AWS Thinkbox Discussion Forums

Volumetric rendering of a PRT with density information?

Hi there.

I’m trying to render a smoke simulation, it comes from Houdini in PRT format, I made sure that the density information is in there, and now I want to render it as smoke using that density information, is there a way?

Am I missing something in the export process, I export density, Id, Velocity adn I think that’s it.

Also my workflow is to advect a bunch of particles using the Pyro simulation as a field, if you knw any other better workflow, please tell me, I’m all ears :slight_smile:
(BTW I can’t afford Stoke for now, just in case you were saying to me that I can use OpenVDB if I have stoke, but I don’t :slight_smile: )

Cheers!

I don’t see why not. Can you post the steps you tried, and where it failed?
Also, can you upload one PRT (or send it to me directly) so I can take a look at the channel data?
You can use the Particle Data Viewer to do the same and confirm the Density values make sense.

That should be enough. Note that the PRT file format expects the ID to be spelled all capitals, but it is not needed for rendering (unless you want to interpolate curved motion between frames when retiming a PRT Loader), so it should not matter.

I’ll post one here as soon as I finish with a simulation that has my computer cluttered.

There was no “failure” per se, the thing was that the result rendering the PRT that has the density is the same result as simulating the PRT that has no density value, so in the end I don’t have the same fine detail I have in the Houdini version, I know it won’t be exactly the same, but at least I was expecting it to be as detailed as the Houdini version.

I’ll post a PRT file and a picture compairing both situations, one with the PRT with density and another one removing the PRT channel.

Do you plan to add OVDB support for KrakatoaMX? (not just Stoke)

Thanks for the answer Bobo.

Cheers.

If you are advecting particles through the field, you are most likely not picking up any densities from the field. This would also be the case with advecting particles through a FumeFX in Max - you can sample the velocity, but the density does not copy over.

If you DO have the Density on the Houdini particles, and you are saving it out, and you know the values are not 1.0 for each particle, then that would be a different situation…

That’s why we want to peek into the content of the PRT files you saved.

I am not sure what the plans for OpenVDB in Krakatoa are. It would make sense though…

The simulation is going to last a bit, so here is one of the PRT files, now I’m not achieving any good result at all.

dropbox.com/s/aooi7guzh0adi … 9.rar?dl=0

Hope you can explain it to me, is the only way I have to render Houdini volumes in max right now.

Cheers.

I downloaded and looked at your file.
It has GOOD Density data in a range somewhere between 0.0 and 5.0.
However, given the size of the data, it requires a very low Final Pass Density Scale, somewhere around .00002 or 2.0 Exponent -5, to produce a nicely smooth rendering:

Here is the same, but with Voxel rendering, Voxel size 0.1, Filter Radius 1:

I used Magma to override the Density to 1.0 for every particle, basically destroying the data coming from Houdini, just for comparison using the same Final Pass Density settings. In this case, the only density factor is the number of particles in a given area of space, without each particle bringing its own variable influence:

Here is the same rendered with the default Final Pass Density of 5.0 Exponent -1 (0.5) for comparison - as you can see, the particles are too dense and blocking light immediately. The Alpha of each particle would be already around 1.0 even before combining with all other particles in the vicinity:

Just to visualize the Density and Velocity data channels, I created a Magma that sets the Color using a Blue/Green/Red gradient (Blue being low, Red being high), normalized approximately to the min-max range of the channel:


I forgot to tell you that it has to be scaled by a 1000% to have the correct scale, sorry for this.

Anyway, the smoke should be a lot more dense, I’ll try to prepare a simpler example and upload the full PRT sequence and a render preview done in mantra to compare, thanks for your help Bobo.

Cheers!

Here is Final Pass Density 5.0 Exponent -4 for comparison, rendered in Voxel mode with Spacing 0.1 and Filter Size 2.
Probably not what you are after, just showing how tweaking the Density Scale affects the result.


And this is the previous Final Pass Density of 2.0 Exponent -5 as a reference:

In Houdini/Mantra, did you render the field, or the particles?

Privacy | Site terms | Cookie preferences