最长的一帧学习 part4

2024-08-30 20:20
文章标签 学习 最长 一帧 part4

本文主要是介绍最长的一帧学习 part4，希望对大家解决编程问题提供一定的参考价值，需要的开发者们随着小编来一起学习吧！

文章目录

  • 九、osgUtil:: SceneView::draw ()
    • part1 初始化osg::State类的GL库函数,判断是否使用场景视图初始化访问器（SceneView::setInitVisitor）
    • part2 针对非立体显示的后续操作
      • part2.1 osgUtil::RenderStage::drawPreRenderStages
      • part2.2 osgUtil::RenderStage::draw 当前渲染台（即渲染树的根节点）的绘制
        • osgUtil::RenderStage::runCameraSetUp
        • osgUtil::RenderStage::copyTexture

九、osgUtil:: SceneView::draw ()

part1 初始化osg::State类的GL库函数,判断是否使用场景视图初始化访问器（SceneView::setInitVisitor）

  • State类保存了所有的OpenGL状态和属性参数；
  • 除此之外，State类还负责从当前系统平台的OpenGL链接库中获取函数的地址，这也是我们第一次执行场景绘制之前的必备工作，所用函数为State::initializeExtensionProcs。
  • 如果用户设置了场景视图初始化访问器（SceneView::setInitVisitor），那么draw函数第一次执行时将使用这个访问器遍历场景树，相关代码位于SceneView::init函数中。
void SceneView::draw()
{if (_camera->getNodeMask()==0) return;osg::State* state = _renderInfo.getState();// we in theory should be able to be able to bypass reset, but we'll call it just in case.//_state->reset();state->setFrameStamp(_frameStamp.get());if (_displaySettings.valid()){state->setDisplaySettings(_displaySettings.get());}//从当前系统平台的OpenGL链接库中获取函数的地址state->initializeExtensionProcs();osg::get<ContextData>(state->getContextID())->newFrame(state->getFrameStamp());// SceneView::setInitVisitor在其中if (!_initCalled) init();// note, to support multi-pipe systems the deletion of OpenGL display list// and texture objects is deferred until the OpenGL context is the correct// context for when the object were originally created.  Here we know what// context we are in so can flush the appropriate caches.//将所有已经标记为要删除的节点或者Drawable对象统一从场景和内存中删除，执行flushDeletedGLObjects函数if (_requiresFlush){double availableTime = 0.005;flushDeletedGLObjects(availableTime);}// assume the draw which is about to happen could generate GL objects that need flushing in the next frame._requiresFlush = _automaticFlush;。。。
}

part2 针对非立体显示的后续操作

void SceneView::draw()
{
。。。else{//  切换立体声关闭时，需要恢复绘制缓冲区。/*设置渲染台（RenderStage）的读/写缓存（通常包括GL_NONE，GL_FRONT_LEFT，GL_FRONT_RIGHT，GL_BACK_LEFT，GL_BACK_RIGHT， GL_FRONT，GL_BACK，GL_LEFT，GL_RIGHT，GL_FRONT_AND_BACK以及GL_AUX辅助缓存），其中的值是根据摄像机的setDrawBuffer和setReadBuffer函数来设定的。*/if( 0 == ( _camera->getInheritanceMask() & DRAW_BUFFER ) ){_renderStage->setDrawBuffer(_camera->getDrawBuffer());_renderStage->setReadBuffer(_camera->getDrawBuffer());}if( 0 == ( _camera->getInheritanceMask() & READ_BUFFER ) ){_renderStage->setReadBuffer(_camera->getReadBuffer());}_localStateSet->setAttribute(getViewport());//确保颜色掩码的每个颜色通道都是被激活的（使用osg::ColorMask）。if (_resetColorMaskToAllEnabled){// ensure that all color planes are active.osg::ColorMask* cmask = static_cast<osg::ColorMask*>(_localStateSet->getAttribute(osg::StateAttribute::COLORMASK));if (cmask){cmask->setMask(true,true,true,true);}else{cmask = new osg::ColorMask(true,true,true,true);_localStateSet->setAttribute(cmask);}_renderStage->setColorMask(cmask);}// 执行“前序渲染”渲染台的绘制（RenderStage::drawPreRenderStages）// bog standard draw._renderStage->drawPreRenderStages(_renderInfo,previous);//执行当前渲染台（即渲染树的根节点）的绘制（RenderStage::draw），无疑这是场景绘制的核心部分_renderStage->draw(_renderInfo,previous);}// 在结束了渲染树的绘制之后，SceneView::draw函数还负责恢复所有的OpenGL状态（使用State::popAllStateSets函数），判断是否在绘制过程中出现了OpenGL绘图命令错误，并将错误信息打印出来// re apply the default OGL state.state->popAllStateSets();state->apply();#if 0if (_camera->getPostDrawCallback()){(*(_camera->getPostDrawCallback()))(*_camera);}
#endifif (state->getCheckForGLErrors()!=osg::State::NEVER_CHECK_GL_ERRORS){if (state->checkGLErrors("end of SceneView::draw()")){// go into debug mode of OGL error in a fine grained way to help// track down OpenGL errors.state->setCheckForGLErrors(osg::State::ONCE_PER_ATTRIBUTE);}}// #define REPORT_TEXTURE_MANAGER_STATS
#ifdef REPORT_TEXTURE_MANAGER_STATStom->reportStats();bom->reportStats();
#endif// OSG_NOTICE<<"SceneView  draw() DynamicObjectCount"<<getState()->getDynamicObjectCount()<<std::endl;}
。。。
}

