Illusion Fields

Now the interesting thing about the following 4 images is....

Using an offset->dilate->edge detect process, gets pretty close to the fluvial eroded mountain ranges, but subtle differences in scale are dramatic and...

and now the original using a 'V' profile curve...

These are great. This is probably absurd, but is there a way to introduce color to this fluvial simulation?

Perhaps just a direct mapping from the B/W result to a 5 color gradient, black=dark green, then light green, brown, beige, white=white. So it looks topographic.

Hey, awesome! Thank you. This is like the awkward moment when student becomes teacher I guess. I absolutely think your Landform filter is the best; and every one is either HU or EP, OMG.

So it turns out that refining things so iteration isn't necessary in the erosion process is possible. It just took a lot of tinkering and small steps forward, and luck finding the perfect scale. The source is a perlin cells mix (a better multifractal one is in the works). Then the offset->dilate->edge detect process is roughly:

Elevation - Angle - Slope (Sphinx)
https://www.filterforge.com/forum/read...MID=112546

Self Scaling2 (Indigo Ray)
https://www.filterforge.com/forum/read...MID=123039

Landform (Emme)
https://www.filterforge.com/filters/12449.html

and with 7-Color Gradient Profile...

Wow! The last two panoramas are wonderful!

and with some post processing goodness, ooh la la...

The following 3 with some good old 'spy vs. spy' style. Surprisingly all use the same source. Shown without AA and with.

and in repeat, aaaaaaa

The 11 following to riddle Sphinx. The only hint this time being this image and the suggestion of some smudging.

Some quick and dirty fun with Maze 2D and a new iterative checker percentile blend thingy.

Cat paths plus some erosion = cat tunnels?

G2

G4

was found

Other pipe methods using scripting for the continuity test. Continuity without overlap:

Here is the method using table.insert(T,x), where the red initialization square seeds all other positions.

1. Scan neighbor to the left to fill 1st row

2. Scan neighbor to the top to fill first column

3. Then scan current row left and previous row top for all other positions.

I'm genuinely amazed at what you just did.

A thread of great inspiration. I particularly like those mayan-looking glyphs. Actually just about everything else here too.

Too many ways to count how your posts and filters continue to inspire me ThreeDee. Thank you.

I'm still working on procedural faces, feathers, and space ships, stay tuned.


To save my fellow FFers some agony, here are two graphics of the matrices needed for the continuity tests. Initially it was just the top and left tests to fill top to bottom, but now here are the all around tests for the outward spiral in progress.

Firstly showing the one adjacent test:

Outwardly spiraling continuity test √

Mayan maze glyphs a comin'... Introducing some cyclic bias by resetting every call to generate a new random number once per 4 sides...