This repository has been archived on 2023-10-09. You can view files and clone it. You cannot open issues or pull requests or push a commit.
Files
blender-archive/source/blender/freestyle/intern/view_map/CulledOccluderSource.cpp

280 lines
11 KiB
C++
Raw Normal View History

/*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU 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 General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software Foundation,
* Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
*/
/** \file
* \ingroup freestyle
* \brief Class to define a cell grid surrounding the projected image of a scene
*/
Optimized view map calculation by Alexander Beels. * View map calculation has been intensively optimized for speed by means of: 1) new spatial grid data structures (SphericalGrid for perspective cameras and BoxGrid for orthographic cameras; automatically switched based on the camera type); 2) a heuristic grid density calculation algorithm; and 3) new line visibility computation algorithms: A "traditional" algorithm for emulating old visibility algorithms, and a "cumulative" algorithm for improved, more consistent line visibility, both exploiting the new spatial grid data structures for fast ray casting. A new option "Raycasting Algorithm" was added to allow users to choose a ray casting (line visibility) algorithm. Available choices are: - Normal Ray Casting - Fast Ray Casting - Very Fast Ray Casting - Culled Traditional Visibility Detection - Unculled Traditional Visibility Detection - Culled Cumulative Visibility Detection - Unculled Cumulative Visibility Detection The first three algorithms are those available in the original Freestyle (the "normal" ray casting was used unconditionally, though). The "fast" and "very fast" ray casting algorithms achieve a faster calculation at the cost of less visibility accuracy. The last four are newly introduced optimized options. The culled versions of the new algorithms will exclude from visibility calculation those faces that lay outside the camera, which leads to a faster view map construction. The unculled counterparts will take all faces into account. The unculled visibility algorithms are useful when culling affects stroke chaining. The recommended options for users are the culled/unculled cumulative visibility algorithms. These options are meant to replace the old algorithms in the future. Performance improvements over the old algorithms depend on the scenes to be rendered. * Silhouette detection has also been considerably optimized for speed. Performance gains by this optimization do not depend on scenes. * Improper handling of error conditions in the view map construction was fixed.
2011-03-14 00:36:27 +00:00
#include "CulledOccluderSource.h"
#include "../geometry/GridHelpers.h"
#include "BKE_global.h"
Attempt to fix a potential name conflict between Freestyle and the compositor. A crash in the Freestyle renderer was reported by Ton on IRC with a stack trace below. Note that #2 is in Freestyle, whereas #1 is in the compositor. The problem was observed in a debug build on OS X 10.7 (gcc 4.2, openmp disabled, no llvm). ---------------------------------------------------------------------- Program received signal EXC_BAD_ACCESS, Could not access memory. Reason: 13 at address: 0x0000000000000000 [Switching to process 72386 thread 0xf303] 0x0000000100c129f3 in NodeBase::~NodeBase (this=0x10e501c80) at COM_NodeBase.cpp:43 43 delete (this->m_outputsockets.back()); Current language: auto; currently c++ (gdb) where #0 0x0000000100c129f3 in NodeBase::~NodeBase (this=0x10e501c80) at COM_NodeBase.cpp:43 #1 0x0000000100c29066 in Node::~Node (this=0x10e501c80) at COM_Node.h:49 #2 0x000000010089c273 in NodeShape::~NodeShape (this=0x10e501c80) at NodeShape.cpp:43 #3 0x000000010089910b in NodeGroup::destroy (this=0x10e501da0) at NodeGroup.cpp:61 #4 0x00000001008990cd in NodeGroup::destroy (this=0x10e5014b0) at NodeGroup.cpp:59 #5 0x00000001008990cd in NodeGroup::destroy (this=0x114e18da0) at NodeGroup.cpp:59 #6 0x00000001007e6602 in Controller::ClearRootNode (this=0x114e19640) at Controller.cpp:329 #7 0x00000001007ea52e in Controller::LoadMesh (this=0x114e19640, re=0x10aba4638, srl=0x1140f5258) at Controller.cpp:302 #8 0x00000001008030ad in prepare (re=0x10aba4638, srl=0x1140f5258) at FRS_freestyle.cpp:302 #9 0x000000010080457a in FRS_do_stroke_rendering (re=0x10aba4638, srl=0x1140f5258) at FRS_freestyle.cpp:600 #10 0x00000001006aeb9d in add_freestyle (re=0x10aba4638) at pipeline.c:1584 #11 0x00000001006aceb7 in do_render_3d (re=0x10aba4638) at pipeline.c:1094 #12 0x00000001006ae061 in do_render_fields_blur_3d (re=0x10aba4638) at pipeline.c:1367 #13 0x00000001006afa16 in do_render_composite_fields_blur_3d (re=0x10aba4638) at pipeline.c:1815 #14 0x00000001006b04e4 in do_render_all_options (re=0x10aba4638) at pipeline.c:2021 ---------------------------------------------------------------------- Apparently a name conflict between the two Blender modules is taking place. The present commit hence intends to address it by putting all the Freestyle C++ classes in the namespace 'Freestyle'. This revision will also prevent potential name conflicts with other Blender modules in the future. Special thanks to Lukas Toenne for the help with C++ namespace.
