I have a simple shader code.
I pass in two image texture (NOTE: One is samplerExternalOES and other one is sampler2D).
The first texture 'sTexture' is the original image i get from a camera frame.
The second texture 'sTexture2' is a mask i get from the cpu.
The shader is as follows:
uniform samplerExternalOES sTexture;
uniform sampler2D sTexture2;
varying vec2 v_TexCoord;
void main(void)
{
vec4 originalrgb = vec4((texture2D(sTexture, v_TexCoord).rgb), 1.0);
vec4 floodfillimage = vec4((texture2D(sTexture2, v_TexCoord).rgb), 1.0);
/*Code To Colour Input Image with Blue Tint Color*/
vec4 c = vec4(0.0, 1.0, 1.0, 1);
vec4 inputColor = vec4((texture2D(sTexture, v_TexCoord).rgb), 1.0);
float average = (inputColor.r + inputColor.g + inputColor.b) / 3.0;
vec4 grayscale = vec4(average, average, average, 1.0);
vec4 colorizedOutput = grayscale * c ;
/*Code To mix original image with blue coloured based on another floodfilledimage passed in */
gl_FragColor = mix(originalrgb.rgba, colorizedOutput.rgba, floodfillimage.r);
}
The error i recieve is glerror 1282, which means GL_INVALID_OPERATION. I've debugged and found out this happens on the mix function line.
NOTE:
If I change the last line to gl_FragColor = mix(originalrgb.rgba, colorizedOutput.rgba, 0.5);, it works.
So, why is it that the code panics when i do floodfillimage.r?
Thank You.
EDIT
I've tested both textures passed in (ie, i just rendered them to gl_FragColor) and they both are the image they're suppose to be
The shader looks legal to me and compiles successfully with various offline compilers, so I suspect you have hit a driver bug on the device you are using.
Related
I am currently programming an application for image processing. To achieve the needed performance, I have to use the GPU to compute the camera input, more specifically use OpenGL ES 2.0.
With the help of this project (https://github.com/yulu/ShaderCam) I achieved to pass the image to the pipeline and do simple operations with the fragment shader (like inverting colors etc).
My knowledge of GLSL, fragment shaders and vertex shaders is fairly limited but I am aware of pipeline constraints and what the two shaders do in the pipeline.
So - formulating the problem - I would like to calculate the average color of a row in my received image and return it (per row) to my application.
I read here https://stackoverflow.com/a/13866636/8038866 that this is generally possible, however I can't seem to find out the following things:
1 (edit: SOLVED by simply passing the w and h of my texture to the vertex and fragment shader):
Knowing where the row ends (and having that information in the fragment shader). For this I assume that I would have to pass the width of the picture to the vertex shader and from there to the fragment shader, right?
2.: How to calculate the average the color values of each row in the fragment shader and then pass them to the application. If I understand it correctly - the fragment shader only excutes the code per pixel, so I am not sure how to achieve this.
Here are the two very basic shaders
vertex shader:
uniform mat4 uTransformM;
uniform mat4 uOrientationM;
uniform vec2 ratios;
attribute vec2 aPosition;
varying vec2 vTextureCoord;
void main(){
gl_Position = vec4(aPosition, 0.0, 1.0);
vTextureCoord = (uTransformM * ((uOrientationM * gl_Position + 1.0)*0.5)).xy;
gl_Position.xy *= ratios;
}
fragment shader:
#extension GL_OES_EGL_image_external : require
precision mediump float;
uniform samplerExternalOES sTexture;
varying vec2 vTextureCoord;
void main(){
gl_FragColor = texture2D(sTexture, vTextureCoord);
//calc mean per row and pass it back
}
I am very thankful for every advice or help you can provide.
