PDA

View Full Version : Another "reconstructing position from depth" thread (can't get an example to work)



aejtgl
11-08-2017, 03:02 PM
First of all, sorry for making another one of these threads (I know there are tons of them already), but I haven't been able to find any which could help me figure out what's wrong with my code. I've been trying to get the method shown here (https://www.khronos.org/opengl/wiki/Compute_eye_space_from_window_space#Optimized_meth od_from_XYZ_of_gl_FragCoord) to work for so long that I've run out of things to try.

First I tried to just implement that into my existing shader code, but as I couldn't get that to work after many hours, I tried using the barebones example (which had a few minor details I had to change to make it compile) and I'm still getting the exact same result.

Vertex shader:


#version 330 core
//Vertex shader
//Half the size of the near plane {tan(fovy/2.0) * aspect, tan(fovy/2.0) }
uniform vec2 halfSizeNearPlane;

layout (location = 0) in vec3 clipPos;
//UV for the depth buffer/screen access.
//(0,0) in bottom left corner (1, 1) in top right corner
layout (location = 1) in vec2 texCoord;

out vec3 eyeDirection;
out vec2 uv;

void main()
{
uv = texCoord;

eyeDirection = vec3((2.0 * halfSizeNearPlane * texCoord) - halfSizeNearPlane , -1.0);
gl_Position = vec4(clipPos.xy, 0, 1);
}


Fragment shader:


#version 330 core
//Fragment shader
layout (location = 0) out vec3 fragColor;
in vec3 eyeDirection;
in vec2 uv;

uniform mat4 persMatrix;
uniform vec2 depthrange;

uniform sampler2D depth;

vec4 CalcEyeFromWindow(in float windowZ, in vec3 eyeDirection, in vec2 depthrange)
{
float ndcZ = (2.0 * windowZ - depthrange.x - depthrange.y) / (depthrange.y - depthrange.x);
float eyeZ = persMatrix[3][2] / ((persMatrix[2][3] * ndcZ) - persMatrix[2][2]);
return vec4(eyeDirection * eyeZ, 1);
}

void main()
{
vec4 eyeSpace = CalcEyeFromWindow(texture(depth, uv).x, eyeDirection, depthrange);
fragColor = eyeSpace.rbg;
}



Where my camera settings are:
float fov = glm::radians(60.0f);
float aspect = 800.0f / 600.0f;

And my uniforms equal:
uniform mat4 persMatrix = glm::perspective(fov, aspect, 0.1f, 100.0f)
uniform vec2 halfSizeNearPlane = glm::vec2(glm::tan(fov/2.0) * aspect, glm::tan(fov/2.0))
uniform vec2 depthrange = glm::vec2(0.0f, 1.0f)

uniform sampler2D depth is a GL_DEPTH24_STENCIL8 texture which has depth values from an earlier pass (if I linearize it and set fragColor = vec3(linearizedZ), it shows up like it should, so nothing seems wrong there).


The resulting image I'm getting doesn't at all look like it does when I save position in the G-buffer or when I use the inverse matrix transforms to reconstruct position:
https://i.imgur.com/s3dzzOU.png
(rendering a single 15x15x15 cube)

Is there any obvious mistake in my code? To me everything looks sound, but it feels like there has to be some small obvious mistake somewhere.

Thanks!