Z-Test

THE POST BELOW IS MORE THAN 5 YEARS OLD. RELATED SUPPORT INFORMATION MIGHT BE OUTDATED OR DEPRECATED

On 01/08/2003 at 05:35, xxxxxxxx wrote:

User Information:
Cinema 4D Version:   8.100 
Platform:      
Language(s) :     C++  ;

---------
Hi
i wanted to upgrade my Pointcollapse Tool with a "only visible Points" mode
so i have the problem of checking whether a point is visible.
 
Does someone know if there is a way to use the Z-buffer for a visibility check.
BaseDraw::TestPointZ(..) seems only to check if the Point is within the nearclip/farclip range.
I could also use the RayCollider to check if there are polys in front of my point.  Hope this is not too slow.
Michael

THE POST BELOW IS MORE THAN 5 YEARS OLD. RELATED SUPPORT INFORMATION MIGHT BE OUTDATED OR DEPRECATED

On 01/08/2003 at 06:05, xxxxxxxx wrote:

Hi Michael,
Z-Buffer...hmm, doubt this would be useable but not sure.
With TestPointZ you are right, it doesn´t check for visibility somehow (maybe it´s buggy?)
Cannot help you with the RayCollider though for I haven´t used it. Mikael will be back on Monday, I guess he can tell you more then.
Good luck! (and let me know how you worked it out

THE POST BELOW IS MORE THAN 5 YEARS OLD. RELATED SUPPORT INFORMATION MIGHT BE OUTDATED OR DEPRECATED

On 02/08/2003 at 15:35, xxxxxxxx wrote:

A Z buffer test is for most purposes not a viable option - it's simply too imprecise. Also think of some gfx cards that do not allow to read GL Z values (though they have become rare).
The best way is to send in rays to detect which points are visible and which not.

THE POST BELOW IS MORE THAN 5 YEARS OLD. RELATED SUPPORT INFORMATION MIGHT BE OUTDATED OR DEPRECATED

On 03/08/2003 at 00:36, xxxxxxxx wrote:

thank you guys
....but, now i would really like to try that z-test thing out       I can hardly believe that it is too imprecise. Since it is precise enougth for drawing.
so i would do someting like 
visible = ( Abs(p.z - buffer[p.x][p.y]) < small_value ) 
i think it would be enougth to filter out points that are far behind the visible surface.
 
And for the ray test, there are two problems:  there are some strange perspective projetions where i don't know where to start the rays.  And i have to test every point against every poly... That is no good unless the point-polys-test is O(log n)
i will think some more about it..

THE POST BELOW IS MORE THAN 5 YEARS OLD. RELATED SUPPORT INFORMATION MIGHT BE OUTDATED OR DEPRECATED

On 04/08/2003 at 04:26, xxxxxxxx wrote:

ok i'm using the ray shooting method now.   It works quite well unless the mesh is not too big.

THE POST BELOW IS MORE THAN 5 YEARS OLD. RELATED SUPPORT INFORMATION MIGHT BE OUTDATED OR DEPRECATED

On 04/08/2003 at 17:18, xxxxxxxx wrote:

yes, I believe it is. But if it works..why not ;o)
IMO there should be a convenience function in the SDK that returns only the visible points, just like we can get the selected, hidden points.

THE POST BELOW IS MORE THAN 5 YEARS OLD. RELATED SUPPORT INFORMATION MIGHT BE OUTDATED OR DEPRECATED

On 07/08/2003 at 15:58, xxxxxxxx wrote:

you could try the C4DObjectList class
this seems to be the fastest way to get a z distance to a given point on screen.
you can use it  something like this
AutoAlloc<C4DObjectList> mylist;
SelectionListCreate(doc, NULL, bd, x, y, NULL, mylist);
you will need to convert your world coords into screen  coords.
this will put a list of all objects that can be found at the screen coord x,y into mylist
of course the first object is the nearest and you can read its z distance.
the advantage of this is that you dont need to consider which objects are in the scene or how many.
spline objects are passed as planes , so you should check through the list for the first valid polygon based object.
im not sure if this is the right method for you plugin , but its a very fast and  easy way to get z depth values and can be abused in many ways.

THE POST BELOW IS MORE THAN 5 YEARS OLD. RELATED SUPPORT INFORMATION MIGHT BE OUTDATED OR DEPRECATED

On 08/08/2003 at 00:16, xxxxxxxx wrote:

that is not necessary for my plugin since it should only consider "self" occlusion of the active object. Similar to the live selection tool.

THE POST BELOW IS MORE THAN 5 YEARS OLD. RELATED SUPPORT INFORMATION MIGHT BE OUTDATED OR DEPRECATED

On 08/08/2003 at 04:28, xxxxxxxx wrote:

You can't access C4D's Z buffer. There are lots of reasons it's not very practical to do so: the Z buffer may be in unititialized state, the Z scaling is different for the various display types, some GL buffers don't support Z access etc.
Finally, as beaker said, its so imprecise. If C4D would use this for selection, you couldn't work with it. (All according to the developers.)

THE POST BELOW IS MORE THAN 5 YEARS OLD. RELATED SUPPORT INFORMATION MIGHT BE OUTDATED OR DEPRECATED

On 08/08/2003 at 05:33, xxxxxxxx wrote:

aha.  then i was wrong i guess. but never mind. i build in ray shooting anyway. It was very easy with the raycollider.  It's slow with large meshes but guess i can limit the number of point's that must be checked... in a future version.