2013-04-09 00:46:49 +00:00
namespace Freestyle {
CulledOccluderSource::CulledOccluderSource(const GridHelpers::Transform &t,
WingedEdge &we,
ViewMap &viewMap,
bool extensiveFEdgeSearch)
: OccluderSource(t, we), rejected(0), gridSpaceOccluderProsceniumInitialized(false)
Optimized view map calculation by Alexander Beels. * View map calculation has been intensively optimized for speed by means of: 1) new spatial grid data structures (SphericalGrid for perspective cameras and BoxGrid for orthographic cameras; automatically switched based on the camera type); 2) a heuristic grid density calculation algorithm; and 3) new line visibility computation algorithms: A "traditional" algorithm for emulating old visibility algorithms, and a "cumulative" algorithm for improved, more consistent line visibility, both exploiting the new spatial grid data structures for fast ray casting. A new option "Raycasting Algorithm" was added to allow users to choose a ray casting (line visibility) algorithm. Available choices are: - Normal Ray Casting - Fast Ray Casting - Very Fast Ray Casting - Culled Traditional Visibility Detection - Unculled Traditional Visibility Detection - Culled Cumulative Visibility Detection - Unculled Cumulative Visibility Detection The first three algorithms are those available in the original Freestyle (the "normal" ray casting was used unconditionally, though). The "fast" and "very fast" ray casting algorithms achieve a faster calculation at the cost of less visibility accuracy. The last four are newly introduced optimized options. The culled versions of the new algorithms will exclude from visibility calculation those faces that lay outside the camera, which leads to a faster view map construction. The unculled counterparts will take all faces into account. The unculled visibility algorithms are useful when culling affects stroke chaining. The recommended options for users are the culled/unculled cumulative visibility algorithms. These options are meant to replace the old algorithms in the future. Performance improvements over the old algorithms depend on the scenes to be rendered. * Silhouette detection has also been considerably optimized for speed. Performance gains by this optimization do not depend on scenes. * Improper handling of error conditions in the view map construction was fixed.
2011-03-14 00:36:27 +00:00
{
cullViewEdges(viewMap, extensiveFEdgeSearch);
Optimized view map calculation by Alexander Beels. * View map calculation has been intensively optimized for speed by means of: 1) new spatial grid data structures (SphericalGrid for perspective cameras and BoxGrid for orthographic cameras; automatically switched based on the camera type); 2) a heuristic grid density calculation algorithm; and 3) new line visibility computation algorithms: A "traditional" algorithm for emulating old visibility algorithms, and a "cumulative" algorithm for improved, more consistent line visibility, both exploiting the new spatial grid data structures for fast ray casting. A new option "Raycasting Algorithm" was added to allow users to choose a ray casting (line visibility) algorithm. Available choices are: - Normal Ray Casting - Fast Ray Casting - Very Fast Ray Casting - Culled Traditional Visibility Detection - Unculled Traditional Visibility Detection - Culled Cumulative Visibility Detection - Unculled Cumulative Visibility Detection The first three algorithms are those available in the original Freestyle (the "normal" ray casting was used unconditionally, though). The "fast" and "very fast" ray casting algorithms achieve a faster calculation at the cost of less visibility accuracy. The last four are newly introduced optimized options. The culled versions of the new algorithms will exclude from visibility calculation those faces that lay outside the camera, which leads to a faster view map construction. The unculled counterparts will take all faces into account. The unculled visibility algorithms are useful when culling affects stroke chaining. The recommended options for users are the culled/unculled cumulative visibility algorithms. These options are meant to replace the old algorithms in the future. Performance improvements over the old algorithms depend on the scenes to be rendered. * Silhouette detection has also been considerably optimized for speed. Performance gains by this optimization do not depend on scenes. * Improper handling of error conditions in the view map construction was fixed.
2011-03-14 00:36:27 +00:00
// If we have not found any visible FEdges during our cull, then there is nothing to iterate
// over. Short-circuit everything.
valid = gridSpaceOccluderProsceniumInitialized;
Optimized view map calculation by Alexander Beels. * View map calculation has been intensively optimized for speed by means of: 1) new spatial grid data structures (SphericalGrid for perspective cameras and BoxGrid for orthographic cameras; automatically switched based on the camera type); 2) a heuristic grid density calculation algorithm; and 3) new line visibility computation algorithms: A "traditional" algorithm for emulating old visibility algorithms, and a "cumulative" algorithm for improved, more consistent line visibility, both exploiting the new spatial grid data structures for fast ray casting. A new option "Raycasting Algorithm" was added to allow users to choose a ray casting (line visibility) algorithm. Available choices are: - Normal Ray Casting - Fast Ray Casting - Very Fast Ray Casting - Culled Traditional Visibility Detection - Unculled Traditional Visibility Detection - Culled Cumulative Visibility Detection - Unculled Cumulative Visibility Detection The first three algorithms are those available in the original Freestyle (the "normal" ray casting was used unconditionally, though). The "fast" and "very fast" ray casting algorithms achieve a faster calculation at the cost of less visibility accuracy. The last four are newly introduced optimized options. The culled versions of the new algorithms will exclude from visibility calculation those faces that lay outside the camera, which leads to a faster view map construction. The unculled counterparts will take all faces into account. The unculled visibility algorithms are useful when culling affects stroke chaining. The recommended options for users are the culled/unculled cumulative visibility algorithms. These options are meant to replace the old algorithms in the future. Performance improvements over the old algorithms depend on the scenes to be rendered. * Silhouette detection has also been considerably optimized for speed. Performance gains by this optimization do not depend on scenes. * Improper handling of error conditions in the view map construction was fixed.
