I'm trying to implement volumetric billboards in OpenGL 3.3+ as described here
and video here.
The problem I'm facing now (quite basic) is: how do I render a 3D object to a 3D texture (as described in the paper) efficiently? Assuming the object could be stored in a 256x256x128 tex creating 256*256*128*2 framebuffers (because it's said that it should be rendered twice at each axis: +X,-X,+Y,-Y,+Z,-Z) would be insane and there are too few texture units to process that many textures as far as I know (not to mention the amount of time needed).
Does anyone have any idea how to deal with something like that?
A slice of 3D texture can be directly attached to the current framebuffer. So, create a frame buffer, a 3D texture and then do rendering like:
glFramebufferTexture3D( GL_FRAMEBUFFER, Attachment, GL_TEXTURE_3D,
TextureID, 0, ZSlice );
...render to the slice of 3D texture...
So, you need only 1 framebuffer that will be iterated by the number of Z-slices in your target 3D texture.
Related
I use OpenGL 3.2 to render shadow maps. For this, I construct a framebuffer that renders to a depth texture.
To attach the texture to the framebuffer, I use:
glFramebufferTexture2D( GL_FRAMEBUFFER, GL_DEPTH_ATTACHMENT, GL_TEXTURE_2D, shdw_texture, 0 );
This works great. After rendering the light view, my GLSL shader can sample the depth texture to solve visibility of light.
The problem I am trying to solve now, is to have many more shadow maps, let's say 50 of them. In my main render pass I don't want to be sampling from 50 different textures. I could use an atlas, but I wondered: could I pass all these shadow maps as slices from a 2D texture array?
So, somehow create a GL_TEXTURE_2D_ARRAY with a DEPTH format, and bind one layer of the array to the framebuffer?
Can framebuffers be backed for DEPTH by a texture array layer, instead of just a depth texture?
In general, you need to distinguish whether you want to create a layered framebuffer (see Layered Images) or whether you want to attach a single layer of a multilayered texture to a framebuffer.
Use glFramebufferTexture3D to attach a layer of a 3D texture (TEXTURE_3D) or array texture to a framebuffer or use glFramebufferTextureLayer to attach a layer of a three-dimensional or array texture to the framebuffer. In either case the last argument specifies the layer of the texture.
Layered attachments can be attached with glFramebufferTexture. See Layered rendering.
I have a 512X512 texture which holds a number of images that i want to use in my application. After adding the image data to the texture i save the texture coords for the individual images. Later i apply these on some quads that i am drawing. The texture has mipmapping activated.
When i take a screenshot of the rendered scene at exactly the same instance in two different runs of the applications, i notice that there are differences in the image only among those quads textured using this mipmapped texture. Can mipmapping cause such an issue?
My best guess is that it has to do with precisions in your shader. Check out this problem that I had (and fought with for a while) and my solution:
opengl texture mapping off by 5-8 pixels
It probably is a combination of mimapping's automatic scaling of your texture atlas and the precision hints in your shader code.
Also see the other linked question:
Why is a texture coordinate of 1.0 getting beyond the edge of the texture?
Wanting to decal multiple irregular textures onto a curved surface (mesh with xyz vertices and uv specified at each). I am loading the mesh from a model file, and don't have any a priori knowledge of the surface... all we know is that it will have a "reasonable" uv mapping. Want to select a few uv regions and apply textures to them. Each region is specified by a bounding poly in uv coordinates. Don't know the equivalent xyz poly in this case, or I think the answer would be simple.
We have this working for flat surfaces and also simple cylindrical surfaces (which we approximate as a series of flat stripes, smoothed by choosing the normal as averages). In both cases we know a unique mapping from uv to xyz so we set up the stencil buffer to limit drawing to the desired uv region by drawing the equivalent xyz poly to the stencil buffer ahead of binding a texture and drawing the real surface.
We are also using rgba transparency within the textures when decaling those onto the surface. Typically each textured region is a small rotated rectangle so we draw the four vertices to the stencil buffer, then use the texture matrix to rotate that, and use the rgba transparency within the texture to ensure only the right part of the texture is applied. This all works nicely.
Would like to reuse our working code, but now apply these textures to an arbitrary curved surface/mesh. We are loading and drawing these models, and can already apply textures to whole faces [ie uv goes from (0,0) to (1,1) ]. Now we want to extend this and apply "placed" textures to regions of each surface.
Thought it might be possible draw the uv poly to the stencil buffer directly, not even knowing the equivalent xyz poly... then all the existing code would work. Perhaps could use some trick like a frame buffer object, and do the initial draw of the stencil poly to that, then using that as the stencil during the "real" draw of the curved surface mesh. Would that be a good approach? Or is there a better way?
Any advice or url links to relevant samples welcome...
