Trying to include a bit more vex in my workflow as a "shallow end" to grasp code a bit more.

This works as expected,
@Cd = {.1, .2, .3};


This doesn't work,
@Cd = {rand(@ptnum), rand(@ptnum+5), rand(@ptnum+10)};


but this does work,
@Cd = set(rand(@ptnum), rand(@ptnum+1), rand(@ptnum+2));


I know what the manual says about set So what's the difference between using set and the first example that makes the middle one fail?

cheers,

Andi.

Hi Grendizer, did you ever find out how to make this work?


cheers,

A>

I've been using object merge with simple animated objects no problem….. Now I need to go near constraints (shudder) and it all falls apart.

I setup a simple test scene that replicates the elements of a setup I'm doing. A simple ground geo and a small grid that's raycast to the ground and drives around. I Then point constrain a cube to the grid, and create a null that scene data references the world space of that cube, so I now have an null that moves and rotates that I can parent a car to… That works, but I want a copy of the car object merged into my ground object so I can do stuff to the ground (after the grids raycast obviously) but no matter what I do the object merge just isn't working as expected.

If the display flag is on the object merge then the object stays at 0.0.0, if the template flag is on the object merge it behaves as expected.

Constraints in Houdini seem very different to the simple tools in other packages and of course in the grand houdini world i'm a newbie. I know I can bake the constraints out but that goes against the whole concept of procedural.

Should it work? Am I missing something fundamental?

Thanks tinyhawkus, I'll see where I get with it.


cheers,

A>

Julien, did you ever get this to work?

A>

Sadly no Jumax, the end result for me being a poor quality sim. I don't really think the small scale liquids are on the side fx radar really if all the disaster movie crews don't need it….

doctorBob you lifesaver,

thank you so much.

A>

So I have lots of things at object level and lots of things in each object with lots of keyframes. Now I want to edit all the fcurves in one go so I need all the keyframed parameters in the animation editor. Do I really have to go into all 60 objects, select all the keyframed nodes and peg them in the animation mixer? Or am I as per usual missing something…..



cheers,

Andi.

I think opchange in an hscript window might do it. I've bulk changed file paths before with it.

Thanks CYTE,

glad to know i'm not going mad. Simple enough to workaround…


cheers,

A>

Hi all,

when I dive into a dopnet that's inside a geo node everything in the viewport disappears, (yes, show all objects is set) but a dopnet at the object level behaves as expected. I'm confused to say the least. Tried a new vieport, no dice, tried v18 instead 17.5, no dice. Can't find any preferences for it…. stumped.


cheers,

A>

“where things are” in houdini isn't a simple question… One habit I've got into is driving a transform node srt with a null at object level using relative referencing, which won't work if there's any hierarchy involved, so evolving from that i managed to get world position and rotation for an object, (hierachy or not) by using the Reference>scene data where you can access the world position and rotation of an obj level object. That has to happen inside and obj level “control object” (at 0,0,0)and then the values can be referenced outside by anything in the usual way.

It certainly works for me in this specific instance, but - given every time i have an aha moment it's followed 3 weeks later by an actually i've no idea what's going moment, I wondering what could be wrong with my setup. I know information about world space is something people request regularly so i'm dubious it was that simple for me to stumble across.

I've referenced the values back on to the node i'm interested in for handiness.

Hip attached

cheers,

A>

Scale is an interesting point as I've always tried to work to real world units so my particle separation is heading towards 0.005 or lower. Never tried working at a hacked scale as I've no idea which parameters need tweaking to accommodate it apart from gravity obviously. Is it a smarter move to simulate at an optimum size and scale down the end surface?

I seem to have problems when simulating anything with that has a finer thinner quality to it ending up with volume loss, or clinging to surfaces. I've uploaded a simple test to illustrate what i'm currently trying to create. A garden feature where the water bubbles out the top of a pole and flows down the outside. Simple but the water needs to be the star of the idea.

