/**************************************************************************\ * * This file is part of the Coin 3D visualization library. * Copyright (C) 1998-2007 by Systems in Motion. All rights reserved. * * This library is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * ("GPL") version 2 as published by the Free Software Foundation. * See the file LICENSE.GPL at the root directory of this source * distribution for additional information about the GNU GPL. * * For using Coin with software that can not be combined with the GNU * GPL, and for taking advantage of the additional benefits of our * support services, please contact Systems in Motion about acquiring * a Coin Professional Edition License. * * See http://www.coin3d.org/ for more information. * * Systems in Motion, Postboks 1283, Pirsenteret, 7462 Trondheim, NORWAY. * http://www.sim.no/ sales@sim.no coin-support@coin3d.org * \**************************************************************************/ /*! \class SoCamera SoCamera.h Inventor/nodes/SoCamera.h \brief The SoCamera class is the abstract base class for camera definition nodes. \ingroup nodes To be able to view a scene, one needs to have a camera in the scene graph. A camera node will set up the projection and viewing matrices for rendering of the geometry in the scene. This node just defines the abstract interface by collecting common fields that all camera type nodes needs. Use the non-abstract camera node subclasses within a scene graph. The ones that are default part of the Coin library are SoPerspectiveCamera and SoOrthographicCamera, which uses the two different projections given by their name. Note that the viewer components of the GUI glue libraries of Coin (SoXt, SoQt, SoWin, etc) will automatically insert a camera into a scene graph is none has been defined. It is possible to have more than one camera in a scene graph. One common trick is for instance to use a second camera to display static geometry or overlay geometry (e.g. for head-up displays ("HUD")), as shown by this example code: \code #include #include #include int main(int argc, char ** argv) { QWidget * mainwin = SoQt::init(argv[0]); SoSeparator * root = new SoSeparator; root->ref(); // Adds a camera and a red cone. The first camera found in the // scene graph by the SoQtExaminerViewer will be picked up and // initialized automatically. root->addChild(new SoPerspectiveCamera); SoMaterial * material = new SoMaterial; material->diffuseColor.setValue(1.0, 0.0, 0.0); root->addChild(material); root->addChild(new SoCone); // Set up a second camera for the remaining geometry. This camera // will not be picked up and influenced by the viewer, so the // geometry will be kept static. SoPerspectiveCamera * pcam = new SoPerspectiveCamera; pcam->position = SbVec3f(0, 0, 5); pcam->nearDistance = 0.1; pcam->farDistance = 10; root->addChild(pcam); // Adds a green cone to demonstrate static geometry. SoMaterial * greenmaterial = new SoMaterial; greenmaterial->diffuseColor.setValue(0, 1.0, 0.0); root->addChild(greenmaterial); root->addChild(new SoCone); SoQtExaminerViewer * viewer = new SoQtExaminerViewer(mainwin); viewer->setSceneGraph(root); viewer->show(); SoQt::show(mainwin); SoQt::mainLoop(); delete viewer; root->unref(); return 0; } \endcode */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include // for FLT_EPSILON #if COIN_DEBUG #include #endif // COIN_DEBUG /*! \enum SoCamera::ViewportMapping Enumerates the available possibilities for how the render frame should map the viewport. */ /*! \var SoSFEnum SoCamera::viewportMapping Set up how the render frame should map the viewport. The default is SoCamera::ADJUST_CAMERA. */ /*! \var SoSFVec3f SoCamera::position Camera position. Defaults