Doing some subdiv modelin in houdini and I am in the tweaking stages. It is getting difficult because everytime I try to edit the surface it takes so long to update the downstream templated subdivision sop that i cant work with the edit sop.

The only way around it is to set update for changes only. But that is like playing marco polo editing….

Would there be a way to either have an update option for downstream stuff, so that it only updates the template when you finish changes? Maybe even multi threaded one so cpu 1 gets the current operation and the second cpu handles downstream templated processing?

If you're at that point and esp. for organic modelling, it might be better just to just add an edit sop at the bottom and tweak from there instead.

riiiiiiight………… I'll just open modo instead :wink:

set up a function key alias so it does the opset -f on yourTemplateNode
(alt+Shit+a)

something like:

opcf your modelingSOPLocation;setenv flip =`if ($flip, 0,1)`; opset -f $flip yourTemplateNode

i do this stuff all the time with complicated networks when updating becomes a pain.
-k

I think that Warren's RFE is valid. So much so that we have already started work on viewport refresh delays for inactive views.

In H7, there is an option in the Preferences dialog 3D Viewports page to turn on “Delay Inactive Viewport Redraws”. This option currently only works for a single viewport that is split in to two, three or four parts. It does not work for other viewport panes or floating viewport panes.

If this option did work for floating viewports, then you could rmb on your subdivide SOP and open a Preview Window and it would only update when there was available cycles. I would sure love to have that when I model!

This is unfinished work in Houdini.

Another method that we could employ would be to have a viewport rendering option to see the subd limit surface directly in the viewport without using the subdivide SOP.

how about adding subdivision surfaces all the way then?
-k

Intelligent update of a preview window sounds the most feasible option. Though eventually full realtime subd would be good.

Thanks jeff…

ps. That alias thing is kinda handy ill have to remember that one. :wink:

Changed the flags in there from -f to -t or -d that made it even more handy!

ps. Is there a way to make it so when using the -d flag that it changes from smooth wireframe shaded with points on to smooth shaded with no points?

in the textport
help viewdisplay

that should give all the bells and whistles.
-k

Just bring up a viewport and do a render region then with render as subd.

Three problems with that:

1) The geometry still updates behind the render region so you don't get any feedback.
2) You have to hit the Re-Render button each tiny tweak you make.
3) Getting the wireframe to render over the subd surface is yet another layer to manage.

The more I sub-d model the more I realize how much feedback you need when trying to keep your poly cage as light as possible. You need a quick sub-d update to do that. If you don't, might as well make the cage dense and whittle away at it.

I am still working my cranial sub-d tool but it isn't going that good.

Hey Jeff, what's your cranial sub-d tool? you building yourself a new brain-box!