Draw YUV image using GLKit openGL - iOS - ios

I want to render a yuv image on an iOS device. I presume it can be achieved using openGL. (Actually I have to render multiple such images in succession)
What I understand is that GLKit is an abstraction that iOS created, in which there is a GLKView which will have and handle the render buffer. I am currently trying to use a GLKViewController and frame update is being done successfully with desired fps. This I conform by using glClear function call.
Now the task is to render an image on the view.
There is a class GLKBaseEffect which will have basic shaders. I can't figure out what properties to set, so I just create it and call prepareToDraw before each render.
There is a class for handling textures, GLKTextureLoader, but it appears to me that it only works with Quartz images, i.e., yuv420 planar can't be loaded into the texture using this class.
// create the texture
GLuint texture;
glGenTextures(1, &texture);
glBindTexture(GL_TEXTURE_2D, texture);
glTexImage2D(GL_TEXTURE_2D, 0, GL_RGBA8, width, height, 0, GL_RGBA, GL_UNSIGNED_BYTE, [imageNSData bytes]);
glFramebufferTexture2D(GL_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, texture, 0);
I use this code for generating a texture and binding it, but I don't really know what I am trying to do here. And whatever it is, it doesn't bring up any image on screen, and I don't know what to do next.
I have not created any shaders, assuming baseEffect will have something.
And this https://developer.apple.com/library/ios/documentation/3DDrawing/Conceptual/OpenGLES_ProgrammingGuide/WorkingwithEAGLContexts/WorkingwithEAGLContexts.html#//apple_ref/doc/uid/TP40008793-CH103-SW8 tells me that I'll have to use EAGLayer to render images on screen.
Can I use GLKit to render images? If YES, do we have any sample code or tutorial that wouldn't use GLKTextureLoader class (I couldn't find any)? If NO, is there a similar tutorial for render using EAGLLayer (I have not explored about it till now) ?

It sounds like you're really asking about a few different topics here:
How to draw with GLKView vs CAEAGLLayer
How GLKBaseEffect and GLKTextureLoader fit into OpenGL ES drawing in general
How to draw a texture once you have one
How to render YUV image data
I'll try to address each in turn...
GLKView is just fine for any OpenGL ES drawing you want to do -- it does everything that the older documentation you linked to does (setting up framebuffers, CAEAGLLayers, etc) for you so you don't have to write that code. Inside the GLKView drawing method (drawRect: or glkView:drawInRect:, depending on whether you're drawing in a subclass or delegate), you write the same OpenGL ES drawing code you would for CAEAGLLayer (or any other system).
You can use GLKView, GLKTextureLoader, and GLKBaseEffect independently of each other. If you want to write all your own drawing code and use your own shaders, you can draw in a GLKView without using GLKBaseEffect. (You can even mix and match GLKBaseEffect and your own stuff, like you see when you create a new Xcode project with the OpenGL Game template.) Likewise, GLKTextureLoader loads image data and spits out the name you'll need for binding it for drawing, and you can use that regardless of whether you're drawing it with GLKBaseEffect.
Once you get a texture, whether via GLKTextureLoader or reading/decoding the data yourself and providing it to glTexImage2D, there are three basic steps to drawing with it:
Bind the texture name with glBindTexture.
Draw some geometry to be textured (using glDrawArrays, glDrawElements, or similar)
Have a fragment shader that looks up texels and outputs colors.
If you just want to draw an image that fills your view, just draw a quad (two triangles). Here's the code I use to set up a vertex array object with one quad when I want to draw fullscreen:
typedef struct __attribute__((packed)) {
GLKVector3 position;
GLKVector2 texcoord;
} Vertex;
static Vertex vertexData[] = {
{{-1.0f, 1.0f, -1.0f}, {0.0f, 0.0f}},
{{-1.0f, -1.0f, -1.0f}, {0.0f, 1.0f}},
{{ 1.0f, 1.0f, -1.0f}, {1.0f, 0.0f}},
{{ 1.0f, -1.0f, -1.0f}, {1.0f, 1.0f}},
};
glGenVertexArraysOES(1, &_vertexArray);
glBindVertexArrayOES(_vertexArray);
glGenBuffers(1, &_vertexBuffer);
glBindBuffer(GL_ARRAY_BUFFER, _vertexBuffer);
glBufferData(GL_ARRAY_BUFFER, sizeof(vertexData), vertexData, GL_STATIC_DRAW);
glEnableVertexAttribArray(GLKVertexAttribPosition);
glVertexAttribPointer(GLKVertexAttribPosition, 3, GL_FLOAT, GL_FALSE, sizeof(Vertex), (GLvoid *)offsetof(Vertex, position));
glEnableVertexAttribArray(GLKVertexAttribTexCoord0);
glVertexAttribPointer(GLKVertexAttribTexCoord0, 2, GL_FLOAT, GL_FALSE, sizeof(Vertex), (GLvoid *)offsetof(Vertex, texcoord));
glBindVertexArrayOES(0);
Then, to draw it:
glBindVertexArrayOES(_vertexArray);
glDrawArrays(GL_TRIANGLE_STRIP, 0, 4);
That's just the geometry-drawing part. Combine this with a GLKBaseEffect -- whose transform property is the default identity transform, and whose texture2d0 property is set up with the name of a texture you've loaded via GLKTextureLoader or other means -- and you'll get a view-filling billboard with your texture on it.
Finally, the YUV part... for which I'll mostly punt. Where are you getting your YUV texture data? If it's from the device camera, you should look into CVOpenGLESTexture/CVOpenGLESTextureCache, as covered by this answer. Regardless, you should be able to handle YUV textures using the APPLE_rgb_422 extension, as covered by this answer. You can also look into this answer for some help writing fragment shaders to process YUV to RGB on the fly.

