Відео

Sure, You'd need to have a wetmap simulation where it starts and fades away completely within a set-time frame, then use two time offset nodes in seperate branches to offset your wetmap to half the time of the overall simulation in both directions. So if it's 100 frames then you'd have an offset with $F+50 and another with $F-50. Then you can use a point wrangle to combine it all together. Put the main branch in the first input and your offsets into the 2nd and 3rd then use the point expression like: @wet += point(1,"wet",@ptnum); @wet += point(2,"wet",@ptnum); Should nicely add it all together and loop then for you.

@@KevRyanCG ah great info! Thank you for sharing!

;)

I'm afraid I don't but when I made this tutorial it began as a windy forest scene I was messing with that I must return to at some point. Though the tutorial I was looking at for that is right here - ua-cam.com/video/NMlYf_T9Nb0/v-deo.html My specific tutorial was for the process of making all your instances in Houdini as merged packed objects so I could instance them, like for debris.

ua-cam.com/video/WOCcT93D0rk/v-deo.html actually this one might be easier to follow. I'm going to have a go of it today.

​@@KevRyanCGthank you. I took advantage of having a long weekend and just completed part 3. It seems to be a decent approach, although looking in the comments section there are other folks who think that perhaps it's a bit cumbersome. But I have to say It's very effective. Certainly better than what I've done on my own. There's another gentleman named Josh Weber, who has put out a tutorial regarding animated instances although his tutorial has very poor audio. It's more of a casual mention than an actual structured tutorial. Difficult to follow / interpret.

​@@KevRyanCGha, that's on my list too!

@@tomallen5837 Man I've been on this all day and I can't really figure it out right. Best I did was make an instance set-up in Sops and referencing that in and creating native instances. There's some method with Value Clips that I just couldn't get to work, though I saw it on a sped up video on vimeo and it seems to be the way to do it. Sadly the artist didn't divulge too much more information. I'm going to look into it further then I'll make a video once I find out.

Yeah it's the kind've thing you'd use on particle sims quite a bit, on the volume sop there's a tab that lets you create a volume from your camera which you can use to group points and then delete. Potentially saving on memory at rendertime. But it's also worth looking up fromNDC and creating frustrums from boxes too. I had covered it in this tutorial but clipped it for time. Again, it's incredibly handy stuff.

Ha I don't know, I've been thought to use Houdini in a very script heavy way so my knowledge of constraints isn't up there. Even for a look at I'm finding the position of the object, getting a vector direction and converting that into matrix rotation. Why? Flexing I guess.

Thank you! Yeah it's just handy to have things at SOP level sometimes, and being able to extract and create the euler parameters is good to know too.

@@KevRyanCG Indeed, sir, thank you once again

HA! There is, I said it right at the start, with constraints and stuff. But this is more about learning things via an elongated and silly process that can be applied to other situations. Matrices and Cracktransform tend to pop up when doing any sort of serious Houdini work.

You're very welcome :)

Oh thanks! I'll have a look at that myself, I was having trouble with adding wind having the smoke just getting eaten by the colliders as well.

substeps wouldve fixed it too most likely but if you dont need pixel accurate colliders its not necessary to do over the fill interior

@@SwipeLess For really big geo I just manually adjust it to what I needed by using the interior/exterior voxel band instead of fill interior. It worked better and way faster than adding more substeps. Adding substeps can help, but didnt solve the problem no many how many I added.

@@evanrudefx good that it worked, using volume colliders too with vdb from polygons etc sdf name "collider" and having same voxel amount as the sim itself will be the most accurate way

what if we need to different material for each types of instance object

Yeah that's where I started but it kept hitting walls with the newer Sparse Solvers and so it needed some extra tweaking to get working. I stand on the shoulders of giants! Hell I even did an Entagma Tutorial today for living particles.

Well done! @@KevRyanCG

Ya know, I've not really tried the OpenCL in simulations, my card is a bit out of date and I find it chugs along a bit. But I'll look into it, thanks!

@@KevRyanCG if your GPU memory is low (mine is 12G), you can limit the loop range, size of the simulation, or voxel size , just for quick tweak. And then switch to sparse to cache.

hi, how to do that ?

It’s Houdini