part2.1 osgUtil::RenderStage::drawPreRenderStages

void RenderStage::drawPreRenderStages(osg::RenderInfo& renderInfo,RenderLeaf*& previous)
{if (_preRenderList.empty()) return;//cout << "Drawing prerendering stages "<<this<< "  "<<_viewport->x()<<","<< _viewport->y()<<","<< _viewport->width()<<","<< _viewport->height()<<std::endl;for(RenderStageList::iterator itr=_preRenderList.begin();itr!=_preRenderList.end();++itr){itr->second->draw(renderInfo,previous);}//cout << "Done Drawing prerendering stages "<<this<< "  "<<_viewport->x()<<","<< _viewport->y()<<","<< _viewport->width()<<","<< _viewport->height()<<std::endl;
}

part2.2 osgUtil::RenderStage::draw 当前渲染台（即渲染树的根节点）的绘制

void RenderStage::draw(osg::RenderInfo& renderInfo,RenderLeaf*& previous)
{if (_stageDrawnThisFrame) return;if(_initialViewMatrix.valid()) renderInfo.getState()->setInitialViewMatrix(_initialViewMatrix.get());// push the stages camera so that drawing code can query itif (_camera.valid()) renderInfo.pushCamera(_camera.get());_stageDrawnThisFrame = true;if (_camera.valid() && _camera->getInitialDrawCallback()){// if we have a camera with a initial draw callback invoke it.// 执行摄像机的初始化回调（Camera::setInitialDrawCallback）_camera->getInitialDrawCallback()->run(renderInfo);}// note, SceneView does call to drawPreRenderStages explicitly// so there is no need to call it here.drawPreRenderStages(renderInfo,previous);if (_cameraRequiresSetUp || (_camera.valid() && _cameraAttachmentMapModifiedCount!=_camera->getAttachmentMapModifiedCount())){// 运行摄像机设置（RenderStage::runCameraSetUp）runCameraSetUp(renderInfo);}//	为了保证各个图形处理线程之间不会产生冲突，这里对当前调用的图形设备指针（GraphicsContext）做了一个检查。// 如果发现正在运行的图形设备与渲染台所记录的当前设备（RenderStage::_graphicsContext）不同的话，// 则转换到当前设备，避免指定渲染上下文时（GraphicsContext::makeCurrent）出错osg::State& state = *renderInfo.getState();osg::State* useState = &state;osg::GraphicsContext* callingContext = state.getGraphicsContext();osg::GraphicsContext* useContext = callingContext;osg::OperationThread* useThread = 0;osg::RenderInfo useRenderInfo(renderInfo);RenderLeaf* saved_previous = previous;// _graphicsContext是当前渲染台记录的gc，参数renderInfo// 需要转到当前渲染台记录的gcif (_graphicsContext.valid() && _graphicsContext != callingContext){// show we release the context so that others can use it?? will do so right// now as an experiment.callingContext->releaseContext();// OSG_NOTICE<<"  enclosing state before - "<<state.getStateSetStackSize()<<std::endl;useState = _graphicsContext->getState();useContext = _graphicsContext.get();useThread = useContext->getGraphicsThread();useRenderInfo.setState(useState);// synchronize the frame stampsuseState->setFrameStamp(const_cast<osg::FrameStamp*>(state.getFrameStamp()));// map the DynamicObjectCount across to the new windowuseState->setDynamicObjectCount(state.getDynamicObjectCount());useState->setDynamicObjectRenderingCompletedCallback(state.getDynamicObjectRenderingCompletedCallback());if (!useThread){previous = 0;useContext->makeCurrent();// OSG_NOTICE<<"  nested state before - "<<useState->getStateSetStackSize()<<std::endl;}}unsigned int originalStackSize = useState->getStateSetStackSize();// 执行摄像机的绘制前回调（Camera::setPreDrawCallback）if (_camera.valid() && _camera->getPreDrawCallback()){// if we have a camera with a pre draw callback invoke it._camera->getPreDrawCallback()->run(renderInfo);}bool doCopyTexture = _texture.valid() ?(callingContext != useContext) :false;if (useThread){
#if 1ref_ptr<osg::BlockAndFlushOperation> block = new osg::BlockAndFlushOperation;useThread->add(new DrawInnerOperation( this, renderInfo ));// 专用于绘制工作useThread->add(block.get());//强制在绘制结束之后方能继续执行线程的其它Operation对象// wait till the DrawInnerOperations is complete.block->block();doCopyTexture = false;#elseuseThread->add(new DrawInnerOperation( this, renderInfo ), true);doCopyTexture = false;
#endif}else{//单线程模型drawInner( useRenderInfo, previous, doCopyTexture);if (useRenderInfo.getUserData() != renderInfo.getUserData()){//	将当前使用的转为当前渲染台记录的renderInfo.setUserData(useRenderInfo.getUserData());}}if (useState != &state){// reset the local State's DynamicObjectCountstate.setDynamicObjectCount(useState->getDynamicObjectCount());useState->setDynamicObjectRenderingCompletedCallback(0);}// now copy the rendered image to attached texture.if (_texture.valid() && !doCopyTexture){if (callingContext && useContext!= callingContext){// make the calling context use the pbuffer context for reading.callingContext->makeContextCurrent(useContext);}// 实现摄像机的RTT（纹理烘焙）方式，将场景拷贝到用户指定的纹理对象中copyTexture(renderInfo);}// 执行摄像机的绘制后回调（Camera::setPostDrawCallback）if (_camera.valid() && _camera->getPostDrawCallback()){// if we have a camera with a post draw callback invoke it._camera->getPostDrawCallback()->run(renderInfo);}// 如果是单线程，使用glFlush刷新所有OpenGL管道中的命令，并释放当前渲染上下文（GraphicsContext::releaseContext）。// 释放的是上文中转到当前渲染设备记录的gcif (_graphicsContext.valid() && _graphicsContext != callingContext){useState->popStateSetStackToSize(originalStackSize);if (!useThread){// flush any command left in the useContex's FIFO// to ensure that textures are updated before main thread commenses.glFlush();useContext->releaseContext();}}if (callingContext && useContext != callingContext){// restore the graphics context.previous = saved_previous;// OSG_NOTICE<<"  nested state after - "<<useState->getStateSetStackSize()<<std::endl;// OSG_NOTICE<<"  enclosing state after - "<<state.getStateSetStackSize()<<std::endl;callingContext->makeCurrent();}// 执行“后序渲染”渲染台的绘制（RenderStage::drawPostRenderStages）// render all the post draw callbacksdrawPostRenderStages(renderInfo,previous);// 执行摄像机的绘制结束回调（Camera::setFinalDrawCallback）if (_camera.valid() && _camera->getFinalDrawCallback()){// if we have a camera with a final callback invoke it._camera->getFinalDrawCallback()->run(renderInfo);}// pop the render stages camera.if (_camera.valid()) renderInfo.popCamera();// clean up state graph to make sure RenderLeaf etc, can be reusedif (_rootStateGraph.valid()) _rootStateGraph->clean();
}

