Displacement from PhoenixFDOceanTex applied on a simulation mesh

[obeida_zakzak1]

We are trying to blend a FLIP simulation with an ocean generated at render time in 3dsMax/Vray.

The ocean is done using a PhoenixFDOceanTex combined with a noise for displacement scaling, applied on a plane using a Vray Displacement Modifier.
The simulation is done in Houdini, exported as alembic, and loaded as Vray Proxy in 3dsMax. This simulation has flat borders, so it can be displaced to match the ocean.

We want to mask out the non-flat area of the simulation so it won’t be displaced.

We did the following :

• bake the velocity as vector attribute, so it works as a mask
• this mask is loaded in a vertex color, connected to a gradient ramp
• we combined ocean’s noise with the velocity mask (ramp’s output)
• this will be the displacement scaling for the simulation’s mesh

We expected to have a displacement on flat areas only, where velocity is low (mask is supposed to be at 1 in these areas). However, it does displace the whole mesh. Any ideas on the workflow ?

Without displacement :



With displacement :



Mask :



PhoenixFDOceanTex on FlipSIM :



Gradient ramp setup :

[Joelaff]

I’d need more info. But perhaps try clamp output on your mask in the output tab or with an output node.

Or possibly look at a ramp mask in object Y space (or is it Z?) where you would mask out anything with a value <> 0 vertically.

[obeida_zakzak1]

hello Joelaff, thanks for your answer, could you tell me what kind of info you want to know ?

I tried with a falloff in Local-Z / Local-Y, the displacement is still visible on the non-flat area.

In our case it would be more accurate to mask out flat areas using the velocity mask rather than a fallof on a single axis.

[Joelaff]

I guess I mean I would have to play with a scene, but sadly I don’t have any time right now.

You tried clamping the output?

Also, perhaps you need the masked area to have Value of 0.5 rather than 0. Maybe use your mask as a mask with a composite node to composite a middle grey values on top to make the masked areas 0.5.

I’m sure the Phoenix team will chime in when they see this. (They are faster in the Phoenix parts of the forum.)

[obeida_zakzak1]

Clamping the output of the mask doesn't seem to change the result (it's normalized in alembic geometry between 0 and 1)

I just posted on the Phoenix part of the forum hoping they have more ideas about it.

[Joelaff]

Cool. Hopefully they will have some ideas. I wonder if show map on back in the Gradient ramp would help.

Also, since you are combining the noise map after the Gradient ramp aren’t you ending up with non-zero values in your masked areas? Or is that a multiply operation in the composite node?