opengl - rendering pipeline and transformations -

So I was studying the rendering pipeline, especially the change, when I came to this figure:

I believe the path of homogeneous clip space from camera space is t And L (or at the top) shader if we are using one), but I'm not sure that the calculation of light is done after or after the launch (e I Andari that they can be first directed to the world) must be divided by us w to pass the NDC. Is it still considered between T & L calculation or is it related to any other phase? Lastly, the last conversion, viewport / screen coordinates from NDS: it is fine before clipping and rasterization, am I right?

I do not know how you expand the word "T & L" from Programmable Pillments Given the object space, going from clip to space to clip space is entirely the work of the programmer, and it is done in the peak shield, as any intermediate is defining the space as the place of the world space or the eyes .

> Perspective segmentation, clipping, and viewport transfers are still defined in the middle of the peak shader and before dynamics (functions related to rasterization in relation to pixel raster of framebuffers) are determined. Note that there are also other stages like the primitive assembly.

But I'm not sure that the calculation of light is done after or after the launch (honestly I think the world is already synced).

You have to differentiate things here: in which space is calculated in the light, and in which pipeline phase. As space: Traditionally, light was calculated in the eyes. But with programmable pipeline, you can do this in any place you can come up with. Usually, the eyes or the world space is used. Generally, you do not want to clip it in space because the perspective will change the angle and distort the angles.

As a pipeline phase: The classic version was Gaurad Shading or "Per-Vertical Light", and it was done in the top processing phase (hence "End L" for that phase "End L") Was there. With modern GPUs, lights are usually per piece / pixels - in pieces / pixel shaders that are invented after the rasterization. But still it is used for some non-malformed space such as eye or world space calculation.