Fundamentals of AR Optics: How FOV And Eyebox Influence Your Visual Experience

FOV (Field of View) and Eyebox are arguably the two most fundamental parameters in the optical systems of AR glasses; however, many people are not entirely clear on how exactly they influence the virtual imagery we perceive. In this post, we will provide a brief overview of these two core parameters—explaining the principles behind them, the factors that affect them, and their overall impact on the user experience.

FOV (Field of View) is a parameter that describes the size of the virtual imagery as perceived by the user.

The field of view (FOV) is determined by both the optical engine and the waveguide.

Assuming the entire image from the optical engine is displayed, if a waveguide with a 30° FOV is paired with an optical engine that has a 20° FOV, the FOV perceived by the user will be 20°. Conversely, if an optical engine with a 30° FOV is paired with a waveguide that has a 20° FOV, the FOV perceived by the user will be slightly larger than 20°; however, any part of the image extending beyond the 20° range may be cut off.

A common question arises: is the FOV the same as the image size?

The answer is no—they are not equivalent. The field of view (FOV) of an optical system (comprising an optical engine and a waveguide) is determined by two parameters: the FOV itself and the virtual image distance. If the FOV remains constant, a greater virtual image distance results in a perceived larger image. For instance, consider a fixed projector: its FOV is constant, yet moving the projection screen closer to the projector makes the image smaller, while moving it further away makes the image larger.

It is also worth noting that the FOV of an optical engine can often be dynamically adjusted by controlling the resolution of the image rendered via software.

Take, for example, the 0.15cc monochrome green optical engine—the most widely used type in the industry—which features a standard FOV of 30° and a resolution of 640×480. If the center point is held constant and the resolution of the displayed image is reduced, the effective FOV of the optical engine will shrink accordingly; this is because, with the optical engine’s lens remaining unchanged, the number of pixels per degree of FOV stays constant.

This leads to another question: when using the same MicroLED display panel, is it possible to pair it with waveguides designed for different FOVs without sacrificing pixel count?

The answer is yes; it simply requires a redesign of the optical engine lens.

However, since re-developing an optical lens entails high costs and a long development cycle, most manufacturers currently continue to use standard 0.15cc optical engines, achieving product differentiation by reducing the number of horizontal and vertical pixels. As for the benefits of reducing pixel counts—that is a topic for another time.

The following are the common resolutions and equivalent fields of view (FOV) currently achieved by scaling down from a standard 0.15cc optical engine:

"Eyebox" is a parameter that describes the effective viewing area within which the entire image can be seen. For those unfamiliar with such products, the concept of an "eyebox" may seem somewhat abstract; it is not a tangible, physical object, but rather an invisible region existing in space.

Here is how I usually explain it: as long as your eyes are positioned within the "eyebox" area, the virtual image you see remains complete. Imagine you are wearing AR glasses and shift them upward; the virtual image will stay intact within a certain range. However, once you reach a specific boundary, the virtual image will begin to disappear—starting from the bottom edge—which indicates that you have reached the limit of the eyebox. This applies across all four directions—up, down, left, and right—collectively forming a rectangular region known as the "eyebox."

The size of the eyebox is determined solely by the grating design of the waveguide itself; essentially, the grating region of the waveguide functions to replicate the image projected by the optical engine.

Suppose the image projected by the optical engine reaches the waveguide’s input-coupling region as a circular area with a diameter of 4 mm. To see the entire displayed image, your eye must be positioned precisely within this small area. The waveguide acts to replicate this circular area along both the X and Y axes. If it is replicated three times along the X-axis and twice along the Y-axis, the output from the waveguide becomes an array of 3×2 circles, each 4 mm in diameter. In this case, we can roughly consider the eyebox of the waveguide to be a rectangular area measuring 12×8 mm.

It is clear what must be considered when defining the "eyebox" parameter: a larger area ensures that the full image remains visible even if the headset shifts or moves in front of the user’s eyes. However, since the capabilities of a waveguide are inherently limited, increasing the number of replicated exit pupils makes it progressively more difficult to maintain consistent performance and uniformity across each region.

Therefore, when designing AR headsets, it is essential to strike a balance: define an appropriately sized eyebox to safeguard optical performance, while simultaneously optimizing the device’s industrial design and structural configuration to keep it positioned stably relative to the user’s eyes—thereby minimizing the risk of the image being cut off or lost.

Source: AR Optical Product Notes