Related

OpenGL ES 2.0 - GL_POINTS and Stencil Buffer

I am trying to draw a texture in OpenGL ES 2.0 using GL_POINTS by applying a stencil buffer. The stencil buffer should come from a texture. I am rendering the results to another texture and then presenting the texture to screen. Here is my code for rendering to texture:
//Initialize buffers, initialize texture, bind frameBuffer
.....
glClearStencil(0);
glClear (GL_STENCIL_BUFFER_BIT);
glColorMask( GL_FALSE, GL_FALSE, GL_FALSE, GL_FALSE );
glEnable(GL_STENCIL_TEST);
glStencilFunc(GL_ALWAYS, 1, 1);
glStencilOp(GL_KEEP, GL_KEEP, GL_REPLACE);
glBindTexture(GL_TEXTURE_2D, stencil);
glUseProgram(program[PROGRAM_POINT].id);
glDrawArrays(GL_POINTS, 0, (int)vertexCount);
glColorMask(GL_TRUE, GL_TRUE, GL_TRUE, GL_TRUE);
glStencilFunc(GL_NEVER, 0, 1);
glStencilOp(GL_REPLACE, GL_KEEP, GL_KEEP);
glBindTexture(GL_TEXTURE_2D, texture);
glUseProgram(program[PROGRAM_POINT].id);
glDrawArrays(GL_POINTS, 0, (int)vertexCount);
glDisable(GL_STENCIL_TEST);
....
//Render texture to screen
The result I am getting is just my texture being drawn without any masking applied from the stencil. I had a few questions regarding this issue:
Is is possible to use a stencil buffer with GL_POINTS?
Is is possible to use a stencil buffer when rendering to a texture?
Does the stencil texture have to have any special properties (solid colour, internal format...etc)?
Are there any apparent mistakes with my code?
This is the result I am looking for:
UPDATE:
My problem, as pointed out by the selected answer, was primarily that I did not attach the stencil to the stencil attachment of the FBO:
glFramebufferRenderbuffer(GL_FRAMEBUFFER, GL_STENCIL_ATTACHMENT,
GL_RENDERBUFFER, stencilBufferId);
I did not know that it was required when rendering to a texture. Secondly I was not using the proper stencil test.
glStencilFunc(GL_EQUAL, 1, 1);
glStencilOp(GL_KEEP, GL_KEEP, GL_KEEP);
Did the job.
Addressing your questions in order:
Is is possible to use a stencil buffer with GL_POINTS?
Yes. The stencil test is applied to all fragments, no matter of the primitive type rendered. The only case where you write to the framebuffer without applying the stencil test is with glClear().
Is is possible to use a stencil buffer when rendering to a texture?
Yes. However, when you render to a texture using an FBO, the stencil buffer of your default framebuffer will not be used. You have to create a stencil renderbuffer, and attach it to the stencil attachment of the FBO:
GLuint stencilBufferId = 0;
glGenRenderbuffers(1, &stencilBufferId);
glBindRenderbuffer(GL_RENDERBUFFER, stencilBufferId);
glRenderbufferStorage(GL_RENDERBUFFER, GL_STENCIL_INDEX8, width, height);
glFramebufferRenderbuffer(GL_FRAMEBUFFER, GL_STENCIL_ATTACHMENT,
GL_RENDERBUFFER, stencilBufferId);
Does the stencil texture have to have any special properties (solid colour, internal format...etc)?
OpenGL ES 2.0 does not have stencil textures. You have to use a renderbuffer as the stencil attachment, as shown in the code fragment above. GL_STENCIL_INDEX8 is the only format supported for renderbuffers that can be used as stencil attachment. ES 3.0 supports depth/stencil textures.
Are there any apparent mistakes with my code?
Maybe. One thing that looks slightly odd is that you never really apply a stencil test in the code that is shown. While you do enable the stencil test, you only use GL_ALWAYS and GL_NEVER for the stencil function. As the names suggest, these functions either always or never pass the stencil test. So you don't let fragments pass/fail depending on the stencil value. I would have expected something like this before the second draw call:
glStencilFunc(GL_EQUAL, 1, 1);
glStencilOp(GL_KEEP, GL_KEEP, GL_KEEP);
This would only render the fragments where the current stencil buffer value is 1, which corresponds to the fragments you rendered with the previous draw call.

