Modelling Theory
1. Subdivision Surface Modelling:
Subdivision modelling is a modelling technique for making high resolution models, by manipulating a lower-resolution “cage” (The cage is the low density model mesh that we modeled which becomes a control cage that can be edited when using subdivisions) model and using software to subdivide for a smoother surface.
Subdividing increases the number of vertices on the model, making curves more rounded. Usually each subdivision level increases the vertex count 4x the original count. A “high-poly hard-surface” means creating a mechanical/constructed models, which are used to bake into “Normal Maps” (A normal map is commonly used to fake high-resolution details on a low-resolution model) and other types of textures, which are in turn used on lower-resolution game-friendly models.
(Note: the red, green and blue channels of the image are used to control the direction of each pixel’s normal)
There are plenty of ways to subdivide models. “3Ds Max” uses mostly the “Meshsmooth” or “Turbosmooth modifiers”, while Maya uses mostly the smooth Mesh Preview (3 Key) or Subdiv Proxy
Although Maya’s “Subdivision surface” system is a specialized case which allows different subdivision levels in different areas; this is different from most other modeling tools which divide the whole model instead.
2. Low Poly Topology:
Topology is the layout of a model, how the vertices and edges are placed to create the mesh surface. A good topology is essential if you want fast framerates (realtime) and good deformation (both realtime and pre-rendered). For realtime rendering, bad topology can also create rendering problems.
Vertices and edges need to be in certain places to give good bending/compressing/stretching when the model is skinned or morphed. Topology is important where the model has to deform the most: crotch/hips/butt, shoulders/armpits, mouth corners/cheeks, knees, elbows, hands/fingers. There are multiple types of topology focusing on different parts of the human form (or any form generally similar).
Body Topology:
Face Topology:
Limb Topology:
Shoulder Topology:
In-game models have to be optimized in order to have great silhouettes, they have to be defined “Edge-loops” (A series of connected edges on the surface of the object that runs completely around an object and end up at the starting point) for better deformation. “Vertex Normals” (are invisible lines pointing out of a 3D model’s surface, at each vertex. The game compares the normals against the lights direction) have to minimized for better shading, and allow for good “UV seams” (are pairs of numbers stored in the vectors of a mesh, the numbers are used to stretch a 2D texture onto a 3D mesh).
Edge loops:
Vertex Normals:
UV seams:
3. Texture Baking:
Texture baking (or Map Baking) is the process of transferring details from one model to another. The baking tool starts a certain distance out from the model (Usually a low-resolution model for game use), and casts rays inwards towards another model (A high-resolution model sculpt). When a ray intersects the 2nd model, it records the surface detail and saves that into a texture map, using the first model’s texture coordinates. There are many characteristics that can be baked, from isolated attributes (ambient occlusion, normals, vertex colours, etc) to multiple, combined characteristics including materials,textures, and lighting all baked into a single texture. Here are examples:
4. Mood Board:
1 . Panasonic TV
This Panasonic T.V is a good design choice as it has detail ranging from small buttons to curves and bevels across its edges. the material variation is limited to mostly to plastic surrounding the T.V itself, there is the glass screen that displays itself at the face of the T.V although limited to the dull grey colour screen only. However despite all this the T.V is a good design choice overall because of its shape and detail.
2. Painting Easel
.
The painting easel is a long simple wooden design with many hinges and straight lines. The material is predominantly wooden but with metal hinges giving the design more variety. Overall the design isn’t all that interesting, although the little specks of detail like the scratches on the wood and paint splodges covered in certain parts give the design a unique feeling.
3. Lamp
The lamp is a complex design with multiple parts and hinges connecting it together, the overall detail is great as there are points of interest like the bulb and the stand. The overall material is metal with some wires and cables connected. The lamp is a good design as it showcases good detail and an interesting design which gives for a good challenge as it varies between looking simple but being complex in design overall.
4. Speakers
The speaker has a cylindrical shape with knobs and cables around the speakers, the knobs have some small details surrounding it but overall simple detail as most the detail is circular so its not much of a challenge to accomplish. The asset is not heavily detailed but it makes it up by having a unique shape, I think the design is interesting but does not have enough a lot of detail to work on.
