in Painter, at least getting them blocked out in some way, I'm going to get started on some world textures in Designer.

Stay tuned for...

FLOORS

 

 

Final Game Scene - Week 2

October 28, 2020

 More texturing this week, and we made some progress on our overall scene plan. Work slowed down a little bit, but we are still ahead.

Mainly, I finished texturing this pylon, once Tyler finished modeling it

Paul modeled this vase, and I got started on texturing it. I made a high poly version and a more intricate ID map for texturing

 

Final Game Scene - Week 1

October 21, 2020

 

For this project, we chose to make the Dreaming City from Destiny 2. Our team lead, Nathan, has played a lot of this game. He asked us what kind of environment we preferred to make, and I said anything with purple. We found a place with lots of purple.

First thing we did was gather a TON of reference material. Nathan ran through the map taking hundreds of screenshots, while Tyler, Paul and I watched on a live stream. I keep this reference sheet floating on my screen while texturing each of the props.

My assigned role is the texture artist. I love making textures and models, and I find it incredibly relaxing to just make textures. Tyler quickly started churning out models and I got to texturing them just as quickly. This was before our blockout and map were created, so our focus so far has been on props rather than architectural pieces.

We started our unity project, but have not done very much on it so far. Nathan put together a basic blockout and drew up an annotated map.

 

Asset list:  

 

Towers:  The Problem with Polys

 October 12, 2020

 

 

My very first impulse was to make fire towers. Most others went for a sci-fi or fantasy theme, but somehow this is where I ended up. I loved Firewatch, and I love the kind of solitude that fire towers represent. The plan was to make fire towers for some sort of tower defense game centered around... bears, I guess? Or global warming. One of the two. Maybe fire AND bears. Probably that.

 

 

 Here you can see the reference images I gathered. The plan from the start was to make one support structure, and three different rooms to go on top. I ended up using the designs from the bottom two and top left images, but with some changes for style and to make them conform to the tower structure.

 

Reference images

 

My biggest struggle throughout this project was with the polycount. It is so tempting for me to make stylized and efficient low poly models, so it was hard to want to get get my tower close to 5,000 tris. At first I wanted to use a flat card with holes cut out in the alpha channel for all the railings. Eventually I found that it was worth it to model them, but I kept the cards for the LOD model. With some clever UV work and baking, I was able to use the same space on the texture sheet for the 3d and flat versions of the rails.

In the end, the towers were still under the limit by over 1,000. If I put more time into it, I could have added more detail. I would like to have added furniture to the interiors of the towers, but whoops, that never happened.

 

 

 

 

These are my texture maps. Each of them has an alpha channel, which you can see in the parts that show up totally black (except for the tree trunk, which just turned out that dark after baked lighting)

 

These are the cute lil' props that I spent most of the the final day of this project working on. These were a whole lot of fun to make because I really enjoy finding clever ways to use limited numbers of polygons and small texture sizes. The limit for each prop was 500 tris, so I was VERY happy when making these.

 

 

 I think they came out quite nice in the end. I was pressed for time when modeling the tree and rocks, and struggled to come up with ideas for what to make. I am happy with how they all turned out, but I would have liked to go back and refine the tree textures some more. The rocks were actually a last minute addition. It was late enough at night that I briefly considered leaving out the fourth prop entirely. I'm glad I didn't.

 

Overall, I am content with how this turned out. I could have done some things differently, but I stand by my work.

 

 

 

 

Sci-Fi Hallway: Disasters & Cookies

September 27, 2020

 

Making this hallway was a bit of a wild experience. I learned some things, had a lot of fun, and probably poured too many hours into it. When we started the assignment, I sort of went on autopilot and started building this... before reading the brief carefully. You should really read the brief, always. In the end, I am quite proud of the final result, even if it may only match a slightly generous interpretation of the specifications.

 

I spent most of my summer modeling and texturing, and it has become second nature to me at this point. I have done so much baking (of mesh maps). So, so much baking. I was so ready for this.
 

I poked around on pinterest for some reference images. I have been collecting reference material for a Half-Life mod I am working on, and it fit the theme of this assignment perfectly. From the start, my plan has been to rework the meshes and textures and port everything back to the Half-Life engine, after finishing the assignment. The goal was to do all of this before submitting, that way I could record a second walkthrough of the other version for fun, but, um, I had higher priorities to worry about.

 

 

First thing I did was open up Blender, not Maya. In fact, I did not use Maya at all until the end. I took my initial reference image, which was rather low resolution, and strategized how to break up the pieces. I got the basics of the design from this image, then took it in my own direction as I went along.

 

My first versions were basic and repetitive. I made a small section of the hallway with UVs and color IDs, then slapped an array modifier on it so it would tile into the distance. I rely heavily on color ID maps in texturing, which allows me to get started on texturing early, and continue to freely work on my UV layouts. 

The idea was to settle on a basic and consistent design convention, get started on the materials, and then move on to make things interesting once I felt satisfied with the foundation.

 

I completely forgot that the scene needed to have doors until more than halfway through the project. I was in a discord call with my friend Nathan, talking about his scene. He mentioned doors, which reminded me that I should probably read the directions for this assignment. So you can thank Nathan for this scene having any kind of door.

 

 
At some point, I decided that my scene had to be a disaster, much like its author. Something was going to go wrong. There had to be red lighting, alarm sounds, and generally an anxiety-producing situation. Being underwater with pipes bursting, evidence of flooding, and locked doors seemed stressful enough. A sequence I would love to put the player through as a level designer. 

 

I found myself casually listening to alarm sounds on a loop while working on this

 

Modeling went very, very smoothly. I made extensive use of Blender's mirror and array modifiers, which enabled me to work semi-procedurally. I used vertex colors for my color IDs, instead of creating separate materials. This was the way I originally learned to create color IDs, and I think it spared me from some of the confusion others have had.

Modular sections with color IDs

Final arrangement in unity

The biggest challenge I encountered had to do with lighting. I had a clear idea of what I wanted for the lighting. There were going to be three different light styles: standard, airlock, and emergency. Area lights seemed to be the best fit, since they gave me the shadows I wanted.

It worked beautifully in Blender, but in Unity, area lights require baked lightmaps. For all the lamps using area lights, I also included a soft, realtime point light as a fill light.

Blender test renders

That part was not much of a challenge, it just meant that I had to spend extra time baking my lightmaps.

The difficult part was when I got to my flickering lights. I wrote a simple script to create a light flickering effect, and I am quite satisfied with how it turned out. I used this on the broken emergency light, but my reliance on area lights came back to bite me. I toyed around with various arrangements before settling on a dim area light, paired with a realtime point light that had the flickering script attached.

The broken light at the end of the corridor was where things got more interesting. Area lights were obviously out of the question, because the light needed to rotate AND flicker. Point lights were not going to give me the effect I wanted, so I had to settle for a spot light. I played around with it, and was a little bit satisfied with how it was looking. I wanted to give the light a mask texture so that it would be shaped like a square, and then I discovered cookies, which let me do just that.

this cookie is for lighting, not for eating

Playable build: LINK