Aliasing is an effect that occurs when a high-frequency signal is undersampled and, as a result, appears erroneously as a lower-frequency signal in the sampled version. Although aliasing can occur in different contexts, such as signal processing and audio, it is especially well known in the field of computer graphics.
In graphics, aliasing manifests itself visually as “jaggies” or “jagged” edges on rendered objects, especially on slanted or curved edges. It is the result of trying to render fine details with a limited number of pixels. For example, when attempting to render a diagonal line on a pixel display, the line appears as a series of steps instead of a smooth line.
On the other hand,“anti-aliasing” (often abbreviated as “AA“) is a technique to counteract aliasing and reduce the “staircase” or “jaggies” effect that can appear on the edges of rendered objects. These jagged edges are the result of the discrete nature of pixels on a screen and the representation of slanted or curved edges in that pixel mesh.
The goal of anti-aliasing is to soften these jagged edges by introducing pixels of intermediate hues between the object’s edge colors and the background, creating a more gradual transition and, therefore, a visually smoother and more pleasing appearance
Types of Aliasing
As we have mentioned, aliasing is an effect that occurs in different contexts. Therefore we can distinguish different types of aliasing:
- Spatial aliasing: occurs when the details of an image exceed the resolution of the screen. It is the type most commonly associated with “jagged edges” in computer graphics.
- Temporal aliasing: Appears in video sequences or animations when an object moves quickly and appears to jump from one place to another due to the low sample rate (or low frame rate).
- Audio aliasing: Appears when an audio signal is sampled at a lower rate than necessary. The Nyquist sampling theorem states that a signal must be sampled at least twice its highest frequency to be reproduced correctly. If this criterion is not met, high frequencies can “alias” and appear as lower frequencies, distorting the sound.
There are a large number of anti-aliasing techniques, here are just a few of the best known:
- MSAA (Multisample Anti-Aliasing): This is one of the most common techniques and works by taking multiple samples (or “samples”) within each pixel and calculating the final color of the pixel based on an average of these samples. For example, in 4x MSAA, 4 samples are taken per pixel.
- SSAA (Supersample Anti-Aliasing): This technique involves rendering the scene at a much higher resolution than the desired final resolution and then downsampling it. This can produce excellent results, but is computationally expensive.
- FXAA (Fast Approximate Anti-Aliasing): This is a post-processing technique that looks for jagged edges in the rendered image and smooths them out. It is less accurate than MSAA or SSAA, but it is also less performance intensive.
- TXAA (Temporal Anti-Aliasing): Combines MSAA with a temporal filter to reduce jitter in moving scenes. It is a more advanced technique and is mainly used in video games and real-time applications.
- MLAA (Morphological Anti-Aliasing): Based on post-processing, this technique looks for jagged edge patterns and smooths them. It is similar in concept to FXAA, but uses different techniques to detect and smooth edges.
If you want to know other terms and concepts related to design, don’t miss our definitive glossary of graphic design, where you will find everything you need to know to become an expert in the field.