可见场景绘制时总共会执行五种不同时机下调用的摄像机回调（尤其注意回调时机与渲染上下文的关系），根据我们的实际需要，可以选择在某个回调中执行OpenGL函数（初始化与结束回调时不能执行）或者自定义代码。

osgUtil::RenderStage::runCameraSetUp
  • 实现了一个场景渲染过程中可能非常重要的功能，即纹理烘焙（Render To Texture，RTT），或者称之为“渲染到纹理”。RTT技术意味着我们可以将后台实时绘制得到的场景图像直接作为另一个场景中对象的纹理，从而实现更加丰富的场景表达效果。
  • RTT实现的基本步骤为：（1）首先创建一个“渲染纹理”（Render Texture），例如FBO帧缓存对象，像素缓存对象等；（2）设置它为图形设备的渲染目标（Render Target）；（3）将“渲染纹理”绑定到一个纹理或图片对象上；（4）此时图形设备的渲染将在后台进行，其结果将直接体现在所绑定的纹理对象上。
  • “渲染目标”的设定即通过Camera::getRenderTargetImplementation函数实现。其可用的传入参数中，FRAME_BUFFER表示帧缓存，可以适用于较广泛的硬件平台上（例如屏幕）；而FRAME_BUFFER_OBJECT表示FBO对象，它可以用来实现离屏渲染（Offscreen Rendering）的工作，其渲染结果不会体现在图形窗口中。
  • 渲染目标和绑定纹理的方法：纹理对象texture与场景绘制的帧缓存（渲染目标）绑定在一起
// 如果将渲染目标设置为FBO，实际上意义为，将设置为渲染目标的渲染纹理（FBO）绑定到相机的纹理附件(attach,attachment)上，
// 结果将直接体现在所绑定的纹理对象/附件上
osg::Texture2D* texture = new osg::Texture2D;
camera->setRenderTargetImplementation( osg::Camera::FRAME_BUFFER );
camera->attach( osg::Camera::COLOR_BUFFER, texture );

我们既可以将texture的内容保存成图片，作为场景的截图；也可以将纹理绑定到某个物体上，实现纹理烘焙的效果。这里osgprerender是一个很好的例子，使用附加参数–fb，–fbo，–pbuffer，–window等可以充分了解不同渲染目标实现的过程及其差异。

  • 绑定到摄像机的实际纹理或者图片，在Camera类中均使用Camera::Attachment结构体来保存

RenderStage::runCameraSetUp则反复遍历名为Camera::BufferAttachmentMap的映射表，检索并设置那些与颜色缓存（COLOR_BUFFER），深度缓存（DEPTH_BUFFER）等相对应的Attachment对象。
关于纹理细节可以参考这里。

