Ok trying to get this sorted out in my head. What I'm trying to do is to make a system where I can zoom in to an object with the camera, .001 units away from it, and it will be a normal size.

There are lots of these objects, maybe 10 million etc. I want to zoom out to a zoom of around 1000 units away, and the pattern that all of these objects will form will be the original object.

Anyone have any ideas with copying etc that can make this a completely seamless process. Meaning that I should be able to zoom in and out of this “loop” an arbitrary amount of iterations from micro to macro in any direction, meaning getting bigger or getting smaller.

I've managed to get one level, but the problem is that since the object will be in a slightly different place there will be a “jump” any ideas?

no replies!! HELP! what kinda math do I have to do here How do I get the centroid to match the zoom at an “octave” of x6. Has anyone had to do anything similar?

I'm thinking I need some sort of method where I align pts vertically along the zoom-out axis, and then those are copy stamped with the correct level of detail according to how close the camera is, but not sure

Sorry I'm a bit confused about the problem. Are all the objects different sizes and therefore you don't know how to get each one to just fill the scene when you have the camera really close, or are you looking for a smooth way to automatically move the camera back and forth, or something else completely?

For the first problem the distance of the camera from the object centroid should I think just be a proportion of the maximum bounding box size. The proportion will be the same for all objects just manually select a value that looks good.

For the second problem you might want to check out this link -
http://forums.odforce.net/index.php?showtopic=7052&hl=dual+quaternion [forums.odforce.net]

It shows a nice way to smoothly blend between two positions and rotations. i sure there are other ways that are easier for you but this gives very nice results.

yeah looking for a good way to move the camera from a distance of .001 to 1000 and getting pretty much the same image. I want to be able to do this indefinatly. The big problem here is that the “loop” is in size/camera distance and not in time.

So, what I'm leaning toards right now is a system that basically attaches itself to the camera and will switch out single objects with million-pices object for certain ranges.

Sounds simle enough but the logic is really tricky; so if anyone has done anything similar I'd like to know about it

So your just looking for LOD switching? Can't you do that in the polyreduce sop. Or the LOD sop if its just displaying things which is the problem. Still slightly confused :?