2011-03-14 00:36:27 +00:00
if (valid && !testCurrent()) {
next();
}
Optimized view map calculation by Alexander Beels. * View map calculation has been intensively optimized for speed by means of: 1) new spatial grid data structures (SphericalGrid for perspective cameras and BoxGrid for orthographic cameras; automatically switched based on the camera type); 2) a heuristic grid density calculation algorithm; and 3) new line visibility computation algorithms: A "traditional" algorithm for emulating old visibility algorithms, and a "cumulative" algorithm for improved, more consistent line visibility, both exploiting the new spatial grid data structures for fast ray casting. A new option "Raycasting Algorithm" was added to allow users to choose a ray casting (line visibility) algorithm. Available choices are: - Normal Ray Casting - Fast Ray Casting - Very Fast Ray Casting - Culled Traditional Visibility Detection - Unculled Traditional Visibility Detection - Culled Cumulative Visibility Detection - Unculled Cumulative Visibility Detection The first three algorithms are those available in the original Freestyle (the "normal" ray casting was used unconditionally, though). The "fast" and "very fast" ray casting algorithms achieve a faster calculation at the cost of less visibility accuracy. The last four are newly introduced optimized options. The culled versions of the new algorithms will exclude from visibility calculation those faces that lay outside the camera, which leads to a faster view map construction. The unculled counterparts will take all faces into account. The unculled visibility algorithms are useful when culling affects stroke chaining. The recommended options for users are the culled/unculled cumulative visibility algorithms. These options are meant to replace the old algorithms in the future. Performance improvements over the old algorithms depend on the scenes to be rendered. * Silhouette detection has also been considerably optimized for speed. Performance gains by this optimization do not depend on scenes. * Improper handling of error conditions in the view map construction was fixed.
2011-03-14 00:36:27 +00:00
}
CulledOccluderSource::~CulledOccluderSource()
{
}
Optimized view map calculation by Alexander Beels. * View map calculation has been intensively optimized for speed by means of: 1) new spatial grid data structures (SphericalGrid for perspective cameras and BoxGrid for orthographic cameras; automatically switched based on the camera type); 2) a heuristic grid density calculation algorithm; and 3) new line visibility computation algorithms: A "traditional" algorithm for emulating old visibility algorithms, and a "cumulative" algorithm for improved, more consistent line visibility, both exploiting the new spatial grid data structures for fast ray casting. A new option "Raycasting Algorithm" was added to allow users to choose a ray casting (line visibility) algorithm. Available choices are: - Normal Ray Casting - Fast Ray Casting - Very Fast Ray Casting - Culled Traditional Visibility Detection - Unculled Traditional Visibility Detection - Culled Cumulative Visibility Detection - Unculled Cumulative Visibility Detection The first three algorithms are those available in the original Freestyle (the "normal" ray casting was used unconditionally, though). The "fast" and "very fast" ray casting algorithms achieve a faster calculation at the cost of less visibility accuracy. The last four are newly introduced optimized options. The culled versions of the new algorithms will exclude from visibility calculation those faces that lay outside the camera, which leads to a faster view map construction. The unculled counterparts will take all faces into account. The unculled visibility algorithms are useful when culling affects stroke chaining. The recommended options for users are the culled/unculled cumulative visibility algorithms. These options are meant to replace the old algorithms in the future. Performance improvements over the old algorithms depend on the scenes to be rendered. * Silhouette detection has also been considerably optimized for speed. Performance gains by this optimization do not depend on scenes. * Improper handling of error conditions in the view map construction was fixed.
2011-03-14 00:36:27 +00:00
bool CulledOccluderSource::testCurrent()
{
if (valid) {
// The test for gridSpaceOccluderProsceniumInitialized should not be necessary
return gridSpaceOccluderProsceniumInitialized &&
GridHelpers::insideProscenium(gridSpaceOccluderProscenium, cachedPolygon);
}
return false;
Optimized view map calculation by Alexander Beels. * View map calculation has been intensively optimized for speed by means of: 1) new spatial grid data structures (SphericalGrid for perspective cameras and BoxGrid for orthographic cameras; automatically switched based on the camera type); 2) a heuristic grid density calculation algorithm; and 3) new line visibility computation algorithms: A "traditional" algorithm for emulating old visibility algorithms, and a "cumulative" algorithm for improved, more consistent line visibility, both exploiting the new spatial grid data structures for fast ray casting. A new option "Raycasting Algorithm" was added to allow users to choose a ray casting (line visibility) algorithm. Available choices are: - Normal Ray Casting - Fast Ray Casting - Very Fast Ray Casting - Culled Traditional Visibility Detection - Unculled Traditional Visibility Detection - Culled Cumulative Visibility Detection - Unculled Cumulative Visibility Detection The first three algorithms are those available in the original Freestyle (the "normal" ray casting was used unconditionally, though). The "fast" and "very fast" ray casting algorithms achieve a faster calculation at the cost of less visibility accuracy. The last four are newly introduced optimized options. The culled versions of the new algorithms will exclude from visibility calculation those faces that lay outside the camera, which leads to a faster view map construction. The unculled counterparts will take all faces into account. The unculled visibility algorithms are useful when culling affects stroke chaining. The recommended options for users are the culled/unculled cumulative visibility algorithms. These options are meant to replace the old algorithms in the future. Performance improvements over the old algorithms depend on the scenes to be rendered. * Silhouette detection has also been considerably optimized for speed. Performance gains by this optimization do not depend on scenes. * Improper handling of error conditions in the view map construction was fixed.
