get data from back in time

[forton]

Hi,

I've created some targetpoints for my strands that set themselves to the closest location of some objects, when another object is getting nearer the targets jump to the other object.
So far no problem, but what I want to do now is having a smooth interpolation between position A and position B, so they travel in -let say 10 frames- from A to B, this would be easy if I could store their positions from their last 10 frames and interpolate, in fact I've managed this with a bit of a hack (see pic), but at best the result is unpredictable.

Now my question: is there a correct way to store values when on certain frames and retrieve them on other frames when needed?
Is there something like "get data from 20 frames ago"?
I want to use this in a non simulated ice tree, but if I have to I'll use a simulated one.

Thanks for your suggestions!

Wim

[owei]

Hi!

If I get you correctly, this tree should do what you want. Just store the data you need in an seperate array and set the new data to the correct position/frame in that array. Element index is equal frame number. Hope this helps!

Cheers,
oli

[forton]

Thanks Oli,

This seems to work already much better, still two problems I have to fix, the values are not stored if it's a non-simulated ice tree, so I think I'm forced to use simulated instead

Another one is this: I want to have particles at the end of my strands, so I ad points to another pointcloud from this one, if I play everything is fine but when I render the points are like one frame behind the strands end, strange thing is that in another renderpass (with a simple raylength shader on everything) the ends have no lag.
I tried creating the particles into a non simulated, simulated (doesn't work because new points are added on top of the other) and post-simulated pointcloud, same problem.

Anyone knows the correct way to do this?

[ChrB]

Do you cache your particles before rendering them? If not, recommended to do so and remember to set the play mode to standard in your simulation time control. More here: http://softimage.wiki.softimage.com/xsi ... lation.htm

[forton]

Thanks guys, caching is the way to go, tried it before and it seemed overly complex, the big isue i have with softimage and many other software is how dificult it is to find some decent info and -even better- examples on how to do things, luckily there's things like vimeo and this forum to get things going!

this is what I get for the moment:



Cheers!

[forton]

PS: the only thing that's animated is the ball

[owei]

Wow! Nice one!!
Yes, chaching is the solution for non simulated stuff..(I over read the "non"..sorry)..
If you want to stuff like this in non simulated tree, you will have to do all the calculations on your "own" (positions, velocities etc.) based on the current frame. I guess, that would be possible, but quite unhandy to built ;)

How did you achieve the dynamics on the strands..? Which tool did you use? Looks nice ;)

cheers,
oli

[forton]

The basic workflow is this:
create a string of strands oriented to the target (closest position on targetgeo from randomized startposition)
then I did some deformations with some f-curves (using the strandratio) depending on where and what I wanted to achieve:
Beginning and end blend with the normals of the source and target object, in the middle there's some animated turbulence around value, no dynamics there.

Arrays used to be a bit of chinese for me (that's how we say it in dutch), but good examples can bring you a long way!

Cheers, Wim

[olivier jeannel]

This is beautifull !
It's cool to have an xsi studio based in belgium !