5. Remote
As was the “Painting Easel” the remote also has unique marks that add to its detail, thus making the design more interesting. The material consist of plastic and rubber buttons with a small bit of glass for the infrared light. The asset is fairly detailed with buttons of various shapes and sizes, with also the shape of the remote being curved and smoothed, this makes the overall design look manageable to model as it has good detail and shape.
6. Fire Extinguisher
The fire extinguisher has many details to its base shape and handle, there is a lot of edges and hinges that have its own detail on them. The material is overall metal with other small materials like plastic. the asset is very interesting as the design has so many kinds of shapes that make for a good model, the detail is also heavily embedded with the shape itself. The overall design is great but could prove difficult over the process of designing.
Final Model Choice:
I’ve decided that my final choice is the “Remote”. I feel like the design is not to complicated or to easy, as the design is fairly small but very detailed. The remote has some unique markings and scratches around the design making it feel more natural and real. I need to break down certain key parts of the remote to greater improve the final outcome for example, the markings are in certain spots and not in order, the buttons are in different sizes and there is a small glass bulb for the infrared light, the face of the remote has a round ditch where the middle buttons are placed and finally the back is curved which could prove for a challenge. Overall I believe i can accomplish this design fairly simply if i work on areas that could prove difficult.
Process
Evaluation
High-Poly Evaluation:
When designing my high poly model of the remote, I came across many problems with geometry and positioning, for example beveling would create gaps of geometry that I had to find by routing out the vertice or edge causing the problem. This occurred quite often during the process however with learning I managed to understand where certain areas of the model should be beveled and others not. This was a great challenge for me as I had to use new tools like “Chamfer”, “Beveling”, “Circlize” and “Boolems + Difference” that made making the model 10x easier for me instead of manually handcrafting certain edges or buttons, for example the light bulb on the top was a great issue that I overcame by using the chamfer tool that created a 4 way split into vertices then combined with circlize i was able to make a circle on the top of the model.
Although I am immensely proud of myself for overcoming these obstacles and making a high poly model, there are some things I wished to improve. For example I wanted to change how the back of the model looked slightly, on the real remote there is a curved casing on the remote whereas on my model the casing is rectangular and less curved overall. There was also a problem with the middle buttons as they were slightly spaced away from each other than they are here on the model.
Overall I am happy with the outcome of my high poly model.
Low-Poly Evaluation:
The low poly process was fairly simple and quick as all I did was trace the high poly model using the x-ray option, however initially was making an entirely separate model which was not accurate to the high poly model. There isn’t much detail to the low poly model as I had to sacrifice more detail for time. That is one of the regrets I had when making the low poly is that I added some detail to the buttons. Although the baking process made it so that final look wasn’t entirely ruined due to the limited amount of detail the low poly had. There were only 84 tries on the low poly even though the maximum amount was 10000 tries. UVing was also quite simple as the cuts and unfolds were not very complicated due to the lack of detail on it.
Overall I am happy that I didn’t have to waste time on the low poly and was able to move on to UVing and baking process with ease.
Baking Evaluation:
The baking process went better than I expected with the high poly model baking quite well onto the low poly even with the lack of detail on the low poly model. One nitpick I will have is that I wanted to make the buttons 3D due to time left for this project I made a quicker and safer alternative that resulted in a pretty successful bake. There are a few areas that could have work of improvements like the top area where the light bulb is and the sides as well, but they became quite unnoticeable once textured.
Overall the baking went better than I expected but could have some improvement.
Overall Evaluation:
Overall I am quite pleased with my model, there are many areas that could have work for improvement if I had spent more time on them instead of using shortcuts due to time. However, I am pleased with attempting to create a real-life item into a 3D modeled and fully textured. This is personally an achievement for me as I learned through this making process how to use so many tools and shortcuts to create a rendered model. I am glad about the final outcome of my model.
Kimathi Mugambi Cov Games 