2011-03-14 00:36:27 +00:00
}
bool CulledOccluderSource::next()
{
while (OccluderSource::next()) {
if (testCurrent()) {
++rejected;
return true;
}
}
if (G.debug & G_DEBUG_FREESTYLE) {
std::cout << "Finished generating occluders. Rejected " << rejected << " faces." << std::endl;
}
return false;
Optimized view map calculation by Alexander Beels. * View map calculation has been intensively optimized for speed by means of: 1) new spatial grid data structures (SphericalGrid for perspective cameras and BoxGrid for orthographic cameras; automatically switched based on the camera type); 2) a heuristic grid density calculation algorithm; and 3) new line visibility computation algorithms: A "traditional" algorithm for emulating old visibility algorithms, and a "cumulative" algorithm for improved, more consistent line visibility, both exploiting the new spatial grid data structures for fast ray casting. A new option "Raycasting Algorithm" was added to allow users to choose a ray casting (line visibility) algorithm. Available choices are: - Normal Ray Casting - Fast Ray Casting - Very Fast Ray Casting - Culled Traditional Visibility Detection - Unculled Traditional Visibility Detection - Culled Cumulative Visibility Detection - Unculled Cumulative Visibility Detection The first three algorithms are those available in the original Freestyle (the "normal" ray casting was used unconditionally, though). The "fast" and "very fast" ray casting algorithms achieve a faster calculation at the cost of less visibility accuracy. The last four are newly introduced optimized options. The culled versions of the new algorithms will exclude from visibility calculation those faces that lay outside the camera, which leads to a faster view map construction. The unculled counterparts will take all faces into account. The unculled visibility algorithms are useful when culling affects stroke chaining. The recommended options for users are the culled/unculled cumulative visibility algorithms. These options are meant to replace the old algorithms in the future. Performance improvements over the old algorithms depend on the scenes to be rendered. * Silhouette detection has also been considerably optimized for speed. Performance gains by this optimization do not depend on scenes. * Improper handling of error conditions in the view map construction was fixed.
2011-03-14 00:36:27 +00:00
}
void CulledOccluderSource::getOccluderProscenium(real proscenium[4])
{
for (unsigned int i = 0; i < 4; ++i) {
proscenium[i] = gridSpaceOccluderProscenium[i];
}
Optimized view map calculation by Alexander Beels. * View map calculation has been intensively optimized for speed by means of: 1) new spatial grid data structures (SphericalGrid for perspective cameras and BoxGrid for orthographic cameras; automatically switched based on the camera type); 2) a heuristic grid density calculation algorithm; and 3) new line visibility computation algorithms: A "traditional" algorithm for emulating old visibility algorithms, and a "cumulative" algorithm for improved, more consistent line visibility, both exploiting the new spatial grid data structures for fast ray casting. A new option "Raycasting Algorithm" was added to allow users to choose a ray casting (line visibility) algorithm. Available choices are: - Normal Ray Casting - Fast Ray Casting - Very Fast Ray Casting - Culled Traditional Visibility Detection - Unculled Traditional Visibility Detection - Culled Cumulative Visibility Detection - Unculled Cumulative Visibility Detection The first three algorithms are those available in the original Freestyle (the "normal" ray casting was used unconditionally, though). The "fast" and "very fast" ray casting algorithms achieve a faster calculation at the cost of less visibility accuracy. The last four are newly introduced optimized options. The culled versions of the new algorithms will exclude from visibility calculation those faces that lay outside the camera, which leads to a faster view map construction. The unculled counterparts will take all faces into account. The unculled visibility algorithms are useful when culling affects stroke chaining. The recommended options for users are the culled/unculled cumulative visibility algorithms. These options are meant to replace the old algorithms in the future. Performance improvements over the old algorithms depend on the scenes to be rendered. * Silhouette detection has also been considerably optimized for speed. Performance gains by this optimization do not depend on scenes. * Improper handling of error conditions in the view map construction was fixed.