  • osgUtil::RenderStage中：
//osgUtil::RenderStage
std::map< osg::Camera::BufferComponent, Attachment> _bufferAttachmentMap;enum BufferComponent{DEPTH_BUFFER,STENCIL_BUFFER,PACKED_DEPTH_STENCIL_BUFFER,COLOR_BUFFER,COLOR_BUFFER0,COLOR_BUFFER1 = COLOR_BUFFER0+1,COLOR_BUFFER2 = COLOR_BUFFER0+2,COLOR_BUFFER3 = COLOR_BUFFER0+3,COLOR_BUFFER4 = COLOR_BUFFER0+4,COLOR_BUFFER5 = COLOR_BUFFER0+5,COLOR_BUFFER6 = COLOR_BUFFER0+6,COLOR_BUFFER7 = COLOR_BUFFER0+7,COLOR_BUFFER8 = COLOR_BUFFER0+8,COLOR_BUFFER9 = COLOR_BUFFER0+9,COLOR_BUFFER10 = COLOR_BUFFER0+10,COLOR_BUFFER11 = COLOR_BUFFER0+11,COLOR_BUFFER12 = COLOR_BUFFER0+12,COLOR_BUFFER13 = COLOR_BUFFER0+13,COLOR_BUFFER14 = COLOR_BUFFER0+14,COLOR_BUFFER15 = COLOR_BUFFER0+15};struct Attachment{osg::ref_ptr<osg::Image>                _image;GLenum                                  _imageReadPixelFormat;GLenum                                  _imageReadPixelDataType;};
  • osg::camera中：
//osg::camera中
typedef std::map< BufferComponent, Attachment> BufferAttachmentMap;
BufferAttachmentMap	_bufferAttachmentMap;//camera成员
struct Attachment{Attachment():_internalFormat(GL_NONE),_level(0),_face(0),_mipMapGeneration(false),_multisampleSamples(0),_multisampleColorSamples(0) {}int width() const{if (_texture.valid()) return _texture->getTextureWidth();if (_image.valid()) return _image->s();return 0;};int height() const{if (_texture.valid()) return _texture->getTextureHeight();if (_image.valid()) return _image->t();return 0;};int depth() const{if (_texture.valid()) return _texture->getTextureDepth();if (_image.valid()) return _image->r();return 0;};GLenum              _internalFormat;ref_ptr<Image>      _image;ref_ptr<Texture>    _texture;unsigned int        _level;unsigned int        _face;bool                _mipMapGeneration;unsigned int        _multisampleSamples;unsigned int        _multisampleColorSamples;};void Camera::attach(BufferComponent buffer, GLenum internalFormat)
{switch(buffer){case DEPTH_BUFFER:if(_bufferAttachmentMap.find(PACKED_DEPTH_STENCIL_BUFFER) != _bufferAttachmentMap.end()){OSG_WARN << "Camera: DEPTH_BUFFER already attached as PACKED_DEPTH_STENCIL_BUFFER !" << std::endl;}break;case STENCIL_BUFFER:if(_bufferAttachmentMap.find(PACKED_DEPTH_STENCIL_BUFFER) != _bufferAttachmentMap.end()){OSG_WARN << "Camera: STENCIL_BUFFER already attached as PACKED_DEPTH_STENCIL_BUFFER !" << std::endl;}break;case PACKED_DEPTH_STENCIL_BUFFER:if(_bufferAttachmentMap.find(DEPTH_BUFFER) != _bufferAttachmentMap.end()){OSG_WARN << "Camera: DEPTH_BUFFER already attached !" << std::endl;}if(_bufferAttachmentMap.find(STENCIL_BUFFER) != _bufferAttachmentMap.end()){OSG_WARN << "Camera: STENCIL_BUFFER already attached !" << std::endl;}break;default:break;}_bufferAttachmentMap[buffer]._internalFormat = internalFormat;
}void Camera::attach(BufferComponent buffer, osg::Texture* texture, unsigned int level, unsigned int face, bool mipMapGeneration,unsigned int multisampleSamples,unsigned int multisampleColorSamples)
{_bufferAttachmentMap[buffer]._texture = texture;_bufferAttachmentMap[buffer]._level = level;_bufferAttachmentMap[buffer]._face = face;_bufferAttachmentMap[buffer]._mipMapGeneration = mipMapGeneration;_bufferAttachmentMap[buffer]._multisampleSamples = multisampleSamples;_bufferAttachmentMap[buffer]._multisampleColorSamples = multisampleColorSamples;
}
  • RenderStage::runCameraSetUp：

