How can I create a buffer in (video) memory to draw to using OpenGL?

Question

OpenGL uses two buffers, one is used to display on the screen, and the other is used to do rendering. They are swapped to avoid flickering. (Double buffering.)

Is it possible to create another 'buffer' in (I assume video memory), so that drawing can be done elsewhere. The reason I ask is that I have several SFML Windows, and I want to be able to instruct OpenGL to draw to an independent buffer for each of them. Currently I have no control over the rendering buffer. There is one for EDIT: ALL (not each) window. Once you call window.Display(), the contents of this buffer are copied to another buffer which appears inside a window. (I think that's how it works.)

Answer 1

The term you're looking for is "off-screen rendering". There are two methods to do this with OpenGL.

The one is by using a dedicated off-screen drawable provided by the underlying graphics layer of the operating system. This is called a PBuffer. A PBuffer can be used very much like a window, that's not mapped to the screen. PBuffers were the first robust method to implement off-screen rendering using OpenGL; they were introduced in 1998. Since PBuffers are fully featured drawables a OpenGL context can be attached to them.

The other method is using an off-screen render target provided by OpenGL itself and not by the operating system. This is called a Framebuffer Object. FBOs require a fully functional OpenGL context to work. But FBOs can not provide the drawable a OpenGL context requires to be attached to, to be functional. So the main use for FBOs is to render intermediate pictures to them, that are later used when rendering on screen visible pictures. Luckily for an FBO to work, the drawable the OpenGL context is bound to may be hidden. So you can use a regular window that's hidden from the user can be used.

If your desire is pure off-screen rendering, a PBuffer still is a very viable option, especially on GLX/X11 (Linux) where they're immediately available without having to tinker with extensions.

Answer 2

查看帧缓冲区对象（FBO） 。

Answer 3

If you have a third buffer you lose the value of double buffering. Double buffer works because you are changing the pointer to the pixel array sent to the display device. If you include the third buffer you'll have to copy into each buffer.

I haven't worked with OpenGL in a while but wouldn't it serve better to render into a texture (bitmap). This lets each implementation of OpenGL choose how it want's to get that bitmap from memory into the video buffer for the appropriate region of the screen.