2011-03-14 00:36:27 +00:00
}
static inline real distance2D(const Vec3r &point, const real origin[2])
{
return ::hypot((point[0] - origin[0]), (point[1] - origin[1]));
Optimized view map calculation by Alexander Beels. * View map calculation has been intensively optimized for speed by means of: 1) new spatial grid data structures (SphericalGrid for perspective cameras and BoxGrid for orthographic cameras; automatically switched based on the camera type); 2) a heuristic grid density calculation algorithm; and 3) new line visibility computation algorithms: A "traditional" algorithm for emulating old visibility algorithms, and a "cumulative" algorithm for improved, more consistent line visibility, both exploiting the new spatial grid data structures for fast ray casting. A new option "Raycasting Algorithm" was added to allow users to choose a ray casting (line visibility) algorithm. Available choices are: - Normal Ray Casting - Fast Ray Casting - Very Fast Ray Casting - Culled Traditional Visibility Detection - Unculled Traditional Visibility Detection - Culled Cumulative Visibility Detection - Unculled Cumulative Visibility Detection The first three algorithms are those available in the original Freestyle (the "normal" ray casting was used unconditionally, though). The "fast" and "very fast" ray casting algorithms achieve a faster calculation at the cost of less visibility accuracy. The last four are newly introduced optimized options. The culled versions of the new algorithms will exclude from visibility calculation those faces that lay outside the camera, which leads to a faster view map construction. The unculled counterparts will take all faces into account. The unculled visibility algorithms are useful when culling affects stroke chaining. The recommended options for users are the culled/unculled cumulative visibility algorithms. These options are meant to replace the old algorithms in the future. Performance improvements over the old algorithms depend on the scenes to be rendered. * Silhouette detection has also been considerably optimized for speed. Performance gains by this optimization do not depend on scenes. * Improper handling of error conditions in the view map construction was fixed.
2011-03-14 00:36:27 +00:00
}
static inline bool crossesProscenium(real proscenium[4], FEdge *fe)
{
Vec2r min(proscenium[0], proscenium[2]);
Vec2r max(proscenium[1], proscenium[3]);
Vec2r A(fe->vertexA()->getProjectedX(), fe->vertexA()->getProjectedY());
Vec2r B(fe->vertexB()->getProjectedX(), fe->vertexB()->getProjectedY());
Optimized view map calculation by Alexander Beels. * View map calculation has been intensively optimized for speed by means of: 1) new spatial grid data structures (SphericalGrid for perspective cameras and BoxGrid for orthographic cameras; automatically switched based on the camera type); 2) a heuristic grid density calculation algorithm; and 3) new line visibility computation algorithms: A "traditional" algorithm for emulating old visibility algorithms, and a "cumulative" algorithm for improved, more consistent line visibility, both exploiting the new spatial grid data structures for fast ray casting. A new option "Raycasting Algorithm" was added to allow users to choose a ray casting (line visibility) algorithm. Available choices are: - Normal Ray Casting - Fast Ray Casting - Very Fast Ray Casting - Culled Traditional Visibility Detection - Unculled Traditional Visibility Detection - Culled Cumulative Visibility Detection - Unculled Cumulative Visibility Detection The first three algorithms are those available in the original Freestyle (the "normal" ray casting was used unconditionally, though). The "fast" and "very fast" ray casting algorithms achieve a faster calculation at the cost of less visibility accuracy. The last four are newly introduced optimized options. The culled versions of the new algorithms will exclude from visibility calculation those faces that lay outside the camera, which leads to a faster view map construction. The unculled counterparts will take all faces into account. The unculled visibility algorithms are useful when culling affects stroke chaining. The recommended options for users are the culled/unculled cumulative visibility algorithms. These options are meant to replace the old algorithms in the future. Performance improvements over the old algorithms depend on the scenes to be rendered. * Silhouette detection has also been considerably optimized for speed. Performance gains by this optimization do not depend on scenes. * Improper handling of error conditions in the view map construction was fixed.
2011-03-14 00:36:27 +00:00
return GeomUtils::intersect2dSeg2dArea(min, max, A, B);
Optimized view map calculation by Alexander Beels. * View map calculation has been intensively optimized for speed by means of: 1) new spatial grid data structures (SphericalGrid for perspective cameras and BoxGrid for orthographic cameras; automatically switched based on the camera type); 2) a heuristic grid density calculation algorithm; and 3) new line visibility computation algorithms: A "traditional" algorithm for emulating old visibility algorithms, and a "cumulative" algorithm for improved, more consistent line visibility, both exploiting the new spatial grid data structures for fast ray casting. A new option "Raycasting Algorithm" was added to allow users to choose a ray casting (line visibility) algorithm. Available choices are: - Normal Ray Casting - Fast Ray Casting - Very Fast Ray Casting - Culled Traditional Visibility Detection - Unculled Traditional Visibility Detection - Culled Cumulative Visibility Detection - Unculled Cumulative Visibility Detection The first three algorithms are those available in the original Freestyle (the "normal" ray casting was used unconditionally, though). The "fast" and "very fast" ray casting algorithms achieve a faster calculation at the cost of less visibility accuracy. The last four are newly introduced optimized options. The culled versions of the new algorithms will exclude from visibility calculation those faces that lay outside the camera, which leads to a faster view map construction. The unculled counterparts will take all faces into account. The unculled visibility algorithms are useful when culling affects stroke chaining. The recommended options for users are the culled/unculled cumulative visibility algorithms. These options are meant to replace the old algorithms in the future. Performance improvements over the old algorithms depend on the scenes to be rendered. * Silhouette detection has also been considerably optimized for speed. Performance gains by this optimization do not depend on scenes. * Improper handling of error conditions in the view map construction was fixed.