to <0,0,1>. */ /*! \var SoSFRotation SoCamera::orientation Camera orientation specified as a rotation value from the default orientation where the camera is pointing along the negative z-axis, with "up" along the positive y-axis. E.g., to rotate the camera to point along the X axis: \code mycamera->orientation.setValue(SbRotation(SbVec3f(0, 1, 0), M_PI / 2.0f)); \endcode For queries, e.g. to get the current "up" and "look at" vectors of the camera: \code SbRotation camrot = mycamera->orientation.getValue(); SbVec3f upvec(0, 1, 0); // init to default up vector camrot.multVec(upvec, upvec); SbVec3f lookat(0, 0, -1); // init to default view direction vector camrot.multVec(lookat, lookat); \endcode */ /*! \var SoSFFloat SoCamera::aspectRatio Aspect ratio for the camera (i.e. width / height). Defaults to 1.0. */ /*! \var SoSFFloat SoCamera::nearDistance Distance from camera position to the near clipping plane in the camera's view volume. Default value is 1.0. Value must be larger than 0.0, or it will not be possible to construct a valid viewing volume (for perspective rendering, at least). If you use one of the viewer components from the So[Xt|Qt|Win|Gtk] GUI libraries provided by Systems in Motion, they will automatically update this value for the scene camera according to the scene bounding box. Ditto for the far clipping plane. \sa SoCamera::farDistance */ /*! \var SoSFFloat SoCamera::farDistance Distance from camera position to the far clipping plane in the camera's view volume. Default value is 10.0. Must be larger than the SoCamera::nearDistance value, or it will not be possible to construct a valid viewing volume. Note that the range [nearDistance, farDistance] decides the dynamic range of the Z-buffer in the underlying polygon-rendering rasterizer. What this means is that if the near and far clipping planes of the camera are wide apart, the possibility of visual artifacts will increase. The artifacts will manifest themselves in the form of flickering of primitives close in depth. It is therefore a good idea to keep the near and far clipping planes of your camera(s) as closely fitted around the geometry of the scene graph as possible. \sa SoCamera::nearDistance, SoPolygonOffset */ /*! \var SoSFFloat SoCamera::focalDistance Distance from camera position to center of scene. */ /*! \fn void SoCamera::scaleHeight(float scalefactor) Sets a \a scalefactor for the height of the camera viewport. What "viewport height" means exactly in this context depends on the camera model. See documentation in subclasses. */ /*! \fn SbViewVolume SoCamera::getViewVolume(float useaspectratio = 0.0f) const Returns total view volume covered by the camera under the current settings. This view volume is not adjusted to account for viewport mapping. If you want the same view volume as the one used during rendering, you should do something like this: \verbatim SbViewVolume vv; float aspectratio = myviewport.getViewportAspectRatio(); switch (camera->viewportMapping.getValue()) { case SoCamera::CROP_VIEWPORT_FILL_FRAME: case SoCamera::CROP_VIEWPORT_LINE_FRAME: case SoCamera::CROP_VIEWPORT_NO_FRAME: vv = camera->getViewVolume(0.0f); break; case SoCamera::ADJUST_CAMERA: vv = camera->getViewVolume(aspectratio); if (aspectratio < 1.0f) vv.scale(1.0f / aspectratio); break; case SoCamera::LEAVE_ALONE: vv = camera->getViewVolume(0.0f); break; default: assert(0 && "unknown viewport mapping"); break; } \endverbatim Also, for the CROPPED viewport mappings, the viewport might be changed if the viewport aspect ratio is not equal to the camera aspect ratio. See SoCamera::getView() to see how this is done. */ /*! \fn void SoCamera::viewBoundingBox(const SbBox3f & box, float aspect, float slack) Convenience method