source: downloads/ogre/OgreMain/include/OgreShadowCameraSetupFocused.h @ 33

/*
-----------------------------------------------------------------------------
This source file is part of OGRE
(Object-oriented Graphics Rendering Engine)
For the latest info, see http://www.ogre3d.org/

Copyright (c) 2006 Torus Knot Software Ltd
Copyright (c) 2006 Matthias Fink, netAllied GmbH <matthias.fink@web.de>                                                         
Also see acknowledgements in Readme.html

This program is free software; you can redistribute it and/or modify it under
the terms of the GNU Lesser General Public License as published by the Free Software
Foundation; either version 2 of the License, or (at your option) any later
version.

This program is distributed in the hope that it will be useful, but WITHOUT
ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with
this program; if not, write to the Free Software Foundation, Inc., 59 Temple
Place - Suite 330, Boston, MA 02111-1307, USA, or go to
http://www.gnu.org/copyleft/lesser.txt.

You may alternatively use this source under the terms of a specific version of
the OGRE Unrestricted License provided you have obtained such a license from
Torus Knot Software Ltd.
-----------------------------------------------------------------------------
*/
#ifndef __ShadowCameraSetupFocused_H__
#define __ShadowCameraSetupFocused_H__

#include "OgrePrerequisites.h"
#include "OgreShadowCameraSetup.h"
#include "OgrePolygon.h"
#include "OgreConvexBody.h"


namespace Ogre {

        class ConvexBody;

        /** Implements the uniform shadow mapping algorithm in focused mode.
        @remarks
                Differs from the default shadow mapping projection in that it focuses the
                shadow map on the visible areas of the scene. This results in better
                shadow map texel usage, at the expense of some 'swimming' of the shadow
                texture on receivers as the basis is constantly being reevaluated.
        @note
                Original implementation by Matthias Fink <matthias.fink@web.de>, 2006.
        */
        class _OgreExport FocusedShadowCameraSetup : public ShadowCameraSetup
        {
        protected:
                /** Transform to or from light space as defined by Wimmer et al.
                @remarks
                Point and spot lights need to be converted to directional lights to enable a 1:1 
                light mapping. Otherwise a directional light may become a point light or a point 
                sink (opposite of a light source) or point/spot lights may become directional lights
                or light sinks. The light direction is always -y.
                */
                static const Matrix4 msNormalToLightSpace;
                static const Matrix4 msLightSpaceToNormal;

                /** Temporary preallocated frustum to set up a projection matrix in 
                ::calculateShadowMappingMatrix()
                */
                Frustum* mTempFrustum;

                /** Temporary preallocated camera to set up a light frustum for clipping in ::calculateB.
                */
                Camera* mLightFrustumCamera;
                mutable bool mLightFrustumCameraCalculated;

                /// Use tighter focus region?
                bool mUseAggressiveRegion;

                /** Internal class holding a point list representation of a convex body.
                */
                class _OgreExport PointListBody
                {
                        Polygon::VertexList mBodyPoints;
                        AxisAlignedBox          mAAB;

                public:
                        PointListBody();
                        PointListBody(const ConvexBody& body);
                        ~PointListBody();

                        /** Merges a second PointListBody into this one.
                        */
                        void merge(const PointListBody& plb);

                        /** Builds a point list body from a 'real' body.
                        @remarks
                        Inserts all vertices from a body into the point list with or without adding duplicate vertices.
                        */
                        void build(const ConvexBody& body, bool filterDuplicates = true);

                        /** Builds a PointListBody from a Body and includes all the space in a given direction.
                        @remarks
                        Intersects the bounding box with a ray from each available point of the body with the given
                        direction. Base and intersection points are stored in a PointListBody structure.
                        @note
                        Duplicate vertices are not filtered.
                        @note
                        Body is not checked for correctness.
                        */
                        void buildAndIncludeDirection(const ConvexBody& body, 
                                const AxisAlignedBox& aabMax, const Vector3& dir);

                        /** Returns the bounding box representation.
                        */
                        const AxisAlignedBox& getAAB(void) const;       

                        /** Adds a specific point to the body list.
                        */
                        void addPoint(const Vector3& point);

                        /** Adds all points of an AAB.
                        */
                        void addAAB(const AxisAlignedBox& aab);

                        /** Returns a point.
                        */
                        const Vector3& getPoint(size_t cnt) const;

                        /** Returns the point count.
                        */
                        size_t getPointCount(void) const;

                        /** Resets the body.
                        */
                        void reset(void);

                };

                // Persistent calculations to prevent reallocation
                mutable ConvexBody mBodyB;
                mutable PointListBody mPointListBodyB;
                mutable PointListBody mPointListBodyLVS;

        protected:
                /** Calculates the standard shadow mapping matrix.
                @remarks
                Provides the view and projection matrix for standard shadow mapping.
                @note
                You can choose which things you want to have: view matrix and/or projection 
                matrix and/or shadow camera. Passing a NULL value as parameter ignores the
                generation of this specific value.
                @param sm: scene manager
                @param cam: currently active camera
                @param light: currently active light
                @param out_view: calculated uniform view shadow mapping matrix (may be NULL)
                @param out_proj: calculated uniform projection shadow mapping matrix (may be NULL)
                @param out_cam: calculated uniform shadow camera (may be NULL)
                */
                void calculateShadowMappingMatrix(const SceneManager& sm, const Camera& cam, 
                        const Light& light, Matrix4 *out_view, 
                        Matrix4 *out_proj, Camera *out_cam) const;

