Hi Alex, A quick reply as haven't yet gone through your post with a fine tooth comb to work out exactly what you are doing and what you understand.
As general note, osgvolume is an example and a testbed. For much of the osgvolume development work I have used the osgvolume example to test out the new alogorithm development. This dual purpose means that there is lots of code in there for testing out a wide range of capabilities of osgVolume rather than illustrating how to use it in the simplistic way. For you own work it may be useful to lean the use the existing osgvolume example from the command line as this could allow you to try out various options. There are various keyboard/mouse capabilities in osgvolume so you can test things interactively when you run osgvolume. Run osgvolume --help to get a list of command line options. With your data I'd suggest trying to load the data directly with osgvolume. I would also recommend using either RayTraceTechnique or the MultipassTechnique (found in OSG-3.4) as these are both what osgVolume is designed out. The FixedFunctionTechnique is a fallback for when shaders aren't supported by drivers/hardware. Almost all modern graphics cards can handle shaders+3d texturing so there shouldn't be any need to fallback to FixedFunctionTechnique. If your datra is full colour already then there is no need to apply a transfer function unless you specifically want to remap the colours. With full colour data the alpha value is used to determine the intensity/alpha blending of the ray so if the alpha value is 1.0 then you'll get a solid cube. The osgvolume has code for computing alpha values for data. Robert. On 14 December 2015 at 21:47, Alex Taylor <[email protected]> wrote: > For context, I come from an image processing and signal processing > background, so computer graphics and OSG are both new to me. > > I'm trying to use osgVolume for volume rendering some data. The only real > example code I can find out there is the osgvolume example that ships with > OSG. I'm a bit confused about the usage model with the various > VolumeTechnique classes. > > I'm working with GL_RGBA uint8 data. In my code, I define a function > createTexture3D that creates an osg::Image object. The image is defined as: > > * osg::ref_ptr<osg::Image> image_3d = new osg::Image;* > * image_3d->allocateImage(nmSAFE_CAST(int, s_nearestPowerOfTwo),* > * nmSAFE_CAST(int, t_nearestPowerOfTwo),* > * nmSAFE_CAST(int, r_nearestPowerOfTwo),* > * desiredPixelFormat, GL_UNSIGNED_BYTE);* > > When I started with a simple example, FixedFunctionTechnique rendered my > data as I expected. See fixedFunctionTechnique.png attached. > > *osg::ref_ptr<osgVolume::VolumeTile> tile = new osgVolume::VolumeTile;* > * volume->addChild(tile.get());* > > * osg::ref_ptr<osg::Image> image_3d = > createTexture3D(data,xfer_table);* > > * osg::ref_ptr<osgVolume::ImageLayer> layer = new > osgVolume::ImageLayer(image_3d.get());* > > * tile->setLayer(layer.get());* > > * tile->setVolumeTechnique(new > osgVolume::FixedFunctionTechnique());* > > * // FixedFunctionTechnique turns on GL_LIGHTING, which breaks > the color rendering.* > * osg::StateSet* stateset = volume->getOrCreateStateSet();* > * stateset->setMode(GL_LIGHTING,osg::StateAttribute::OFF | > osg::StateAttribute::OVERRIDE);* > > * // Our original implementation positioned the bbox [-0.5,0.5] > in each dimension.* > * // FixedFunctionTechnique applies the locator matrix to the a > unit cube [0 1] in each dimension.