[osg-users] What's the difference between put rttcamera under root node and viewer?

[John Green]

Hi,
I'm trying to implement a rearview mirror in my car simulation project.So I need a rtt camera.And my scene consists of the osgearth node,some building models nodes,and a speed tree node.
I find two way to build a rtt camera,one is add it to a scene node,using addchild method;the other is add it to the viewer,using addslave method.
When the rttcamera is under a scene node,the osgearth node and the models nodes are rendered correctly,but the speedtree node is disapeared.(The speedtree node is rendered by its selfdefined drawable with opengl).
And when the rttcaera is added to the viewer,with useMastersSceneData been set true,the whole scene can be rendered correctly,however there are still some nodes which should not be seen in the rearview mirror,like the visual aid lines , points or texts.If I set the useMastersSceneData false,and add the osgearth node,builds nodes  and speedtree node into the rttcamera manually,then the buildings and trees are fine.But the osgearth's textures are disappeared,and the whole terrain turn into black.
So i am confused with the mechanism of the two kinds of rttcameras.I want to know the differences between them,and why the osgearth terrain turn into black when using a "addslave" rttcamera.
Thank you![/b]

------------------
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=65726#65726