                /** Calculates the intersection bodyB.
                @remarks
                The intersection bodyB consists of the concatenation the cam frustum clipped 
                by the scene bounding box followed by a convex hullification with the light's 
                position and the clipping with the scene bounding box and the light frustum:
                ((V \cap S) + l) \cap S \cap L (\cap: convex intersection, +: convex hull 
                operation).
                For directional lights the bodyB is assembled out of the camera frustum 
                clipped by the scene bounding box followed by the extrusion of all available 
                bodyB points towards the negative light direction. The rays are intersected 
                by a maximum bounding box and added to the bodyB points to form the final 
                intersection bodyB point list.
                @param sm: scene manager
                @param cam: currently active camera
                @param light: currently active light
                @param sceneBB: scene bounding box for clipping operations
                @param out_bodyB: final intersection bodyB point list
                */
                void calculateB(const SceneManager& sm, const Camera& cam, const Light& light, 
                        const AxisAlignedBox& sceneBB, PointListBody *out_bodyB) const;

                /** Calculates the bodyLVS.
                @remarks
                Calculates the bodyLVS which consists of the convex intersection operation 
                affecting the light frustum, the view frustum, and the current scene bounding
                box is used to find suitable positions in the viewer's frustum to build the 
                rotation matrix L_r. This matrix is applied after the projection matrix L_p to 
                avoid an accidental flip of the frustum orientation for views tilted with 
                respect to the shadow map.
                @param scene: holds all potential occluders / receivers as one single bounding box
                of the currently active scene node
                @param cam: current viewer camera
                @param light: current light
                @param out_LVS: intersection body LVS (world coordinates)
                */
                void calculateLVS(const SceneManager& sm, const Camera& cam, const Light& light,
                        const AxisAlignedBox& sceneBB, PointListBody *out_LVS) const;

                /**     Returns the projection view direction.
                @remarks
                After the matrix L_p is applied the orientation of the light space may tilt for
                non-identity projections. To prevent a false shadow cast the real view direction
                is evaluated and applied to the light matrix L.
                @param lightSpace: matrix of the light space transformation
                @param cam: current viewer camera
                @param bodyLVS: intersection body LVS (relevant space in front of the camera)
                */
                Vector3 getLSProjViewDir(const Matrix4& lightSpace, const Camera& cam, 
                        const PointListBody& bodyLVS) const;

                /** Returns a valid near-point seen by the camera.
                @remarks
                Returns a point that is situated near the camera by analyzing the bodyLVS that
                contains all the relevant scene space in front of the light and the camera in
                a point list array. The view matrix is relevant because the nearest point in
                front of the camera should be determined.
                @param viewMatrix: view matrix of the current camera
                @param bodyLVS: intersection body LVS (relevant space in front of the camera) 
                */
                Vector3 getNearCameraPoint_ws(const Matrix4& viewMatrix, 
                        const PointListBody& bodyLVS) const;

                /** Transforms a given body to the unit cube (-1,-1,-1) / (+1,+1,+1) with a specific 
                shadow matrix enabled.
                @remarks
                Transforms a given point list body object with the matrix m and then maps its
                extends to a (-1,-1,-1) / (+1,+1,+1) unit cube
                @param m: transformation matrix applied on the point list body
                @param body: contains the points of the extends of all valid scene elements which 
                are mapped to the unit cube
                */
                Matrix4 transformToUnitCube(const Matrix4& m, const PointListBody& body) const;

                /** Builds a view matrix.
                @remarks
                Builds a standard view matrix out of a given position, direction and up vector.
                */
                Matrix4 buildViewMatrix(const Vector3& pos, const Vector3& dir, const Vector3& up) const;

        public:
                /** Default constructor.
                @remarks
                Temporary frustum and camera set up here.
                */
                FocusedShadowCameraSetup(void);

                /** Default destructor.
                @remarks
                Temporary frustum and camera destroyed here.
                */
                virtual ~FocusedShadowCameraSetup(void);

                /** Returns a uniform shadow camera with a focused view.
                */
                virtual void getShadowCamera(const SceneManager *sm, const Camera *cam, 
                        const Viewport *vp, const Light *light, Camera *texCam) const;

                /** Sets whether or not to use the more agressive approach to deciding on
                        the focus region or not.
                @note
                        There are 2 approaches that can  be used to define the focus region,
                        the more aggressive way introduced by Wimmer et al, or the original
                        way as described in Stamminger et al. Wimmer et al's way tends to 
                        come up with a tighter focus region but in rare cases (mostly highly
                        glancing angles) can cause some shadow casters to be clipped 
                        incorrectly. By default the more aggressive approach is used since it
                        leads to significantly better results in most cases, but if you experience
                        clipping issues, you can use the less agressive version.
                @param aggressive True to use the more agressive approach, false otherwise.
                */
                void setUseAggressiveFocusRegion(bool aggressive) { mUseAggressiveRegion = aggressive; }

                bool getUseAggressiveFocusRegion() const { return mUseAggressiveRegion; }

        };


}

#endif