Super thick pyro? (sand blowing layer)

[caskal]

Hey magicians, I'm trying to create a thick layer of pyro (for a sand blowing effect) but so far I can't get the desired result, I got a decent view from distance but the camera is close to the floor so it gets bloby, any advice? this is the ref I'm trying to achieve

And these are my tests so far

 

Any advice will be awesome, thanks in advance!

 

[Atom]

Those test look great!

 

I thought I saw something a short while back posted here about increasing density based upon camera distance. Maybe think about how to implement that or locate that thread. Also, do you have dissipation enabled? If so, try turning it off so the smoke never decays.

It looked like you have a moving bar across the terrain. So density is derived uniformly (plus noise) from the volume of the bar. What if you scattered points inside the bar then stamped density from those points. Then place a sphere around the camera  and group the bar scatter points based upon proximity via. Take those groups of points and multiply their density by 10 or some amount to make them more dense when near the camera. Just a thought about how to possibly implement that...

Edited August 17, 2018 by Atom

[DarkMatter]

Hey Caskal,

From what I can see it looks like you would gain a lot by increasing the Pyro resolution. Perhaps you could use a clustered approach for this to get max rez.

Also a really important part in achieving detailed Pyro sims is to start with a nice detailed source. Don't be shy going high rez on your source volume. Ideally you want your source to already look like it has enough detail/resolution before you even sim it.

A good workflow for generating a high rez source is to use a dense set of points from which you convert to a high rez volume for sourcing. 

Fluid Source SOP is pretty slow when you give it lots of points or going high rez so I suggest using 'VDB from points' or even faster 'Volume Rasterize Particles'. You wont have all the bells and whistles like Fluid Source, but you can use a Volume VOP or Volume wrangle to add some noise or copy over some of the data from your points if you need to. You will see huge performance gains using this technique over Fluid Source which will allow you to crank the rez of your source volume.

Lemme know if you need an example scene and I will throw something together.

Good luck!

 

[Atom]

I came across the density falloff via geo proximity file.

ap_densityFalloffFromGeo_081718.hiplc

[caskal]

5 minutes ago, Atom said:

I came across the density falloff via geo proximity file.

ap_densityFalloffFromGeo_081718.hiplc

You are a life saver, been struggling with this since yesterday, testing this now, thank you so much!

[caskal]

11 hours ago, DarkMatter said:

Hey Caskal,

From what I can see it looks like you would gain a lot by increasing the Pyro resolution. Perhaps you could use a clustered approach for this to get max rez.

Also a really important part in achieving detailed Pyro sims is to start with a nice detailed source. Don't be shy going high rez on your source volume. Ideally you want your source to already look like it has enough detail/resolution before you even sim it.

A good workflow for generating a high rez source is to use a dense set of points from which you convert to a high rez volume for sourcing. 

Fluid Source SOP is pretty slow when you give it lots of points or going high rez so I suggest using 'VDB from points' or even faster 'Volume Rasterize Particles'. You wont have all the bells and whistles like Fluid Source, but you can use a Volume VOP or Volume wrangle to add some noise or copy over some of the data from your points if you need to. You will see huge performance gains using this technique over Fluid Source which will allow you to crank the rez of your source volume.

Lemme know if you need an example scene and I will throw something together.

Good luck!

 

Hey @DarkMatter thanks a lot for the tips. I'm realizing that having fun with houdini is a totally different thing that working on real projects  tons of challenges, but learning a lot, specially when it comes to resolve big size stuff and how to optimize it. Didn't know about vdb from points and vol. rast for getting better results than fluid source, great to know!

@Atom thanks for that file, I think I'm getting there, I figured out how to clip the pyro container using a bounding from camera so doesn't simulate the entire stuff, although on the end we see almost all the landscape but probably I can multiply the size by the time inc to get it bigger at the end. Now I'm trying to multiply the density from the camera distance with I think analyzing your file a little more will get me there, so far I tried this on my own but of course dont work  

 

And here's the pyro container following the camera

 

Thanks a lot guys, will check Atom's file to figure out how to multiply the bounding box that follows the camera with the pyro density, cheers!

 

[caskal]

I think I have it!

[Atom]

That is more of a forum file I found and reposted, I am more of a wrangle person than a VOP person. But you can drop down a constant node inside the VOP and connect it to the 3rd multiply input. This will give you a way to overall boost the density. It may be more useful to assign points to group, rather than apply a new attribute to all points. then you can drop down a new wrangle node and just boost the density in those regions. Although fading distance does allow a form of falloff.

 

Edited August 17, 2018 by Atom

[caskal]

@Atom thanks for the tips! I hope I can get into wrangles someday :D, I was having issues for boost the overall density so this will be handy. also Tim Van Helsdingen gave me a hand over facebook with a handy thread here on odforce for getting a falloff 

 So I think everything is optimized, now I'll run some tests, thanks again for your help, cheers!