Sign in to follow this  

OpenGL Handling resolutions and optimizing? [2D game]

This topic is 2132 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

So I've got my nice little 2D game, and I'd like to ask what's the best method for handling resolutions?

Setting glViewPort to my new width and height scales the graphics well enough, but the window size doesn't change, but if I create a new window (and reload the textures) while setting my view port, the graphics don't scale, just remain in the same position with the same size. And if I create a new window without setting the view port the graphics just shift positions (and they do scale) but it doesn't seem like it's working right.

Also, would I ever need to manually scale up the graphics? Because I'm using box2d for the physics and it'd be annoying to have to destroy all box2d objects and create new, scaled versions. Just playing with the view port scales up and down the graphics without any collision problems, just need to get the window to change without any problems as well.

[url=""]My resolution changing class.[/url]

[b]Now question 2: Optimization[/b]

The game runs quite fine at 60 fps normally, but I decided to do some stress tests and tossed about 400 box2d objects with textures, PNG, with alpha and resized falling around, and ones that exit the screen are destroyed and new ones are created (Could have probably done some object pooling I think)

Now capping the FPS at 200, it runs at 60 fps on my above average computer, and about 30 on below average computers. Now I imagined openGL could render far more than 400 moving textures, especially since it can do all that 3D magic, so I'm guessing I'm doing something wrong?

Are there any optimization tips with openGL or "gotchas" ? For example when I first implemented openGL I was creating and destroying textures every time I drew, so that was incredibly laggy. Now I create the texture once, then bind and draw every frame.

Any tips on either of these questions would be appreciated, thanks!

Share this post

Link to post
Share on other sites
1) If you want your graphics to scale when you change resolution you need to change the projection matrix aswell(Which you appear to be doing) and render using a fixed scale that is independant of the resolution. (This should basically make things appear smaller when you increase the resolution)

The rendering shouldn't affect the game simulation in any way. (You can change the scale of objects manually using glScale) (or for newer opengl versions, by modifying the transformation matrix accordingly)

First the big OpenGL one: Immediate mode is "slow", if you're using glbegin / glend you are using immediate mode. (the immediate mode overhead for 400 quads however is fairly negligable and most likely isn't the source of your performance problems).

Secondly, Avoid rebinding textures, state changes are expensive (The driver may or may not ignore redundant statechanges but there is still a cost to attempting them), Try to sort the quads based on which texture they have and render all quads using textureA first, then all quads with textureB etc, only rebind texture when you switch group, combining multiple texture into an atlas (and offsetting the texture coordinates accordingly for the objects) can let you greatly cut down on these changes.

For Box2D:
Try simply moving the objects that fall off the screen to the top instead of destroying them and recreating them, it is quite likely that this is one of the major performance hogs, You could also try running the simulation without actually rendering anything to see how slow it gets when you eliminate the rendering aspect.

Share this post

Link to post
Share on other sites
Those are some really good tips for optimization SimonForsman, I'll try them out as soon as I can. I'm guessing this means making a sprite sheet for animation is better than having a sequence of PNG images? (Even if the sprite sheets are as big as 2000x2000 ?)

About resolutions, so you're saying the way I'm doing it is working the way it should? And that I should append a function to scale the graphics using glScale? (Also, shouldn't the graphics get larger as we increase the resolution?)

Share this post

Link to post
Share on other sites
[quote name='OmarShehata' timestamp='1331473279' post='4921117']
Those are some really good tips for optimization SimonForsman, I'll try them out as soon as I can. I'm guessing this means making a sprite sheet for animation is better than having a sequence of PNG images? (Even if the sprite sheets are as big as 2000x2000 ?)

About resolutions, so you're saying the way I'm doing it is working the way it should? And that I should append a function to scale the graphics using glScale? (Also, shouldn't the graphics get larger as we increase the resolution?)

No, increasing resolution means you get more pixels on screen, if your resolution is 100x100 (very low resolution) and you got an image that is 50x50 that image will occupy 25% of the screen (or window), if you increase the resolution to 200x200 then the 50x50 image will only occupy 6.25% of the screen/window (Thus if you're using fullscreen then increasing resolution will make the graphics appear smaller, if you're using a non fullscreen window then the graphics will keep the same size relative to the monitor but not relative to the window)

If what you want to do is to have objects always occupy the same portion of the screen so that they become larger when you make the window larger then the easiest way is to pass a fixed width and height to your projection matrix (glOrtho doesn't have to take the width and height of the window as parameters, you can pass it for example 800x600 all the time and it will scale the graphics for you so that a 400x300 object always covers 25% of the window) (This will however cause the graphics to stretch if you change the aspect ratio)

