konstantin magnus

Merge all objects, pack them after connectivity and lay them out with a for loop which sums up all widths plus the spacing. This detail wrangle code is not sufficient but should give you an idea: float space = chf('space'); float shift = 0.0; for(int i = 0; i < npoints(0); i++){ float bounds[] = primintrinsic(0, 'packedbounds', i); float width = bounds[1] + abs(bounds[0]); vector pos = set(shift, 0.0, 0.0); setpointattrib(0, 'P', i, pos, 'add'); shift += width + space; } distribute_objects.hiplc

I am trying to fit a 2d mesh between 4 curves. Pretty much what the skin SOP does but potentially with a custom mesh. vector bbox = relbbox(0, v@P); function vector blerp(vector a, b, c, d; float x, y){ vector lr = a * (1.0 - x) + b * (x); vector tb = c * (1.0 - y) + d * (y); return 0.5 * (lr + tb); } vector p0 = primuv(1, 'P', 0, bbox.z); vector p1 = primuv(1, 'P', 1, bbox.x); vector p2 = primuv(1, 'P', 2, bbox.z); vector p3 = primuv(1, 'P', 3, bbox.x); v@P = blerp(p0, p2, p1, p3, bbox.x, bbox.z); How can I smoothly interpolate the point positions so the borders actually touch the curves? blerp.hiplc

Unwrap the mesh (right) before deformation (left): Assign zigzag UVs to the curve (right) and transfer positions from the mesh based on both UVs (left): spiral_mesh.hipnc

Isn't just boolean with 'resolve self-intersections' and divide with 'removed shared edges' sufficient? curve_boolean.hiplc

It should be sufficient to simply replace bgeo by obj in the code, I think.

Hi Ryuji, you can inset a polygon or polyline with these steps: vector next = normalize( point(0, 'P', (@ptnum + 1) % @numpt) - v@P ); vector prev = normalize( v@P - point(0, 'P', (@ptnum - 1) % @numpt) ); vector avg = normalize(next + prev); vector up = normalize( abs(cross(prev, avg)) ); vector in = cross(avg, up); float dist = dot(next, avg); vector dir = in / dist; v@P += dir * chf('extrude'); Kim Goossens discusses this here: https://www.youtube.com/watch?v=tnDqwcNG20Y

Wouldnt that be possible with Python already or what exactly do you have in mind?

Hi Mani, they all work except no.1, hence the thread. Petz` scripts (variants 3 to 5) write bgeo-files to the location where your houdini file is located. You can tell by: filePath = hou.hscriptExpandString("$HIP") Variant 2 was mine, but I found it visually too wasteful for just saving. Variant 3 uses some Python code inside the file node to increment the file names. Just click on the label 'Geometry file' to see it. 2 and 3 are related but Petz` variant is more compact. Variant 4 packs the fragments which makes handling and in this case exporting much easier. getEmbeddedGeometry() inside hou.PackedGeometry class seems crucial here. Variant 5 utilizes a Python SOP with two loops: First assigns the segment number to each primitive and the second deletes all non-related primitives before writing out all left-overs. Just look in your folder: You should get a list of objects ranging from part_0.bgeo to part_24.bgeo. The nodes just dont overwrite each others files, I think.

Petz solved this issue a few weeks ago:

Hi Adam, Akira Saito has probably voronoi-fractured the shell and minimized each surface. Reminds me of what BMW has concepted with 'GINA' or of a Zeppelin for that matter. Here is an exemplary workflow: fracture_smooth.hiplc

Hi Jonas, I attached a simple scene that does shading based on the dot product of the surface normal and the normalized light direction. It includes a SOP-variant in VEX and a VOP-shader. Also Yancy Lindquist recently published an extensive tutorial on shading concepts: If all you want is a lambertian toon shader you can also watch mine: light_angle.hip

Thank you again @petz for posting a mathematical solution. I am currently trying to apply it to another mesh (fig. 1). Unfortunately most of the points get collapsed (fig. 2). Is there anything I am missing out? Meanwhile I was trying to come up with a simple solution on my own (fig. 3), which is not unifying the distances between the points along a circle but is averaging out their distance towards the center. Unfortunately I would not quite know how to flatten my points on an oblique plane based on their average normal (fig. 4). human.hiplc

It is better to paint on a black surface right away, because that way all color channels are initially set to 0. After painting you should promote colors to primitives, because this is what primitive extrude is working with. If however you have chosen to paint red on white, the red channel will be set to 1 everywhere. Red = 1,0,0 and White = 1, 1, 1. Thats why the green or blue channel is more distinctive here: Set your primitive wrangle to f@r = 1.0 - @Cd.g; so its able to tell where red is being used. Inside the primitive extrude set 'distance' to something larger than 0 and set 'distance scale' to r. exturde by color_corr.hiplc

Set a normal node to 'points' and put this in an attribute wrangle: if( max( abs(v@N) ) < 0.9 ){ int nbs[] = neighbours(0, @ptnum); foreach(int nb; nbs){ if(nb > @ptnum){ vector nml = point(0, 'N', nb); setedgegroup(0, 'outer_edges', @ptnum, nb, max(abs(nml)) < 0.9); } } } box_edges.hiplc

Leaving out the constraint about no PC2 point used twice, it would be just this between both point clouds: int connections = 10; for(int i = 0; i < connections; i++){ int pt_PC2 = int( rand(@ptnum + 45 * i) * npoints(1) ); vector pt_pos = point(1, 'P', pt_PC2); int pt_add = addpoint(0, pt_pos); int prim_add = addprim(0, 'polyline', @ptnum, pt_add); } With sorting out points however you will probably need a sequentiell approach using arrays. connect_point_clouds.hiplc