i have a 2k x 1k image with randomly placed "target" pixels. these pixels are pure red.
in a frag/pixel shader, for each pixel that is not red (target color), i need to find the distance to the closest red pixel. i'll use this distance value to create a gradient.
i found this answer, which seems the closest to my problem ---
Finding closest non-black pixel in an image fast ---
but it's not glsl specific.
i have the option to send my red target pixels into the frag shader as a texture buffer array. but i think it would be cleaner if i didn't need to.
A shader cannot read and write to the same texture because that would introduce too many constraints and complexities about the sequence of execution and would make caching much more difficult. So you're talking about sending some data about the red pixels in and getting the distance information out.
Fragment shaders run in parallel and it's much more expensive to perform random-access texture reads than to read from a location that is known outside of the shader, primarily due to pipelining considerations. The pre-programmable situation where sampling coordinates are known at vertices and then interpolated across the face of the geometry is still the most optimal way to access a texture.
So, writing a shader that, for each pixel, did a search outwards for a red pixel would be extremely inefficient. It's definitely possible, doing much the algorithm you link to, but probably not the smartest way around.
Ideally you'd phrase things the other way around and use some sort of accumulation. So:
clear your output buffer to its maximal values;
for each red location:
for every output fragment, work out the distance from the location;
check what distance is already stored for that fragment;
if the new distance is less than that stored, replace the stored version.
The easiest way to do that in OpenGL is likely going to be to use a depth buffer, because that has the per-fragment steps (2) and (3) implemented directly in hardware.
So for each each fragment you're going to calculate the distance from the current red fragment. You're going to output that as depth. When you're finished with all red dots you can use the depth buffer as input to a shader that outputs appropriate colours.
To avoid 2000 red spots turning into a 2000-pass drawing algorithm which would quickly run up against memory bandwidth, you'll probably want to write a single shader that does a large number of red dots at once and outputs a single depth value.
You should check GL_MAX_UNIFORM_LOCATIONS to find out how many uniforms you can push at once. It's guaranteed to be at least 1024 on recent versions of desktop OpenGL. You'll probably want to generate your shader dynamically.
Why don't you gaussian blur the whole image with a large radius, but at each iteration keep adding the red pixels back into the equation at full intensity so they bleed out. The red channel of the final blur would be your distance values - the higher values are closer to the red pixels. It's an approximation, but then you can make use of heavily optimised blur shaders.
Related
I have a application that will encode data for bearing and intensity using 32 bits. My fragment shader already decodes the values and then sets the color depending on bearing and intensity.
I'm wondering if it's also possible, via shader, to change the size (and possibly shape) of the drawn pixel.
As an example, let's say we have 4 possible values for intensity, then 0 would cause a single pixel to be drawn, 1 would draw a 2x2 square, 2 a 4x4 square and 3 a circle with a radius of 6 pixels.
In the past, we had to do all this on the CPU side and I was hoping to offload this job to the GPU.
No, fragment shaders cannot affect the "size" of the data they write. Once something has been rasterized into fragments, it doesn't really have a "size" anymore.
If you're rendering GL_POINTS primitives, you can change their size from the vertex shader. As for point sizes, it's rather difficult to ensure that a particular point covers an exact number of fragments.
The first thing that came into my mind is doing something similiar to blur technique, but instead of bluring the texture, we use it to look for neighbouring texels with a range to check if it has the intensity above 1.0f. if yes, then set the current texel color to for example red.
If you're using a fbo that is 1:1 in comparison to window size, use 1/width and 1/height in texture coordinates to get approximately 1 pixel (well not exactly because it is not a pixel but texel, just nearly)
Although this work just fine, the downside of this is it is very expensive as it will have n^2 complexity and probably some branching.
Edit: after thinking awhile this might not work for size with even number
TL;DR I'm computing a depth map in a fragment shader and then trying to use that map in a vertex shader to see if vertices are 'in view' or not and the vertices don't line up with the fragment texel coordinates. The imprecision causes rendering artifacts, and I'm seeking alternatives for filtering vertices based on depth.
