Deep Dream Generator

I have been playing around with an online graphics-processing tool called Deep Dream Generator. It allows you two take two images, and remake the first image in the style of the second image. It’s really easy to use, and can produce some stunning results. (See the many examples on the tool’s homepage.) Considering that all you have to do is click a few buttons, I am amazed by the quality and accuracy of the generated output. Here is a link to my profile on the site; and here are some of the better images I was able to create by on my own:

Fractal keyboard clutter

I added two new fractal images to my Artwork & Design page. The first is an updated version of the older kaleidoscopic flame image, except with three axes of symmetry instead of two. The second is an updated version of one of the Yin and Yang fractals, except with added levels of recursion. The idea for the scaling (versus rotation) animation on the latter image was gotten from this YouTube video. Click on the images below to view the animations.

Yin & Yang Fractal 3

For the VGKD project, I renamed the project from “Video Game Keyboard Diagrams” to “Video Game Keyboard Database”, fixed a bug causing the submission form export to be garbled, added the German Apple keyboard, made it so you can fade or hide the key labels, and added a new set of bindings for League of Legends which was submitted to me by Alex Shafer. I also fixed a bug in the cube snake script that was preventing cube segments from being positioned with the correct z-index.

Lastly, I have been working on getting the procedural POV-Ray script “clutter.mcr” by Jonathan Rafael Ghiglia functioning in order to place trees and shrubbery around Datsville. The script takes several grayscale images or pigments as input. One is for the terrain height, called a “heightmap” or “heightfield”. Another is called a “probability map”, and determines the probability that an item will be placed at a particular location. I have configured the script such that the probabilities are determined by the altitude and slope of the terrain. I would also like to alter the probabilities based on a location’s proximity to moisture, soil type, etc. but that lies beyond my current abilities.

Here are the results so far. The spheres are supposed to represent deciduous trees, and the cones are supposed to represent conifers. Both will eventually be replaced with (much smaller and much more numerous) LEGO equivalents. I plan on adding other types of plants and trees as well.

More Wikipedia illustrations

I have been busy creating illustrations for Wikipedia. They can be seen on my user page at Wikimedia Commons. Most of them involve the spherical or elliptical coordinate systems somehow.

I also created a series of sectional drawings comparing the internal geometries of the various “Color Spheres” of (from left to right) Johannes Itten, Albert H. Munsell, and Philipp Otto Runge; as well as a fourth sphere created for comparison based on the spherical coordinate system. It is amazing how many ways there are to draw a sphere!

And I updated a couple of existing images that show how the color-making attribute saturation can be determined from the chroma and value parameters of the Munsell Color System. (Though this definition of saturation is a bit different than the one used in the HSL color space I am accustomed to.)

Lastly, I finished an old scene featuring my interpretation of Larry Niven’s Ringworld. It is a giant ring encircling a far away sun, with oceans, continents and Earth-like weather embedded into it. Here’s the latest render:

The source code for these images can be found in my new GitHub repository which I just created for this purpose.

Microsoft Azure and colony ship panoramic render

I decided to open an Microsoft Azure account, and use the $200 in free points for new members to render a better version of my colony ship panoramic scene [edit: click here for an external view]. I was still unable to add atmospheric media and radiosity inside the station habitat. They simply take too long to render, even with dual 8-core CPUs and 32GiB of RAM. If I could afford several virtual machines linked together to form a render farm, maybe it would be feasible to enable them. But that would cost a lot more than the points I have available, I think.


Google Cloud has a similar offer for new customers, but limit you to a single CPU with 4 cores and 8 threads until you “upgrade” your subscription. So I did not find it very useful.

Color space plots

With lots of help from Christoph Lipka at the POV-Ray newsgroup, I created four more images for Wikipedia having to do with color spaces. They show the sRGB color gamut projected within the bounds of the CIELAB and CIELUV color spaces. Both the rectangular and cylindrical models are depicted. You can see all the renders, as well as some animations, on my Wikimedia Commons page.


In other news…

In other news, I fixed my GeoGebra applets so they work again in browsers that no longer support Java, such as Google Chrome. View them here.

I’ve put a lot of work lately in my Colony Ship POV-Ray model. The latest version can be viewed at DeviantArt, here.

I’ve also updated several of my illustrations for Wikipedia. The switch from POV-Ray version 3.6 to version 3.7 required changes to how pigments are defined among other things.

Lastly, I updated ToEE d20 Inspector to fix some bugs related to how skill points are calculated and distributed. They should no longer be wildly inaccurate. You can find the program listed here.

Proudly powered by WordPress. Theme: isometricland by Michael Horvath based on GPLv2 or later.