?? cvcamera.cpp
字號:
}
}
up.normalize();
SetLoc(*(GxVec3f*)&loc)->SetUp(*(GxVec3f*)&up)->SetDir(*(GxVec3f*)&dir);
}
#if (defined(CH_USE_RLAB) || defined(CH_USE_D3D))
ChQvPCameraRenderData * ChQvPCameraRenderData::Select()
{
m_pRC->LockScene();
// Make this camera the current one for the RC
if(m_boolNew)
{
// We only need to compute all this stuff if we are using this camera for the first time
// Otherwise, we remember from the last time selected
GxVec3f up, yAxis(0.0f, 1.0f, 0.0f); // starting loc for look dirs
GxVec3f dir, dira(0.0f, 0.0f, -1.0f); // will rotate later
GxVec3f loc, axis;
float angle, heightAngle, height;
float nearDistance, farDistance;
// Get the node's data; be safe about it
if(m_boolPerspective)
{
QvPerspectiveCamera *pCamera = (QvPerspectiveCamera*)((QvNode*)m_pCamera);
Qv2Gx(pCamera->position, loc);
Qv2Gx(pCamera->orientation.axis, axis);
angle = pCamera->orientation.angle;
heightAngle = pCamera->heightAngle.value;
nearDistance = pCamera->nearDistance.value;
farDistance = pCamera->farDistance.value;
}
else
{
QvOrthographicCamera *pCamera = (QvOrthographicCamera*)((QvNode*)m_pCamera);
Qv2Gx(pCamera->position, loc);
Qv2Gx(pCamera->orientation.axis, axis);
angle = pCamera->orientation.angle;
height = pCamera->height.value;
nearDistance = pCamera->nearDistance.value;
farDistance = pCamera->farDistance.value;
}
angle = -angle;
ChNrViewport view = m_pRC->GetViewport();
ChNrFrame cameraFrame;
ChNrViewportGetCamera(view, cameraFrame);
// Inventor puts the camera in transformed world (model) coordinates,
// RLAB's camera is in the base frame. Apply our stack transform to the
// frame, then apply the camera loc and direction.
// Remember VRML is RHS, but RLab is LHS
GxTransform3Wf stackMat = m_transform;
ChNrFrameAddTransform(cameraFrame, CombineReplace, *(stackMat.GetMatrix())); // TODO make an intermediate frame for this
GxTransform3Wf rotMat(axis, -angle);
GxTransform3Wf cameraMat = rotMat; // up/down has nothing to do with the transform jwd 12/7/95
dir = cameraMat * dira;
dir.normalize();
if(m_pRC->GetViewerMode() != walk )
{
up = cameraMat * yAxis;
}
else // We're walking; keep camera level
{ // This computation is not pure VRML,
// but it keeps camera manipulations much cleaner for
// walking situations. The camera acts more camera-like
// and less airplane-like.
GxVec3f n = yAxis.cross(dir); // normal to plane defined by (dir, yaxis)
if (n.dot(n) > SMIDGEON)
{
up = dir.cross(n);
}
else
{
// special case; looking straight up or down
up = cameraMat * yAxis;
}
if (up.dot(up) < SMIDGEON)
{
up = yAxis;
}
}
up.normalize();
SetLoc(*(GxVec3f*)&loc)->SetUp(*(GxVec3f*)&up)->SetDir(*(GxVec3f*)&dir);
SetStartLoc(*(GxVec3f*)&loc)->SetStartUp(*(GxVec3f*)&up)->SetStartDir(*(GxVec3f*)&dir);
m_boolNew = false;
D3DRelease(cameraFrame);
}
SetDirty(false);
SetRenderer();
m_pRC->CreateHeadlight();
// ChNrFrameAddLight(m_pRC->GetFrame(), RLCreateLightRGB(ChNrLightAmbient, RLVal(0.2), RLVal(0.2), RLVal(0.2)));
// Now let's save the matrices into the render data for later use
#if 0
memcpy(pRenderData->m_cameraTransform, G3dGetModelCamMatrix(hGC), sizeof(pRenderData->m_cameraTransform));
memcpy(pRenderData->m_invCameraTransform, G3dGetInverseModelCamMatrix(hGC), sizeof(pRenderData->m_invCameraTransform));
G3dSetActiveStack(hGC, G3DT_CAM_CLIP);
GxTransformF_t mat;
G3dGetTransform(hGC, mat);
pRenderData->m_camClipTransform = GxTransform3Wf(mat);
G3dSetActiveStack(hGC, G3DT_MODEL);
#else
//#pragma message("Saving of camera transforms not done.")