Background. I am very loosely attempting to implement a scheme outlined in this paper (http://dash.harvard.edu/handle/1/4138746). The idea is to represent arbitrary virtual objects as lots of tangent discs. While they wanted to replace triangles in some graphics card of the future, I'm implementing this on conventional cards; my discs are just fans of triangles ("Discs") around center points ("Points").
This is targeting WebGL.
The strategy I intend to use, similar to what's done in the paper, is:
Render the Discs in a Depth-Only pass.
In a second (or more) pass, compute what's visible based solely on which Points are "visible" - ie their depth is <= the depth from the Depth-Only pass at that x and y.
I believe the authors of the paper used a gaussian blur on top of the equivalent of a GL_POINTS render applied to the Points (ie re-using the depth buffer from the DepthOnly pass, not clearing it) to actually render their object. It's hard to say: the process is unfortunately a one line comment, and I'm unsure of how to duplicate it in WebGL anyway (a naive gaussian blur will just blur in the background pixels that weren't touched by the GL_POINTS call).
Instead, I'm hoping to do something slightly different, by rerendering the discs in a second pass instead as cones (center of disc becomes apex of cone, think "close the umbrella") and effectively computing a voronoi diagram on the surface of the object (ala redbook http://www.glprogramming.com/red/chapter14.html#name19). The idea is that an output pixel is the color value of the first disc to reach it when growing radiuses from 0 -> their natural size.
The crux of the problem is that only discs whose centers pass the depth test in the first pass should be allowed to carry on (as cones) to the 2nd pass. Because what's true at the disc center applies to the whole disc/cone, I believe this requires evaluating a depth test at a vertex or object level, and not at a fragment level.
Since WebGL support for accessing depth buffers is still poor, in my first pass I am packing depth info into an RGBA Framebuffer in a fragment shader. I then intended to use this in the vertex shader of the second pass via a sampler2D; any disc center that was closer than the relative texture2D() lookup would be allowed on to the second pass; otherwise I would hack "discarding" the vertex (its alpha would be set to 0 or some flag set that would cause discard of fragments associated with the disc/cone or etc).
This actually kind of worked but it caused horrendous z-fighting between discs that were close together (very small perturbations wildly changed which discs were visible). I believe there is some floating point error between depth->rgba->depth. More importantly, though, the depth texture is being set by fragment texel coords, but I'm looking up vertices, which almost certainly don't line up exactly on top of relevant texel coordinates; so I get depth +/- noise, essentially, and the noise is the issue. Adding or subtracting .000001 or something isn't sufficient: you trade Type I errors for Type II. My render became more accurate when I switched from NEAREST to LINEAR for the depth texture interpolation, but it still wasn't good enough.
How else can I determine which disc's centers would be visible in a given render, so that I can do a second vertex/fragment (or more) pass focused on objects associated with those points? Or: is there a better way to go about this in general?
I've been trying to utilize the techniques in Eric Penner's "Shader Amortization using
Pixel Quad Message Passing" from GPU Pro 2, Chapter VI.2. The basic idea is that modern GPU's process fragment shaders in 2x2 fragment quads, and you can use ddx() and ddy() to get the value of some_var at all four fragments as long as the following hold:
Your GPU supports high-quality derivatives
You know which fragment you're processing (top-left, top-right, bottom-left, bottom-right)
This opens up a lot of opportunities for fragment shader optimization (like distributing texture fetches over a 2x2 pixel quad) that you'd need Compute Shaders to beat.
My problem is this:
I can't deterministically detect which fragment I'm processing. Ideally, each fragment block would start at even-numbered output pixel coords like (0, 0), (2, 0), ... (1024, 1024), ..., so you'd just need to check whether the output pixel x and y coords are even or odd to know which fragment you're currently processing. The method Penner uses in the book assumes this works...but it seems to be going wrong for me.
