http://pastebin.com/vkTJt0sT
I am trying to render an image similiar to the left one and having problem with shadows+reflections.
Right now, only the shadow code is open for showing the problem.
As you can see, the red ball should be shadowed near the green one, but the pixels all get messed up for some weird reason. When I close the shadow part of the code, it renders the red ball normally without shadows.
I think the root of this problem is also affecting reflections. hope you guys can give me some tips; I’m losing it.
Given that your left image shows cancer, this is a classic case of the shadow ray hitting the object off which it was reflected. When hit-testing a shadow ray, you need to exclude the surface that generated the ray. Just pass the source object into your shadow function, and ignore it.
This method only works for convex shapes. If you have shapes that do self-shadow (a torus, for example), you need to be more general. The usual approach is to define an epsilon (floating-point error tolerance) and ignore any intersection points that are nearer than that.
The other approach is to detect which side of a surface you hit. You should not self-shadow on a sphere because the ray is being cast in the same general direction as the surface normal (ie the dot product of the outgoing ray and the surface normal is positive) - this should not be counted as a shadow.
Solved:
There was an algorithmic problem which isn't easy to explain,
Another method is: Basicly I make a check if the cosine angle is bigger than 0.0001, if it is bigger then this I don't shadow it
Related
I just implemented a light system in my engine. In the following screenshot you can see a light (the yellow square) in action:
Take into account that on top of the light illuminating the scenario, its also implemented a FOV which will occlude anything outside your field of view. Thats why the left part of the shadow seems so off.
As you can see the light's shadows are pretty "hard", as they won't even illuminate one bit of the area outside its direct reach.
In order to make the lights look better, I applied a filter to them, which pretty much limits the range to be illuminated, and also iluminate the area slightly within this limit:
In the big yellow circle you can see how the area is illuminated even if no direct light reaches it.
This solution however comes with some undesirable side effects. As you can see in the following screenshot, even if no light at all reaches an area, it will be illuminated if its too close to the light source:
I was wondering if there is any way to achieve what im trying to do by using shaders properly.
The main problem that I encounter comes from how I draw these shadows.
1) First I take the structures within the light's range.
At this point, I'm working with vertex, as they define the area of the shadow casting items:
2) Then, for each of these objects, I calculate the shadow they cast individually:
The shadow they cast is done by the CPU, by calpulating projections for each vertex of the body.
3) Then the GPU draws these shapes into a texture to compose the final shadow:
The problem I find is that making this difuse shadow effect, I need the final shadow. If I were to calculate the diffused shadows in step 2), a gap of light would appear between solids B and C.
But if I difuse the shadows in step 3, I no longer have the vertex information, as all the info I have are the 3 local textures added up together in one final texture.
So, is there anyway to achieve this? My first idea would be to pass a varying to the fragmentshader to calculate how much light comes into the dark area, but since I'll be processing this info on the final shadow, which has no vertex information, I'm completely lost about what approach I should use to make this.
I might be completely wrong on this approach, since I have very very limited experience with shaders.
Here is an example of what I have right now, and what I desire:
What I have: Plain illumination withing the light radius (which causes the light to clip through walls.
What I want: Shadow is more intense the farther away it is from where the light ends.
I think that no matter what you do, you're going to have to calculate the light falloff differently for the areas of the scene that don't have direct visibility to the original light source. I'm not sure how you're actually applying the shadow volume to the scene, but you might be able to do something like classify the edges as you're generating the shadow volume (i.e. did they come from a wall or from the line from the light to a corner?), and treat them accordingly.
In short I don't think there's any clever shader-specific trick you can pull off to fix this problem. It's a limitation of the algorithm you're using, so you need to improve the algorithm to take into account the different nature of the edges in your shadow volume.
--- Edit ---
Ok I think your best bet is going to be to create two shadow volumes (or rather shadow areas since we're working in 2D). One will be exactly as you have now, the other will be smaller -- you'll want to exclude the areas that are in "soft" shadow. Then you'll have three categories in your fragment shader: total shadow, soft shadow, and unshadowed.
