Flag respecting parent movement of flag pole without using physics

[anupamd]

Ok so Im trying to have a flag that is waved around (think of it like a checkered flag on a race car finish line).

I dont want to use cloth, and I dont want to use particles. I think I have a technique that will work.

Basically I need to look at the point position at a previous frame and weight that point position by say the U value of the Flag UVs. That way the base sticks to the pole and the tip "lags" behind a few frames.

My problem is how do I get access to point position from a previous frame or from "N frames back"? Once I have that, I can multiply that by a weighting value and Im pretty much done.

Any suggestions or samples would be appreciated.

[J.Santos]

This might do the trick for you:

http://www.sidefx.com/docs/houdini9.5/nodes/sop/spring

If you scroll down you'll notice that the Spring SOP has an example with a flag...

Hope this helps!

[glfloat]

Thanks for the response.

The spring sop uses simulation.. its basically a spring solver in sop level.

The technique Im talking about is basically a lag or delay on the lookup of the point position a few frames back based on a weighting value. With that technique there is no simulation required and is very fast.

Its a common technique used in motion graphics.

I suppose my question boils down to how to look up position of points n-frames back. The point expression does not allow you to get the point position at a certain frame. So I'd have to use chf or something else.

Any other suggestions?

[Marc]

lag CHOP would be my first thought.

[old school]

If you want to use the previous frame's state to derive the current frame's state, no matter how you slice it, it's a simulation.

This opens up the door to the Spring SOP as mentioned above.

CHOPs as Marc mentioned is a bit of a different beast but too caches the flag pole's animation where you can derive motion on the geometry. If you use CHOPs in timeslice mode, then it too becomes like a simulation where those CHOPs that require to look a bit in to the past (Lag CHOP, Spring CHOP, etc.) will do so but not force the entire simulation to cook.

There is the Solver SOP that will give you direct access to the previous frame's motion but again this is a simulation where the current frame is dependent on the previous frame's state.