How are these images made?

Occasionally people ask what kind of software I use or what kind of process I follow to make these images.

It's a little difficult to answer that questions for much of my fractal work -- the reality is that I fiddle with this and muck with that and then I look at it and I say "maybe if I changed the background more to something like that stuff I did in that other picture but more in a direction of that or, wait, maybe I should put another inflection point there and, hey that looks cool; maybe with some of that scaly texture on it...." and in the end I get something I kinda like.

Really a lot more perspiration than inspiration.

At the core, many of my images tend to use one of the fractal engines out there -- Ultrafractal (for much of my older stuff), flam3, Apophysis, Chaoscope ... the basic shape you see usually comes from an equation. The rest boils down to choices from location in the fractal landscape through things like light and color and shade to subtleties of composition and emphasis.

Starting around '11 I usually wrote this all in GLSL to run directly as a shader on the GPU because I'm impatient; and also because this gives me much more fine-grained control over everything that happens in every pixel.

Then there's a bunch of self-created filters; mostly composed from the netpbm library, but I've also written the one or other ImageMagic-based tool. There is sometimes a certain amount of GIMP work that is put on top of it all.

Experimenting until I like it. A little math, a little art, a little idle play...

Now any one image in particular may adhere to this general description more-or-less closely. An image might, for example just simply come out of the fractal engine just perfectly the way I like it. That is rare, but happens. Luck. Or otherwise (much more often, I fear) I can spend days, sometimes weeks fiddling with some concept and never actually manage to get it where I'd like it to be.

Resolution

This question has come up often enough to warrant putting it here:

The concept of "resolution" really only applies to a printout, not the image itself, since the "resolution" is always fixed by the monitor. Consequently, these images are not rendered at any one "resolution": An image that is 1280 pixels wide, for example, is 17.7 inches wide if you think of it as 72ppi (72 pixels per inch) or it's 10 inches wide if you take it to be 128ppi or 4.25 inches wide if you think of it as 300 ppi -- "resolution" only enters the picture if/when you want to print them, i.e. impose a size. Before that, they are simply "1280 by 1024 pixels" without any fixed notion of the size of a pixel (or the number of pixels per inch).

Some image processing software (including PSP and PS if I'm not mistaken) simply always set the "resolution" to 72 when there is none hardcoded in the image.

Now if you'd like to make largerish prints at a high resolution, then you may want to have renders with more pixels. If you'd like to print something in 4x6 inches at 300ppi then you'd need a 1200x1800 pixel image. But the same image could then be rendered at 6x9 inches at 200ppi. The images here are 1280x1024 (the older stuff) through 1920x1200 (the most recent things) and would thus print at 5x4 to 7.5x4.7 inches if printed at 256ppi.

On the other hand "dpi", i.e. dots per inch, is really a printer quality statement and says how many droplets of ink the printer can put down in a given area. But since a printer usually only has three or four base colors, it will need several drops to mix some intermediate hue. If you have a 600dpi printer and you print at 300ppi then each pixel is approximated by four droplets of ink, which makes for a somewhat coarse representation of subtle shades.

(This question came up sp often, in fact, that I expanded on it here...)

HTML

This page was written in early 2001 in fairly vanilla XHTML1.0 which ought to be both XML and HTML-4 parsable. It contains a single small chunk of CSS, mostly for the liquid layout (change the width of your browser window to see that at work) and the occasional closable little boxes on the left side here.

There are a few lines of Javascript for closing (and reopening) these little textboxes and a non-Javascript capable browser will simply have all of these open all the time. I tested this page back then in IE-6, Konqueror-3.1, Mozilla-1. It works just fine in Lynx 2.8 under Linux and Sun/Solaris (albeit without the thumbnail images) and Netscape6 under Solaris. And as of this writing (Jan 2017) it looks like it's aged quite well.

The intent here is of course to create something that actually works for everybody. Should you come across a browser (of this century) that has problems with this page, please shoot me a line so that I can have a look at it.

I do not like "HTML editors" and similar abominations that produce monstrous files that are not in the least compliant with the World Wide Web Consortium's standards and look horrible (or don't work at all) with every browser except the exact version of the exact browser the author was using.

It is not my intent to switch carreers at this point, but I'm certainly available for small web-layout jobs if you like what you see here.