Let’s call the parent of the target element the “branch element” and the actual target element the “leaf element.” We will also refer to the viewport as the “root.”

The branch element’s transform must be recalculated relative to the root’s transform. The leaf element’s transform must be recalculated relative to the branch element’s transform. Perhaps there are multiple leaf elements that we want to leave unchanged. One solution is to apply the calculations that would occur between the root and branch element to the leaf element, as well as the branch and leaf element calculations. In this scenario, we take two passes to compute the target leaf element’s transform, one with the root and branch element’s transforms, and one with the branch and *the original* target leaf element’s transform.

Sample algorithm:
Create a copy of the target leaf element’s transform.
Apply root to branch calculations to the copied transform.
Apply branch to leaf calculations using the original leaf element’s transform to the copied transform.
Set the target leaf element’s transform to the copied transform.
Force render if needed.

Hope this helps,
Melih

Any help would be much appreciated…

Thanks

callLater(fileRef.download, new Array(urlRequest));

Different Flex SDK/Flash IDE version and Flash Player version combinations can produce different results… Try creating a unit test for the process(es), or isolating the code in a clean project and debugging it with traces.

Thanks a lot !

I am posting the answer to your question so that others will have the answer in case they have the same question.

The math for projection in papervision 2.0 is a matter of finding the perspective of the camera object. Once this has been achieved, one can project a 3D object to the screen by doing some basic math on the coordinates of the 3D object in camera space.

The following equation assumes the coordinates are relative to the coordinate space of the camera. This can be achieved by multiplying the objects world transform with the camera transform, but I believe there is a public var for 3D objects in papervision that has this transform already computed. The trick is to look out for optimizations. The public var camera transform may not be updated unless that object has been rendered. I believe this property is screen, but this may have changed or may change with the upcoming version of papervision.

Assume in the following code snippet that cam is the camera object, object is the 3D object, and screen is the objects transform in camera space.

var perspective = cam.focus*cam.zoom/screen.z;
var x = object.x*perspective;
var y = object.y*perspective;

That’s it!

The math for this is bit more involved when using a projection matrix.