I found a way that does the trick for me. The idea is to calculate the mean for only one row of pixels and then later in the application to get this line with
glReadPixels( GLint x, GLint y, GLsizei width, GLsizei height, GLenum format, GLenum type, GLvoid * data);
Here is my fragment shader (notice that the width of the surface is required as well):
#extension GL_OES_EGL_image_external : require
precision mediump float;
uniform samplerExternalOES sTexture;
varying float width;
varying vec2 vTextureCoord;
void main(){
vec4 accumulatedRGB = texture2D(sTexture, vec2(0,vTextureCoord.y));
if(vTextureCoord.x < 0.50 && vTextureCoord.x > 0.499){ //small enough to only cover one line
for(float i=1.0;i<=width;++i)
{
float xPosOnTexture = i/width;
vec4 current = texture2D(sTexture, vec2(xPosOnTexture,vTextureCoord.y));
accumulatedRGB += current;
}
vec4 mean = accumulatedRGB/width;
gl_FragColor = vec4(mean.rgb , mean.a);//avg color for one line
}
else{
gl_FragColor = vec4(0.0,0.0,0.0,0.0);//rest of the screen
}
}
I have written this simple shader to overlay texture over another (base) texture -
varying highp vec2 textureCoordinate;
varying highp vec2 textureCoordinate2;
uniform sampler2D inputImageTexture;
uniform sampler2D inputImageTexture2;
void main()
{
mediump vec4 base = texture2D(inputImageTexture, textureCoordinate);
mediump vec4 overlay = texture2D(inputImageTexture2, textureCoordinate2);
mediump float ra = (overlay.a) * overlay.r + (1.0 - overlay.a) * base.r;
mediump float ga = (overlay.a) * overlay.g + (1.0 - overlay.a) * base.g;
mediump float ba = (overlay.a) * overlay.b + (1.0 - overlay.a) * base.b;
gl_FragColor = vec4(ra, ga, ba, 1.0);
}
Issue - This works except for one issue. If the overlay image is smaller than the base image, the outside region of overlay image gives alpha value of 1.0, i.e overlay.a == 1.0. Due to this the base image is clipped by overlay image. The region outside overlay appears as black.
I am new to opengl, and was expecting that outside its bounds, the texture's alpha should appear as 0.0? How can I fix my shader code to achieve desired behaviour? Or do I need to modify my graphics pipeline?
EDIT Vertex shader below-
attribute vec4 inputTextureCoordinate2;
varying vec2 textureCoordinate;
varying vec2 textureCoordinate2;
void main()
{
gl_Position = pos;
textureCoordinate = uv;
textureCoordinate2 = inputTextureCoordinate2.xy;
}
I was expecting that outside its bounds, the texture's alpha should appear as 0.0
How are you sampling the texture outside of its bounds? When sampling a texture, the uv coordinates should range from 0 to 1. If the coordinates are outside of this range, then one of two things will happen:
If GL_CLAMP_TO_EDGE is set, then the cooridnate will be clamped to the (0, 1) range, and you'll sample an edge pixel
If GL_REPEAT is set, then the fractional part of the coordinate will be taken, and you'll sample somewhere in the middle of the texture
See the docs on glTexParameter for more details.
If your use case is simply overlaying images, perhaps you should try writing a pixel shader.
Set the viewport to the base image dimensions and draw a quad from (-1, 1).
Your fragment shader will now operate on each pixel, known as a texel. Get the texel with gl_FragCoord
Sample the base and overlay by texel e.g. using texelFetch
If the texel is outside of the overlay, set the overlay's rgba values to 0
For example
//fragment shader
uniform ivec2 overlayDim;
uniform sampler2D baseTexture;
uniform sampler2D overlayTexture;
void main() {
vec2 texelf = floor(gl_FragCoord.xy);
ivec2 texel = (int(texelf.x), int(texelf.y));
vec4 base = texelFetch(baseTexture, texel, 0);
vec4 overlay = texelFetch(overlayTexture, texel, 0);
float overlayIsValid = float(texel.x < overlayDim.x && texel.y < overlayDim.y);
overlay *= overlayIsValid;
//rest of code
}
What happens if you sample outside the range of the texture is controlled by the value you set for GL_TEXTURE_WRAP_S and GL_TEXTURE_WRAP_T using glTexParameteri().
In full OpenGL, you could set the value to GL_CLAMP_TO_BORDER, set the border color to a value with alpha 0.0, and be done with it. But texture borders are not available in OpenGL ES 2.0 (the option is introduced in ES 3.2, but not in earlier versions).
Without this, I can think of two options:
If you have control over the texture data, you could set a one pixel border to transparent values. The GL_CLAMP_TO_EDGE then gives you a transparent value when sampling outside the range.
Check the range in the fragment shader.
The fragment shader code for the second option could look something like this (untested):
mediump vec3 col = texture2D(inputImageTexture, textureCoordinate).xyz;
if (all(greaterThan(textureCoordinate2, vec2(0.0))) &&
all(lessThan(textureCoordinate2, vec2(1.0))))
{
mediump vec3 overlay = texture2D(inputImageTexture2, textureCoordinate2).xyz;
col = mix(col, overlay, overlay.a);
}
gl_FragColor = vec4(col, 1.0);
Compared to your original code, also note the use of vector operations. Whenever there is a good way of operating on vectors, it will make the code simpler. It will also make the job of the optimizer easier for GPUs with vector operations.