for setting up the camera definition to cover the given bounding \a box with the given \a aspect ratio. Multiplies the exact dimensions with a \a slack factor to have some space between the rendered model and the borders of the rendering area. If you define your own camera node class, be aware that this method should \e not set the orientation field of the camera, only the position, focal distance and near and far clipping planes. */ /*! \enum SoCamera::StereoMode Enumerates the possible stereo modes. */ /*! \var SoCamera::MONOSCOPIC No stereo. */ /*! \var SoCamera::LEFT_VIEW Left view. */ /*! \var SoCamera::RIGHT_VIEW Right view. */ SO_NODE_ABSTRACT_SOURCE(SoCamera); /*! Constructor. */ SoCamera::SoCamera() { SO_NODE_INTERNAL_CONSTRUCTOR(SoCamera); SO_NODE_ADD_FIELD(viewportMapping, (ADJUST_CAMERA)); SO_NODE_ADD_FIELD(position, (0.0f, 0.0f, 1.0f)); SO_NODE_ADD_FIELD(orientation, (SbRotation(SbVec3f(0.0f, 0.0f, 1.0f), 0.0f))); SO_NODE_ADD_FIELD(nearDistance, (1.0f)); SO_NODE_ADD_FIELD(farDistance, (10.0f)); SO_NODE_ADD_FIELD(aspectRatio, (1.0f)); SO_NODE_ADD_FIELD(focalDistance, (5.0f)); SO_NODE_DEFINE_ENUM_VALUE(ViewportMapping, CROP_VIEWPORT_FILL_FRAME); SO_NODE_DEFINE_ENUM_VALUE(ViewportMapping, CROP_VIEWPORT_LINE_FRAME); SO_NODE_DEFINE_ENUM_VALUE(ViewportMapping, CROP_VIEWPORT_NO_FRAME); SO_NODE_DEFINE_ENUM_VALUE(ViewportMapping, ADJUST_CAMERA); SO_NODE_DEFINE_ENUM_VALUE(ViewportMapping, LEAVE_ALONE); SO_NODE_SET_SF_ENUM_TYPE(viewportMapping, ViewportMapping); this->stereomode = MONOSCOPIC; this->stereoadjustment = 0.1f; this->balanceadjustment = 1.0f; } /*! Destructor. */ SoCamera::~SoCamera() { } // Doc in superclass. void SoCamera::initClass(void) { SO_NODE_INTERNAL_INIT_ABSTRACT_CLASS(SoCamera, SO_FROM_INVENTOR_1); SO_ENABLE(SoGLRenderAction, SoFocalDistanceElement); SO_ENABLE(SoGLRenderAction, SoGLProjectionMatrixElement); SO_ENABLE(SoGLRenderAction, SoViewVolumeElement); SO_ENABLE(SoGLRenderAction, SoGLViewingMatrixElement); SO_ENABLE(SoGLRenderAction, SoCullElement); SO_ENABLE(SoGetBoundingBoxAction, SoFocalDistanceElement); SO_ENABLE(SoGetBoundingBoxAction, SoProjectionMatrixElement); SO_ENABLE(SoGetBoundingBoxAction, SoViewVolumeElement); SO_ENABLE(SoGetBoundingBoxAction, SoViewingMatrixElement); SO_ENABLE(SoRayPickAction, SoFocalDistanceElement); SO_ENABLE(SoRayPickAction, SoProjectionMatrixElement); SO_ENABLE(SoRayPickAction, SoViewVolumeElement); SO_ENABLE(SoRayPickAction, SoViewingMatrixElement); SO_ENABLE(SoCallbackAction, SoFocalDistanceElement); SO_ENABLE(SoCallbackAction, SoProjectionMatrixElement); SO_ENABLE(SoCallbackAction, SoViewVolumeElement); SO_ENABLE(SoCallbackAction, SoViewingMatrixElement); SO_ENABLE(SoGetPrimitiveCountAction, SoFocalDistanceElement); SO_ENABLE(SoGetPrimitiveCountAction, SoProjectionMatrixElement); SO_ENABLE(SoGetPrimitiveCountAction, SoViewVolumeElement); SO_ENABLE(SoGetPrimitiveCountAction, SoViewingMatrixElement); SO_ENABLE(SoAudioRenderAction, SoListenerPositionElement); SO_ENABLE(SoAudioRenderAction, SoListenerOrientationElement); SO_ENABLE(SoAudioRenderAction, SoListenerDopplerElement); SO_ENABLE(SoAudioRenderAction, SoListenerGainElement); } /*! Reorients the camera so that it points towards \a targetpoint. The positive y-axis is used as the up vector of the camera, unless the new camera direction is parallel to this axis, in which case the positive z-axis will be used instead. */ void SoCamera::pointAt(const SbVec3f & targetpoint) { SbVec3f dir = targetpoint - this->position.getValue(); if (dir.normalize() == 0.0f) { #if COIN_DEBUG SoDebugError::postInfo("SoCamera::pointAt", "targetpoint == camera position."); #endif // debug return; } SbVec3f up(0.0f, 1.0f, 0.0f); // use 0,1,0 as the up vector unless direction and