Antialiasing is the next big thing in tech – or, at least, it’s a sensationalist way to start off an article, in spite of being pretty much meaningless and untrue. Antialiasing is a very, very powerful tool, however, one that’s often holding up the very structure of a device’s graphics, and that’s no exaggeration.
In this article, you can expect to learn what it is, why it’s needed, the different modes, its relationship with transparency and gamma correction, as well as clearing up some common misconceptions.
What is Antialiasing?
We like to point this out a lot, but even the best gaming PC is nothing without the best graphics card. The ones who’ll best understand are the brave soldiers who’ve been stuck with the struggles of using an old computer that can barely hold its own teeny-weeny weight, let alone run high-quality graphics video games.
In such a situation, you get jagged edges – blocky pixels and blurred boxes of colors as opposed to actual seamless edges of high-def images. These are called jaggies, and they occur due to the staircase effect (a confusing term, also close to something used in medical circles), which we’ll explain in very simple terms.
This is because any image is essentially made out of extremely tiny RGB dots. When an image is blown up, you can essentially see the smaller pixels that would’ve been unnoticeable at a small size. If you zoomed in on any screen or got really close to it, you could be able to make out the actual individual pixels.
Antialiasing is something that counters this. It’s a complex method of “coloring” in missing pixels when there aren’t enough. Antialiasing accomplishes in a dynamic and smart way, run by powerful computer algorithms. This way, hard edges become smoother; there are now hundreds of the “same” pixels where there were previously 10.
A key component of antialiasing is gamma correction. While the edges of a black object might look to be “half” black and white, or somewhat transparent, it’s much more complex than that. Accounting for gamma can’t be done manually, but that’s the power of antialiasing at the pixel level.
An image can be made sharper through antialiasing, especially where higher resolution (or the increased processing power required to do this) isn’t possible; in many ways, reducing the capability gap between PCs and consoles.
The Two Antialiasing Modes
While there are many antialiasing methods, they can broadly be summed up into two modes.
There are many types of antialiasing, some exclusive to certain processors. There’s supersampling anti-aliasing (SSAA), multisample anti-aliasing (MSAA), CSAA and EQAA, MLAA and FXAA, temporal anti-aliasing (TXAA), and enhanced subpixel morphological anti-aliasing (SMAA).
Largely, however, we can divide this into two modes of antialiasing: spatial antialiasing and post-process antialiasing.
Spatial antialiasing uses samples to “color in” a low-resolution image, resulting in the birth of new pixels that previously weren’t there before. Gamma correction antialiasing is adjacent to this in terms of the idea, whereas transparency antialiasing is closer to post-process antialiasing.
It sounds futuristic, but the idea is fairly simple: spatial antialiasing works by blowing up an image (full of jaggies) to a new resolution, which now has a hundred times more jaggies, but at least there’s an approximation of what colors are filling up what pixels, and how the variation is spread. Color samples are taken, and the image is brought back to its original size. Now the jaggies can be smoothed out by using the color samples to blend all the pixels together much better – infuse a new depth of colors to every pixel – averaged from the blown-up pixels with the color samples.
Post-process antialiasing, on the other hand, is a process that’s much faster, a bit more intuitive, and requires a lot less processing power as compared to spatial antialiasing.
Blurring is the key component of this method. Each rendered pixel is blurred, and then the color contrast between two pixels is compared to determine the edge of a polygon. Wherever two similar pixels are found (and hence, found to be the same polygon’s pixels), they’re further blurred, in proportion to their contrast. This softens the jaggies, essentially making it less noticeable to the naked eye that there are actively low-render low-quality visuals. Of course, wherever there’s a need for textures to be detailed and lighting to be dynamic, especially in video games, the images can become too noticeably blurry.
The purpose of this article has been to prime you with all the information you’d want to understand the need for antialiasing. For many, a natural question may now arise, one of HDR vs. antialiasing, the answer to which again lies in the need; both can be utilized at the same time, with antialiasing in the graphics card, and HDR in the game itself!