#endif
m_pRC->UnlockScene();
return this;
}
ChQvPCameraRenderData * ChQvPCameraRenderData::SetRenderer(bool boolNow)
{
if(!boolNow)
{
m_boolRendererDirty = true;
}
else if(m_boolRendererDirty)
{
m_boolRendererDirty = false;
m_pRC->LockScene();
// Set this camera's settings into the RC, viewport, etc.
float heightAngle, height;
float nearDistance, farDistance;
// Get the node's data; be safe about it
if(m_boolPerspective)
{
QvPerspectiveCamera *pCamera = (QvPerspectiveCamera*)((QvNode*)m_pCamera);
heightAngle = pCamera->heightAngle.value;
nearDistance = pCamera->nearDistance.value;
farDistance = pCamera->farDistance.value;
}
else
{
QvOrthographicCamera *pCamera = (QvOrthographicCamera*)((QvNode*)m_pCamera);
height = pCamera->height.value;
nearDistance = pCamera->nearDistance.value;
farDistance = pCamera->farDistance.value;
}
ChNrViewport view = m_pRC->GetViewport();
ChNrFrame cameraFrame;
ChNrViewportGetCamera(view, cameraFrame); // Release is well below here
ChNrViewportSetProjection(view, m_boolPerspective ? ProjectPerspective : ProjectOrthographic );
// Inventor puts the camera in transformed world (model) coordinates,
// RLAB's camera is in the base frame. Apply our stack transform to the
// frame, then apply the camera loc and direction.
// Remember VRML is RHS, but RLab is LHS
GxTransform3Wf stackMat = m_transform;
ChNrFrameAddTransform(cameraFrame, CombineReplace, *(stackMat.GetMatrix())); // TODO make an intermediate frame for this
ChNrFrameAddScale(cameraFrame, CombineBefore, 1., 1., -1.); // ??? --
GxTransform3Wf theTransform;
GxVec3f dir = m_dir, up = m_up;
dir.z() = -dir.z();
up.z() = -up.z();
theTransform.LookAt( dir, up, false); // --
ChNrFrameAddTranslation(cameraFrame, CombineBefore, m_loc.x(), m_loc.y(), -m_loc.z()); // z alrady scaled - ???
ChNrFrameAddTransform(cameraFrame, CombineBefore, *(theTransform.GetMatrix()));
#if 0
ChNrFrameAddTransform(cameraFrame, CombineReplace, *(stackMat.GetMatrix())); // TODO make an intermediate frame for this
GxTransform3Wf theTransform;
GxVec3f dir = -m_dir, up = -m_up;
theTransform.LookAt( dir, up, false); // --
ChNrFrameAddTranslation(cameraFrame, CombineBefore, m_loc.x(), m_loc.y(), m_loc.z()); // z alrady scaled - ???
ChNrFrameAddTransform(cameraFrame, CombineBefore, *(theTransform.GetMatrix()));
ChNrFrameAddScale(cameraFrame, CombineBefore, 1., 1.,1.); // ??? --
#endif
// Massage the front and back planes to enclose the whole scene if possible.
// If the version is 1.0 and nearDistance is 1.0, ignore it
float fNear = nearDistance;
if(true && fNear == 1.0) // for >= vrml 1.1
{
fNear = .01;
}
float fFar = 8000.;
//const float maxZResolution = 32000.; // Assuming ?? bits signed
const float maxZResolution = 8000.; // Assuming ?? bits signed
ChMazeWnd * pWnd = (ChMazeWnd *)(m_pRC->GetWnd());
if(pWnd->GetBounds())
{
// Compute based on scene bounds, and
// multiply in a little fudge factor for luck
fFar = 1.5 * pWnd->GetBounds()->GetFarDistance(m_loc.x(), m_loc.y(), m_loc.z());
//fFar *= 40.; // Work in the front of the buffer; we get less problems that way
fFar *= 15.; // Work in the front of the buffer; we get less problems that way
}
float minNear = fFar / maxZResolution; // lessen zbuffer roundoff error
fNear = max(fNear, minNear);
ChNrViewportSetFront(view, fNear);
ChNrViewportSetBack(view, fFar);
m_right = ChNrViewportGetWidth(view); /* right */
m_left = 0.0f; /* left*/
m_top = ChNrViewportGetHeight(view); /* top */
m_bottom = 0.0f; /* bottom */
// RL uses largest dimension to match to "field", so if this is
// width, adjust accordingly, since VRML is height-based.