How to input video (frames) into a GLSL shader

I'm trying to do video processing using GLSL. I'm using OpenCV to open a video file up and take each frame as a single image an then I want to use each frame in a GLSL shader
What is the best/ideal/smart solution to using video with GLSL?
Reading From Video
VideoCapture cap("movie.MOV");
Mat image;
bool success = cap.read(image);
if(!success)
{
printf("Could not grab a frame\n\7");
exit(0);
}
Image to Texture
GLuint tex;
glGenTextures(1, tex);
glActiveTexture(GL_TEXTURE0);
glBindTexture(GL_TEXTURE_2D, tex);
glTexImage2D(GL_TEXTURE_2D, 0, GL_RGB, image.cols, image.rows, 0,
GL_BGR, GL_UNSIGNED_BYTE, image.data);
glUniform1i(glGetUniformLocation(shaderProgram, "Texture"), 0);
What needs to be in my render while loop?
Do I need to recompile/reattach/relink my shader every time? Or Once my shader is created and compiled and I use glUseProgram(shaderProgram) can I keep sending it new textures?
The current loop I've been using to render a texture to the screen is as follows. How Could I adapt this to work with video? Where would I need to make my calls to update the texture being used in the shader?
while(!glfwWindowShouldClose(window))
{
glClearColor(0.0f, 0.0f, 0.0f, 1.0f);
glClear(GL_COLOR_BUFFER_BIT);
glViewport(0,0,512,512);
glBindFramebuffer(GL_READ_FRAMEBUFFER, frameBuffer);
glReadBuffer(GL_COLOR_ATTACHMENT0);
glDrawElements(GL_TRIANGLES, 6, GL_UNSIGNED_INT, 0);
glBindFramebuffer(GL_DRAW_FRAMEBUFFER, 0);
glBlitFramebuffer(0, 0, image.cols, image.rows, 0, 0, image.cols, image.rows, GL_COLOR_BUFFER_BIT, GL_LINEAR);
glfwSwapBuffers(window);
glfwPollEvents();
}
Let's clarify a few things that needs to happen before the loop:
Set the pixel storage mode with glPixelStorei();
Generate only one texture with glGenTextures(), because at every iteration of the loop its content will be replaced with new data;
Compile the shader with glCompileShader(), use glCreateShader() to create a shader object, invoke glCreateProgram()to create a program, call glAttachShader() to attach the shader object to the new program, and finally glLinkProgram() to make everything ready to go.
That said, every iteration of the loop must:
Clear the color and depth buffer;
Load the modelview matrix with the identity matrix;
Specify the location where the drawing is going to happen glTranslatef();
Retrieve a new frame from the video;
Enable the appropriate texture target, bind it and then transfer the frame to the GPU with glTexImage2D();
Invoke glUseProgram() to activate your GLSL shader;
Draw a 2D face using GL_QUADS or whatever;
Disable the program with glUseProgram(0);
Disable the texture target with glDisable(GL_TEXTURE_YOUR_TEXTURE_TARGET);
This is more or less what needs to be done.
By the way: here's my OpenCV/OpenGL/Qt application that retrieves frames from the camera and displays it in a window. No shaders, though.
Good luck!
You don't need to use a framebuffer to send textures to the shader. Once you've got texture 0 selected as the active texture and 0 set as the value of the uniform sampler2D in your shader, every time you call glBindTexture(), it will set the sampler2D to whichever texture you've specified in the function parameter. So no, you don't need to relink or recompile your shader each time you want to change texture.

