Ok, I have a renderer class which has all kinds of special functions called by the rest of the program:
DrawBoxFilled
DrawText
DrawLine
About 30 more...
Each of these functions calls glBegin/glEnd separably, which I know can be very inefficiently(its even deprecated). So anyways, I am planning a total rewrite of the renderer and I need to know the most efficient ways to set up the functions so that when something calls it, it draws it all at once, or whatever else it needs to do so it will run most efficiently. Thanks in advance :)
The efficient way to render is generally to use VBO's (vertex buffer objects) to store your vertex data, but that is only really meaningful if you are rendering (mostly) static data.
Without knowing more about what your application is supposed to render, it's hard to say how you should structure it. But ideally, you should never draw individual primitives, but rather draw the contents (a subset) of a vertexbuffer.
The most efficient way is not to expose such low-level methods at all. Instead, what you want to do is build a scene graph, which is a data structure that contains a representation of the entire scene. You update the scene graph in your "update" method, then render the whole thing in one go in your "render" method.
Another, slightly different approach is to re-build the entire scene graph each frame. This has the advantage that once the scene graph is composed, it doesn't change. So you can call your "render" method on another thread while your "update" method is going through and constructing the scene for the next frame at the same time.
Many of the more advanced effects are simply not possible without a complete scene graph. You can't do shadow mapping, for instance (which requires you to render the scene multiple times from a different angle), you can't do deferred rendering, it also makes anything which relies on sorted draw order (e.g. alpha-blending) very difficult.
From your method names, it looks like you're working in 2D, so while shadow mapping is probably not high on your feature list, alpha-blending deferred rendering might be.
Related
I am currently working on a Graphics Engine for a 2D game with OpenGL, but I have issues implementing a viable solution for partial transparency, since opaque objects need to be drawn first, and transparent objects from back to front.
Specifically, I want to be able to pass render commands in an arbitrary order, and let the Engine handle the sorting.
My current approach is that I simply collect all vertices and textures of each transparent object in a sorted list, and at the and of each frame, I simply draw them from this list.
Although I get correct results, this is obviously not a viable solution, due to a lot of copy instructions. In case I have a few thousand partially transparent particles, this approach does not work at all due to its low performance.
I did not find any other way to implement this for 2D Graphics, so my question is: What is the most common approach to this ? Are there any sources where I can read more about this topic?
I use glBlendFunc(GL_SOURCE_ALPHA, GL_ONE_MINUS_SRC_ALPHA) by the way.
I am making a simple STG engine with OpenGL (To be exact, with LWJGL3).In this game, there can be several different types of items(called bullet) in one frame, and each type can have 10-20 instances.I hope to find an efficient way to render it.
I have read some books about modern OpenGL and find a method called "Instanced Rendering", but it seems only to work with same instances.Should I use for-loop to draw all items directly for my case?
Another question is about memory.Should I create an VBO for each frame, since the number of items is always changing?
Not the easiest question to answer but I'll try my best anyways.
An important property of OpenGL is that the OpenGL context is always bound to a single thread. So every OpenGL-method has to be called within that thread. A common way of dealing with this is using Queuing.
Example:
We are using Model-View-Controller architecture.
We have 3 threads; One to read input, one to handle received messages and one to render the scene.
Here OpenGL context is bound to rendering thread.
The first thread receives a message "Add model to position x". First thread has no time to handle the message, because there might be another message coming right after and we don't want to delay it. So we just give this message for the second thread to handle by adding it to second thread's queue.
Second thread reads the message and performs the required tasks as far as it can before OpenGL context is required. Like reads the Wavefront (.obj)-file from the memory and creates arrays from the received data.
Our second thread then queues this data to our OpenGL thread to handle. OpenGL thread generates VBOs and VAO and stores the data in there.
Back to your question
OpenGL generated Objects stay in the context memory until they are manually deleted or the context is destroyed. So it works kind of like C, where you have to manually allocate memory and free it after it's no more used. So you should not create new Objects for each frame, but reuse the data that stays unchanged. Also when you have multiple objects that use the same model or texture, you should just load that model once and apply all object specific differences on shaders.
Example:
You have an environment with 10 rocks that all share the same rock model.
You load the data, store it in VBOs and attach those VBOs into a VAO. So now you have a VAO defining a rock.
