Optimising a card setup

I have a shot where I’m trying to fake a dust trail on the ground from a crashing fighterjet.

Sorry, I can’t put up my script so I’ll describe it as best as I can.

My basic set up is –

2d rendering of ground and also one of the jet,
Bring in camera and proxy geo from 3DSMax (Oh, a huge kudos to whoever wrote that Nuke’Em script, what a lifesaver!)
use stock footage of dust to put onto cards and line up in 3d space.

Is there any limit to the amount of cards to use?

I have 5 sequences of dust and I’ve done some retiming on them (to slow them down) and a few other things (grade, blur etc) and then rerendered them and then I’ve plugged them straight into a large amount of cards. I have about 50 cards and some of them I have edited using bicubics to loose a bit of the flat feel. Should I project instead of just pluggin them in? That would be a hell of a lot of cameras and I would think that would slow the script down even more.

My script was starting to get a bis messy so I started to group the cards to keep track of them and then plug several scene nodes into one main scene before the scanline render node. Is this bad? Should I just use one scene node? Or use a copy (or a clone) of the camera for each smaller scene?

Also, I have used Retime nodes to offset the starting of each of the dust sequences for each card, should I use a timeoffset node instead?

I have to try and optimise this scene as it keeps bloody crashing on me and I really, REALLY, want to finish this shot tomorrow as it’s driving me mad! 😡 (And no finished shot will mean no beer for me tomorrow night, and by god after this week I could do with a few!)

Phew!

Any advice would be much appreciated.

No Responses to “Optimising a card setup”

Post a Comment