2011-03-14 00:36:27 +00:00
}
static inline bool insideProscenium(real proscenium[4], const Vec3r &point)
{
return !(point[0] < proscenium[0] || point[0] > proscenium[1] || point[1] < proscenium[2] ||
point[1] > proscenium[3]);
Optimized view map calculation by Alexander Beels. * View map calculation has been intensively optimized for speed by means of: 1) new spatial grid data structures (SphericalGrid for perspective cameras and BoxGrid for orthographic cameras; automatically switched based on the camera type); 2) a heuristic grid density calculation algorithm; and 3) new line visibility computation algorithms: A "traditional" algorithm for emulating old visibility algorithms, and a "cumulative" algorithm for improved, more consistent line visibility, both exploiting the new spatial grid data structures for fast ray casting. A new option "Raycasting Algorithm" was added to allow users to choose a ray casting (line visibility) algorithm. Available choices are: - Normal Ray Casting - Fast Ray Casting - Very Fast Ray Casting - Culled Traditional Visibility Detection - Unculled Traditional Visibility Detection - Culled Cumulative Visibility Detection - Unculled Cumulative Visibility Detection The first three algorithms are those available in the original Freestyle (the "normal" ray casting was used unconditionally, though). The "fast" and "very fast" ray casting algorithms achieve a faster calculation at the cost of less visibility accuracy. The last four are newly introduced optimized options. The culled versions of the new algorithms will exclude from visibility calculation those faces that lay outside the camera, which leads to a faster view map construction. The unculled counterparts will take all faces into account. The unculled visibility algorithms are useful when culling affects stroke chaining. The recommended options for users are the culled/unculled cumulative visibility algorithms. These options are meant to replace the old algorithms in the future. Performance improvements over the old algorithms depend on the scenes to be rendered. * Silhouette detection has also been considerably optimized for speed. Performance gains by this optimization do not depend on scenes. * Improper handling of error conditions in the view map construction was fixed.
2011-03-14 00:36:27 +00:00
}
void CulledOccluderSource::cullViewEdges(ViewMap &viewMap, bool extensiveFEdgeSearch)
{
// Cull view edges by marking them as non-displayable.
// This avoids the complications of trying to delete edges from the ViewMap.
// Non-displayable view edges will be skipped over during visibility calculation.
// View edges will be culled according to their position w.r.t. the viewport proscenium (viewport
// + 5% border, or some such).
// Get proscenium boundary for culling
real viewProscenium[4];
GridHelpers::getDefaultViewProscenium(viewProscenium);
real prosceniumOrigin[2];
prosceniumOrigin[0] = (viewProscenium[1] - viewProscenium[0]) / 2.0;
prosceniumOrigin[1] = (viewProscenium[3] - viewProscenium[2]) / 2.0;
if (G.debug & G_DEBUG_FREESTYLE) {
cout << "Proscenium culling:" << endl;
cout << "Proscenium: [" << viewProscenium[0] << ", " << viewProscenium[1] << ", "
<< viewProscenium[2] << ", " << viewProscenium[3] << "]" << endl;
cout << "Origin: [" << prosceniumOrigin[0] << ", " << prosceniumOrigin[1] << "]" << endl;
}
// A separate occluder proscenium will also be maintained, starting out the same as the viewport
// proscenium, and expanding as necessary so that it encompasses the center point of at least one
// feature edge in each retained view edge. The occluder proscenium will be used later to cull
// occluding triangles before they are inserted into the Grid. The occluder proscenium starts out
// the same size as the view proscenium
GridHelpers::getDefaultViewProscenium(occluderProscenium);
// XXX Freestyle is inconsistent in its use of ViewMap::viewedges_container and
// vector<ViewEdge*>::iterator. Probably all occurrences of vector<ViewEdge*>::iterator should be
// replaced ViewMap::viewedges_container throughout the code. For each view edge
ViewMap::viewedges_container::iterator ve, veend;
for (ve = viewMap.ViewEdges().begin(), veend = viewMap.ViewEdges().end(); ve != veend; ve++) {
// Overview:
// Search for a visible feature edge
// If none: mark view edge as non-displayable
// Otherwise:
// Find a feature edge with center point inside occluder proscenium.
// If none exists, find the feature edge with center point closest to viewport origin.
// Expand occluder proscenium to enclose center point.
// For each feature edge, while bestOccluderTarget not found and view edge not visible
bool bestOccluderTargetFound = false;
FEdge *bestOccluderTarget = NULL;
real bestOccluderDistance = 0.0;
FEdge *festart = (*ve)->fedgeA();
FEdge *fe = festart;
// All ViewEdges start culled
(*ve)->setIsInImage(false);
// For simple visibility calculation: mark a feature edge that is known to have a center point
// inside the occluder proscenium. Cull all other feature edges.
do {
// All FEdges start culled
fe->setIsInImage(false);
// Look for the visible edge that can most easily be included in the occluder proscenium.