I've a very rudimentary grasp of what flip is doing… gleaned from Steve Knippings cg circuit flip series, but I've no idea how to debug flips behaviour, or go beyond the basic setups. I know for sure you don't master flips in H with the occasional sim over 12 months, but I've come from bashing stuff out quickly in Softimage and Realflow to taking a fair bit longer in H and the boss is tapping his foot!


cheers,

Andi

Looking for thoughts and opinions on this.. From the horses mouth would be amazing too.

I'm wondering if there are plans to do anything on a grander scale for liquid simulations (not really biblical floods or oceans but smaller scale stuff)? Does anyone think it needs work? I'm looking for solutions to my issue(s) with it not looking to grumble about it not working.

I have to say first up I've only been using H for 12 months so I'm a total beginner with all the problems that brings so the question comes from that perspective. Liquids or rather flip seems to be the only part of H I've not been able to get in any way comfortable with. My experience of liquids pre H is small scale uncomplicated stuff done in realflow, laterly using dyverso. I never used hybrido as the sph solver was fast enough for my needs and I could turn simple stuff out without issue. Flip in H seems like a different beast and I've really got nowhere with it. A fair few voices suggest that flip has things it's good for and things it's not and that seems to be backed up up by the kind of tutorials and masterclasses that are available.

Looking at doing some small garden water feature designs recently and after a few weeks if fidling around with flip I've felt like I should just do it in Realflow to get it done but that seems like a massive cop out and I don't have time or money to use multiple bits of software. Using the result of following a tutorial is a good start but after that I've only to breathe on a setup/parameter and it instantly turns to garbage and advice from forums seems to be unwieldingly varied.

So, more questions than my inital one… Do people who really know about it think liquids need looking at? Is a third party solution a necessary evil at this time? or Is it really just time and experience that one needs to be able to achieve quick, simple small scale liquids.

cheers,

Andi

Thanks mrCatfish,

that works perfectly, vops rather than vex and wrangles is much more familiar given my softimage history but somethings tells me i'll need to grasp vex properly at some point as i've zero coding experience whatsoever. Am I right in thinking that groups are actually just attributes under the hood?

Solver sop seems like a handy node if I just just figure out how to incorporate it properly. Here's my scenario:

I've a a long strip of points travelling along in X (like rocks on a conveyor) and each rock has a float attribute that's a random value between 0 and 1. When each point hits a certain position in X the random value gets added to every frame which is where I though the solver sop would come in. Then at a further position in X the counting stops, and a delete node gets rid of all the points who's random value is above a threshold.

I was using a group by bounding volume to determine where the counting should occur but I can't combine the counting bit inside the solver sop with the grouping bit that comes before.

How do I get the adding wrangle in the solver sop to use the group that's changing frame to frame.

Alternatively what other techniques would people use to achieve the same result?

cheers,

A>

I though it was me and I was just breaking something but I get the same behaviour with a simple test scene. I'm viewing the values in my pyro voxels using a volume slice on the import_pyrofileds node and getting some wild density or heat values. There's the normal range I would expect 0 to 1 or suchlike, and then I'm getting values like 4.10663e-26 or -3.07336e-0 which is confusing.

Are these numbers anything to worry about? Is Houdini, and for that matter redshift ignoring them?



cheers,

Andi.

Hi,

I'm using flip to show some water feature ideas, so it a couple of metres in scale. I've turned on viscosity just to stop the water moving so quickly but kept the value very low - 0.001, and turned on slip on surface and set the slip scale at .95. Once slip scale goes above about .8 i start getting little pops in the sim. I've checked for interpenetration between the emitter and the obstacle and it seems like there is plenty of space but my flip debugging skills end there sadly…

I assumed that turning viscosity on and keeping it low was appropriate as all liquids must have a viscosity value and I wanted the water to interact more with the surface rather than glide so effortlessly over it.

I've included the hiplc file in case anyone is bored and wants to take a look.


cheers,

Andi.

Thanks Michael…. Infinite ways to skin a Houdini cat, and every day learn something new.


cheers,


A.

Thanks Benny, that certainly works. To me it appears to need some convoluted thinking to have reached that solution myself but then I've only had 12 months experience at the simplest end of Houdini so my brain is still working in oldschool soft mode. Now I'm out of that hole I can crack on.

cheers,

A>