Looking at your code again i think the resolution problem you have is simply that you're not using a fullscreen window (And thus its not really a problem as it is working as intended), (If you change the resolution of a normal window you are really changing its size, the actual resolution will still be whatever the desktop resolution is), You shouldn't have to use glScale unless you want to scale the graphics (scaling graphics is useful if you want two objects to use the same mesh but have different sizes).

a 2000x2000 sprite sheet shouldn't be a huge problem for modern graphicscards, older ones however will struggle though:

1) non power of two textures can be slower on old hardware (a 2048x2048 texture will often be faster than a 2000x2000 texture) and square textures are often faster than rectangular textures. newer hardware are better at dealing with odd texture sizes but can still take a serious performance hit.

2) a 2000x2000 texture might be too large for older GPUs to render using hardware acceleration or even at all . (Most new consumer cards will happily handle 4096x4096 while professional cards can often do 8192x8192 textures without problems) , you can create a proxy texture using GL_PROXY_TEXTURE_2D to see if the OpenGL implementation can use a texture of a given size, (For proxy textures the width and height will be set to 0 if the driver is unable to deal with it), there is no guarantee that it will be hardware accelerated though).

Also, you can have a sequence of png images and combine them into one or more larger textures on loadtime if you want. (This can allow you to make optimized texture atlases for each level that only includes the images you need for that specific level, or you can group the most common objects for a level into one atlas).

If for example you use sprite animations where each object always is on the same frame (so that when objectX is on frame1 objectY will also always be on frame1) it could be beneficial to keep all first frames in one texture, all second frames in another, etc (That way you only have to rebind the texture once at the start of each frame and then possibly once more to draw all non animated objects)

Share this post

Link to post
Share on other sites

I've already done a few benchmarks on the stuff you're discussing right now.

First, by compiling all of your textures into one mega-texture (or at least relevant textures), you'll help yourself out a lot. I found that by putting all of my heroes into one giant sprite sheet (which is done at run-time from several sources), you can increase your frame rate a bit. Switching between texture contexts is a somewhat expensive operation, from what I've been told.

I try to compile all textures into relevant layers -- i.e. tiles/world graphics in one might suit your needs; heroes in another; maybe a "monster" one for regional monsters that appear in different areas. Just depends on what you're using. But you can knock down all of the differing calls to glBindTexture() to your sprite sheet once for each layer. i.e. Draw World; Draw Heroes; Draw Monsters. You'd glBindTexture three times.

By doing this (I have a few more layers than that), I jumped my frame-rate pretty drastically. Somewhere around 20-25%. I was rendering a lot of objects, though.

Another note is that the engine I'm working on runs fine on an old netbook (Intel Celeron low voltage processor) with 2048x2048 compiled sprite sheets just fine. The engine also ran fine on an Intel Pentium 4, which supported up to that 2048x2048 texture. These machines did *not* have dedicated GPU's. Of course, I limited how much of the screen was drawn (about 1/2 the real estate -- from 800x600 to 640x480) to maintain frame rates. However, I'm still getting some good frame-rates, so I may jump it back up after using Drawing Arrays. Food for thought.

Share this post

Link to post
Share on other sites
Thanks for the help so far everyone! I really really appreciate it!

I'm still fiddling with the resolution stuff. I've passed a fixed height and width to my glOrtho function as SimonForsman suggested, (960,540 as it's my minimum resolution) and everything's working. The problem is when I go into full screen. Sometimes there will be black strips on the left and right, which makes sense if the game's resolution doesn't match the screen resolution right? The game's aspect ratio is always 16:9. So far so good.

Now testing on a 5:4 monitor, the results are pretty haywire. Pic:

Left: 960x540

The middle picture I think is exactly how it should be working. The game's screen and graphics are not stretched, and there is a black bar at the top (should be half at the top, half at the bottom if I figure out how to center the game screen).

The picture on left isn't stretched, which is good, but it doesn't scale enough. I still don't really think that's a problem.

The problem is with the picture on the right. The game's screen is now stretched. Why would it be stretched? There's no reason for it to be stretched, the aspect ratio hasn't changed.

I'd really like to understand exactly what glOrtho, glViewport and full screen work and why things are doing what they're currently doing. I know glOrtho is the camera, and glViewport is like, the window in which I render things. So my keeping glOrtho a constant size and width, the graphics should never scale right? Or do I need to keep my view port constant as well? But then I'd see a larger area as the screen gets larger...which I don't want. The glViewport should always be 16:9, so why does it stretch?