if (!bestOccluderTargetFound) {
// If center point is inside occluder proscenium,
if (insideProscenium(occluderProscenium, fe->center2d())) {
// Use this feature edge for visibility deterimination
fe->setIsInImage(true);
expandGridSpaceOccluderProscenium(fe);
// Mark bestOccluderTarget as found
bestOccluderTargetFound = true;
bestOccluderTarget = fe;
}
else {
real d = distance2D(fe->center2d(), prosceniumOrigin);
// If center point is closer to viewport origin than current target
if (bestOccluderTarget == NULL || d < bestOccluderDistance) {
// Then store as bestOccluderTarget
bestOccluderDistance = d;
bestOccluderTarget = fe;
}
}
}
// If feature edge crosses the view proscenium
if (!(*ve)->isInImage() && crossesProscenium(viewProscenium, fe)) {
// Then the view edge will be included in the image
(*ve)->setIsInImage(true);
}
fe = fe->nextEdge();
} while (fe != NULL && fe != festart && !(bestOccluderTargetFound && (*ve)->isInImage()));
// Either we have run out of FEdges, or we already have the one edge we need to determine
// visibility Cull all remaining edges.
while (fe != NULL && fe != festart) {
fe->setIsInImage(false);
fe = fe->nextEdge();
}
// If bestOccluderTarget was not found inside the occluder proscenium,
// we need to expand the occluder proscenium to include it.
if ((*ve)->isInImage() && bestOccluderTarget != NULL && !bestOccluderTargetFound) {
// Expand occluder proscenium to enclose bestOccluderTarget
Vec3r point = bestOccluderTarget->center2d();
if (point[0] < occluderProscenium[0]) {
occluderProscenium[0] = point[0];
}
else if (point[0] > occluderProscenium[1]) {
occluderProscenium[1] = point[0];
}
if (point[1] < occluderProscenium[2]) {
occluderProscenium[2] = point[1];
}
else if (point[1] > occluderProscenium[3]) {
occluderProscenium[3] = point[1];
}
// Use bestOccluderTarget for visibility determination
bestOccluderTarget->setIsInImage(true);
}
}
// We are done calculating the occluder proscenium.
// Expand the occluder proscenium by an epsilon to avoid rounding errors.
const real epsilon = 1.0e-6;
occluderProscenium[0] -= epsilon;
occluderProscenium[1] += epsilon;
occluderProscenium[2] -= epsilon;
occluderProscenium[3] += epsilon;
// For "Normal" or "Fast" style visibility computation only:
// For more detailed visibility calculation, make a second pass through the view map, marking all
// feature edges with center points inside the final occluder proscenium. All of these feature
// edges can be considered during visibility calculation.
// So far we have only found one FEdge per ViewEdge. The "Normal" and "Fast" styles of visibility
// computation want to consider many FEdges for each ViewEdge. Here we re-scan the view map to
// find any usable FEdges that we skipped on the first pass, or that have become usable because
// the occluder proscenium has been expanded since the edge was visited on the first pass.
if (extensiveFEdgeSearch) {
// For each view edge,
for (ve = viewMap.ViewEdges().begin(), veend = viewMap.ViewEdges().end(); ve != veend; ve++) {
if (!(*ve)->isInImage()) {
continue;
}
// For each feature edge,
FEdge *festart = (*ve)->fedgeA();
FEdge *fe = festart;
do {
// If not (already) visible and center point inside occluder proscenium,
if (!fe->isInImage() && insideProscenium(occluderProscenium, fe->center2d())) {
// Use the feature edge for visibility determination
fe->setIsInImage(true);
expandGridSpaceOccluderProscenium(fe);
}
fe = fe->nextEdge();
} while (fe != NULL && fe != festart);
}
}
// Up until now, all calculations have been done in camera space.
// However, the occluder source's iteration and the grid that consumes the occluders both work in
// gridspace, so we need a version of the occluder proscenium in gridspace. Set the gridspace
// occlude proscenium
Optimized view map calculation by Alexander Beels. * View map calculation has been intensively optimized for speed by means of: 1) new spatial grid data structures (SphericalGrid for perspective cameras and BoxGrid for orthographic cameras; automatically switched based on the camera type); 2) a heuristic grid density calculation algorithm; and 3) new line visibility computation algorithms: A "traditional" algorithm for emulating old visibility algorithms, and a "cumulative" algorithm for improved, more consistent line visibility, both exploiting the new spatial grid data structures for fast ray casting. A new option "Raycasting Algorithm" was added to allow users to choose a ray casting (line visibility) algorithm. Available choices are: - Normal Ray Casting - Fast Ray Casting - Very Fast Ray Casting - Culled Traditional Visibility Detection - Unculled Traditional Visibility Detection - Culled Cumulative Visibility Detection - Unculled Cumulative Visibility Detection The first three algorithms are those available in the original Freestyle (the "normal" ray casting was used unconditionally, though). The "fast" and "very fast" ray casting algorithms achieve a faster calculation at the cost of less visibility accuracy. The last four are newly introduced optimized options. The culled versions of the new algorithms will exclude from visibility calculation those faces that lay outside the camera, which leads to a faster view map construction. The unculled counterparts will take all faces into account. The unculled visibility algorithms are useful when culling affects stroke chaining. The recommended options for users are the culled/unculled cumulative visibility algorithms. These options are meant to replace the old algorithms in the future. Performance improvements over the old algorithms depend on the scenes to be rendered. * Silhouette detection has also been considerably optimized for speed. Performance gains by this optimization do not depend on scenes. * Improper handling of error conditions in the view map construction was fixed.