You generate 10 rock entities that all have position, rotation and scale. When rendering, you first bind the shader, then bind the model and texture, then loop through the stone entities and for each stone entity you bind that entity's position, rotation and scale (usually stored in a transformationMatrix) and render.
bind shader
load values to shader's uniform variables that don't change between entities.
bind model and texture (as those stay the same for each rock)
for(each rock in rocks){
load values to shader's uniform variables that do change between each rock, like the transformation.
render
}
unbind shader
Note: You don't need to unbind/bind shader each frame if you only use one shader. Same goes for VAO's and every other OpenGL object as well. So the binding will also stay over each rendering cycle.
Hope this will help you when getting started. Altho I would recommend some tutorial that might have a bit more context to it.
I have read some books about modern OpenGL and find a method called
"Instanced Rendering", but it seems only to work with same
instances.Should I use for-loop to draw all items directly for my
case?
Another question is about memory.Should I create an VBO for each
frame, since the number of items is always changing?
These both depend on the amount of bullets you plan on having. If you think you will have less than a thousand bullets, you can almost certainly push all of them to a VBO each frame and upload and your end users will not notice. If you plan on some obscene amount, then don't do this.
I would say that you should write everything each frame because it's the simplest to do right now, and if you start noticing performance issues then you need to look into instancing or some other method. When you get to "later" you should be more comfortable with OpenGL and find out ways to optimize it that won't be over your head (not saying it is over your head right now, but more experience can only help make it less complex later on).
Culling bullets not on the screen either should be on your radar.
If you plan on having a ridiculous amount of bullets on screen, then you should say so and we can talk about more advanced methods, however my guess is that if you ever reach that limit on today's hardware then you have a large ambitious game with a zoomed out camera and a significant amount of entities on screen, or you are zoomed up and likely have a mess on your screen anyways.
20 objects is nothing. Your program will be plenty fast no matter how you draw them.
When you have 10000 objects, then you'll want to ask for an efficient way.
Until then, draw them whichever way is most convenient. This probably means a separate draw call per object.
I have written an entity component system for my game (C++). I have then refactored my render system to work with Entities / RenderComponents rather than some virtual drawable interface. Their are some classes for which I don't think it makes too much sense to force them to be a component. One of those classes is the map.
My map class consists of a tiled terrain class and some other data (not important). The tiled terrain class manages multiple layers in form of (what is at the moment) the TiledTerrainLayer class. Before refactoring the render system I simply inherited from Drawable and Transformable to enable this class to be drawn by the render system. Now it is required to be an entity with at least a TransformComponent and some RenderComponent.
Now, the TiledTerrainLayerRenderComponent should really only own the vertices and a reference of the texture and maybe a flag for whether it has been created yet. The TiledTerrainComponent would then own the list of tile indecies as well as tile and map size.
Now my problem is that when I set a tile (using something like a SetTile(size_t tileIndex, const Position & pos) method, I also have to update texture coordinates of the vertex array.
I am generally fine with one component requiring another component. For example the SpriteRenderComponent requires a TransformComponent and I am also fine with one component accessing the information of another. E.g. the GetBoundingBox() method uses the position of the transform component.
What I want to avoid is two components 'cross-referencing' each other like it would be the case with the TiledTerrainComponent (TTC) and TiledTerrainRenderComponent. (TTRC) (The TTRC gets the TTC's tileIndexList to create itself and the TTC calls the TTRC's UpdateVertices() method when its SetTile() method is called.
Lastly, I am aware that components should mainly be data. I have only added methods that directly get or modify that data such as SetTile() or GetTexture(). Would a system be viable in the case described above and if yes how would it look like?
It sounds like all you need here is a Dirty Flag.
When you change tile index, size, or other properties on your Tiled Terrain, you do not immediately phone the Tiled Renderer to update its vertices (after all, you might have many tile updates yet to come this frame — it could be wasteful to recalculate your vertices every time)
Instead, the Tiled Terrain renderer just sets its internal hasBeenModifiedSinceLastUse flag to true. It doesn't need to know about the Renderer at all.
Next, when updating your Tiled Renderer just prior to drawing, you have it ask its Tiled Terrain whether it's been updated since the last draw (you could even query a list of updates if you want to target the changes). If so, you update the vertices in one big batch, for better code & data locality.
In the process, you reset the modified flag so that if there are no updates on subsequent frames you can re-use the last generated set of vertices as-is.
Now your dependency points only one way — the renderer depends on the tile data, but the tile data has no knowledge of the rendering apart from maintaining its flag.
I am working on a 2d game. Imagine a XY plane and you are a character. As your character walks, the rest of the scene comes into view.
Imagine that the XY plane is quite large and there are other characters outside of your current view.
Here is my question, with opengl, if those objects aren't rendered outside of the current view, do they eat up processing time?