up vector are parallel if (SbAbs(dir.dot(up)) >= (1.0f - FLT_EPSILON)) up.setValue(0.0f, 0.0f, 1.0f); this->lookAt(dir, up); } /*! Reorients the camera so that it points towards \a targetpoint, using \a upvector as the camera up vector. \COIN_FUNCTION_EXTENSION */ void SoCamera::pointAt(const SbVec3f & targetpoint, const SbVec3f & upvector) { SbVec3f dir = targetpoint - this->position.getValue(); if (dir.normalize() == 0.0f) { #if COIN_DEBUG SoDebugError::postInfo("SoCamera::pointAt", "targetpoint == camera position."); #endif // debug return; } this->lookAt(dir, upvector); } // FIXME: should collect common code from the two viewAll() methods // below. 20010824 mortene. /*! Position the camera so that all geometry of the scene from \a sceneroot is contained in the view volume of the camera, while keeping the camera orientation constant. Finds the bounding box of the scene and calls SoCamera::viewBoundingBox(). A bounding sphere will be calculated from the scene bounding box, so the camera will "view all" even when the scene is rotated, in any way. The \a slack argument gives a multiplication factor to the distance the camera is supposed to move out from the \a sceneroot mid-point. A value less than 1.0 for the \a slack argument will therefore cause the camera to come closer to the scene, a value of 1.0 will position the camera as exactly outside the scene bounding sphere, and a value larger than 1.0 will give "extra slack" versus the scene bounding sphere. */ void SoCamera::viewAll(SoNode * const sceneroot, const SbViewportRegion & vpregion, const float slack) { SoGetBoundingBoxAction action(vpregion); action.apply(sceneroot); SbBox3f box = action.getBoundingBox(); #if COIN_DEBUG && 0 // debug SoDebugError::postInfo("SoCamera::viewAll", "bbox: <%f %f %f>, <%f %f %f>\n", box.getMin()[0], box.getMin()[1], box.getMin()[2], box.getMax()[0], box.getMax()[1], box.getMax()[2]); SoDebugError::postInfo("SoCamera::viewAll", "viewportregion, windowsize: <%f, %f>, <%d, %d>\n", vpregion.getViewportSize()[0], vpregion.getViewportSize()[1], vpregion.getWindowSize()[0], vpregion.getWindowSize()[1] ); #endif // debug // Only check for "flagged" emptiness and don't use // SbBox3f::hasVolume(), as we *can* handle flat boxes (in all // dimensions). if (box.isEmpty()) { return; } this->viewBoundingBox(box, this->aspectRatio.getValue(), slack); } /*! Position the camera so all geometry of the scene in \a path is contained in the view volume of the camera. Finds the bounding box of the scene and calls SoCamera::viewBoundingBox(). */ void SoCamera::viewAll(SoPath * const path, const SbViewportRegion & vpregion, const float slack) { SoGetBoundingBoxAction action(vpregion); action.apply(path); SbBox3f box = action.getBoundingBox(); // Only check for "flagged" emptiness and don't use // SbBox3f::hasVolume(), as we *can* handle flat boxes (in all // dimensions). if (box.isEmpty()) { return; } this->viewBoundingBox(box, this->aspectRatio.getValue(), slack); } /*! Based in the SoCamera::viewportMapping setting, convert the values of \a region to the viewport region we will actually render into. */ SbViewportRegion SoCamera::getViewportBounds(const SbViewportRegion & region) const { SbViewportRegion vp = region; switch (this->viewportMapping.getValue()) { case CROP_VIEWPORT_FILL_FRAME: case CROP_VIEWPORT_LINE_FRAME: case CROP_VIEWPORT_NO_FRAME: { float vpaspect = region.getViewportAspectRatio(); float camaspect = this->aspectRatio.getValue(); if (vpaspect > camaspect) { vp.scaleWidth(camaspect / vpaspect); return vp; } else if (vpaspect < camaspect) { vp.scaleHeight(vpaspect / camaspect); } } break; default: // do nothing break; } return vp; } // Doc in superclass. void SoCamera::GLRender(SoGLRenderAction * action) { SoState * state = action->getState(); SbViewportRegion vp; SbViewVolume vv; this->getView(action, vv, vp, FALSE); SbMatrix affine, proj; if (vv.getDepth() == 0.0f || vv.getWidth() == 0.0f || vv.getHeight() == 0.0f) { // Handle empty scenes. affine = proj = SbMatrix::identity(); } else { if (this->stereomode != MONOSCOPIC) { SbViewVolume copyvv = vv; SbMatrix dummy; float offset = this->stereoadjustment * 0.5f; if (this->stereomode == LEFT_VIEW) offset = -offset; SbVec3f r = vv.getProjectionDirection().cross(vv.getViewUp()); (void) r.normalize(); // get the current camera transformation/size vv.getMatrices(affine, proj); affine = affine.inverse(); float nearv, farv, left, right, top, bottom; nearv = vv.getNearDist(); farv = nearv + vv.getDepth(); right = vv.getWidth() * 0.5f; left = -right; top = vv.getHeight() * 0.5f; bottom = -top; // create a skewed frustum float focaldist = this->focalDistance.getValue() * this->balanceadjustment; if (focaldist < nearv) focaldist = nearv; left -= offset * nearv / focaldist; right -= offset * nearv / focaldist; vv.frustum(left,right,bottom,top,nearv,farv); // transform the skewed view volume to the same location as the original vv.transform(affine); // translate to account for left/right view affine.setTranslate(r * offset); vv.transform(affine); // read out the stereo view volume vv.getMatrices(affine, proj); } else { vv.getMatrices(affine, proj); } SbBool identity; const SbMatrix & mm = SoModelMatrixElement::get(state, identity); if (!identity) { affine.multRight(mm.inverse()); vv.transform(SoModelMatrixElement::get(state)); } SoCullElement::setViewVolume(state, vv); } SoViewVolumeElement::set(state, this, vv); if (action->getNumPasses() > 1) { SbVec3f jittervec; this->jitter(action->getNumPasses(), SoGLRenderPassElement::get(state), vp, jittervec); SbMatrix m; m.setTranslate(jittervec); proj.multRight(m); } SoProjectionMatrixElement::set(state, this, proj); SoViewingMatrixElement::set(state, this, affine); SoFocalDistanceElement::set(state, this, this->focalDistance.getValue()); } // Documented in superclass. void SoCamera::audioRender(SoAudioRenderAction *action) { SoState * state = action->getState(); SbBool setbylistener; setbylistener = SoListenerPositionElement::isSetByListener(state); if ((! setbylistener) && (! this->position.isIgnored())) { SbVec3f pos, worldpos; pos = this->position.getValue(); SoModelMatrixElement::get(action->getState()).multVecMatrix(pos, worldpos); SoListenerPositionElement::set(state, this, worldpos, FALSE); #if COIN_DEBUG && 0 float x, y, z; worldpos.getValue(x, y, z); SoDebugError::postInfo("SoCamera::audioRender","listenerpos (%0.2f, %0.2f, %0.2f)\n", x, y, z); #endif // debug } else { #if COIN_DEBUG && 0 SoDebugError::postInfo("SoCamera::audioRender","ignoring listenerpos\n"); #endif // debug } setbylistener = SoListenerOrientationElement::isSetByListener(state); if ((! setbylistener) && (! this->orientation.isIgnored())) { SbBool mmidentity; SbRotation r; SbMatrix m = SoModelMatrixElement::get(state, mmidentity); if (!mmidentity) { SbVec3f t; SbVec3f s; SbRotation so; m.getTransform(t, r, s, so); r *= this->orientation.getValue(); } else { r = this->orientation.getValue(); } SoListenerOrientationElement::set(state, this, r, FALSE); } // Set view volume. This is needed for LOD nodes to work properly. SbViewportRegion vp; SbViewVolume vv; this->getView(action, vv, vp, FALSE); if (! (vv.getDepth() == 0.0f || vv.getWidth() == 0.0f || vv.getHeight() == 0.0f) ) { SbBool identity; const SbMatrix & mm = SoModelMatrixElement::get(state, identity); if (!identity) vv.transform(mm); } SoViewVolumeElement::set(state, this, vv); } // Doc in superclass. void SoCamera::getBoundingBox(SoGetBoundingBoxAction * action) { SoCacheElement::invalidate(action->getState()); SoCamera::doAction(action); } /*! Picking actions can be triggered during handle event action traversal, and to do picking we need to know the camera state. \sa SoCamera::rayPick() */ void SoCamera::handleEvent(SoHandleEventAction * action) { SbViewportRegion vp; SbViewVolume vv; this->getView(action, vv, vp, FALSE); SoViewVolumeElement::set(action->getState(), this, vv); } /*! "Jitter" the camera according to the current rendering pass (\a curpass), to get an antialiased rendering of the scene when doing multipass rendering. */ void SoCamera::jitter(int numpasses, int curpass, const SbViewportRegion & vpreg, SbVec3f & jitteramount) const { int vpsize[2]; vpsize[0] = vpreg.getViewportSizePixels()[0]; vpsize[1] = vpreg.getViewportSizePixels()[1]; coin_viewvolume_jitter(numpasses, curpass, vpsize, (float*) jitteramount.getValue()); } // Documented in superclass. Overridden to set up the viewing and // projection matrices. void SoCamera::doAction(SoAction * action) { SoState * state = action->getState(); SbViewportRegion vp; SbViewVolume vv; this->getView(action, vv, vp, FALSE); SbMatrix affine, proj; if (vv.getDepth() == 0.0f || vv.getWidth() == 0.0f || vv.getHeight() == 0.0f) { // Handle empty scenes. affine = proj = SbMatrix::identity(); } else { vv.getMatrices(affine, proj); SbBool identity; const SbMatrix & mm = SoModelMatrixElement::get(state, identity); if (!identity) { vv.transform(mm); affine.multRight(mm.inverse()); } } SoViewVolumeElement::set(state, this, vv); SoProjectionMatrixElement::set(state, this, proj); SoViewingMatrixElement::set(state, this, affine); SoFocalDistanceElement::set(state, this, this->focalDistance.getValue()); } // Doc in superclass. void SoCamera::callback(SoCallbackAction * action) { SoCamera::doAction(action); } // Documented in superclass. void SoCamera::rayPick(SoRayPickAction * action) { // Overridden to calculate the coordinates of the ray within the // current camera settings. SoCamera::doAction(action); // We need to check for a non-empty view volume, as caused by scene // graphs with no geometry in them. SbViewVolume vv = this->getViewVolume(1.0f); if (vv.getDepth() != 0.0f && vv.getWidth() != 0.0f && vv.getHeight() != 0.0f) { action->computeWorldSpaceRay(); } } // Documented in superclass. void SoCamera::getPrimitiveCount(SoGetPrimitiveCountAction * action) { // The number of primitives used to render a shape can change // according to the shape's distance to the camera, so we need to // override this method from the superclass to modify the traversal // state settings for the camera view. SoCamera::doAction(action); } // // private method which calculates view volume, and calculates // new viewport region if viewportMapping requires this. // The state is updated with the new viewport, not with the // new view volume. // void SoCamera::getView(SoAction * action, SbViewVolume & resultvv, SbViewportRegion & resultvp, const SbBool considermodelmatrix) { SoState * state = action->getState(); // need to test if vp element is enabled. SoGetPrimitiveCountAction // does not enable this element, although I think it should (to get // correct SCREEN_SPACE complexity handling). pederb, 2001-10-31 SbBool usevpelement = state->isElementEnabled(SoViewportRegionElement::getClassStackIndex()); if (usevpelement) { resultvp = SoViewportRegionElement::get(state); } else { // just set it to some value. It's not important as the current // action does not support viewports. resultvp = SbViewportRegion(256, 256); } float aspectratio = resultvp.getViewportAspectRatio(); int vpm = this->viewportMapping.getValue(); SbBool adjustvp = FALSE; switch (vpm) { case CROP_VIEWPORT_FILL_FRAME: