I wanted to do one of these wireheads, but with a flip fluid oozing between the wires. Not unsurprisingly this needs such a resolution for the flip fluid that getting the sim right doesn’t quite fit into a daily project, but at least here’s something.
Today, a straightforward implementation of Inigo Quilez’s texture bombing shader. It’s a vop node, so you can easily use it as a building block in a material. I will be doing just that the next time I build an übermaterial.
The header image is iq’s “method1”, here is “method2”:
and for comparison, regular tiling:
I wanted to test the new nested dielectrics of rsd’s great Physhader. In the end it got too slow with the number of reflection/refraction bounces I needed, so I ended up using the unnested one instead.
Just some relaxing knob twiddling. When a TD models, a TD models a torus. This one is actually fog, with a glass skeleton.
I had never used the Sensor VOP family and wanted to do something with them. Here is a “floating point” Game of Life. Every cell creates a sensor panorama of its surroundings, moves to the direction of most neighbors and gets a “strength” value according to how much stuff there is in its vicinity (how much of the solid angle is blocked). If the strength is much or too little the cell loses volume, otherwise it gains volume. If a cell gets too small, it dies; if too big, it divides in two.
Way back I studied meteorology for a few courses, and ever since then I’ve wanted to do something meteorology-related. For a start I wanted to just download some data and get it to show up. I was hoping I could recreate data snapshots from different points in time by just advecting data from a previous timestep. Data here is with 3h intervals. That doesn’t seem to work too well, I suppose there are too many confounding factors (surface geometry, sun’s energy, vertical advection etc). Or, *gasp*, I’m doing something wrong 🙂
Getting good data is a bit of an effort – lots of trawling through archaic-feeling websites..
Fun start anyway, I will probably revisit this some other day. This could be taken a lot further..
data volume snapshots (37M)
This was a vision of my visually talented wife, sparks travelling up a tree and blossoming at the ends of the branches. It’s based on an earlier project of mine at eetu’s lab.
It looks better as a proper video, check out the mp4.
Tonight’s exercise was to replicate Valdez Araujo’s nifty fluid image trick. With an added twist of creating the particles on the fly and matching by id attribute.
Keeping up with the theme, I wanted to try alternate ways of finding the paint stroke directions. Here I radially sample around the current point, finding the direction to which there is most change. The paint stroke direction should be perpendicular to the found vector. Kinda sorta works but clearly would need further development, but, hey, daily hip 🙂
Advecting points through the vector field gives some funky squiggles:
daily.hip 