I found the issue in my code. I had
GLES20.glClearColor(0, 0, 0, 1);
in my code. Changing it to -
GLES20.glClearColor(0, 0, 0, 0);
fixed the issue.
Also as mentioned by #Reto, I have changed my fragment shader to use vector operations for optimisation.
void main()
{
mediump vec4 overlay = texture2D(inputImageTexture2, textureCoordinate2);
mediump vec3 col = texture2D(inputImageTexture, textureCoordinate).xyz;
col = mix(col, overlay.xyz, overlay.a);
gl_FragColor = vec4(col, 1.0);
}
Here is my code.It runs well on PC/Windows,but jagged on Android 4.42 when I magnify the image.
#ifdef GL_ES
precision highp float;
#endif
varying vec4 v_fragmentColor;
varying vec2 v_texCoord;
uniform float u_width; //width of image
uniform float u_height; //height of image
void main()
{
float texelSizeX = 1.0/u_width;
float texelSizeY = 1.0/u_height;
//four pixels' color
vec4 p0q0 = texture2D(CC_Texture0, v_texCoord);
vec4 p1q0 = texture2D(CC_Texture0, v_texCoord + vec2(texelSizeX, 0));
vec4 p0q1 = texture2D(CC_Texture0, v_texCoord + vec2(0, texelSizeY));
vec4 p1q1 = texture2D(CC_Texture0, v_texCoord + vec2(texelSizeX , texelSizeY));
//bilinear interpolation
float a = fract(v_texCoord.s * u_width);
float b = fract(v_texCoord.t * u_height);
vec4 color_q0 = mix( p0q0, p1q0, a );
vec4 color_q1 = mix( p0q1, p1q1, a );
vec4 color = mix( color_q0, color_q1, b);
gl_FragColor = v_fragmentColor * color;
}
I'm sorry that I cannot post pictures. I debug the code well with VS2012, and the image seems smooth.
But when I run the program on Android, the image is full of jag. I don't know why.
Obvious question: why are you doing bilinear filtering in your shader and not just using the built-in hardware bilinear filtering? I'm sure there's a good reason, but telling us that might help you avoid a lot of questions along the lines of "have you set your filtering mode appropriately?"
That being said, it's likely to be a precision problem. You probably want to round v_texCoord to the exact sampling site as I'd guess that you have GL_NEAREST filtering set, to disable the hardware bilinear filtering, but due to precision problems e.g. v_texCoord + vec2(texelSizeX, 0) is then sampling the same texel rather than the next one along when v_texCoord is close to 0, or possibly the sample taken at v_texCoord is the next texel along when it's close to 1, or something along those lines.
OpenGL considers the centre of a texel to be its location. So if you were in 1d you could do something like:
r_texCoord.x = v_texCoord.x - mod(v_texCoord.x, 1.0/u_width) + 0.5/u_width;
Or if you were happy to use integral texture coordinates rather than the normal OpenGL [0.0, 1.0) range then you could simplify slightly because floor (and indeed ceil) already knows how to move you to an integral boundary:
(floor(v_texCoord.x) + 0.5) / u_width
... both of which are dependent reads so performance will suffer quite a bit.
In OpenGL ES 2.0 for Android, I am drawing a sphere. The sphere appears on the screen as a circle, so I need to add lighting. When I added lighting, instead of it being smooth, like it would be in real life, it is a fine line between light and dark, as shown here:
However, I want it to look like this, where the shading is much smoother and blended:
Here is my vertex shader code:
uniform mat4 u_Matrix;
uniform vec3 u_VectorToLight;
attribute vec4 a_Position;
attribute vec3 a_Color;
attribute vec3 a_Normal;
varying vec3 v_Color;
void main() {
v_Color = a_Color;
vec3 scaledNormal = a_Normal;
scaledNormal = normalize(scaledNormal);
float diffuse = max(dot(scaledNormal, u_VectorToLight), 0.0);
v_Color *= diffuse;
float ambient = 0.1;
v_Color += ambient;
gl_Position = u_Matrix * a_Position;
}
And my fragment shader:
precision mediump float;
varying vec3 v_Color;
void main() {
gl_FragColor = vec4(v_Color, 1.0);
}
The normal is calculated by getting the vector from the center of the sphere to the point on the sphere, then normalizing it (giving it a length of 1)
Here is how I set the colors:
vertices[offset++] = Color.red(color);
vertices[offset++] = Color.green(color);
vertices[offset++] = Color.blue(color);
Where color is 0xffea00.