void RenderStage::runCameraSetUp(osg::RenderInfo& renderInfo)
{_cameraRequiresSetUp = false;if (!_camera) return;OSG_INFO<<"RenderStage::runCameraSetUp(osg::RenderInfo& renderInfo) "<<this<<std::endl;_cameraAttachmentMapModifiedCount = _camera->getAttachmentMapModifiedCount();osg::State& state = *renderInfo.getState();osg::Camera::RenderTargetImplementation renderTargetImplementation = _camera->getRenderTargetImplementation();osg::Camera::RenderTargetImplementation renderTargetFallback = _camera->getRenderTargetFallback();osg::Camera::BufferAttachmentMap& bufferAttachments = _camera->getBufferAttachmentMap();_bufferAttachmentMap.clear();// compute the required dimensionsint width = static_cast<int>(_viewport->x() + _viewport->width());int height = static_cast<int>(_viewport->y() + _viewport->height());int depth = 1;for(osg::Camera::BufferAttachmentMap::iterator itr = bufferAttachments.begin();itr != bufferAttachments.end();++itr){width = osg::maximum(width,itr->second.width());height = osg::maximum(height,itr->second.height());depth = osg::maximum(depth,itr->second.depth());}// OSG_NOTICE<<"RenderStage::runCameraSetUp viewport "<<_viewport->x()<<" "<<_viewport->y()<<" "<<_viewport->width()<<" "<<_viewport->height()<<std::endl;// OSG_NOTICE<<"RenderStage::runCameraSetUp computed "<<width<<" "<<height<<" "<<depth<<std::endl;// attach images that need to be copied after the stage is drawn.for(osg::Camera::BufferAttachmentMap::iterator itr = bufferAttachments.begin();itr != bufferAttachments.end();++itr){// if one exist attach image to the RenderStage.if (itr->second._image.valid()){osg::Image* image = itr->second._image.get();GLenum pixelFormat = image->getPixelFormat();GLenum dataType = image->getDataType();if (image->data()==0){if (pixelFormat==0) pixelFormat = itr->second._internalFormat;if (pixelFormat==0) pixelFormat = _imageReadPixelFormat;if (pixelFormat==0) pixelFormat = GL_RGBA;if (dataType==0) dataType = _imageReadPixelDataType;if (dataType==0) dataType = GL_UNSIGNED_BYTE;}_bufferAttachmentMap[itr->first]._imageReadPixelFormat = pixelFormat;_bufferAttachmentMap[itr->first]._imageReadPixelDataType = dataType;_bufferAttachmentMap[itr->first]._image = image;}if (itr->second._texture.valid()){osg::Texture* texture = itr->second._texture.get();osg::Texture1D* texture1D = 0;osg::Texture2D* texture2D = 0;osg::Texture2DMultisample* texture2DMS = 0;osg::Texture3D* texture3D = 0;osg::TextureCubeMap* textureCubeMap = 0;osg::TextureRectangle* textureRectangle = 0;if (0 != (texture1D=dynamic_cast<osg::Texture1D*>(texture))){if (texture1D->getTextureWidth()==0){texture1D->setTextureWidth(width);}}else if (0 != (texture2D = dynamic_cast<osg::Texture2D*>(texture))){if (texture2D->getTextureWidth()==0 || texture2D->getTextureHeight()==0){texture2D->setTextureSize(width,height);}}else if (0 != (texture2DMS = dynamic_cast<osg::Texture2DMultisample*>(texture))){if (texture2DMS->getTextureWidth()==0 || texture2DMS->getTextureHeight()==0){texture2DMS->setTextureSize(width,height);}}else if (0 != (texture3D = dynamic_cast<osg::Texture3D*>(texture))){if (texture3D->getTextureWidth()==0 || texture3D->getTextureHeight()==0 || texture3D->getTextureDepth()==0 ){// note we dont' have the depth here, so we'll heave to assume that height and depth are the same..texture3D->setTextureSize(width,height,height);}}else if (0 != (textureCubeMap = dynamic_cast<osg::TextureCubeMap*>(texture))){if (textureCubeMap->getTextureWidth()==0 || textureCubeMap->getTextureHeight()==0){textureCubeMap->setTextureSize(width,height);}}else if (0 != (textureRectangle = dynamic_cast<osg::TextureRectangle*>(texture))){if (textureRectangle->getTextureWidth()==0 || textureRectangle->getTextureHeight()==0){textureRectangle->setTextureSize(width,height);}}}}if (renderTargetImplementation==osg::Camera::FRAME_BUFFER_OBJECT){osg::GLExtensions* ext = state.get<osg::GLExtensions>();bool fbo_supported = ext->isFrameBufferObjectSupported;if (fbo_supported){OSG_INFO<<"Setting up osg::Camera::FRAME_BUFFER_OBJECT"<<std::endl;OpenThreads::ScopedLock<OpenThreads::Mutex> lock(*(_camera->getDataChangeMutex()));osg::ref_ptr<osg::FrameBufferObject> fbo = new osg::FrameBufferObject;osg::ref_ptr<osg::FrameBufferObject> fbo_multisample;bool colorAttached = false;bool depthAttached = false;bool stencilAttached = false;unsigned samples = 0;unsigned colorSamples = 0;// This is not a cut and paste error. Set BOTH local masks// to the value of the Camera's use render buffers mask.// We'll change this if and only if we decide we're doing MSFBO.unsigned int renderBuffersMask = _camera->getImplicitBufferAttachmentRenderMask(true);unsigned int resolveBuffersMask = _camera->getImplicitBufferAttachmentRenderMask(true);if (ext->isRenderbufferMultisampleSupported()){for(osg::Camera::BufferAttachmentMap::iterator itr = bufferAttachments.begin();itr != bufferAttachments.end();++itr){osg::Camera::Attachment& attachment = itr->second;samples = maximum(samples, attachment._multisampleSamples);colorSamples = maximum(colorSamples, attachment._multisampleColorSamples);}if (colorSamples > samples){OSG_NOTIFY(WARN) << "Multisample color samples must be less than or ""equal to samples. Setting color samples equal to samples." << std::endl;colorSamples = samples;}if (samples){fbo_multisample = new osg::FrameBufferObject;// Use the value of the Camera's use resolve buffers mask as the// resolve mask.