Which OpenGL ES 1.1 alpha blending configuration to use on iOS?

If you want to achieve a blending of textures with transparency (like PNG) that is similar to UIKit, how do you configure OpenGL ES 1.1 appropriately?
I found:
glEnable(GL_BLEND);
glBlendFunc(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA);
glDisable(GL_ALPHA_TEST);
But there is also:
glEnable(GL_BLEND);
glBlendFunc(GL_ONE, GL_ONE);
You should use glBlendFunc(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA);. The other one has no respect for the alpha channel at all in this case, it will simply sum up the source and destination colours for you.
There is one thing you should note though. This will work great on colours but your destination alpha channel will not be correct. In most cases you do not use it but if you wish to extract the image from buffer with the alpha channel as well you will need it, same goes for using FBO and reuse texture with transparency. In this case you should draw the alpha channel separately using glBlendFunc(GL_ONE_MINUS_DST_ALPHA, GL_ONE). This means doubling the draw calls (at least I can't think of a better solution without shaders). To draw to colour only you have to set glColorMask(GL_TRUE, GL_TRUE, GL_TRUE, GL_FALSE) and to draw alpha only use glColorMask(GL_FALSE, GL_FALSE, GL_FALSE, GL_TRUE)
To sum up:
glEnable(GL_BLEND);
glColorMask(GL_TRUE, GL_TRUE, GL_TRUE, GL_FALSE);
glBlendFunc(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA);
//draw the scene
glColorMask(GL_FALSE, GL_FALSE, GL_FALSE, GL_TRUE);
glBlendFunc(GL_ONE_MINUS_DST_ALPHA, GL_ONE);
//draw the scene
glDisable(GL_ALPHA_TEST);
glDisable(GL_BLEND);

