I'm currently working through a set of tutorials on Android OpenGL ES (1.1) and feel like I'm starting to get a grasp of how the vertices and textures work, along with some sprite animation. As I understand, the only primitives here are points, straight lines, and triangles.
I'm now trying to create a simple curve and really don't know where to start.
I want the curve to be drawn dynamically to represent something like a beam deflection like this where I could input a force and have the curve change.
Is it something I would create with a line loop or triangle fan with a ton of vertices? Or perhaps a texture that I then manipulate?
Any input or a point in the right direction is much appreciated, thanks.
I can recommend this blog post http://blog.uncle.se/2012/02/opengl-es-tutorial-for-android-part-ii-building-a-polygon/ sadly the original source returns a 404. Hopefully the link provides the same quality of information. Anyhow, a good read for openGL.
You have the general idea. Whatever you do must be made of lines, points or triangles. You can generate all the numbers for any pseudo-curve however you want, but you're always going to be passing the resulting vertices to OPENGL then connecting those with lines and triangles.
Related
I've been programming for about a couple years now, and I know the syntax of different languages pretty well, but I've spent most of my time making programs in basic Lua. I have played around with game developing with Java and C#, with DirectX and OpenGl, but most times I don't even know terms that others may know really well such as 'shaders' and 'VBOs'. I've heard of them, but I really don't know what they mean, so please explain to me like I'm five if you have a answer.
I'm trying to make a simple Android game using Open GL ES 2.0, and, I've seen many tutorials explaining how to draw one triangle, and not many. For my case, I have a world full of triangles, its simply an array of true and false's, where if there is a true a triangle will be drawn in that position, if false the triangle won't be drawn. I've made it so it will only draw triangle's until they are off screen, but I realized one problem still, and that is that rendering is still super slow, takes about 2 seconds to render one frame.
At the moment, I currently 1 triangle class, and when it checks if a triangle is supposed to belong where it is, it will create vertex points for that triangle and pass them into the triangle class, then creating a new triangle object with those updated vertices. Now I can see how wrong this can go, but it's honestly the only thing I can come up with, with my very-limited knowledge on OpenGL ES
What I am looking for is to draw all of these triangles, in their correct position, with the same color and size, in a very efficient way, so that it wouldn't take about 2 seconds to draw one frame.
If anyone has a solution, Thank you.
What you are looking for is for instanced draw calls that basically means that you are going to send one VBO (Vertex Buffer Object - this holds all the information about the vertex of a mesh) and an array of MVPM(Model View Projection Matrices - These hold all the transformations information of every draw you want to do of that VBO) to the shader, all this in one single draw call.
unfortunately for us (i'm using right now OpenGL Es 2.0 as well) there is no way provided by opengl to do this.
Any way, there is this article about fake that but i think is too much trouble.
In other words ir are pretty much trapped with a draw call per mesh.
I am working on an Android OpenGL ES tutorial and it says: "defining triangles is pretty easy in OpenGL, but what if you want to get a just a little more complex? Say, a square? There are a number of ways to do this, but a typical path to drawing such a shape in OpenGL ES is to use two triangles drawn together"
http://developer.android.com/training/graphics/opengl/shapes.html
Why is the typical path to drawing such a shape to use two triangles instead of drawing the four corner coordinates of the square?
Graphics cards and rendering options are only really coded to render triangles, and not anything harder. The idea is that every other shape anyone could ever think of can be duplicated or approximated with triangles, sometimes billions or trillions of them. When you see GPUs being compared, sometimes you hear "maximum polygons on the screen" or something similar. They really mean triangles, but polygons sounds cooler. Triangles are simple enough to create, but provide fantastic utility. They don't need to have ordered points, which is a huuuge help.
The tl;dr answer is that GPUs render triangles really well, so much so they don't bother knowing how to render much else.
OpenGL ES 2.0 is implemented in a project that I have been working on with a couple shader components that define what a texture should look like after modifications from a Bitmap. The SurfaceView will only ever have a single image in it for my project.
While doing several different approaches and looking through code in the past 24 hours, just hoping for a quick response or two from the community. Not looking for solutions, I'll do that research.
It sounds as though since we are using shaders, that in order to do scaling and movements in the texture based on touch events, that I will have have to use the Matrix utilities and OpenGL translations or movements with the camera to get the same effect as what is currently done within an ImageView. Would this be the appropriate approach? Perhaps even modify the shader code so that I have some additional input variables?
I don't believe that I can use anything on the Android side that would get the same effect, such as modifying the canvas of the SurfaceView or altering dimensions of the UI in some other fashion that would achieve the same effect?
Thanks. Again, solutions for zooming and moving around aren't necessary, just trying to get a grasp on intermixing OpenGL and Android appropriately for the task.
Why does it seem that several elements in 1.0 are easier than 2.0; ease of use should improve between releases.
Yes. You will need to use an ortho projection and adjust the extents to zoom. See this link here. To pan, you can simply use a glTranslatef.
If you would like to do this entirely in the pixel shader, you can use the texture matrix stack with glScalef and glTranslatef.
I have a 3D cube created using GL_TRIANGLE_STRIP.Is it possible to draw points(using GL_POINTS) or a triangle (using GL_TRIANGLE) on/inside my 3D Cube?How could that be achieved?
If you want to draw something directly of the face of another object (by using the exact same vertex coordinates), you will need to use glPolygonOffset to prevent stitching. There is a chapter in the Red Book that explains it.
If by inside you mean to draw something in the volume of the cube, than there is nothing stopping you. You just need to get the alpha values and blending right to actually see through the cube. Look for some generic tutorial on transparency in OpenGL.
But maybe I'm horribly mistaken and what you are looking for a textures.
If I understand you correctly you could just generate the appropriate texture with the points and apply it to the cube.
Hey all,
I'm writing a little 3D engine for android to better learn the platform and opengl. I hope to eventually use it as a base for little 3D games.
I'm trying to implement lighting right now, and generally following the NeHe opengl android port tutorials. I have a simple spinning cube and 1 light that should not change position. Upon rendering my scene on a device the light appears to "dim" and re-lighten as cube rotates.
This is a swf video of the behavior:
http://drop.io/obzfq4g
The code for my "engine" is located here: http://github.com/mlasky/Smashout-Android-3D-Engine/
The relevant bits are:
http://github.com/mlasky/Smashout-Android-3D-Engine/blob/master/src/com/supernovamobile/smashout/SmashoutGLRenderer.java
this is where I'm initializing opengl and setting up my scene.
and
http://github.com/mlasky/Smashout-Android-3D-Engine/blob/master/src/com/supernovamobile/smashout/SMMesh.java
this is the code for actually rotating / drawing the cube mesh.
I wish I could formulate a better question; but I'm very stuck/confused and can't even think of an intelligent question to ask besides "does anyone know what might cause the kind of behavior I'm seeing?"
Thanks for any help you can provide.
Edit: Also the slowness / choppyness of the animation in the video is a result of the screencap software. It's smooth throughout the whole rotation on the emu.
From the video it looks like your normals are incorrect.
For an axis-aligned cube the normals in cube.xml seem off. They should be axis-aligned, not whatever they are. Where did you get the model?
This answer may also be related if mVNormalsBuffer is empty.
You are using
gl.glNormalPointer(3, GL10.GL_FLOAT, mVNormalsBuffer);
wrong.
It should be:
gl.glNormalPointer(GL10.GL_FLOAT,0, mVNormalsBuffer);