// 可以自定义，也可以从camera._displaySettings中获得resolveBuffersMask = _camera->getImplicitBufferAttachmentResolveMask(true);}}for(osg::Camera::BufferAttachmentMap::iterator itr = bufferAttachments.begin();itr != bufferAttachments.end();++itr){osg::Camera::BufferComponent buffer = itr->first;osg::Camera::Attachment& attachment = itr->second;// 往帧缓冲对象中设置各种附件if (attachment._texture.valid() || attachment._image.valid())fbo->setAttachment(buffer, osg::FrameBufferAttachment(attachment));// 或先渲染到renderbufferelsefbo->setAttachment(buffer, osg::FrameBufferAttachment(new osg::RenderBuffer(width, height, attachment._internalFormat)));if (fbo_multisample.valid()){GLenum internalFormat = attachment._internalFormat;if (!internalFormat){switch (buffer){case Camera::DEPTH_BUFFER:internalFormat = GL_DEPTH_COMPONENT24;break;case Camera::STENCIL_BUFFER:internalFormat = GL_STENCIL_INDEX8_EXT;break;case Camera::PACKED_DEPTH_STENCIL_BUFFER:internalFormat = GL_DEPTH_STENCIL_EXT;break;// all other buffers are color buffersdefault:// setup the internal format based on attached texture if such exists, otherwise just default formatif (attachment._texture)internalFormat = attachment._texture->getInternalFormat();elseinternalFormat = GL_RGBA;break;}}fbo_multisample->setAttachment(buffer,osg::FrameBufferAttachment(new osg::RenderBuffer(width, height, internalFormat,samples, colorSamples)));}if (buffer==osg::Camera::DEPTH_BUFFER) depthAttached = true;else if (buffer==osg::Camera::STENCIL_BUFFER) stencilAttached = true;else if (buffer==osg::Camera::PACKED_DEPTH_STENCIL_BUFFER){depthAttached = true;stencilAttached = true;}else if (buffer>=osg::Camera::COLOR_BUFFER) colorAttached = true;}if (!depthAttached){// If doing MSFBO (and therefore need two FBOs, one for multisampled rendering and one for// final resolve), then configure "fbo" as the resolve FBO, and When done// configuring, swap it into "_resolveFbo" (see line 554). But, if not// using MSFBO, then "fbo" is just the render fbo.// If using MSFBO, then resolveBuffersMask// is the value set by the app for the resolve buffers. But if not using// MSFBO, then resolveBuffersMask is the value set by the app for render// buffers. In both cases, resolveBuffersMask is used to configure "fbo".if( resolveBuffersMask & osg::Camera::IMPLICIT_DEPTH_BUFFER_ATTACHMENT ){fbo->setAttachment(osg::Camera::DEPTH_BUFFER, osg::FrameBufferAttachment(new osg::RenderBuffer(width, height, GL_DEPTH_COMPONENT24)));depthAttached = true;}if (fbo_multisample.valid() &&( renderBuffersMask & osg::Camera::IMPLICIT_DEPTH_BUFFER_ATTACHMENT ) ){fbo_multisample->setAttachment(osg::Camera::DEPTH_BUFFER,osg::FrameBufferAttachment(new osg::RenderBuffer(width,height, GL_DEPTH_COMPONENT24, samples, colorSamples)));}}if (!stencilAttached){if( resolveBuffersMask & osg::Camera::IMPLICIT_STENCIL_BUFFER_ATTACHMENT ){fbo->setAttachment(osg::Camera::STENCIL_BUFFER, osg::FrameBufferAttachment(new osg::RenderBuffer(width, height, GL_STENCIL_INDEX8_EXT)));stencilAttached = true;}if (fbo_multisample.valid() &&( renderBuffersMask & osg::Camera::IMPLICIT_STENCIL_BUFFER_ATTACHMENT ) ){fbo_multisample->setAttachment(osg::Camera::STENCIL_BUFFER,osg::FrameBufferAttachment(new osg::RenderBuffer(width,height, GL_STENCIL_INDEX8_EXT, samples, colorSamples)));}}if (!colorAttached){if( resolveBuffersMask & osg::Camera::IMPLICIT_COLOR_BUFFER_ATTACHMENT ){fbo->setAttachment(osg::Camera::COLOR_BUFFER, osg::FrameBufferAttachment(new osg::RenderBuffer(width, height, GL_RGB)));colorAttached = true;}if (fbo_multisample.valid() &&( renderBuffersMask & osg::Camera::IMPLICIT_COLOR_BUFFER_ATTACHMENT ) ){fbo_multisample->setAttachment(osg::Camera::COLOR_BUFFER,osg::FrameBufferAttachment(new osg::RenderBuffer(width,height, GL_RGB, samples, colorSamples)));}}fbo->apply(state);// If no color attachment make sure to set glDrawBuffer/glReadBuffer to none// otherwise glCheckFramebufferStatus will fail// It has to be done after call to glBindFramebuffer (fbo->apply)// and before call to glCheckFramebufferStatusif ( !colorAttached ){#if !defined(OSG_GLES1_AVAILABLE) && !defined(OSG_GLES2_AVAILABLE) && !defined(OSG_GLES3_AVAILABLE)setDrawBuffer( GL_NONE, true );state.glDrawBuffer( GL_NONE );#endif}GLenum status = ext->glCheckFramebufferStatus(GL_FRAMEBUFFER_EXT);if (status != GL_FRAMEBUFFER_COMPLETE_EXT){OSG_NOTICE<<"RenderStage::runCameraSetUp(), FBO setup failed, FBO status= 0x"<<std::hex<<status<<std::dec<<std::endl;fbo_supported = false;GLuint fboId = state.getGraphicsContext() ? state.getGraphicsContext()->getDefaultFboId() : 0;ext->glBindFramebuffer(GL_FRAMEBUFFER_EXT, fboId);fbo = 0;// clean up.osg::get<osg::GLRenderBufferManager>(state.getContextID())->flushAllDeletedGLObjects();osg::get<osg::GLFrameBufferObjectManager>(state.getContextID())->flushAllDeletedGLObjects();}else{setDrawBuffer(GL_NONE, false );setReadBuffer(GL_NONE, false );_fbo = fbo;if (fbo_multisample.valid()){fbo_multisample->apply(state);status = ext->glCheckFramebufferStatus(GL_FRAMEBUFFER_EXT);if (status != GL_FRAMEBUFFER_COMPLETE_EXT){OSG_NOTICE << "RenderStage::runCameraSetUp(), ""multisample FBO setup failed, FBO status = 0x"<< std::hex << status << std::dec << std::endl;fbo->apply(state);fbo_multisample = 0;_resolveFbo = 0;// clean up.osg::get<osg::GLRenderBufferManager>(state.getContextID())->flushAllDeletedGLObjects();osg::get<osg::GLFrameBufferObjectManager>(state.getContextID())->flushAllDeletedGLObjects();}else{_resolveFbo.swap(_fbo);_fbo = fbo_multisample;}}else{_resolveFbo = 0;}}}if (!fbo_supported){if (renderTargetImplementation<renderTargetFallback)renderTargetImplementation = renderTargetFallback;elserenderTargetImplementation = osg::Camera::PIXEL_BUFFER_RTT;}}}