Also, what are some approaches to avoid having parts of the scene rendered that aren't in view. If I have a cube that is 1000 units away from my current position, I don't want that object rendered. How could I have opengl not render that.
I guess the easiest approaches is to calculate the position and then not draw that cube/object if it is too far away.
OpenGL faq on "Clipping, Culling and Visibility Testing" says this:
OpenGL provides no direct support for determining whether a given primitive will be visible in a scene for a given viewpoint. At worst, an application will need to perform these tests manually. The previous question contains information on how to do this.
Go ahead and read the rest of that link, it's all relevant.
If you've set up your scene graph correctly objects outside your field of view should be culled early on in the display pipeline. It will require a box check in your code to verify that the object is invisible, so there will be some processing overhead (but not much).
If you organise your objects into a sensible hierarchy then you could cull large sections of the scene with only one box check.
Typically your application must perform these optimisations - OpenGL is literally just the rendering part, and doesn't perform object management or anything like that. If you pass in data for something invisible it still has to transform the relevant coordinates into view space before it can determine that it's entirely off-screen or beyond one of your clip planes.
There are several ways of culling invisible objects from the pipeline. Checking if an object is behind the camera is probably the easiest and cheapest check to perform since you can reject half your data set on average with a simple calculation per object. It's not much harder to perform the same sort of test against the actual view frustrum to reject everything that isn't at all visible.
Obviously in a complex game you won't want to have to do this for every tiny object, so it's typical to group them, either hierarchically (eg. you wouldn't render a gun if you've already determined that you're not rendering the character that holds it), spatially (eg. dividing the world up into a grid/quadtree/octree and rejecting any object that you know is within a zone that you have already determined is currently invisible), or more commonly a combination of both.
"the only winning move is not to play"
Every glVertex etc is going to be a performance hit regardless of whether it ultimately gets rendered on your screen. The only way to get around that is to not draw (i.e. cull) objects which wont ever be rendered anyways.
most common method is to have a viewing frustum tied to your camera. Couple that with an octtree or quadtree depending on whether your game is 3d/2d so you dont need to check every single game object against the frustum.
The underlying driver may do some culling behind the scenes, but you can't depend on that since it's not part of the OpenGL standard. Maybe your computer's driver does it, but maybe someone else's (who might run your game) doesn't. It's best for you do to your own culling.
I'm having a rough time trying to set up this behavior in my program.
Basically, I want it that when a the user presses the "a" key a new sphere is displayed on the screen.
How can you do that?
I would probably do it by simply having some kind of data structure (array, linked list, whatever) holding the current "scene". Initially this is empty. Then when the event occurs, you create some kind of representation of the new desired geometry, and add that to the list.
On each frame, you clear the screen, and go through the data structure, mapping each representation into a suitble set of OpenGL commands. This is really standard.
The data structure is often referred to as a scene graph, it is often in the form of a tree or graph, where geometry can have child-geometries and so on.
If you're using the GLuT library (which is pretty standard), you can take advantage of its automatic primitive generation functions, like glutSolidSphere. You can find the API docs here. Take a look at section 11, 'Geometric Object Rendering'.
As unwind suggested, your program could keep some sort of list, but of the parameters for each primitive, rather than the actual geometry. In the case of the sphere, this would be position/radius/slices. You can then use the GLuT functions to easily draw the objects. Obviously this limits you to what GLuT can draw, but that's usually fine for simple cases.
Without some more details of what environment you are using it's difficult to be specific, but a few of pointers to things that can easily go wrong when setting up OpenGL
Make sure you have the camera set up to look at point you are drawing the sphere. This can be surprisingly hard, and the simplest approach is to implement glutLookAt from the OpenGL Utility Toolkit. Make sure you front and back planes are set to sensible values.
Turn off backface culling, at least to start with. Sure with production code backface culling gives you a quick performance gain, but it's remarkably easy to set up normals incorrectly on an object and not see it because you're looking at the invisible face
Remember to call glFlush to make sure that all commands are executed. Drawing to the back buffer then failing to call glSwapBuffers is also a common mistake.
Occasionally you can run into issues with buffer formats - although if you copy from sample code that works on your system this is less likely to be a problem.
Graphics coding tends to be quite straightforward to debug once you have the basic environment correct because the output is visual, but setting up the rendering environment on a new system can always be a bit tricky until you have that first cube or sphere rendered. I would recommend obtaining a sample or template and modifying that to start with rather than trying to set up the rendering window from scratch. Using GLUT to check out first drafts of OpenGL calls is good technique too.