Converting shader to avoid needing using ofDisableArbTex()

[Eviral]

Hello,

I’m working on a OpenGL2 (version 120) app.
I’m already using some shaders in my app that are using :
#extension GL_ARB_texture_rectangle : enable
and Texture2DRect instead of Texture2D

It means i do not need to use ofDisableArbTex() and i don’t want to use it.
If i use it nothing displays anymore (black texture)

Now i would like to add a shader that doesn’t have this #extension and that is using Texture2D and not Texture2DRect.
This shader is called FrostedGlass.
I have tried to tweak it to make it supports #extension GL_ARB_texture_rectangle : enable

But i can’t make it works.
I have no shader errors but it displays a single color texture instead of my texture image with the frosted glass effect

Original version of FrostedGlass.frag :

#version 120

uniform sampler2D sceneTex; // 0
uniform sampler2D NoiseTex; // 1
uniform float vx_offset = 100.0;
uniform float PixelX = 2.0;
uniform float PixelY = 2.0;
uniform float Freq = 0.115;
uniform float rt_w; // GeeXLab built-in
uniform float rt_h; // GeeXLab built-in

vec4 spline(float x, vec4 c1, vec4 c2, vec4 c3, vec4 c4, \
                vec4 c5, vec4 c6, vec4 c7, vec4 c8, vec4 c9)
{
  float w1, w2, w3, w4, w5, w6, w7, w8, w9;
  w1 = 0.0;
  w2 = 0.0;
  w3 = 0.0;
  w4 = 0.0;
  w5 = 0.0;
  w6 = 0.0;
  w7 = 0.0;
  w8 = 0.0;
  w9 = 0.0;
  float tmp = x * 8.0;
  if (tmp<=1.0) {
    w1 = 1.0 - tmp;
    w2 = tmp;
  }
  else if (tmp<=2.0) {
    tmp = tmp - 1.0;
    w2 = 1.0 - tmp;
    w3 = tmp;
  }
  else if (tmp<=3.0) {
    tmp = tmp - 2.0;
    w3 = 1.0-tmp;
    w4 = tmp;
  }
  else if (tmp<=4.0) {
    tmp = tmp - 3.0;
    w4 = 1.0-tmp;
    w5 = tmp;
  }
  else if (tmp<=5.0) {
    tmp = tmp - 4.0;
    w5 = 1.0-tmp;
    w6 = tmp;
  }
  else if (tmp<=6.0) {
    tmp = tmp - 5.0;
    w6 = 1.0-tmp;
    w7 = tmp;
  }
  else if (tmp<=7.0) {
    tmp = tmp - 6.0;
    w7 = 1.0 - tmp;
    w8 = tmp;
  }
  else 
  {
    //tmp = saturate(tmp - 7.0);
    // http://www.ozone3d.net/blogs/lab/20080709/saturate-function-in-glsl/
    tmp = clamp(tmp - 7.0, 0.0, 1.0);
    w8 = 1.0-tmp;
    w9 = tmp;
  }
  return w1*c1 + w2*c2 + w3*c3 + w4*c4 + w5*c5 + w6*c6 + w7*c7 + w8*c8 + w9*c9;
}

vec3 NOISE2D(vec2 p)
  { return texture2D(NoiseTex,p).xyz; }

void main() 
{ 
  vec2 uv = gl_TexCoord[0].xy;
  vec3 tc = vec3(1.0, 0.0, 0.0);
    
  if (uv.x < (vx_offset-0.005))
  {
    float DeltaX = PixelX / rt_w;
    float DeltaY = PixelY / rt_h;
    vec2 ox = vec2(DeltaX,0.0);
    vec2 oy = vec2(0.0,DeltaY);
    vec2 PP = uv - oy;
    vec4 C00 = texture2D(sceneTex,PP - ox);
    vec4 C01 = texture2D(sceneTex,PP);
    vec4 C02 = texture2D(sceneTex,PP + ox);
    PP = uv;
    vec4 C10 = texture2D(sceneTex,PP - ox);
    vec4 C11 = texture2D(sceneTex,PP);
    vec4 C12 = texture2D(sceneTex,PP + ox);
    PP = uv + oy;
    vec4 C20 = texture2D(sceneTex,PP - ox);
    vec4 C21 = texture2D(sceneTex,PP);
    vec4 C22 = texture2D(sceneTex,PP + ox);
    
    float n = NOISE2D(Freq*uv).x;
    n = mod(n, 0.111111)/0.111111;
    vec4 result = spline(n,C00,C01,C02,C10,C11,C12,C20,C21,C22);
    tc = result.rgb;  
  }
  else if (uv.x>=(vx_offset+0.005))
  {
    tc = texture2D(sceneTex, uv).rgb;
  }
  
  gl_FragColor = vec4(tc, 1.0);
}

Tweaked version of FrostedGlass.frag (not working) :