float fieldHeight;
if (m_boolPerspective)
{
fieldHeight = fNear * tan(heightAngle / 2.);
}
else
{
fieldHeight = height / 2;
}
float aspect = float(ChNrViewportGetHeight(view)) / ChNrViewportGetWidth(view);
if (aspect < 1.0) fieldHeight /= aspect;
ChNrViewportSetField(view, fieldHeight);
D3DRelease(cameraFrame); // matches 'GetCamera'
m_pRC->UnlockScene();
}
return this;
}
void ChQvPCameraRenderData::GetOrientationTransform(GxTransform3Wf &theTransform)
{
GxVec3f dir = m_dir, up = m_up;
dir.z() = -dir.z();
up.z() = -up.z();
theTransform.LookAt( dir, up, false); // --
}
float ChQvPCameraRenderData::GetHither()
{
if(!m_pRC) return .01;
ChNrViewport view = m_pRC->GetViewport();
return ChNrViewportGetFront(view);
}
float ChQvPCameraRenderData::GetYon()
{
if(!m_pRC) return 1.;
ChNrViewport view = m_pRC->GetViewport();
return ChNrViewportGetBack(view);
}
#endif
#if defined(CH_USE_3DR)
void ChQvPCameraRenderData::TransformPoint( const GxVec3f &src, GxVec3f &dst )
{
::TransformPoint( *((PointF_t*)(&src)), *((PointF_t*)(&dst)), m_cameraTransform);
}
void ChQvPCameraRenderData::InverseTransformPoint( const GxVec3f &src, GxVec3f &dst )
{
::TransformPoint( *((PointF_t*)(&src)), *((PointF_t*)(&dst)), m_invCameraTransform);
}
void ChQvPCameraRenderData::TransformNormal( const GxVec3f &src, GxVec3f &dst )
{
PointF_t tmp;
G3dTransformNormalF( ((PointF_t*)(&src)), ((PointF_t*)(&tmp)), m_cameraTransform );
dst.set(tmp.x, tmp.y, tmp.z);
}
void ChQvPCameraRenderData::InverseTransformNormal( const GxVec3f &src, GxVec3f &dst )
{
PointF_t tmp;
G3dTransformNormalF( ((PointF_t*)(&src)), ((PointF_t*)(&tmp)), m_invCameraTransform );
dst.set(tmp.x, tmp.y, tmp.z);
}
#else
//#pragma message("Bunch of camera transform methods Not done!")
#endif
bool ChQvPCameraRenderData::OnSetOrientation(ChRenderContext *pRC, ChApplet * pApplet, ChIVrmlEvent *pEventList)
{
bool boolSuccess = true;
#if defined(CH_VRML_EVENTS)
// Change the orientation value, and dirty the context so it takes effect
float axis[3];
float angle;
ChRotationEvent *pEvent = (ChRotationEvent *)pEventList;
pEvent->GetValue(axis, angle);
string strType;
nodeType typ = nodeType(m_pNode->GetType(strType));
QvSFRotation *pRotation = 0;
if(typ == typeQvOrthographicCamera)
{
pRotation = &(((QvOrthographicCamera *)m_pNode)->orientation);
}
else if(typ == typeQvPerspectiveCamera)
{
pRotation = &(((QvPerspectiveCamera *)m_pNode)->orientation);
}
pRotation->axis[0] = axis[0];
pRotation->axis[1] = axis[1];
pRotation->axis[2] = axis[2];
pRotation->angle = angle;
// Now update the camera settings from the Qv values
if(m_pNode == pRC->GetCurrentCamera())
{
// Stop navigating - If we are spinning, everything screws up
((ChMazeWnd*)(pRC->GetWnd()))->GetSceneNavigator()->SetState(ChVrmlStateTransition::s_start);
}
Synch();
if(m_pNode == pRC->GetCurrentCamera())
{
SetRenderer();
// And set the context dirty
pRC->SetDirty();
}
#endif
return boolSuccess;
}
bool ChQvPCameraRenderData::OnSetPosition(ChRenderContext *pRC, ChApplet * pApplet, ChIVrmlEvent *pEventList)
{
bool boolSuccess = false;
#if defined(CH_VRML_EVENTS)
// Change the position value, and dirty the context so it takes effect
float position[3];
ChVec3fEvent *pEvent = (ChVec3fEvent *)pEventList;
pEvent->GetValue(position);
string strType;
nodeType typ = nodeType(m_pNode->GetType(strType));
QvSFVec3f *pPosition = 0;
if(typ == typeQvOrthographicCamera)
{
pPosition = &(((QvOrthographicCamera *)m_pNode)->position);
}
else if(typ == typeQvPerspectiveCamera)
{
pPosition = &(((QvPerspectiveCamera *)m_pNode)->position);
}
else
{
boolSuccess = false;
}
pPosition->value[0] = position[0];
pPosition->value[1] = position[1];
pPosition->value[2] = position[2];
// Now update the camera settings from the Qv values
if(m_pNode == pRC->GetCurrentCamera())
{
// Stop navigating - If we are spinning, everything screws up
((ChMazeWnd*)(pRC->GetWnd()))->GetSceneNavigator()->SetState(ChVrmlStateTransition::s_start);
}
Synch();
if(m_pNode == pRC->GetCurrentCamera())
{
SetRenderer();
// And set the context dirty
pRC->SetDirty();
}
#endif
return boolSuccess;
}
#if (defined(CH_USE_RLAB) || defined(CH_USE_D3D)) || defined(CH_USE_D3D)
bool ChQvPCameraRenderData::OnTick()
{
// 'move' tick
SetRenderer(true);
return true;
}
#endif
?? 快捷鍵說明
復制代碼
Ctrl + C
搜索代碼
Ctrl + F
全屏模式
F11
切換主題
Ctrl + Shift + D
顯示快捷鍵
?
增大字號
Ctrl + =
減小字號
Ctrl + -