Sierpinski's Calm

Proof Of Concept I

Proof of Concept II

Oily

Final Presentation

For these renders I began by building the pyramid. I did this by creating a single square based equilaterally-sided pyramid. I then arrayed this in 2 dimensions using an offset of the length of a side of the square base. This created a 3x3 square of pyramids. Four of these were then instanced above and aligned so that the center of this layer aligned with the center of the other. A single one of these was then instanced above and aligned so each vertex of its base touched a point of the pyramids in the layer below.

This whole unit of 14 pyramids was then grouped, so that it could be treated as a single (pyramidal) object.

This entire process was repeated 3 times (only twice for the first 2 images), treating each constituate final group as a pyramid.

The result was 14^3=2744 separate base pyramids, i.e 2744x5=13720 faces/vertices.

I set up a simple mental ray lighting system, and also calculated the global illumination pass too. The results (due to the "spongey" nature of the model) appear to have a macroscopic sub surface scattering (SSS) effect.

Bubble Men (The "Knellers")

Conceptual Particle Bridge/Toon Shaders

Bridge Construction - Standard Render

Bridge Construction - Toon Shader

Mental Ray - Ambient Occlusion - Rear View

Quick Particle Test:

Pen....ROCKET - Continuation of Pen Modelling

Extrusion by Polygon

Subsurface scattering material used for flames.

Omni light and glass material used for flames.

Extrusion And Scale

Pen - Hard Edge Polygon Modelled, MR Rendered

Initial Model - Proof of Concept - Modelled with control edges and turbosmooth.

Initial Presentation

Two Pens - Environment Added

Depth of Field Added -further modelling done.

New Environment Test

Colour Corrected Environment Map

 New Layout/Composition

 Environment Added

Perspective Left View - Lens Correction

New Composition

Animation Reel

Kit Elliott - Animation Showreel from Kit Elliott on Vimeo.

Walkway

In a previous post of mine I had a sketchup-exported model of an odd building with a walkway. Digging out some old renders I came across these:

 Looks as if I used a mental ray sunlight haze-driven system paired with a fog effect. In conjunction with these I seem to have used some custom MR materials. New material coming soon.

Sky Models

   3DS Max presents us with 3 different sky models to use: CIE, Haze, and Perez. Note that all the images below have undergone levels correction as a post process.
   At render time, the fastest was CIE, then Haze and Finally Perez. Apparently CIE is designed to be fast and accurate, and Perez is designed to be used in conjunction with real weather data files. It appears to me as if the haze driven model is designed to look accurate, without actually being so.

CIE System

Haze Driven

Perez Model

Solidworks Surface Modeling

 Having gotten used to the Solidworks workflow I decided to move on to something a bit more challenging - surface modelling. For the subject I chose to model a car. I began by sketching out the profile, and then by sketching guides to extrude this profile along (using the boundary surface modifier):

This resulted in a mesh (having 3 dimensions but a thickness of 0). The blue and lines coming out of the mesh, "combs", show the curvature at that particular point. During manufacture this would correspond to the amount of bend that point would have to undergo to achieve the required shape:

 Having brought this mesh into photoview for a test render, I realised that the scale of the mesh was incorrect. To fix this I added a scale modifer to the stack. From there on I added thickness to the surface and modelled a cockpit and adjusted some of the colours. The image below is a screen shot of the finished shell part, and therefore will not be identical in colour or geometry to the render.

This was opened up in photoview for the following render:

But what is a car without wheels? I started a new assembly with the car shell and a new wheel part, which was built seperately:

The wheel was locked in place by using perpendicular, coincident and concentric mating pairs. I only added wheels on the back to begin with. The following renders were calculated in photoview:

Front 3/4

 Right

Rear 3/4

Rear Tyre, Depth of Field

 This is where the problems began. I wanted to incorporate front tyres too, but the current ones looked too small (on both the front and the back). The front of the car also dropped lower than the base of the rear tyre, meaning that I would have to enlarge the rear tyre to sustain the profile. However to do this I would need to alter the initial sketch profile of the surface cage. When I tried this I got many rebuild errors due tot he scale before adding the other features. So in the end I decided to bring the wheel out from the body of the car, resulting in a caricatured/remote controlled look. The front tyre was a copied version of the rear tyre with the scale changed. The wheels were then rigged to the axles, and then an assembly belt was added. This meant that if and wheel was rotated (i.e if it were moving along the ground), the opposite tyre would rotate at the same rate and the other tyre pair would rotate at the correct speed (faster or slower depending on the diameter ratios). The final assembly and belt can be seen here:

Opened in photoview for the following renders:

 Satin Finish Front

Gloss Finish Front

3/4 Black

3/4 Final

More Attempts in Solidworks

In this image I attempted to replicate a windows flag. The flag was constructed using the following steps:

• An equation-driven spline (using y=sin(x) between 0 and 2 pi) was sketched on the front plane.
• This curve was offset, and then the ends were joined.
• This face was then extruded by a distancce of 2 pi [mm].
• Two rectangles were sketched on the top plane, and an extruded cut feature with a through-all distance parameter was applied to this sketch plane.
• This resulted in the part being split into 4 bodies, which I then apllied textures to and rendered in Photoview 360.
  

Here I was experimenting with recording and using macros. I recorded a macro to offset the boundary of the selected face inwards 10mm, and then to extrude cut this new sketch inwards the same amount. The object is meaningless, but in creating was useful practice for becoming a more efficient Solidworks user.

A Quick Word On Post-Processing

Having realised that almost any renderer will churn out a washed out image, I've decided to post-process my renders from now on.
This is done by importing the .jpg file in photoshop (although any program with a histogram will do), the black threshold is then defined in the levels dialogue box either by using the colour droppper or moving the black slider to the right. This is then repeated and adjusted with the white point. This has the effect of squeezing all colours available in to between the two markers, contrasting the image. The curves and gamma are then adjusted to meet my requirements. Here is an example of what this method can be used to do:

Early Days in Solidworks

For these images I build the mesh in Solidworks and rendered in Photoview 360.
This model was not made to represent any object in particular, but came out resembling a mix between a tyre hub, an ash tray, and a cylindrical bearing axle. It was my first attempt in Solidworks, and so it uses most of the basic sketching and featuring tools. To name a few (roughly in order used): circle, polygon, extruded boss, fillet, chamfer, shell, extruded cut, dome, circular array.
The model is not fully dimensioned, as I was just getting used to the interface and malleable workflow of Solidworks. A screenshot showing one of the sketch dimensions is posted above.

Here I used a checkered studio material, and applied Depth-of-field to the PhotoView 360 Camera. The scene is a 3 light set up with a soft HDRI image for an environment. I feel the DOF may be too strong.





Here was my first go at adding materials, a flecked car paint. Environment was altered, and gamma slightly adjusted.




Here the environment was again altered, and I changed the focal length to enable a more realistic DOF in a later render. A bloom of 95% threshold and 4% radius was also applied in post processing.





Here is my final image. The material is shown to be the correct scaling on the mesh and the differences between the filleting and chamfering can be seen on the top faces and edges of the hexagonal extrusion. Diffuse reflections are also well shown. Bloom was added for a more "showcase" image. The render size is 1600x900.