How to perform a box selection?

I'm reading the PhysX tutorials, and I see the following intersections:


1. NxScene::overlapSphereShapes() - test for shapes in the scene intersecting a sphere
2. NxScene::overlapAABBShapes() - test for shapes in the scene intersecting an AABB
3. NxScene::cullShapes() - test for shapes in the scene culled inside a set of planes (you can use this for view frustum culling)

4. NxScene::checkOverlapSphere() - test to see if the sphere intersects any shapes in the scene
5. NxScene::checkOverlapAABB() - test to see if the AABB intersects any shapes in the scene


I'm assuming AABB is Axis Aligned Bounding Box. I want to project a cube from the camera, so that I can make a box selection using PhysX. (Grabbing objects by their Shape and not by Ogre::MovableObject)

Any ideas how to do this?

Via the intersection class my friend, or if you just want a one off piece of code then Scene::getActorsFromRegion and Scene::getActorsFromLastRegion will do the same job.

Thanks for the direction.

For anybody else who wants a quick reference to the code:


SimpleBox* shape = static_cast<SimpleBox*>(mShape);
NxBox box = NxBox(shape->getPose().t, shape->getDimensions(), shape->getPose().M);
mScene->getNxScene()->overlapOBBShapes(box, mShapeFilter, 0, NULL, this);


This is easier, and more in line with NxOgre.

mScene->getActorsFromRegion(mShape);

Has anybody ever done this? The problem with this is that I don't know how to set the dimensions and position of the SimpleBox shape I want to use.

The orientation of the box (I think) is the same as the camera's. Problems:

I use a 2d rectangular box to outline the selection. I can get the dimensions of the 2d box, ie (0.5,0.5), but I don't know how to convert this to Ogre units. Something to do with the camera or frustum...
I'm guessing the 2d width is the x dimensions of the box, and the 2d height is the y dimensions of the box. The z dimensions is the box as it goes into the screen. Not sure what value to use here, I could make something up like 2000.

Any help with this would be greatly appreciated.

Hi,

I also need to perform a multi selection using a rectangle on the screen.

To get the rectangle in 3d world i can use the getCameraToViewportRay from Ogre for the 4 corners of the rectangle.

But how can i now get all the actors AND characters in this region ?

Thank you very much

P.S.: I am using 0.9

no ideas ????????

Characters also use an Actor, though I´m not sure if getActorsFromRegion() filters out the Characteractors as Raycaster does.
The other way would be a manual approach by using the Characters positions.

This is what I use in Cake Bleeding, it's slightly buggy but for the most part works.

if (dragMode == DRAG_NONE && mKeyboard->isKeyDown(KC_LSHIFT) && mMouse->getMouseState().buttonDown(OIS::MB_Left)) {
mDragger->setXY(mGUI->mMousePointer->getLeft(), mGUI->mMousePointer->getTop());
mDragger->show();

dragMode = DRAG_SIZE;
return;
}

if (dragMode == DRAG_SIZE && mMouse->getMouseState().buttonDown(OIS::MB_Left)) {
mDragger->setWH(mGUI->mMousePointer->getLeft(), mGUI->mMousePointer->getTop());
}
else if (dragMode == DRAG_SIZE) {

mDragger->hide();
dragMode = DRAG_NONE;

Ogre::Vector3 pos = mCamera->getDerivedPosition();
Ogre::Quaternion quat = mCamera->getDerivedOrientation();

NxReal width = widthTo3D(mCamera, mDragger->mOverlayContainer->getLeft(), mDragger->mOverlayContainer->getWidth());
NxReal height = heightTo3D(mCamera, mDragger->mOverlayContainer->getTop(), mDragger->mOverlayContainer->getHeight());

std::cout << "Width => " << width << ", Height => " << height << std::endl;
NxReal length = 10;
Ogre::Vector3 direction(width, height, length);

// Place the intersection half way through.
pos += (quat * Vector3(0,0,-length * 0.5f));

if (intersection == 0) {
sbox = new SimpleBox(direction * 0.5f, NxOgre::Pose(pos, quat));
intersection = new Intersection(mScene, sbox);

inode = mSceneMgr->getRootSceneNode()->createChildSceneNode("inode");
Entity* ent = mSceneMgr->createEntity("ient", "cube.1m.mesh");
ent->setMaterialName("Drag");
inode->attachObject(ent);
inode->setPosition(pos);
inode->setOrientation(quat);
inode->setScale(direction);
}
else {
sbox->setDimensions(NxVec3(direction.x, direction.y, direction.z));
sbox->setPose(NxOgre::Pose(pos, quat));

inode->setPosition(pos);
inode->setOrientation(quat);
inode->setScale(direction);
}

intersection->intersect();
for (Actor* actor = intersection->begin(); actor = intersection->next();) {
std::cout << actor->getName() << "\n";
quickSetMaterial(actor, "gloss-red");
}

}

void quickSetMaterial(Actor* actor, const NxString& mat) {
if (actor->getUserData()->RenderPtr == 0)
return;

// Dangerous.....
RenderableSource* rs = actor->getUserData()->RenderPtr;
OgreNodeRenderable* nr = static_cast<OgreNodeRenderable*>(rs->getRenderable());
nr->setMaterial(mat);

}

Ogre::Vector3 screenTo3D(Ogre::Camera *cam, float x = 0.5, float y = 0.5) {
Ogre::Ray ray = cam->getCameraToViewportRay(x, y);
return ray.getOrigin();
}

float widthTo3D(Ogre::Camera *cam, float x0, float x1) {
return (screenTo3D(cam, x0, 0.5) - screenTo3D(cam, x1, 0.5)).length();
}

float heightTo3D(Ogre::Camera *cam, float y0, float y1) {
return (screenTo3D(cam, 0.5, y0) - screenTo3D(cam, 0.5, y0)).length();
}
"mDragger" is just a wrapped BetaGUI window, you don't need to know the implementation.