#version 120

#extension GL_ARB_texture_rectangle : enable


uniform vec2                u_resImg;

uniform sampler2DRect sceneTex; // 0
uniform sampler2DRect NoiseTex; // 1
uniform float vx_offset = 100.0;
uniform float PixelX = 2.0;
uniform float PixelY = 2.0;
uniform float Freq = 0.115;
uniform float rt_w; // GeeXLab built-in
uniform float rt_h; // GeeXLab built-in

vec4 spline(float x, vec4 c1, vec4 c2, vec4 c3, vec4 c4, \
                vec4 c5, vec4 c6, vec4 c7, vec4 c8, vec4 c9)
{
  float w1, w2, w3, w4, w5, w6, w7, w8, w9;
  w1 = 0.0;
  w2 = 0.0;
  w3 = 0.0;
  w4 = 0.0;
  w5 = 0.0;
  w6 = 0.0;
  w7 = 0.0;
  w8 = 0.0;
  w9 = 0.0;
  float tmp = x * 8.0;
  if (tmp<=1.0) {
    w1 = 1.0 - tmp;
    w2 = tmp;
  }
  else if (tmp<=2.0) {
    tmp = tmp - 1.0;
    w2 = 1.0 - tmp;
    w3 = tmp;
  }
  else if (tmp<=3.0) {
    tmp = tmp - 2.0;
    w3 = 1.0-tmp;
    w4 = tmp;
  }
  else if (tmp<=4.0) {
    tmp = tmp - 3.0;
    w4 = 1.0-tmp;
    w5 = tmp;
  }
  else if (tmp<=5.0) {
    tmp = tmp - 4.0;
    w5 = 1.0-tmp;
    w6 = tmp;
  }
  else if (tmp<=6.0) {
    tmp = tmp - 5.0;
    w6 = 1.0-tmp;
    w7 = tmp;
  }
  else if (tmp<=7.0) {
    tmp = tmp - 6.0;
    w7 = 1.0 - tmp;
    w8 = tmp;
  }
  else 
  {
    //tmp = saturate(tmp - 7.0);
    // http://www.ozone3d.net/blogs/lab/20080709/saturate-function-in-glsl/
    tmp = clamp(tmp - 7.0, 0.0, 1.0);
    w8 = 1.0-tmp;
    w9 = tmp;
  }
  return w1*c1 + w2*c2 + w3*c3 + w4*c4 + w5*c5 + w6*c6 + w7*c7 + w8*c8 + w9*c9;
}

vec3 NOISE2D(vec2 p)
  { return texture2DRect(NoiseTex,p).xyz; }

void main() 
{ 
  //vec2 uv = gl_TexCoord[0].xy;
  vec2 uv = gl_TexCoord[0].st / u_resImg;
  
  vec3 tc = vec3(1.0, 0.0, 0.0);
  
  if (uv.x < (vx_offset-0.005))
  {
    float DeltaX = PixelX / rt_w;
    float DeltaY = PixelY / rt_h;
    vec2 ox = vec2(DeltaX,0.0);
    vec2 oy = vec2(0.0,DeltaY);
    vec2 PP = uv - oy;
    vec4 C00 = texture2DRect(sceneTex,PP - ox);
    vec4 C01 = texture2DRect(sceneTex,PP);
    vec4 C02 = texture2DRect(sceneTex,PP + ox);
    PP = uv;
    vec4 C10 = texture2DRect(sceneTex,PP - ox);
    vec4 C11 = texture2DRect(sceneTex,PP);
    vec4 C12 = texture2DRect(sceneTex,PP + ox);
    PP = uv + oy;
    vec4 C20 = texture2DRect(sceneTex,PP - ox);
    vec4 C21 = texture2DRect(sceneTex,PP);
    vec4 C22 = texture2DRect(sceneTex,PP + ox);
    
    float n = NOISE2D(Freq*uv).x;
    n = mod(n, 0.111111)/0.111111;
    vec4 result = spline(n,C00,C01,C02,C10,C11,C12,C20,C21,C22);
    tc = result.rgb; 
  }
  else if (uv.x>=(vx_offset+0.005))
  {
    tc = texture2DRect(sceneTex, uv).rgb;
  }
  
  gl_FragColor = vec4(tc, 1.0);
 
}

I have tried to follow the advices of @pierre_tardif00 but no way

sampler2D / sampler2Drect normalised texcoords

Probably a rather simple question but couldn’t find a confirming answer; I’ve been playing around with textures in shaders and notices that the coordinates coming in through a sampler2D can be sampled with normalised coordinates (0 - 1 // gl_FragCoord.xy / u_resolution.xy). However textures coming in with a sampler2DRect seem to need the actual pixel coordinates of that picture; x;512 y;600 for example. Is this finding correct, and if so, is it custom to still normalise those coordinates in som…

What did i missed ???

Thanks a lot

[Jona]

@Eviral I guess you just need to divide the texcoords through the tex width / height (gl_FragCoord.xy / resolution), so that the range is from 0.0 to 1.0. Thats actually what is written in the example that you posted.

[TimChi]

If coordinates like uv have been normalized earlier in the shader,

  vec2 uv = gl_TexCoord[0].st / u_resImg;

then try “un-normalizing” them whenever sampling a texture:

    tc = texture2DRect(sceneTex, uv * u_resImg).rgb;

@Eviral

I have no shader errors but it displays a single color texture instead of my texture image

The above can happen when sampling only the first pixel of the texture (0.0 - 1.0) using normalized coordinates.

[Eviral]

Thanks a lot !
It works