Unfortunately, my 2x2 fragment quads appear to be starting in nondeterministic places: I've seen them start at (even, even), (even, odd), and (odd, even). I can't remember if I've seen (odd, odd) or not, but anyway, the arrangement seems to depend on a myriad of factors I don't understand, including the output resolution and shader specifics. (I'm testing on an 8800 GTS, in case anyone's wondering.)
Does anyone know what might be causing this nondeterminism or have any documentation on it? I understand there's virtually no official standardization in this area, but I'm more interested in how things work in practice on modern desktop-level GPU's, and I'm hoping there's a way to get this technique to work. If no one knows how to reason about the even/odd start behavior, does anyone know any other way of determining the current fragment's relative location in its 2x2 quad?
Thanks :)
As it turns out, the premise of my question was mostly wrong:
The 2x2 fragment quads DO almost always start on even pixel numbers...as long as the output resolution is even-numbered.
If the output resolution is odd-numbered (a possibility with the underlying program I'm working with), things can get more complicated, for obvious reasons. I don't expect there's any uniformity here across drivers/GPU's/etc. either, but my current tests (which themselves may still be buggy) appear to demonstrate 2x2 pixel quads starting at an odd pixel along the dimension with odd resolution, at least when the odd dimension is horizontal.
All of this weirdness helped obscure my bigger issue: The code I used to detect the fragment's location in the pixel quad was buggy. I tested by setting the texture coordinates equal within a pixel quad (set to the pixel quad center)...or so I thought. However, I calculated the screen coordinates based on a full-screen quad where the uv mapping has the +v axis pointing downward. The screenspace origin starts at the bottom-left, because it's based on the top-right quadrant of Cartesian coordinates, and I accidentally forgot to invert the v-coordinate of the uv offset I used to find the pixel quad center. Many of my nondeterministic observations came from failing to check my assumptions while debugging and misinterpreting things as a result, particularly in combination with odd resolutions.
This was an embarrassing mistake I should have caught a lot sooner, but I figured I'd detail it as a warning to others to always double-check the direction of your vertical axis when you're dealing with opposite-facing coordinate frames. ;)
UPDATE:
I ran across a situation where 2x2 pixel quads started on even pixel numbers even when the resolution was odd. Thanks to the nondeterminism under odd resolutions, I had to work out another solution:
If you're deriving your screen pixel numbers from the uv coords of a fullscreen quad (for post-processing), the fragment location derived from this is only useful for arranging/placing shared samples between fragments, etc., not for the quad-pixel communication itself. You'll need to have screen pixel numbers with respect to the screenspace origin for that. You can derive these from vertex positions, or you can use ddx().x and ddy().y on the uv-based pixel numbers to find out their screen direction and mirror the fragment position in the appropriate direction from there.
Calculate the fragment location based on your screen pixel numbers (with respect to the true screenspace origin) and the assumption 2x2 pixel quads start on even pixels. (If you used uv-based pixel numbers, now is the time to mirror things.)
Do a ddx().x and ddy().y on the fragment location, and if they're negative in either direction, you know the pixel quad starts at an odd pixel number in that direction...so mirror in that direction.
If you calculate two fragment positions, one based on a uv origin and one based on a screen origin, use the uv-based one for reasoning about uv-based sample placement, and use the screen-based one for actually obtaining the values of a variable at neighboring fragments.
Profit.
I'll post a link to my working MIT-licensed code once I release it on Github, along with usage examples (the speedup is unfortunately not what I expected, but whatever ;)). I'm just waiting to get done with a larger shader I'll be uploading along with it.
I'm building a LIDAR simulator in OpenGL. This means that the fragment shader returns the length of the light vector (the distance) in place of one of the color channels, normalized by the distance to the far plane (so it'll be between 0 and 1). In other words, I use red to indicate light intensity and blue to indicate distance; and I set green to 0. Alpha is unused, but I keep it at 1.
Here's my test object, which happens to be a rock:
I then write the pixel data to a file and load it into a point cloud visualizer (one point per pixel) — basically the default. When I do that, it becomes clear that all of my points are in discrete planes each located at a different depth:
I tried plotting the same data in R. It doesn't show up initially with the default histogram because the density of the planes is pretty high. But when I set the breaks to about 60, I get this:
.