PS Have looked at these threads... sort of relevant but not quite the same problem I think...
Binding a stencil render buffer to a frame buffer in opengl
Visualizing the Stencil Buffer to a texture
I am currently looking at some working FBO setup/usage code I have for off-screen shadow mapping, and trying to make it work for this seemingly simpler situation. The bit I'm unclear on is the setup gl calls needed ... I am rather confused about how to set this up. Here's an extract of the hardware shadowing FBO setup with bits chopped out and ?? added... any help on correct sequence here appreciated.
glBindTexture(GL_TEXTURE_2D, tex);
?? not
::glTexImage2D(GL_TEXTURE_2D, 0, GL_DEPTH_COMPONENT32, shadowsize, shadowsize, 0, GL_DEPTH_COMPONENT, GL_UNSIGNED_INT, NULL);
?? but a more normal binding approp to drawing RGBA textures
::glBindTexture(GL_TEXTURE_2D, 0);
glBindFramebufferEXT(GL_FRAMEBUFFER_EXT, m_Framebuffer);
// Attach everything, tell fbo there will be a drawbuffer, unlike shadows tex draw
// ?? use GL_COLOR_ATTACHMENT0_EXT
glDrawBuffer(GL_NONE);
// no color buffer dest...
??wrong glReadBuffer(GL_NONE);
// no color buffer src
?? glFramebufferTexture2DEXT(GL_FRAMEBUFFER_EXT, GL_DEPTH_ATTACHMENT_EXT, GL_TEXTURE_2D, tex, 0);
//??
Note: tex, m_Frambuffer are ints, correctly allocated textureid and framebuffer, think that bit is ok. My main points of confusion are
Seems that code does glBindTexture, glTexImage2D, glBindTexture release to 0: is it correct to release this early?
glDrawBuffer + glReadBuffer calls required?
I'm implementing an algorithm about pencil rendering. First, I should render the model using Phong shading to determine the intensity. Then I should map the texture to the rendered result.
I'm going to do a multipass rendering with opengl and cg shaders. Someone told me that I should try 'render to texture'. But I don't know how to use this method to get the effects that I want. In my opinion, we should first use this method to render the mesh, then we can get a 2D texture about the whole scene. Now that we have draw content to the framebuffer, next we should render to the screen, right? But how to use the rendered texture and do some post-processing on it? Can anybody show me some code or links about it?
I made this tutorial, it might help you : http://www.opengl-tutorial.org/intermediate-tutorials/tutorial-14-render-to-texture/
However, using RTT is overkill for what you're trying to do, I think. If you need the fragment's intensity in the texture, well, you already have it in your shader, so there is no need to render it twice...
Maybe this could be useful ? http://www.ozone3d.net/demos_projects/toon-snow.php
render to a texture with Phong shading
Draw that texture to the screen again in a full screen textured quad, applying a shader that does your desired operation.
I'll assume you need clarification on RTT and using it.
Essentially, your screen is a framebuffer (very similar to a texture); it's a 2D image at the end of the day. The idea of RTT is to capture that 2D image. To do this, the best way is to use a framebuffer object (FBO) (Google "framebuffer object", and click on the first link). From here, you have a 2D picture of your scene (you should check it by saving to an image file that it actually is what you want).
Once you have the image, you'll set up a 2D view and draw that image back onto the screen with an 800x600 quadrilateral or what-have-you. When drawing, you use a fragment program (shader), which transforms the brightness of the image into a greyscale value. You can output this, or you can use it as an offset to another, "pencil" texture.
First example:
You can take a huge rock shaped mesh and put a tiled rock texture all over it.
Now, some places needs to be covered with a grass texture (or other vegetation).
Another example:
Usually, terrain are built from tiled textures. In order to achieve a less "tilly" look, you can apply 4 times bigger (or 16 and so on..) tiled texture on it, and by that you'll gain a nice "random" tiled texture (seen that in the UDK's docs).
Blender (the 3d graphics app) is OpenGL based, and it allows you to assign multiple materials to a single mesh.
How can i do it in my own OpenGL application?
Thanks,
Amir
P.S:
I'm looking for a better solution than rendering 50 tris with tex a and and 3 more tris with tex b.
What you're looking for is called multitexturing. Modern graphics cards have several texture units that each can sample a different texture. When you render your rock you specify vertices that have UV coordinates for each texture you want to render.
In OpenGL you can use glActiveTexture to select your active texture unit so that you can bind a texture to it and use it in subsequent rendering. Your vertices will need additional texture coordinate pairs; one pair per texture you intend to render.
The modern way to do multitexturing is using shaders (GLSL in OpenGL typically). Load and bind each texture to a different texture unit, set your shader uniforms to the value of the texture units (0 for texture unit 0 etc) you're using, sample each texture, and blend using the desired blending function to get your output color.