Steam Roller Hand In
Overview
This was a prop brief in which I had to create a realistic steam roller from provided reference.
There was a 3000 triangle limit and the a maximum texture size of 2048 x 2048
Design
The brief was made to be similar to industry briefs in that the design of the steamroller must be changed slightly to avoid any infringement on the original design.
I planned to change the arrangement of the tertiary shapes, also I planned to remove some decals and add some grills. I tried to look for reference in other heavy machinery to keep the design realistic.
UV Layout
3D Viewer
To make sure I had a high texel density within the 2048 limit a lot of the UV's are double up, for example the roller cylinder itself is split in half and layered on top. That example I feel works well but some of my other choices (like the steps) for reusing UV space may have been better textured uniquely as you can see the repetition.
Textures
PBR using albedo, roughness, metalness, normals, and AO
Beauty Shot
Additional Images
Low poly wireframe
High poly used for baking
Conclusion
When starting out on this project I tried to follow the workflow of my previous high poly bin project that I mentioned in my blogs, I believe this led me to some pitfalls. While I was able to create a nice high poly model, it was not wholly as useful as it should have been.
In future I would definitely be more wary making sure that I only make high poly elements of models that actually need it, although it would be nice to bake an entire model, it is very time consuming and in this case had to be sacrificed. I would also be sure to exaggerate the chamfers of my high poly models so they are more visible after baking, and stay visible through multiple levels of mips, there is no point having smooth edges if they disappear immediately after moving away from an object in game.
I am however very pleased with how the bake came out on the tire, that seems to be the one part of the high poly that was especially worth spending the time on.
I think I had some lessons to learn with reusing UV space as well, there is a fine balance to be had with trying to achieve a high texel density and making sure that the repetition is not noticeable, that's if you are going to overlay UV islands at all.
Given more time I would have liked to pack my UV layout a little better, that part of the process I feel was a little rushed so I had time to actually texture my model, an easy sacrifice to make at the time, but with proper time management, I wouldn't have had to sacrifice it at all.
In conclusion I feel that this project was a great learning experience, I got to try out a whole heap of new software which I'm sure will be very useful in the coming rest of the year, and try out a workflow I had been dabbling in for a while from start to finish. When looking in detail at this project there is a lot I would like to improve, however taking a step back I am satisfied with the model as a whole.