To create the second shadow map, I think you'll want to do something like add a fixed angle to the edges that are created by the light shining around a corner.
--- Edit #2 ---
I think you can solve the problem of the gap between objects B and C by taking the silhouette of the shadow areas and the outer box of the scene. So for each of your shadow areas, you'd find the two outermost points that intersect the outer box, then throw out all the line segments in between and replace them with that portion of the box. I feel like I should be able to name that algorithm but it escapes me at the moment...
Original Scene:
Individual "hard" shadows:
Now union the shadows together:
Finally, trace around the shadows, keeping to the edges of the box. The corners you want to identify are the ones in yellow. As you are tracing around the perimeter of the shadows, you want to cut across the edge of the box until you reach the opposite corner. Not the easiest thing to code but if you've gotten this far I think you can figure it out :)
Alternatively, it might be easier to simply consider the lit areas. For example, you could do something like take the difference of the scene box and the unioned shadowed area This will usually give you multiple polygons; discard everything except the one that contains the light because the rest are the false gaps.
I need to make spherical billboards (i.e., setting depth), but taking into account perspective projection--ideally including off-center frusta.
I wasn't able to find any references to anyone succeeding at this--although there are plenty of explanations as to why standard billboards don't have perspective distortions. Unfortunately, for my application, the lack isn't a cosmetic defect; it's actually important to the algorithm.
I did a bit of investigation on my own:
The math gets pretty messy rather quickly. The obvious approaches don't work: for example, you can't orient the billboard perpendicular to a viewing ray because tangential rays wouldn't intersect the billboard at right angles.
Probably the most promising approach I found was to render the billboard parallel to the near clipping plane, stretching it with a vertex shader into an ellipse. This only handles perturbations along one axis (so e.g. it won't handle spheres rendered in a corner of the view), but the main obstacle is calculating depth correctly; you can't compute it as you would for an undistorted sphere because the "sphere" is occluding itself.
Point of fact, I didn't find a good solution, and I couldn't find anyone who has. Anyone have an idea?
While browsing around not even remotely working on this problem, I stumbled on http://iquilezles.org/www/articles/sphereproj/sphereproj.htm, which is pretty close. The linked tutorial shows how to compute a bounding ellipse for a rasterized sphere; getting the depth (at worst, using a raycast) should be fairly easy to derive.
I'm implementing a deferred lighting mechanism in my OpenGL graphics engine following this tutorial. It works fine, I don't get into trouble with that.
When it comes to the point lights, it says to render spheres around the lights to only pass those pixels throught the lighting shader, that might be affected by the light. There are some Issues with that method concerning cullface and camera position precisely explained here. To solve those, the tutorial uses the stencil-test.
I doubt the efficiency of that method which leads me to my first Question:
Wouldn't it be much better to draw a circle representing the light-sphere?
A sphere always looks like a circle on the screen, no matter from which perspective you're lokking at it. The task would be to determine the screenposition and -scaling of the circle. This method would have 3 advantages:
No cullface-issue
No camereposition-in-lightsphere-issue
Much more efficient (amount of vertices severely reduced + no stencil test)
Are there any disadvantages using this technique?
My second Question deals with implementing mentioned method. The circles' center position could be easily calculated as always:
vec4 screenpos = modelViewProjectionMatrix * vec4(pos, 1.0);
vec2 centerpoint = vec2(screenpos / screenpos.w);
But now how to calculate the scaling of the resulting circle?
It should be dependent on the distance (camera to light) and somehow the perspective view.
I don't think that would work. The point of using spheres is they are used as light volumes and not just circles. We want to apply lighting to those polygons in the scene that are inside the light volume. As the scene is rendered, the depth buffer is written to. This data is used by the light volume render step to apply lighting correctly. If it were just a circle, you would have no way of knowing whether A and C should be illuminated or not, even if the circle was projected to a correct depth.
I didn't read the whole thing, but i think i understand general idea of this method.
Won't help much. You will still have issues if you move the camera so that the circle will be behind the near plane - in this case none of the fragments will be generated, and the light will "disappear"
Lights described in the article will have a sharp falloff - understandably so, since sphere or circle will have sharp border. I wouldn-t call it point lightning...
