By Akrid 9 Comments
I took a bit of a break this week, or at least I didn't pursue with my usual vigor. I cycled through 3 projects, gathering a ton of reference, only to realize that the subjects were far beyond my comprehension, which was a depressing realization. I put a decent amount of time into modelling a Luger P08, but my method was convoluted from the start. I was using far too complicated schematics for reference so I was modelling the frame of the gun outwards, Putting too much effort into a perfect mechanical re-creation as opposed to modelling what matters. Plus, that particular gun is one of the most complicated and unusual, not a great thing to start on. I don't think I'll be coming back to this one.
Since I spent a lot of time re-rendering some old stuff, I guess I'll talk about some of that.
This is one of my earliest projects. I'm really pleased how this one came out. I think I achieved my goal of making a vibrant, cartoonish image. All the textures are 100% procedural, meaning all are generated using parameters within the program. No images - meaning pngs, jpegs etc. - where used as textures. The procedurals here are pretty simplistic, but they've definitely gone a long way in my understanding how and when to use procedurals. Working on this also really helped me glean an understanding of color theory. It helped me to begin to understand how to handle numerous contrasting elements, a bigger hurdle then you may think.
The scene is lit by image-based lighting, a method that uses a high-res HDR image to light the scene. Think of it as a giant sphere around the entire scene with an image applied to it. The image dictates the reflections, as well as working as an auxiliary light source. This scene really lived and died on what image I had applied, the fact that I got those nice reflections and lighting was on the merit of the image I had. It was a happy accident that I managed to find a suitable one.
Next up, a very similar project. Spheres as it turns out are fairly versatile. Generally I really don't want to fall into a pattern of variations on a theme. I really fought to find something different to do than another planetoid image in a row, but eventually I caved and just did it. This one was different enough to present it's own set of challenges anyway. I had to figure out how to snap a plane around a sphere, how to style the grass, and how to render out a background that has a simultaneously unaffected and fitting relationship to the subject. None of which is interesting enough to go into specific detail about right now.
As for the leaves in the wind, there is an invisible piece of geometry vaguely approximating the wind direction. I then used a special shader to randomly spawn, rotate, and scale a single leaf "Prototype" along the surface of the hidden geometry. The same goes for the daisies; there is a hidden second sphere that "replicates" them. I'm interested if at first look you realized that the leaves are supposed to be flying in a channel of wind. I really couldn't figure out a way to convey movement and still keep the non-realistic look of the image. I tried to indicate it with the grass, but I don't think that worked... For lighting, there's orange and blue point lights on opposite sides horizontally, as well as 2 directional lights on the opposite "Poles" of the sphere. Also, I seem to remember planning to put a park bench in to denote it's park-ness. I guess I forgot! I also was pondering with the idea of a person on the path.
There's a couple more I could show, but I guess keeping this to my "Planetoid" projects is a nice encapsulation. I'll show the others on another inevitable week in which I can't bring myself to do anything. Also, neither of these things are the ones I re-rendered as it turns out.
As mentioned I'd love to hear some comments on whether or not you recognized the wind, and maybe some thoughts on how to fix it. And comments in general! I crave validation people.