I've read that in MSAA each pixel's final color is an interpolation of its current color and the triangles color, depending on how many samples the triangle is covering.

That means that when drawing the scene in different orders affects the "current color" I mentioned above. So if I draw the scene in different orders when using MSAA, will the final render image be affected by this?

1 minute ago, Pilpel said:

I've read that in MSAA each pixel's final color is an interpolation of its current color and the triangles color, depending on how many samples the triangle is covering.

No, it stores N colours (e.g. 2xMSAA stores 2 colours) and the "resolve" stage at the end blends them together.

If it stores N times more colours, how is it different that SSAA?

1 hour ago, Pilpel said:

If it stores N times more colours, how is it different that SSAA?

For what I know, in SSAA that color is calculated per each pixel with the image being 4 times higher resolution if we talk about 4x SSAA and then downsampled to your actual resolution.

With MSAA you render the image at your monitor resolution (so you start with 4 time less calculation compared to above) and then for each pixel of that, the same calculated color is shared for each of the 4 sub-pixels based on visibility (using depth buffer, 4 times bigger) and coverage (is the subpixel inside or outside the polygon?), and then the 4 sub-pixels(my example is 4xMSAA) are averanged togeter to get the final color, and all this check and averange is apparently less expensive than actually calculating the real lighting for 4 pixels, since you do 1/4 of the heavy calculation only once compared to SSAA.

Take it as a grain of salt, I am begginner in this.

So SSAA and MMAA have the same memory usage?

40 minutes ago, Pilpel said:

So SSAA and MMAA have the same memory usage?

I bet they do, but I would like to know the informed answer to this as well

You could just jump onto the Wiki page for multisampling 

https://en.wikipedia.org/wiki/Multisample_anti-aliasing

it has a full descriptive comparison between MSAA and SSAA and why MSAA is fairly less resource / computation heavy, it even goes into a brief comparison with FXAA etc as well

I'm not sure if I get it. Does MSAA differ from SSAA in a way that the pixel shader is ran, most of the times, fewer times per pixel?

Quote

In the extreme case where each of the multi sample locations is covered by a different triangle, a different shading computation will be performed for each location and the results then combined to give the final pixel, and the result and computational expense are the same as in the equivalent supersampled image.

Memory usage (number of elements in the color buffer) seems to be the same for both MSAA and SSAA, no?

MSAA doesn’t actually improve on supersampling in terms of rasterization complexity or memory usage. At first glance we might conclude that the only advantage of MSAA is that pixel shader costs are reduced. However this isn’t actually true, since it’s also possible to improve bandwidth usage. Recall that the pixel shader is only executed once per pixel with MSAA. As a result, the same value is often written to all N subsamples of an MSAA render target. GPU hardware is able to exploit this by sending the pixel shader value coupled with another value indicating which subsamples should be written, which acts as a form of lossless compression. With such a compression scheme the bandwidth required to fill an MSAA render target can be significantly less than it would be for the supersampling case.

20 minutes ago, GibbonThatCodes said:

... As a result, the same value is often written to all N subsamples of an MSAA render target.

This only happens inside a triangle, right?