For me this looks like premature optimization... I would certainly just be rendering whole screenquad and do the shading almost as usual, with no special cases to worry about. Don't forget that all the manipulations with opengl state and additional draw operations will also introduce overhead, and it is not clear which one will outscale the other here.
You forgot to do perspective division here
The simplest way to calculate scaling - transform a point on the surface of sphere to screen coords, and calculate vector length. It mst be a point on the border in screen space, obviously.
I have recently implemented soft shadows using voxel cone tracing in OpenGL 4.3 by tracing a cone in the direction of the light and accumulating opacity values.
The key thing that I am trying to resolve or hide is the very voxelized shadowing effect as the occluded surface gets closer to the occluder, as well as hide the clear spots in the shadow due to surface voxelization. I am using low resolution voxels 64x64x64; however, even if I use higher resolution voxels, some of the low-res voxels at a higher mip-map level are still captured in the trace.
So here's what my first idea is: I want to be able to keep the softest parts of the shadow that is furthest away and replace the parts of the shadow that is closer to the occluder with a shadow map. The shadow map will fade as it is further away from each occluder and I will somehow blend it into with the cone traced shadows.
Can anyone think of a way to fade a shadow away based on distance from each object for a shadow-map and then have it blend smoothly into the cone-traced shadow?
Another idea I have would be to somehow ray-trace shadows onto surfaces that are closer to an occluder, but this would probably be too expensive.
Alternatively, I would welcome any other ideas to help improve my soft shadow algorithm.
I've also put up a video to show it in motion:
https://www.youtube.com/watch?v=SUiUlRojBpM
Still haven't found a way to resolve the shadowing issue.
I'm guessing the "clear spot" artifacts are occurring due to large voxel sizes only being partially filled with geometry: "accumulating opacity values.". How many samples are you taking when converting from rasterized pixels to voxels? If the sample volume/voxel volume is small then there could be issues with correctly rendering transparency - there will be noise indicated by lighter areas.
Also, are your voxels' transparency direction dependent? Based on the author's original paper. Directional dependence is important to ensure semi-opaque voxels are rendered correctly.
A quick picture to explain
"for a shadow-map and then have it blend smoothly into the cone-traced shadow?"
This seems like you are kind of shooting yourself in the foot. You get an even larger performance hit and get the disadvantages of both shadow mapping and voxel cone tracing. Voxel cone tracing is expensive but can give nice soft shadows and do global illumination. Shadow mapping is better at doing hard shadows and is faster for smaller scenes, but as you add more geometry you end up redrawing the same stuff multiple times, at least once for each light.
Great work btw. I came across your problem while doing preliminary research for my own DirectX implementation of voxel cone tracing.
[Edit]
I realized that I made a typo in the picture. The picture on the right should be 4/64, rather than 4/16. I forgot about the Z dimension.
Well, in this case you can do it by adding more lights. You can add more lights closer to the original one and then compose the shadow of a light with the shadows of a bunch of closer lights. That is the 'area light' effect.
I have a Spotlight source in OpenGL, pointing towards a texture mapped sphere.
I rotate the lightsource with the sphere, such that if I rotate the sphere to the 'non-light' side, that side should be dark.
The odd part is, the spotlight seems to be shining through my sphere (it's a solid, no gaps between triangles. The light seems to be 'leaking' through to the other side.
Any thoughts on why this is happening?
Screenshots:
Front view, low light to emphasize the problem
Back view, notice the round area that is 'shining through'
Its really hard to tell from the images, but:
Check if GL_LIGHT_MODEL_TWO_SIDE is being set (two sided lighting), but more importantly have a look at the normals of the sphere you are rendering.
Edit: Also - change the background colour to something lighter. Oh and make sure you aren't rendering with alpha blending turned on (maybe its a polygon sorting issue).
OK, I'm a nob - I was specifying my normals, but not calling glEnableClientState(GL_NORMAL_ARRAY). Hence all normals were facing one direction (I think that's the default, no?)
Anyway - a lesson learned - always go back over the basics.