Share this post

Link to post
Share on other sites
Sign in to follow this  

  • Similar Content

    • By _OskaR
      I have an OpenGL application but without possibility to wite own shaders.
      I need to perform small VS modification - is possible to do it in an alternative way? Do we have apps or driver modifictions which will catch the shader sent to GPU and override it?
    • By xhcao
      Does sync be needed to read texture content after access texture image in compute shader?
      My simple code is as below,
      glBindImageTexture(0, texture[0], 0, GL_FALSE, 3, GL_READ_ONLY, GL_R32UI);
      glBindImageTexture(1, texture[1], 0, GL_FALSE, 4, GL_WRITE_ONLY, GL_R32UI);
      glDispatchCompute(1, 1, 1);
      // Does sync be needed here?
      glBindFramebuffer(GL_READ_FRAMEBUFFER, framebuffer);
                                     GL_TEXTURE_CUBE_MAP_POSITIVE_X + face, texture[1], 0);
      glReadPixels(0, 0, kWidth, kHeight, GL_RED_INTEGER, GL_UNSIGNED_INT, outputValues);
      Compute shader is very simple, imageLoad content from texture[0], and imageStore content to texture[1]. Does need to sync after dispatchCompute?
    • By Jonathan2006
      My question: is it possible to transform multiple angular velocities so that they can be reinserted as one? My research is below:
      // This works quat quaternion1 = GEQuaternionFromAngleRadians(angleRadiansVector1); quat quaternion2 = GEMultiplyQuaternions(quaternion1, GEQuaternionFromAngleRadians(angleRadiansVector2)); quat quaternion3 = GEMultiplyQuaternions(quaternion2, GEQuaternionFromAngleRadians(angleRadiansVector3)); glMultMatrixf(GEMat4FromQuaternion(quaternion3).array); // The first two work fine but not the third. Why? quat quaternion1 = GEQuaternionFromAngleRadians(angleRadiansVector1); vec3 vector1 = GETransformQuaternionAndVector(quaternion1, angularVelocity1); quat quaternion2 = GEQuaternionFromAngleRadians(angleRadiansVector2); vec3 vector2 = GETransformQuaternionAndVector(quaternion2, angularVelocity2); // This doesn't work //quat quaternion3 = GEQuaternionFromAngleRadians(angleRadiansVector3); //vec3 vector3 = GETransformQuaternionAndVector(quaternion3, angularVelocity3); vec3 angleVelocity = GEAddVectors(vector1, vector2); // Does not work: vec3 angleVelocity = GEAddVectors(vector1, GEAddVectors(vector2, vector3)); static vec3 angleRadiansVector; vec3 angularAcceleration = GESetVector(0.0, 0.0, 0.0); // Sending it through one angular velocity later in my motion engine angleVelocity = GEAddVectors(angleVelocity, GEMultiplyVectorAndScalar(angularAcceleration, timeStep)); angleRadiansVector = GEAddVectors(angleRadiansVector, GEMultiplyVectorAndScalar(angleVelocity, timeStep)); glMultMatrixf(GEMat4FromEulerAngle(angleRadiansVector).array); Also how do I combine multiple angularAcceleration variables? Is there an easier way to transform the angular values?
    • By dpadam450
      I have this code below in both my vertex and fragment shader, however when I request glGetUniformLocation("Lights[0].diffuse") or "Lights[0].attenuation", it returns -1. It will only give me a valid uniform location if I actually use the diffuse/attenuation variables in the VERTEX shader. Because I use position in the vertex shader, it always returns a valid uniform location. I've read that I can share uniforms across both vertex and fragment, but I'm confused what this is even compiling to if this is the case.
      #define NUM_LIGHTS 2
      struct Light
          vec3 position;
          vec3 diffuse;
          float attenuation;
      uniform Light Lights[NUM_LIGHTS];
    • By pr033r
      I have a Bachelor project on topic "Implenet 3D Boid's algorithm in OpenGL". All OpenGL issues works fine for me, all rendering etc. But when I started implement the boid's algorithm it was getting worse and worse. I read article ( inspirate from another code (here: but it still doesn't work like in tutorials and videos. For example the main problem: when I apply Cohesion (one of three main laws of boids) it makes some "cycling knot". Second, when some flock touch to another it scary change the coordination or respawn in origin (x: 0, y:0. z:0). Just some streng things. 
      I followed many tutorials, change a try everything but it isn't so smooth, without lags like in another videos. I really need your help. 
      My code (optimalizing branch):
      Exe file (if you want to look) and models folder (for those who will download the sources):
      Thanks for any help...

  • Popular Now