The problem is with the range of color values you use. OpenGL operates with color component values in the range [0.0, 1.0]. But you are specifying colors in the range [0, 255] in your Java code.
You have two options to fix this:
Divide the color values you get from the Color class by 255.0f.
Specify the colors with type GL_UNSIGNED_BYTE. To do this, store the values in an array/buffer with element type byte, store those values in a VBO, and then set up the vertex attribute with:
glVertexAttribPointer(attrLoc, 3, GL_UNSIGNED_BYTE, GL_TRUE, stride, 0);
Note the value for the 4th argument. While it does not matter for GL_FLOAT attributes, it is critical that you use GL_TRUE in this case, because the byte values need to be normalized.
My terrain uses shader which itself uses four different textures. It runs fine on windows and linux machines, but on android it gets only ~25FPS on both galaxies. I thought, that textures are the problem, but no, as it appears the problem is with the part where I divide texture coordinates and use frac to get tiled coordinates. Without it, I get 60FPS.
// Material data.
//uniform vec3 uAmbient;
//uniform vec3 uDiffuse;
//uniform vec3 uLightPos[8];
//uniform vec3 uEyePos;
//uniform vec3 uFogColor;
uniform sampler2D terrain_blend;
uniform sampler2D grass;
uniform sampler2D rock;
uniform sampler2D dirt;
varying vec2 varTexCoords;
//varying vec3 varEyeNormal;
//varying float varFogWeight;
//------------------------------------------------------------------
// Name: fog
// Desc: applies calculated fog weight to fog color and mixes with
// specified color.
//------------------------------------------------------------------
//vec4 fog(vec4 color) {
// return mix(color, vec4(uFogColor, 1.0), varFogWeight);
//}
void main(void)
{
/*vec3 N = normalize(varEyeNormal);
vec3 L = normalize(uLightPos[0]);
vec3 H = normalize(L + normalize(uEyePos));
float df = max(0.0, dot(N, L));
vec3 col = uAmbient + uDiffuse * df;*/
// Take color information from textures and tile them.
vec2 tiledCoords = varTexCoords;
//vec2 tiledCoords = fract(varTexCoords / 0.05); // <========= HERE!!!!!!!!!
//vec4 colGrass = texture2D(grass, tiledCoords);
vec4 colGrass = texture2D(grass, tiledCoords);
//vec4 colDirt = texture2D(dirt, tiledCoords);
vec4 colDirt = texture2D(dirt, tiledCoords);
//vec4 colRock = texture2D(rock, tiledCoords);
vec4 colRock = texture2D(rock, tiledCoords);
// Take color information from not tiled blend map.
vec4 colBlend = texture2D(terrain_blend, varTexCoords);
// Find the inverse of all the blend weights.
float inverse = 1.0 / (colBlend.r + colBlend.g + colBlend.b);
// Scale colors by its corresponding weight.
colGrass *= colBlend.r * inverse;
colDirt *= colBlend.g * inverse;
colRock *= colBlend.b * inverse;
vec4 final = colGrass + colDirt + colRock;
//final = fog(final);
gl_FragColor = final;
}
Note: there's some more code for light calculation and fog, but it isn't used. I indicated the line that, when uncommented, causes massive lag. I tried using floor and calculating fractional part manually, but lag is the same. What might be wrong?
EDIT: Now here's what I don't understand.
This:
vec2 tiledCoords = fract(varTexCoords * 2.0);
Runs great.
This:
vec2 tiledCoords = fract(varTexCoords * 10.0);
Runs average on SIII.
This:
vec2 tiledCoords = fract(varTexCoords * 20.0);
Lags...
This:
vec2 tiledCoords = fract(varTexCoords * 100.0);
Well 5FPS is still better than I expected...
So what gives? Why is this happening? To my understanding this shouldn't make any difference. But it does. And a huge one.
I would run your code on a profiler (check Mali-400), but by the looks of it, you are killing the texture cache. For the first pixel computed, all those 4 texture look-ups are fetched but also the contiguous data is also fetched into the texture cache. For the next pixel, you are not accessing data in the cache but looking quite far (10, 20..etc) which completely defies the purpose of such a cache.
This of course a guess, without proper profiling is hard to tell.
EDIT: #harism also pointed you to that direction.