ok I am cooking. I clicked Cloud FX>Sky rig, default bounding box.

its makes a 1x1x1 cube with some lumps in them. thats not really big enough for a sky now surely. ok lets try scaling 10 x 10 x 10 . it says its cooking Op but am still waiting. it just tells me the time elapsed, is there a way to tell how long something is likely to take? if it will ever finish?

my processors are maxed out so its definately doing *something*

ops: sorry if this is stupid question

Probably not, progress bars have no meaning for time.

Some operations report the percentage complete beside their status message in brackets, such as “Convexing (34%)”. But those tend to be more of the exception than the rule.

no progress bars? :shock:

ok well I stopped whatever the heck it was doing after 20 mins full proccessor trundle

made a new cube 10 x 10 x 10 and a sky rig takes microseconds to create.

so first rule of sky rig.

do not scale afterwards

:?

You can always scale afterwards - just scale the ‘box_object1’ not the cloud_from_box_object1'

Scaling the box_object1 takes a less than a second.

thanks Marty, sorry for being a pedant, but that's exactly what I did.

there is one object. A sky rig.

entering that there are 2 nodes connected. A sky_Bounding_box and a sky1

selecting the sky_bounding_box object

in the transform panel changing the scale from 1 to 10 on each axis is really slow, 20 mins+ and never cooks for what should be something very quick, no? I am not scaling the sky I am scaling the bounding_box that comes default with the sky rig

also tried with a box rather than default one. works quick to cook on any scale to start but scale it afterwards… uh uh

its not a big deal as I can just get the correct size box *before* applying.

No probs. The magic ‘speed’ is all in the nodes

Have a look at the sky node in sky_rig and the Voxel Size.

Its 0.1 when you place a box with a size of 10,10,10 and when it's 0.01 when you place a default box of size 1,1,1. When you place a box of size 100,100,100 the Voxel Size is 1

When you post-scale the box you are cubing the amount of voxels, so just divide the amount of Voxels to match your new size.

BTW you can see the amount of Voxels in the viewport by bringing up the Display Option, press ‘d’ whilst the cursor is over the viewport, under the Guides tab select Always on for Geometry information

ha there you go, that is the info I needed!

thanks Marty thats exactly what I wanted to hear, why it is so slow and of course if voxel size is set to a ratio of initial bounding size and does not keep that ratio after creation - which admittedly I assumed it would as it is called a ‘rig’ but hey. will have to improve this rig so the ratio factoring stays intact

8)

Great! - more advanced users consider the shelf tools as starting points to make their own set ups.

another handy feature that I would turn on if I am new user is
press d on your node view >
go to Dependency tab>
turn all the tix box on>
It shows you what node is affecting other nodes.will speed up your learning process and Houdini way. :twisted:

I use them in many cases to see how to do things, then I do them my way… kind of like a tiny snippet tutorial.

MartybNz
Great! - more advanced users consider the shelf tools as starting points to make their own set ups.

Jordi is bang on.

The shelf tools are designed to set up various iconic setups to help guide users to best practices and to get something that performs well for most use cases.

But…

If you are using Houdini, you are after the unique cases of course. This means starting with the shelf tool then tweaking it to suit your needs. Almost everyone uses the shelf tools to try out new areas to create set-ups. Especially with FLIP fluids.

Most other apps have tools that execute scripts and what you end up with is a no-going-back scenario where invariably a few steps end up with a baked result somewhere along the line. That no-going-back fear that simply doesn't exist in Houdini. I haven't had that feeling since 1993 when I started using PRISMS and then Houdini.


In Houdini, a shelf tool just creates nodes in networks with various dependencies as parameter node references or expressions. Nothing is hidden. Nothing is baked. It's there for you to dissect and re-work. You can copy-paste networks to keep original versions around as well. Houdini only evaluates nodes that are doing work up from the display flags. Non-evaluated nodes have a minuscule overhead of about 15k I believe so nothing. No fear for copying and pasting off various versions of nodes.


To see these dependencies you can open up the Network display views (cursor over network and hit the “D” hotkey or press the “eye” icon in the network pane tools) and go to the Dependencies Tab and turn everything on except the last option. This will show you in the network all the dependencies on nodes. Very handy coming from any other app. It was put in there for this very reason. After a couple months of using Houdini, you don't need this any more as you get it and you kind of know where the dependencies are.

You can also RMB on any node tile icon and choose “View Dependencies” to see parent and child dependencies.


As for using volumes, size matters. The shelf tools try to inject some sanity to things by evaluating the bounding box of the incoming geometry and then trying to set up a voxel size that won't be too excessive. Some don't though…


You'll not forget this when wiring up SOPs manually to process volumes. Just tap the Esc key several times so that Houdini can catch the event and stop the process.

jeff
Just tap the Esc key several times so that Houdini can catch the event and stop the process.

That's something i think you'll soon appreciate when coming from Softimage!

CiaranM

jeff
Just tap the Esc key several times so that Houdini can catch the event and stop the process.

That's something i think you'll soon appreciate when coming from Softimage!

Definitely. That and being able to pause and resume a simulation.

-T

we all know what VFX artists do all day…