case CROP_VIEWPORT_LINE_FRAME: case CROP_VIEWPORT_NO_FRAME: resultvv = this->getViewVolume(0.0f); adjustvp = TRUE; break; case ADJUST_CAMERA: resultvv = this->getViewVolume(aspectratio); if (aspectratio < 1.0f) resultvv.scale(1.0f / aspectratio); break; case LEAVE_ALONE: resultvv = this->getViewVolume(0.0f); break; default: assert(0 && "unknown viewport mapping"); break; } if (considermodelmatrix) { SbBool isidentity; const SbMatrix &mm = SoModelMatrixElement::get(state, isidentity); if (!isidentity) resultvv.transform(mm); } if (adjustvp) { float cameraratio = this->aspectRatio.getValue(); if (aspectratio != cameraratio) { SbViewportRegion oldvp = resultvp; if (aspectratio < cameraratio) { resultvp.scaleHeight(aspectratio/cameraratio); } else { resultvp.scaleWidth(cameraratio/aspectratio); } // only draw if this is an SoGLRenderAction if (action->isOfType(SoGLRenderAction::getClassTypeId())) { this->drawCroppedFrame((SoGLRenderAction*)action, vpm, oldvp, resultvp); } if (usevpelement) { SoViewportRegionElement::set(action->getState(), resultvp); } } } } // // private method that draws a cropped frame // void SoCamera::drawCroppedFrame(SoGLRenderAction *action, const int viewportmapping, const SbViewportRegion & oldvp, const SbViewportRegion & newvp) { if (viewportmapping == SoCamera::CROP_VIEWPORT_NO_FRAME) return; if (action->handleTransparency(FALSE)) return; const cc_glglue * glw = cc_glglue_instance(action->getCacheContext()); SoState *state = action->getState(); state->push(); if (viewportmapping == SoCamera::CROP_VIEWPORT_LINE_FRAME) { SoLineWidthElement::set(state, this, 1.0f); } else { // FILL SoDrawStyleElement::set(state, this, SoDrawStyleElement::FILLED); // turn off backface culling SoGLShapeHintsElement::forceSend(state, TRUE, FALSE); } SbVec2s oldorigin = oldvp.getViewportOriginPixels(); SbVec2s oldsize = oldvp.getViewportSizePixels(); glMatrixMode(GL_PROJECTION); // projection matrix will be set later, so don't push glOrtho(oldorigin[0], oldorigin[0]+oldsize[0]-1, oldorigin[1], oldorigin[1]+oldsize[1]-1, -1, 1); glPushAttrib(GL_LIGHTING_BIT| GL_FOG_BIT| GL_DEPTH_BUFFER_BIT| GL_TEXTURE_BIT| GL_CURRENT_BIT); glMatrixMode(GL_MODELVIEW); glPushMatrix(); glLoadIdentity(); glDisable(GL_LIGHTING); glDisable(GL_TEXTURE_2D); if (cc_glglue_has_3d_textures(glw)) { glDisable(GL_TEXTURE_3D); } glDisable(GL_FOG); glDisable(GL_DEPTH_TEST); glColor3f(0.8f, 0.8f, 0.8f); SbVec2s origin = newvp.getViewportOriginPixels(); SbVec2s size = newvp.getViewportSizePixels(); SbVec2s orgsize = oldvp.getViewportSizePixels(); if (size[0] < orgsize[0]) { short minpos = origin[0] - 1; short maxpos = origin[0] + size[0]; if (viewportmapping == SoCamera::CROP_VIEWPORT_LINE_FRAME) { glBegin(GL_LINES); glVertex2s(minpos, oldorigin[1]); glVertex2s(minpos, oldorigin[1]+oldsize[1]); glVertex2s(maxpos, oldorigin[1]); glVertex2s(maxpos, oldorigin[1]+oldsize[1]); glEnd(); } else { glBegin(GL_QUADS); glVertex2s(oldorigin[0], oldorigin[1]); glVertex2s(oldorigin[0], oldorigin[1]+oldsize[1]-1); glVertex2s(minpos, oldorigin[1]+oldsize[1]); glVertex2s(minpos, oldorigin[1]); glVertex2s(maxpos, oldorigin[1]); glVertex2s(maxpos, oldorigin[1]+oldsize[1]-1); glVertex2s(oldorigin[0]+oldsize[0]-1, oldorigin[1]+oldsize[1]-1); glVertex2s(oldorigin[0]+oldsize[0]-1, oldorigin[1]); glEnd(); } } else if (size[1] < orgsize[1]) { short minpos = origin[1] - 1; short maxpos = origin[1] + size[1]; if (viewportmapping == SoCamera::CROP_VIEWPORT_LINE_FRAME) { glBegin(GL_LINES); glVertex2s(oldorigin[0], minpos); glVertex2s(oldorigin[0]+oldsize[0], minpos); glVertex2s(oldorigin[0], maxpos); glVertex2s(oldorigin[0]+oldsize[0], maxpos); glEnd(); } else { glBegin(GL_QUADS); glVertex2s(oldorigin[0], minpos); glVertex2s(oldorigin[0]+oldsize[0]-1, minpos); glVertex2s(oldorigin[0]+oldsize[0]-1, oldorigin[1]); glVertex2s(oldorigin[0], oldorigin[1]); glVertex2s(oldorigin[0], maxpos); glVertex2s(oldorigin[0], oldorigin[1]+oldsize[1]-1); glVertex2s(oldorigin[0]+oldsize[0]-1, oldorigin[1]+oldsize[1]-1); glVertex2s(oldorigin[1]+oldsize[0]-1, maxpos); glEnd(); } } glPopMatrix(); glPopAttrib(); state->pop(); } /*! Sets the stereo mode. */ void SoCamera::setStereoMode(StereoMode mode) { this->stereomode = mode; } /*! Returns the stereo mode. */ SoCamera::StereoMode SoCamera::getStereoMode(void) const { return this->stereomode; } /*! Sets the stereo adjustment. This is the distance between the left and right "eye" when doing stereo rendering. When doing stereo rendering, Coin will render two views, one for the left eye, and one for the right eye. The stereo adjustment is, a bit simplified, how much the camera is translated along the local X-axis between the left and the right view. The default distance is 0.1, which is chosen since it's the approximate distance between the human eyes. To create a nice looking and visible stereo effect, the application programmer will often have to adjust this value. If all you want to do is examine simple stand-alone 3D objects, it is possible to calculate a stereo offset based on the bounding box of the 3D model (or scale the model down to an appropriate size). However, if you have a large scene, where you want to fly around in the scene, and see stereo on different objects as you approach them, you can't calculate the stereo offset based on the bounding box of the scene, but rather use a stereo offset based on the scale of the individual objects/details you want to examine. Please note that it's important to set a sensible focal distance when doing stereo rendering. See setBalanceAdjustment() for information about how the focal distance affects the stereo rendering. \sa setBalanceAdjustment() */ void SoCamera::setStereoAdjustment(float adjustment) { this->stereoadjustment = adjustment; } /*! Returns the stereo adjustment. \sa setStereoAdjustment() */ float SoCamera::getStereoAdjustment(void) const { return this->stereoadjustment; } /*! Sets the stereo balance adjustment. This is a factor that enables you to move the zero parallax plane. Geometry in front of the zero parallax plane will appear to be in front of the screen. The balance adjustment is multiplied with the focal distance to find the zero parallax plane. The default value is 1.0, and the zero parallax plane is then at the focal point. \sa SoCamera::focalDistance */ void SoCamera::setBalanceAdjustment(float adjustment) { this->balanceadjustment = adjustment; } /*! Returns the stereo balance adjustment. \sa setBalanceAdjustment() */ float SoCamera::getBalanceAdjustment(void) const { return this->balanceadjustment; } // Private method that calculates a new orientation based on camera // direction and camera up vector. Vectors must be unit length. void SoCamera::lookAt(const SbVec3f & dir, const SbVec3f & up) { SbVec3f z = -dir; SbVec3f y = up; SbVec3f x = y.cross(z); // recompute y to create a valid coordinate system y = z.cross(x); // normalize x and y to create an orthonormal coord system if ((y.normalize() == 0.0f) || (x.normalize() == 0.0f)) { #if COIN_DEBUG SoDebugError::postInfo("SoCamera::lookAt", "Unable to create a rotation matrix " "(dir = %g %g %g, up = %g %g %g)\n", dir[0], dir[1], dir[2], up[0], up[1], up[2]); #endif // debug return; } // create a rotation matrix SbMatrix rot = SbMatrix::identity(); rot[0][0] = x[0]; rot[0][1] = x[1]; rot[0][2] = x[2]; rot[1][0] = y[0]; rot[1][1] = y[1]; rot[1][2] = y[2]; rot[2][0] = z[0]; rot[2][1] = z[1]; rot[2][2] = z[2]; this->orientation.setValue(SbRotation(rot)); }