* > * // To get the equivalent spatial referencing, apply a > translation of -0.5 to each dimension. * > * osg::ref_ptr<osg::RefMatrix> matrix = new osg::RefMatrix();* > * matrix->makeTranslate(-0.5,-0.5,-0.5);* > > I attempted to modify this simple example to use RayTracedTechnique next. > When I do that, both the color and the opacity of my volume data is off. In > some cases, I see something on the screen (see attached screenshot, in > other cases, I don't see anything). I almost feel like somehow my data is > outside of an expected colormap/alphamap range, but I can't figure out > what's off. In this particular example with an engine block CT dataset and > a specific colormap/alphamap, I see what is shown in the attached > rayTracedTechnique.png. Here is a partial code listing based on the OSG > example: > > > * osg::ref_ptr<osg::Image> image_3d = createTexture3D(data,xfer_table);* > > * osg::ref_ptr<osgVolume::ImageLayer> layer = new > osgVolume::ImageLayer(image_3d.get());* > > * tile->setLayer(layer.get());* > > * osgVolume::SwitchProperty* sp = new > osgVolume::SwitchProperty;* > * sp->setActiveProperty(0);* > > * osg::TransferFunction1D::ColorMap colorMap;* > * for (size_t r = 0; r < 256; ++r){* > * size_t rowOffset = r*4;* > * float red = > static_cast<float>(xfer_table[rowOffset])/255.0;* > * float green = > static_cast<float>(xfer_table[rowOffset+1])/255.0;* > * float blue = > static_cast<float>(xfer_table[rowOffset+2])/255.0;* > * float alpha = > static_cast<float>(xfer_table[rowOffset+3])/255.0;* > * colorMap[r] = osg::Vec4(red,green,blue,alpha);* > * }* > > * osg::ref_ptr<osg::TransferFunction1D> transferFunction = new > osg::TransferFunction1D;* > * transferFunction->assign(colorMap);* > > * float alphaFunc=0.02f;* > * float sampleDensityWhenMoving = 0.02;* > * osgVolume::AlphaFuncProperty* ap = new > osgVolume::AlphaFuncProperty(alphaFunc);* > * osgVolume::SampleDensityProperty* sd = new > osgVolume::SampleDensityProperty(0.005);* > * osgVolume::SampleDensityWhenMovingProperty* sdwm = > sampleDensityWhenMoving!=0.0 ? new > osgVolume::SampleDensityWhenMovingProperty(sampleDensityWhenMoving) : 0;* > * osgVolume::TransparencyProperty* tp = new > osgVolume::TransparencyProperty(1.0);* > * osgVolume::TransferFunctionProperty* tfp = > transferFunction.valid() ? new > osgVolume::TransferFunctionProperty(transferFunction.get()) : 0;* > > * // Standard config from osgVolume example* > * osgVolume::CompositeProperty* cp = new > osgVolume::CompositeProperty;* > * cp->addProperty(ap);* > * cp->addProperty(sd);* > * cp->addProperty(tp);* > * if (sdwm) cp->addProperty(sdwm);* > * // if (tfp) cp->addProperty(tfp);* > > * sp->addProperty(cp);* > * sp->setActiveProperty(0); // For now, always use "Standard" > config* > * layer->addProperty(sp);* > > * tile->setVolumeTechnique(new > osgVolume::RayTracedTechnique());* > > * // FixedFunctionTechnique turns on GL_LIGHTING, which breaks > the color rendering.* > * osg::StateSet* stateset = volume->getOrCreateStateSet();* > * // stateset->setMode(GL_LIGHTING,osg::StateAttribute::OFF | > osg::StateAttribute::OVERRIDE);* > > * // Our original implementation positioned the bbox [-0.5,0.5] > in each dimension.* > * // FixedFunctionTechnique applies the locator matrix to the a > unit cube [0 1] in each dimension.* > * // To get the equivalent spatial referencing, apply a > translation of -0.5 to each dimension.* > * osg::ref_ptr<osg::RefMatrix> matrix = new osg::RefMatrix();* > * matrix->makeTranslate(-0.5,-0.5,-0.5);* > > * tile->setLocator(new osgVolume::Locator(*matrix));* > > I've spent the last few hours toggling settings on/off. I thought I'd see > if anyone here has ideas, or how I would go about debugging why things > aren't rendering correctly. > > I also have a specific question. When working with RGBA data, should I be > specifying a *osgVolume::TransferFunctionProperty*, and should this > property be in the float [0,1] normalized range as shown in the OSG > example, or should it be in a range that matches my data? I'm a bit > confused about whether/why I need to specify the transfer function when my > RGBA texture already knows the color and opacity for each voxel sample. > > Sorry for the very long email. I'd appreciate any help or pointers anyone > can provide. > > Best, > > Alex Taylor > > > _______________________________________________ > osg-users mailing list > [email protected] > http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org > >
_______________________________________________ osg-users mailing list [email protected] http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
