MENOZ

make sure your vector field has the right dimensions. usually i use gas match field instead. also, not sure about the data name of gas analysis to be the same as the vector field, (i didn't open the hip)

it seems to me it's not a problem of voxel density but of the simulation itself. voxels looks to be smaller than your blocks as i can see smaller scale details, also the shape of that mushroom is weirdly geometric. it looks like you are using sparse pyro, you might need to increase the padding or try to use the option "extrapolate velocity into new tiles". https://www.sidefx.com/docs/houdini/nodes/dop/smokesolver_sparse.html#sparse_streakvel

if you have a curve you can check the distance between p1 and the points of the curve. if you resample your curve with lots of points you could select the points within some range from l, the more the points of your curve, the smaller the range you can use and the more precise it will be. You could also find more than one point on the curve that is at distance "l", in that case you should select the one with the "u" parameter closer to whatever "u" value you have at p1. You could also iteratively check if the difference between l and the distance form the world position at u value and p1 is within some range, and progressively move along the u trying to reduce that difference, But i feel this could be slow because you have to call the function to read the u value for each iteration.

SkinBLEND-Menoz.hipnc

https://www.sidefx.com/docs/houdini/copy/tutorial_stamping.html have you tried this?

Hey, what tutorials have you seen? could you give more info on this please?

because houdini writes the value on the geometry after the code is executed, so while you are in the code, you cant' read @pointatr because its value has not been written to the geometry yet.

hello, I've made a hip file showcasing some tests i've made, regarding both sourcing (in the new pyrosolver sop) and solving collisions properly By "solving collisions properly" I mean without density loss due to collision, for that I disabled "correct collision" parameter in order to see what's going inside the collision object. I tried different combinations of settings, the grid is 2d and quite low res, in order to have a decent interaction on my old laptop. I tried on a modern workstation with higher resolution, the problem is generally less apparent but definetly still there. The hip file is commented as best as i can with notes. It took me quite some effort to test meaningful settings and put together the example file. I'm confused by the fact that the more "accurate" solvers and higher supsteps produce a worste collision behaviour. If anyone wants to try it out and see if he can make the collision to solve correctly and avoid density going inside the collision objects i've attacehd the hipfile. beware! even if it seems that on "first interaction" with the smoke the collision works fine ( this is probably the case setting higher grid resolution), let the object oscillate back an forward. cheers! collDebug.hipnc

In the case I have an HDA that does a series of things, for example bend, and draw curves. Inside I have those nodes that performs those operations and have their own view states. Would it be possible to access/merge those to the view state of my asset? I'd like to be able to choose if i want to enter the bend view state, or the stroke view state, from my hda.

you could speed it up a lot separating the capture step from the deforming step. You capture once, which takes lot of time and it would be always the same, and just deform on every frame. I'm referring to point deform's capture and deform parameters.

I'd say it's because it's a function that is only available in shading context. I'm not sure exactly how is this structured, but I know some functions only works for shading, I'd be interesting to know why. https://www.sidefx.com/docs/houdini/vex/functions/index.html

create various smoke caches and place them by hand

I tried using nested dielectrics and absorption. if you force specular of air bubbles you would get those sharp highlights as well. You should increase a bit the render quality maybe, I'm on a laptop. nested.hipnc

you could object merge different stages of your network in other geometries and link those to different viewports. but you can't set 2 distinct display flag in the same network.

-something to simulate feathers, leaves, scales, fast and efficently. Vellum is helping a bit but is struggling a lot with rigid pieces and collisions. needs too many iterations to converge.