OpenGL slows down when rendering nearby objects on top of others

I am writing an iOS app using OpenGL ES 2.0 to render a number of objects to the screen.
Currently, those objects are simple shapes (squares, spheres, and cylinders).
When none of the objects overlap each other, the program runs smoothly at 30 fps.
My problem arises when I add objects that appear behind the rest of my models (a background rectangle, for example). When I attempt to draw a background rectangle, I can only draw objects in front of it that take up less than half the screen. Any larger than that and the frame rate drops to between 15 and 20 fps.
As it stands, all of my models, including the background, are drawn with the following code:
- (void)drawSingleModel:(Model *)model
{
//Create a model transform matrix.
CC3GLMatrix *modelView = [CC3GLMatrix matrix];
//Transform model view
// ...
//Pass matrix to shader.
glUniformMatrix4fv(_modelViewUniform, 1, 0, modelView.glMatrix);
//Bind the correct buffers to openGL.
glBindBuffer(GL_ARRAY_BUFFER, [model vertexBuffer]);
glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, [model indexBuffer]);
glVertexAttribPointer(_positionSlot, 3, GL_FLOAT, GL_FALSE, sizeof(Vertex), 0);
glVertexAttribPointer(_colorSlot, 4, GL_FLOAT, GL_FALSE, sizeof(Vertex), (GLvoid*) (sizeof(float) * 3));
//Load vertex texture coordinate attributes into the texture buffer.
glVertexAttribPointer(_texCoordSlot, 2, GL_FLOAT, GL_FALSE, sizeof(Vertex), (GLvoid*) (sizeof(float) * 7));
glActiveTexture(GL_TEXTURE0);
glBindTexture(GL_TEXTURE_2D, [model textureIndex]);
glUniform1i(_textureUniform, 0);
glDrawElements([model drawMode], [model numIndices], GL_UNSIGNED_SHORT, 0);
}
This code is called from my draw method, which is defined as follows:
- (void)draw
{
glUseProgram(_programHandle);
//Perform OpenGL rendering here.
glBlendFunc(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA);
glEnable(GL_BLEND);
glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);
glEnable(GL_DEPTH_TEST);
glEnable(GL_CULL_FACE);
_camera = [CC3GLMatrix matrix];
//Camera orientation code.
//...
//Pass the camera matrix to the shader program.
glUniformMatrix4fv(_projectionUniform, 1, 0, _camera.glMatrix);
glViewport(0, 0, self.frame.size.width, self.frame.size.height);
//Render the background.
[self drawSingleModel:_background];
//Render the objects.
for(int x = 0; x < [_models count]; ++x)
{
[self drawSingleModel:[_models objectAtIndex:x]];
}
//Send the contents of the render buffer to the UI View.
[_context presentRenderbuffer:GL_RENDERBUFFER];
}
I found that by changing the render order as follows:
for(int x = 0; x < [_models count]; ++x)
{
[self drawSingleModel:[_models objectAtIndex:x]];
}
[self drawSingleModel:_background];
my frame rate when rendering on top of the background is 30 fps.
Of course, the slowdown still occurs if any objects in _models must render in front of each other. Additionally, rendering in this order causes translucent and transparent objects to be drawn black.
I'm still somewhat new to OpenGL, so I don't quite know where my problem lies. My assumption is that there is a slowdown in performing depth testing, and I also realize I'm working on a mobile device. But I can't believe that iOS devices are simply too slow to do this. The program is only rendering 5 models, with around 180 triangles each.
Is there something I'm not seeing, or some sort of workaround for this?
Any suggestions or pointers would be greatly appreciated.
You're running in one of the peculiarities of mobile GPUs: Those things (except the NVidia Tegra) don't do depth testing for hidden surface removal. Most mobile GPUs, including the one in the iPad are tile based rasterizers. The reason for this is to save memory bandwidth, because memory access is actually a power intensive operation. In the power constrained environment of a mobile device reducing required memory bandwidth gains significant battery lifetime.
Tile based renderers split up the viewport into a number of tiles. When sending geometry into it, it is split into the tiles and then for each tile it is intersected with the the geometry already in the tile. Most of the time the tile is covered by only a single primitive. If the incoming primitive happens to be in front of the already present geometry it replaces it. If there's a cutting intersection a new edge is added. Only if a certain threshold of number of edges is reached, that single tile will switch to depth testing mode.
Only at synchronization points the prepared tiles are rasterized, then.
Now it's obvious why overlapping objects reduce rendering performance: The more primitives overlap, the more preprocessing has to be done to setup the tiles.
See "transparency sorting"/"alpha sorting".
I suspect the slowness you're seeing is largely due to "overdraw", i.e. framebuffer pixels being drawn more than once. This is worst when you draw the scene back-to-front, since the depth test always passes. While the iPhone 4/4S/5 may have a beefy GPU, last I checked the memory bandwidth was pretty terrible (I don't know how big the GPU cache is).
If you render front-to-back, the problem is that transparent pixels still write to the depth buffer, causing them to occlude polys behind them. You can reduce this slightly (but only slightly) using the alpha test.
The simple solution: Render opaque polys approximately front-to-back and then transparent polys back-to-front. This may mean making two passes through your scene, and ideally you want to sort the transparent polys which isn't that easy to do well.
I think it's also possible (in principle) to render everything front-to-back and perform alpha testing on the destination alpha, but I don't think OpenGL supports this.

Opengl ES2 clipping mask

I'm looking for a way to mask my entire viewport using a texture. In my case I would like to use a checkered black and white pattern (or any 2 colors) and only show the parts that are black on the scene.
Would the best way to do this be with a cliping mask, a fragment shaders, or an alpha blending. I've seen on SO this post: How to create Stencil buffer with texture (Image) in OpenGL-ES 2.0 which seems similar to what I need, but I don't completely understand what to do with the discard keyword. Would it apply to my situation.
Let's assume you have a checkered texture of black and white squares. First, you'll want to setup the stencil test to draw the mask:
glEnable(GL_STENCIL_TEST);
glDisable(GL_DEPTH_TEST);
glColorMask(GL_FALSE, GL_FALSE, GL_FALSE, GL_FALSE);
glDepthMask(GL_FALSE);
glStencilFunc(GL_ALWAYS, 1, -1);
glStencilOp(GL_KEEP, GL_KEEP, GL_REPLACE);
Next draw the checkered texture. The key step here (where discard comes in), is that you set up your fragment shader to only draw fragments where the checkerboard is black (or flip it for white if you prefer). The discard keyword skips rendering of any pixels that don't match your criteria.
//in the fragment shader
if(sampleColor.r > 0.5) { discard; }
After this rendering step you will have a stencil buffer with a checkerboard image where half of the buffer has a stencil value of 0 and the other half has a stencil value of 1.
Then render normally with stencil test enabled to pass when the stencil buffer is == 1.
glStencilOp(GL_KEEP, GL_KEEP, GL_KEEP);
glStencilFunc(GL_EQUAL, 1, -1);

Resources