I've tried shrinking the distance between the near and far planes, in case it was a precision issue. First I was doing 1–1000, and now I'm at 1–500. It may have decreased the distance between planes, but I can't tell, because it means the camera has to be closer to the object.
Is there something I'm missing? Does this have to do with the fact that I disabled anti-aliasing? (Anti-aliasing was causing even worse periodic artifacts, but between the camera and the object instead. I disabled line smoothing, polygon smoothing, and multisampling, and that took care of that particular problem.)
Edit
These are the two places the distance calculation is performed:
The vertex shader calculates ec_pos, the position of the vertex relative to the camera.
The fragment shader calculates light_dir0 from ec_pos and the camera position and uses this to compute a distance.
Is it because I'm calculating ec_pos in the vertex shader? How can I calculate ec_pos in the fragment shader instead?
There are several possible issues I can think of.
(1) Your depth precision. The far plane has very little effect on resolution; the near plane is what's important. See Learning to Love your Z-Buffer.
(2) The more probable explanation, based on what you've provided, is the conversion/saving of the pixel data. The shader outputs floating point values, but these are stored in the framebuffer, which will typically have only 8bits per channel. For color, what that means is that your floats will be mapped to the underlying 8-bit (fixed width, integer) representation, therefore only possessing 256 values.
If you want to output pixel data as the true floats they are, you should make a 32-bit floating point RGBA FBO (with e.g. GL_RGBA32F or something similar). This will store actual floats. Then, when your data from the GPU, it will return the original shader values.
I suppose you could alternately encode a single float in a vec4 with some multiplication, if you don't have a FBO implementation handy.
OpenGL. Let's say I've drawn one image and then the second one using XOR. Now I've got black buffer with non-black pixels somewhere, I've read that I can use shaders to count black [ rgb(0,0,0) ] pixels ON GPU?
I've also read that it has to do something with OcclusionQuery.
http://oss.sgi.com/projects/ogl-sample/registry/ARB/occlusion_query.txt
Is it possible and how? [any programming language]
If you've got other idea on how to find similarity via OpenGL/GPU - that would be great too.
I'm not sure how you do the XOR bit (at least it should be slow; I don't think any of current GPUs accelerate that), but here's my idea:
have two input images
turn on occlusion query.
draw the two images to the screen (i.e. full screen quad with two textures set up), with a fragment shader that computes abs(texel1-texel2), and kills the pixel (discard in GLSL) if the pixels are the same (difference is zero or below some threshold). Easiest is probably just using a GLSL fragment shader, and there you just read two textures, compute abs() of the difference and discard the pixel. Very basic GLSL knowledge is enough here.
get number of pixels that passed the query. For pixels that are the same, the query won't pass (pixels will be discarded by the shader), and for pixels that are different, the query will pass.
At first I though of a more complex approach that involves depth buffer, but then realized that just killing pixels should be enough. Here's my original though (but the above one is simpler and more efficient):
have two input images
clear screen and depth buffer
draw the two images to the screen (i.e. full screen quad with two textures set up), with a fragment shader that computes abs(texel1-texel2), and kills the pixel (discard in GLSL) if the pixels are different. Draw the quad so that it's depth buffer value is something close to near plane.
after this step, depth buffer will contain small depth values for pixels that are the same, and large (far plane) depth values for pixels that are different.
turn on occlusion query, and draw another full screen quad with depth closer than far plane, but larger than the previous quad.
get number of pixels that passed the query. For pixels that are the same, the query won't pass (depth buffer is already closer), and for pixels that are different, the query will pass. You'd use SAMPLES_PASSED_ARB to get this. There's an occlusion query example at CodeSampler.com to get your started.
Of course all this requires GPU with occlusion query support. Most GPUs since 2002 or so do support that, with exception of some low-end ones (in particular, Intel 915 (aka GMA 900) and Intel 945 (aka GMA 950)).