https://github.com/openscenegraph/OpenSceneGraph/blob/OpenSceneGraph-3.6/src/osgUtil/RenderStage.cpp

https://github.com/openscenegraph/OpenSceneGraph/blob/OpenSceneGraph-3.6/src/osg/FrameBufferObject.cpp

  • 在osg::FrameBufferObject.cpp中：
void FrameBufferObject::apply(State &state) const
{apply(state, READ_DRAW_FRAMEBUFFER);
}
void FrameBufferObject::apply(State &state, BindTarget target) const
{unsigned int contextID = state.getContextID();if (_unsupported[contextID])return;GLExtensions* ext = state.get<GLExtensions>();if (!ext->isFrameBufferObjectSupported){_unsupported[contextID] = 1;OSG_WARN << "Warning: EXT_framebuffer_object is not supported" << std::endl;return;}if (_attachments.empty()){ext->glBindFramebuffer(target, 0);return;}int &dirtyAttachmentList = _dirtyAttachmentList[contextID];GLuint &fboID = _fboID[contextID];if (fboID == 0){ext->glGenFramebuffers(1, &fboID);if (fboID == 0){OSG_WARN << "Warning: FrameBufferObject: could not create the FBO" << std::endl;return;}dirtyAttachmentList = 1;}if (dirtyAttachmentList){// the set of of attachments appears to be thread sensitive, it shouldn't be because// OpenGL FBO handles osg::FrameBufferObject has are multi-buffered...// so as a temporary fix will stick in a mutex to ensure that only one thread passes through here// at one time.static OpenThreads::Mutex s_mutex;OpenThreads::ScopedLock<OpenThreads::Mutex> lock(s_mutex);// create textures and mipmaps before we bind the frame buffer objectfor (AttachmentMap::const_iterator i=_attachments.begin(); i!=_attachments.end(); ++i){const FrameBufferAttachment &fa = i->second;fa.createRequiredTexturesAndApplyGenerateMipMap(state, ext);}}ext->glBindFramebuffer(target, fboID);// enable drawing buffers to render the result to fboif ( (target == READ_DRAW_FRAMEBUFFER) || (target == DRAW_FRAMEBUFFER) ){if (_drawBuffers.size() > 0){GLExtensions *gl2e = state.get<GLExtensions>();if (gl2e && gl2e->glDrawBuffers){gl2e->glDrawBuffers(_drawBuffers.size(), &(_drawBuffers[0]));}else{OSG_WARN <<"Warning: FrameBufferObject: could not set draw buffers, glDrawBuffers is not supported!" << std::endl;}}}if (dirtyAttachmentList){for (AttachmentMap::const_iterator i=_attachments.begin(); i!=_attachments.end(); ++i){const FrameBufferAttachment &fa = i->second;switch(i->first){case(Camera::PACKED_DEPTH_STENCIL_BUFFER):if (ext->isPackedDepthStencilSupported){fa.attach(state, target, GL_DEPTH_ATTACHMENT_EXT, ext);fa.attach(state, target, GL_STENCIL_ATTACHMENT_EXT, ext);}else{OSG_WARN <<"Warning: FrameBufferObject: could not attach PACKED_DEPTH_STENCIL_BUFFER, ""EXT_packed_depth_stencil is not supported!" << std::endl;}break;default:fa.attach(state, target, convertBufferComponentToGLenum(i->first), ext);break;}}dirtyAttachmentList = 0;}}
osgUtil::RenderStage::copyTexture

RenderStage::copyTexture则负责针对FRAME_BUFFER渲染目标，拷贝场景图像到Attachment对象（绑定到摄像机的实际纹理或者图片）中。

这篇关于最长的一帧学习 part4的文章就介绍到这儿，希望我们推荐的文章对编程师们有所帮助！


http://www.chinasem.cn/article/1121835。

相关文章

HarmonyOS学习(七)——UI（五）常用布局总结

HarmonyOS学习(七)——UI（五）常用布局总结

自适应布局 1.1、线性布局（LinearLayout） 通过线性容器Row和Column实现线性布局。Column容器内的子组件按照垂直方向排列，Row组件中的子组件按照水平方向排列。 属性说明space通过space参数设置主轴上子组件的间距，达到各子组件在排列上的等间距效果alignItems设置子组件在交叉轴上的对齐方式，且在各类尺寸屏幕上表现一致，其中交叉轴为垂直时，取值为Vert
阅读更多...
Ilya-AI分享的他在OpenAI学习到的15个提示工程技巧

