celd

If you use raytracing or pbr render engine you can increase stochaistic samples up to like 128 to remove most of the noise without increasing render time too much, otherwise you are gonna need to bring pixel samples up - the value you are using is simply too low. Also your volume does not appear grainy in the viewport because it is being rendered by OpenGL, completely different than mantra.

if (@id % 10 < 3) This divides the id number by 10 and then checks if the value after decimal point is below 3 (modulo operation), if it is, it assigns the particle into a given group. So in practice it will do for example: 1/10 = 0.1, then it takes the 1 and compares if 1<3 and then assigns the group; or 98/10 = 9.8 where 8<3 - no group assignment. if (@id % 10 >= 3 && @id % 10 < 5) This does the same thing except it compares with 2 different values and returns true if both statements are correct, resulting in grouping two tenths of your points.

You could put the following code into a POP Wrangle to separate your particles into 4 groups. if (@id % 10 < 3){ @group_group1 = 1; } if (@id % 10 >= 3 && @id % 10 < 5){ @group_group2 = 1; } if (@id % 10 >= 5 && @id % 10 < 6){ @group_group3 = 1; } if (@id % 10 >= 6){ @group_group4 = 1; } And then you could also replace @id by rand(@id+123) * 10 to get a random seed

I wanted to try a method which works independent of point numbers: int npoints = npoints(0); for (int i = 0; i<npoints; i++){ int group = inpointgroup(0,"end_point",i); if (group){ setpointgroup(0, "end_point",i, 0); break; } } You would put this in a detail wrangle after you create your end_point group and then replace "end_point" with whatever you named it. This probably is not the best method for heavy geometry.

First of all, your noise output value is a float, which means it will put the same value into all 3 fields of the force vector - this is the reason why it goes in a same direction. The reason why AA noise goes both ways (in the same direction) is that the output values are between -0.5 and 0.5 and unified noise output values are between 0 and 1. What you want to do is change the noise signature to 3d input, 3d noise and also potentially fit the output so its for example between -1 and 1.

So it is 64bit but only in chops? seems to me like they misadvertised this new feature.

Hi, I wrote a factorial function in vex but the calculations break down after 12!. I am using Houdini 17.0.416 and I thought vex now can calculate up to 64bit. Is this something that needs to be enabled somewhere or am I doing something wrong? I am trying to do this in a detail attribute wrangle. Thanks

Thanks, I was hoping for this kind of solution, you saved us a lot of time!

Hi, I am rendering pyro with 2 extra image planes, combined lighting "all" with light exports, which gives me 3 passes - the smoke lit by 3 individual lights (direct, enviromental and volume light) and combined emission "all_emission" which gives me just the flames. Now all of this is very useful in compositing process, however it also increases render time to 8-15 times the length of render time without passes. I would expect there would not be any increase of render time since it already is a part of the "beauty" render. Is there any way to get around this? Thanks

use volume gradient for a direction towards the surface and volume sample for a distance to the surface, then add the vector to the points position - use point vop

Cant you simply use polywire?

you need to go to Render/Sampling tab on the object you are rendering and tick Geometry Velocity Blur on, and on the mantra node in Rendering tab tick Allow Motion Blur on you will then have motion blur in the direction of velocity vector v in your render

I will try to answer this with various levels of knowledge I have about your questions: Pump preset introduces velocity into the simulation, you should always look into "SOP to DOP binding" tab, that's where you can see what is actually going on. As you can see in your hip file, it is trying to use a field called "pumpvel" and add it into simulations "vel" field. There is a second field, scalar field, called "pump", which is being added into "pump" field, I believe this scalar field does not actually do anything, however in Masks tab you can see its being used to mask out the vel field, so that the vel field gets multiplied by the pump field, effectively masking it. You can see how these fields are created on the SOP level. The pressure field is the one that you cannot visualize in the smoke object. It is an internal field that pushes against the velocity to make the simulation divergence free. Pressure solve is one of the steps of calculating the solver equation. Now the expand preset does indeed do what you originally wanted to do, that is, add divergence. I admit I am not entirely sure how it fits into the whole picture, but in the end it is not too important to know more than that it introduces divergence into the sim to make it expand. Hope this helps and if anyone wants to fill in the blanks - please do.

the combustion model's "gas released" parameter introduces divergence into the simulation, you can see it when you look at the field visualisation source it into the simulation just like any other field, as Alexander mentioned

some ways to solve this would be either gluing the larger pieces together in the simulation, or using "active" attribute to not sim them at all it is not exactly what you want to do, but it does the trick