Based on the success of my previous animation, I decided to render out a few stills at higher quality with some lighting tweaks and such. In the animation, it was difficult to control the exposure frame by frame so at times it wasn’t compensated for properly, so I was careful to adjust that in these renders. I also applied some post render HDR tweaking in Photoshop to make the caustic patterns really punchy without blowing out the object or backdrop.

After my successful attempt to rectify the earlier issues I was facing with perlin noise, I decided to do some more renders in different states of this model. I became interested in how the object would reflect and refract light and the caustic patterns it would produce, so I did a series of renders experimenting with this.

I tried variants of the intensity of caustic patterns and using different coloured lights to cast through the object to see how they would affect the render. For the most part, there is not a lot of clarity in what the mesh actually looks like (see last image), but I really like the caustic patterns that were generated on the surface and overall, I’d say this was a reasonably successful render.

The other day I got working on a capsule design in grasshopper. The basic idea was to create a series of struts between two spheres of different sizes (the logic is shown in the diagram below).

In order to put this together, we started with two spheres, and populated points across them. Then we used exoskeleton (a grasshopper plugin) in order to generate a network of connecting lines between the two spheres. I also worked in an algorithm which would generate a solid ring around the centre so that the two halves could be pulled apart and not sit as disparate elements. The network was then thickened to generate the geometry.

The third render is one done with very thin struts, the second render shows the wireframe breakdown of the model, and the main render is the thickened mesh with a psychedelic looking caustic pattern result, because why not?

I used several programs to create this overall shape. Firstly, the base geometry was created in rhino’s grasshopper, with the help of the exoskeleton and weaverbird plugins. I then took the model into 3ds max, where I generated the housing for a magnet on each side of the model. From there, I took the model into zbrush, because I’ve found that zbrush handles mesh booleans far better than any other program I’ve come across, especially with its dynamesh feature. Then I took it back into rhino in order to slice the mesh into two chunks and close the holes (unfortunately zbrush wasn’t able to do this part properly). Then it went into maya to fill some small holes in the mesh. From there, I brought it back into zbrush to retopologise it with zremesher, and finally brought it into 3ds Max for rendering.

I took the model I’d done a while ago, and finally got around to 3D printing it as I’d hoped for some time, here’s the result (and compared to the original render).

I was pretty stoked with how the model turned out, but I did change some parameters of the original model such as the thickness, bend, and size so that it could be printed out. As this model was done with a filament based printer, there was also a lot of support that had to be pulled out. If part of a model overhangs the perimeter, the software calculates the support required to produce it, hence there was a lot of support structure both in and outside the final result.

My next idea was to pull out regions from the models and to see what sort of spaces could be created through those volumes. Surprisingly they returned some of the most successful results I’ve had in a while especially in terms of generating some sort of architectural form. Although that said, these are still quite far from any rational form of architecture I am beginning to see potential in it (something which I’ve been struggling with for the last few weeks). But this method is producing some strikingly unique forms which is a plus.

One of the most difficult things I’m trying to reconcile, and a problem with the notion of parametricism in architecture is that parametric architecture has a tendency to reject context, the forms which it produces are unlike anything ever seen in the past hundred years or any time before that, so the tendency for parametric buildings is to attempt to establish their own context, and negate the pre-existing site altogether.

But for now, my concern is with the form, and not yet how they fit into the context, as is why all the renders that I’m doing at the moment (see below) are pure white geometric forms.

In light of some of my recent voxelisation attempts, I decided to render out two frames in 3ds at a higher quality than the others. Unfortunately I could not get the shadow parameters the way I wanted, so they still ended up producing a rather intense shadow on the ground plane.

I also had some issues with scale. As a rule of thumb, models that you’re rendering (or even just models in general) should be modeled to their actual size, this ensures that the lighting acts exactly the way the eye would expect it to and hence adds more believability. But with a model like this, I had no sense of scale for how big these cubes were, whether it be that they were ice cube sized, or maybe the cubes were the size of some form of landscaping item such as a bench. That was probably the most confusing part of this model for me.

This was another by-product of the voxelisation series I discovered while tinkering with the parameters I’d set in my grasshopper file. I really liked the way the cube decomposes through the animation, here are several of the keyframes that I like the best in this short. This model, (unlike the one prior) seems to hold its overall shape a little bit better, but I think that might simply be due to the voxelised nature of the model.

This post ties in with my previous voxelisation series and the next post. This is a short animation which I rendered out of the transformation the mesh goes through before it gets voxelised. What I find really strange but kind of interesting is the rapid change that happens just before the ten second mark, the moment where we completely lose the description of any sort of cube and it undergoes a massive change in form. Some of the mesh characteristics at this moment are also rather intriguing. The mesh for a brief moment has a very fine/refined edge to it, a sort of crispness. This is typically an effect I’ve come across in something like 3ds max where you add several swift loops light up against eachother, but this is done completely algorithmicly, and hence I’m not entirely sure how this detail developed – something to look into in future.