Unless I'm doing this wrong, I think PhysX has a bug with its cullShapes function, regarding filtering. All other ray casting and intersection queries I've done so far works with filtering correctly. Here is my code:


//do volumequery
Ogre::Ray topLeft = camera->getCameraToViewportRay(sides[SelectionRectangle::LEFT], sides[SelectionRectangle::TOP]);
Ogre::Ray topRight = camera->getCameraToViewportRay(sides[SelectionRectangle::RIGHT], sides[SelectionRectangle::TOP]);
Ogre::Ray bottomLeft = camera->getCameraToViewportRay(sides[SelectionRectangle::LEFT], sides[SelectionRectangle::BOTTOM]);
Ogre::Ray bottomRight = camera->getCameraToViewportRay(sides[SelectionRectangle::RIGHT], sides[SelectionRectangle::BOTTOM]);

NxPlane worldPlanes[5];
NxVec3 point1, point2, point3;

// front plane
point1 = NxVec3(topLeft.getPoint(3).x, topLeft.getPoint(3).y, topLeft.getPoint(3).z);
point2 = NxVec3(topRight.getPoint(3).x, topRight.getPoint(3).y, topRight.getPoint(3).z);
point3 = NxVec3(bottomRight.getPoint(3).x, bottomRight.getPoint(3).y, bottomRight.getPoint(3).z);
worldPlanes[0] = NxPlane(point3, point2, point1);

// top plane
point1 = NxVec3(topLeft.getOrigin().x, topLeft.getOrigin().y, topLeft.getOrigin().z);
point2 = NxVec3(topLeft.getPoint(100).x, topLeft.getPoint(100).y, topLeft.getPoint(100).z);
point3 = NxVec3(topRight.getPoint(100).x, topRight.getPoint(100).y, topRight.getPoint(100).z);
worldPlanes[1] = NxPlane(point3, point2, point1);

// left plane
point1 = NxVec3(topLeft.getOrigin().x, topLeft.getOrigin().y, topLeft.getOrigin().z);
point2 = NxVec3(bottomLeft.getPoint(100).x, bottomLeft.getPoint(100).y, bottomLeft.getPoint(100).z);
point3 = NxVec3(topLeft.getPoint(100).x, topLeft.getPoint(100).y, topLeft.getPoint(100).z);
worldPlanes[2] = NxPlane(point3, point2, point1);

// bottom plane
point1 = NxVec3(topLeft.getOrigin().x, topLeft.getOrigin().y, topLeft.getOrigin().z);
point2 = NxVec3(bottomRight.getPoint(100).x, bottomRight.getPoint(100).y, bottomRight.getPoint(100).z);
point3 = NxVec3(bottomLeft.getPoint(100).x, bottomLeft.getPoint(100).y, bottomLeft.getPoint(100).z);
worldPlanes[3] = NxPlane(point3, point2, point1);

// right plane
point1 = NxVec3(topLeft.getOrigin().x, topLeft.getOrigin().y, topLeft.getOrigin().z);
point2 = NxVec3(topRight.getPoint(100).x, topRight.getPoint(100).y, topRight.getPoint(100).z);
point3 = NxVec3(bottomRight.getPoint(100).x, bottomRight.getPoint(100).y, bottomRight.getPoint(100).z);
worldPlanes[4] = NxPlane(point3, point2, point1);

NxU32 nbShapes = mNxScene->getNbDynamicShapes() + mNxScene->getNbStaticShapes();
NxShape** shapes = new NxShape* [nbShapes];
for (NxU32 i = 0; i < nbShapes; i++)
shapes[i] = NULL;

this->getNxScene()->cullShapes(5, worldPlanes, NX_ALL_SHAPES, nbShapes, shapes, NULL, collisionMask);

for (NxU32 i = 0; i < nbShapes; i++)
if(shapes[i])
selectedObjects.push_back(static_cast<Gaia::GameObject*>(shapes[i]->getActor().userData));


The only collision groups I have are '1', '2', and '3'. The mask as used in the code is '4', which is (1 << 2). Unfortunately all shapes are being returned, regardless of my selection size. I'll probably check out the physx forums and post there, unless somebody here spots anything wrong with my code.

I was never sent an activation email, and I can't post on any relevant PhysX forums. Can anybody help me out, or at least post this issue and bring it to their attention? I'm fairly certain there is a bug, where the function "cullShapes" does not apply the collision filter.