2011-03-14 00:36:27 +00:00
}
void CulledOccluderSource::expandGridSpaceOccluderProscenium(FEdge *fe)
{
if (gridSpaceOccluderProsceniumInitialized) {
GridHelpers::expandProscenium(gridSpaceOccluderProscenium, transform(fe->center3d()));
}
else {
const Vec3r &point = transform(fe->center3d());
gridSpaceOccluderProscenium[0] = gridSpaceOccluderProscenium[1] = point[0];
gridSpaceOccluderProscenium[2] = gridSpaceOccluderProscenium[3] = point[1];
gridSpaceOccluderProsceniumInitialized = true;
}
Optimized view map calculation by Alexander Beels. * View map calculation has been intensively optimized for speed by means of: 1) new spatial grid data structures (SphericalGrid for perspective cameras and BoxGrid for orthographic cameras; automatically switched based on the camera type); 2) a heuristic grid density calculation algorithm; and 3) new line visibility computation algorithms: A "traditional" algorithm for emulating old visibility algorithms, and a "cumulative" algorithm for improved, more consistent line visibility, both exploiting the new spatial grid data structures for fast ray casting. A new option "Raycasting Algorithm" was added to allow users to choose a ray casting (line visibility) algorithm. Available choices are: - Normal Ray Casting - Fast Ray Casting - Very Fast Ray Casting - Culled Traditional Visibility Detection - Unculled Traditional Visibility Detection - Culled Cumulative Visibility Detection - Unculled Cumulative Visibility Detection The first three algorithms are those available in the original Freestyle (the "normal" ray casting was used unconditionally, though). The "fast" and "very fast" ray casting algorithms achieve a faster calculation at the cost of less visibility accuracy. The last four are newly introduced optimized options. The culled versions of the new algorithms will exclude from visibility calculation those faces that lay outside the camera, which leads to a faster view map construction. The unculled counterparts will take all faces into account. The unculled visibility algorithms are useful when culling affects stroke chaining. The recommended options for users are the culled/unculled cumulative visibility algorithms. These options are meant to replace the old algorithms in the future. Performance improvements over the old algorithms depend on the scenes to be rendered. * Silhouette detection has also been considerably optimized for speed. Performance gains by this optimization do not depend on scenes. * Improper handling of error conditions in the view map construction was fixed.
2011-03-14 00:36:27 +00:00
}
Attempt to fix a potential name conflict between Freestyle and the compositor. A crash in the Freestyle renderer was reported by Ton on IRC with a stack trace below. Note that #2 is in Freestyle, whereas #1 is in the compositor. The problem was observed in a debug build on OS X 10.7 (gcc 4.2, openmp disabled, no llvm). ---------------------------------------------------------------------- Program received signal EXC_BAD_ACCESS, Could not access memory. Reason: 13 at address: 0x0000000000000000 [Switching to process 72386 thread 0xf303] 0x0000000100c129f3 in NodeBase::~NodeBase (this=0x10e501c80) at COM_NodeBase.cpp:43 43 delete (this->m_outputsockets.back()); Current language: auto; currently c++ (gdb) where #0 0x0000000100c129f3 in NodeBase::~NodeBase (this=0x10e501c80) at COM_NodeBase.cpp:43 #1 0x0000000100c29066 in Node::~Node (this=0x10e501c80) at COM_Node.h:49 #2 0x000000010089c273 in NodeShape::~NodeShape (this=0x10e501c80) at NodeShape.cpp:43 #3 0x000000010089910b in NodeGroup::destroy (this=0x10e501da0) at NodeGroup.cpp:61 #4 0x00000001008990cd in NodeGroup::destroy (this=0x10e5014b0) at NodeGroup.cpp:59 #5 0x00000001008990cd in NodeGroup::destroy (this=0x114e18da0) at NodeGroup.cpp:59 #6 0x00000001007e6602 in Controller::ClearRootNode (this=0x114e19640) at Controller.cpp:329 #7 0x00000001007ea52e in Controller::LoadMesh (this=0x114e19640, re=0x10aba4638, srl=0x1140f5258) at Controller.cpp:302 #8 0x00000001008030ad in prepare (re=0x10aba4638, srl=0x1140f5258) at FRS_freestyle.cpp:302 #9 0x000000010080457a in FRS_do_stroke_rendering (re=0x10aba4638, srl=0x1140f5258) at FRS_freestyle.cpp:600 #10 0x00000001006aeb9d in add_freestyle (re=0x10aba4638) at pipeline.c:1584 #11 0x00000001006aceb7 in do_render_3d (re=0x10aba4638) at pipeline.c:1094 #12 0x00000001006ae061 in do_render_fields_blur_3d (re=0x10aba4638) at pipeline.c:1367 #13 0x00000001006afa16 in do_render_composite_fields_blur_3d (re=0x10aba4638) at pipeline.c:1815 #14 0x00000001006b04e4 in do_render_all_options (re=0x10aba4638) at pipeline.c:2021 ---------------------------------------------------------------------- Apparently a name conflict between the two Blender modules is taking place. The present commit hence intends to address it by putting all the Freestyle C++ classes in the namespace 'Freestyle'. This revision will also prevent potential name conflicts with other Blender modules in the future. Special thanks to Lukas Toenne for the help with C++ namespace.
2013-04-09 00:46:49 +00:00
} /* namespace Freestyle */