Ilya-AI分享的他在OpenAI学习到的15个提示工程技巧

Ilya（不是本人，claude AI）在社交媒体上分享了他在OpenAI学习到的15个Prompt撰写技巧。 以下是详细的内容： 提示精确化:在编写提示时,力求表达清晰准确。清楚地阐述任务需求和概念定义至关重要。例:不用"分析文本",而用"判断这段话的情感倾向:积极、消极还是中性"。 快速迭代:善于快速连续调整提示。熟练的提示工程师能够灵活地进行多轮优化。例:从"总结文章"到"用
阅读更多...
【前端学习】AntV G6-08 深入图形与图形分组、自定义节点、节点动画（下）

【前端学习】AntV G6-08 深入图形与图形分组、自定义节点、节点动画（下）

【课程链接】 AntV G6：深入图形与图形分组、自定义节点、节点动画（下）_哔哩哔哩_bilibili 本章十吾老师讲解了一个复杂的自定义节点中，应该怎样去计算和绘制图形，如何给一个图形制作不间断的动画，以及在鼠标事件之后产生动画。（有点难，需要好好理解） <!DOCTYPE html><html><head><meta charset="UTF-8"><title>06
阅读更多...
学习hash总结

学习hash总结

2014/1/29/   最近刚开始学hash，名字很陌生，但是hash的思想却很熟悉，以前早就做过此类的题，但是不知道这就是hash思想而已，说白了hash就是一个映射，往往灵活利用数组的下标来实现算法，hash的作用：1、判重；2、统计次数；
阅读更多...
poj3261（可重复k次的最长子串）

poj3261（可重复k次的最长子串）

题意：可重复k次的最长子串 解题思路：求所有区间[x,x+k-1]中的最小值的最大值。求sa时间复杂度Nlog(N),求最值时间复杂度N*N，但实际复杂度很低。题目数据也比较水，不然估计过不了。 代码入下： #include<iostream>#include<algorithm>#include<stdio.h>#include<math.h>#include<cstring
阅读更多...
零基础学习Redis(10) -- zset类型命令使用

零基础学习Redis(10) -- zset类型命令使用

zset是有序集合，内部除了存储元素外，还会存储一个score，存储在zset中的元素会按照score的大小升序排列，不同元素的score可以重复，score相同的元素会按照元素的字典序排列。 1. zset常用命令 1.1 zadd  zadd key [NX | XX] [GT | LT]   [CH] [INCR] score member [score member ...]
阅读更多...
poj 3974 and hdu 3068 最长回文串的O（n）解法(Manacher算法)

poj 3974 and hdu 3068 最长回文串的O（n）解法(Manacher算法)

求一段字符串中的最长回文串。 因为数据量比较大，用原来的O（n^2）会爆。 小白上的O(n^2)解法代码：TLE啦~ #include<stdio.h>#include<string.h>const int Maxn = 1000000;char s[Maxn];int main(){char e[] = {"END"};while(scanf("%s", s) != EO
阅读更多...
【机器学习】高斯过程的基本概念和应用领域以及在python中的实例

【机器学习】高斯过程的基本概念和应用领域以及在python中的实例

引言 高斯过程（Gaussian Process，简称GP）是一种概率模型，用于描述一组随机变量的联合概率分布，其中任何一个有限维度的子集都具有高斯分布 文章目录 引言一、高斯过程1.1 基本定义1.1.1 随机过程1.1.2 高斯分布 1.2 高斯过程的特性1.2.1 联合高斯性1.2.2 均值函数1.2.3 协方差函数（或核函数） 1.3 核函数1.4 高斯过程回归（Gauss
阅读更多...
【学习笔记】 陈强-机器学习-Python-Ch15 人工神经网络（1）sklearn

【学习笔记】 陈强-机器学习-Python-Ch15 人工神经网络（1）sklearn

系列文章目录 监督学习：参数方法 【学习笔记】 陈强-机器学习-Python-Ch4 线性回归 【学习笔记】 陈强-机器学习-Python-Ch5 逻辑回归 【课后题练习】 陈强-机器学习-Python-Ch5 逻辑回归（SAheart.csv） 【学习笔记】 陈强-机器学习-Python-Ch6 多项逻辑回归 【学习笔记 及 课后题练习】 陈强-机器学习-Python-Ch7 判别分析 【学
阅读更多...
系统架构师考试学习笔记第三篇——架构设计高级知识（20）通信系统架构设计理论与实践

系统架构师考试学习笔记第三篇——架构设计高级知识（20）通信系统架构设计理论与实践

本章知识考点：         第20课时主要学习通信系统架构设计的理论和工作中的实践。根据新版考试大纲,本课时知识点会涉及案例分析题(25分),而在历年考试中,案例题对该部分内容的考查并不多,虽在综合知识选择题目中经常考查,但分值也不高。本课时内容侧重于对知识点的记忆和理解,按照以往的出题规律,通信系统架构设计基础知识点多来源于教材内的基础网络设备、网络架构和教材外最新时事热点技术。本课时知识
阅读更多...

Copyright China编程 23002807@qq.com

沪ICP备2023033072号-2