mawi

Just a tip from some one watching x amount of reels every month... Tone down or drop the music.

So many ways do do stuff like this. Grouping based on normal direction works good, but maybe projecting the base curve upwards and transfer from there is bett cave.hipnc

Sorry I just dont agree. It can newer be wrong to take a job in the industri while looking for the perfect fit. If I got two exactly equal candidates but one has spent time as render wrangler and been introduced to deadlines, I know who I would pick. Lose time? You got something better to do?

Looks cool. With the long hair, I think the biggest problem is your input. Its som much easier if each "ribbon" is a single connected patch. Then you can run a forEach and copy curves to each patch in uv-space. I`ve made a few of these systems and I always use NURBS. Its really simple to plot out hair if you know the parametric u and v direction. Shout if you need some simple scenes.

Cache the emitter and try partition it into smaller sims. Its advecting thru a velocity field and I dont think anyone will notice if its not perfectly density constrained.

Big/small scenescale? Houdini version? Scaled camera/lights? Engine(pbr, raytrace)? Is there any part of the scene we can look at?

If the data isn´t to big, you can try with a chop filter. Since you filter each channel it might post some problems for non b/w values. chopFilter.hipnc

Sounds like you use the guidegroom SOP on a deforming object, this will regenerate each frame with a new seed. Groom on static rest and deform with the guidedeform SOP. If it still doesn´t work... A .hip would help to solve the problem.

Its about inconsistent point counts in the bubbles collision geometry. You are using a trail-SOP after a popnet that got shifting point count, giving "exploding" velocities on frames were a particle dies. Put the displayflag on the vop after the trail and turn on velocity view. Play the disco To fix this, just kill the trail2 SOP, it doesnt do anything since you apply loads of noise directy after. If you really need that velocity, just check the "use template point attributes" in the copy sop. This will let the bubbles inherit the velocity from the particles.

Cant look at the file. But if you are using a static object as collider make sure you got consistent pointnumbers.

Instead of constant VOP use a parameter VOP. That will automaticly promote parameters to the attribute VOP, then write your expressions in the promoted parameter fields, not inside of VOPs.

The vop is not discarding/ignoring anything in the first input, but your up vector is on the curve in the second. You need to fetch it, just like you have done with P and N inside that vop.

There is no up-vector. You use the primuv vop to fetch N and P, but not up.

Do it with a SOP-solver... ap_cloth_confetti_leaf_101916a.hiplc

I think you are better of leaving the furdirection normalized. You got a pic or a hip file to share with the problem? Regarding the blur... By default the hair shader is created with full opacity at the root. Are you sure nothing is happening if you set the opacity to all white? Ive also noticed that the hair shader is reacting "strange" in some situations with the combination of opacity and diffuse limit above 0.