Why AR Needs High Brightness

The idea of Augmented Reality (AR) is widespread in science fiction, and with good reason: it is a fantastic concept! Integrating the virtual into the real world brings many advantages, from professional use cases to personal use. This new reality is closer than many believe, with the main challenge remaining the display unit's brightness.

VitreaLab GmbH
7 min readMay 11, 2023

After all, even the best rendering and data connections are useless if the shown image is bleached out by the sun while taking a stroll. Unlike enclosed and artificially lit living rooms, the outside world has an enormous dynamic range, going from highly dark (<1 nit) to very bright (> 100,000 nits). Current display technology has a hard time competing with the bright side of this spectrum, as anyone that has tried to use a laptop outside can testify. A typical laptop screen manages to output just ~300 nits, while a white wall behind it illuminated by the sun will be 30x brighter (10,000 nits). While laptops are mainly used indoors, the opposite is true for AR, turning this from a nuisance into a serious problem.

The 10,000 nits target

When using AR glasses, we would like to see the content shown, even in bright conditions. In particular, in professional environments readability under all circumstances is a must; just imagine first responders not finding the victim of an accident because the display is too dim!

Stock image illustrating the challenge of readability in bright environments.

Luckily, the content shown by AR glasses will be symbols and text in most situations, meaning that the contrast required for readability is low. Unlike complex figures or images, which need high contrast to be read, we can still decipher text if the contrast towards the background is only 3:1. Still, given that the ambient illuminance on a sunny day is 3,000 nits, this will require at least 10,000 nits of brightness (more details here).

An important side note is that this brightness should be delivered within an acceptable power and thermal budget, which for AR glasses is very small (<1 W per eye).

AR glasses consist of two main components: the projector and the combiner. The projector generates the image and the combiner overlays it on the view of the real world.

How AR glasses work

In AR glasses, two components play together to bring the image to the eye: the projector and the waveguide combiner. Here we will focus exclusively on combiners that feature pupil replication as they are thought to be the best technology for enabling glasses-like form factors.

Two types of combiners are available to do this job: diffractive (e.g., Dispelix, Waveoptics, Cellid, Lochn Optics) and reflective combiners (e.g., Lumus, Oorym, Lochn Optics). The key specification we are interested in this article is the brightness that they can deliver to the viewer, i.e., how many of the photons emitted by the projector will they successfully guide towards the eye.

The combiner specification we can use to calculate the brightness at the eye is nits per lumens. Lumens represent the number of photons generated by the projector and coupled into the input aperture of the combiner, and nits are the brightness the eye perceives when the photons leave the combiner. Of course, a significant factor to consider is the field of view because if the photons have to be distributed over a larger angle, the resulting brightness will be reduced. For example, while at 30° FOV, a brightness of 1,000 nits can be generated by injecting 1 lm, at 50° FOV, it will only yield 300 nits. This is a consequence of growing etendue, and increasing FOV will always reduce brightness.

Typical waveguide-combiner layout with a small in-coupling grating where the projector injects the image and a large out-coupling grating where the light is emitted towards the eye.

Taking as reference a Dispelix waveguide combiner with 30° FOV, a brightness of 310 nits can be delivered for every lumen injected. Achieving 10,000 nits would therefore require a projector capable of providing 32 lumens into the input pupil of the combiner. Similarly, a Cellid waveguide can achieve ~450 nits/lm (though at 60° FOV), requiring 22 lm to achieve 10,000 nits.

Laser LCoS for high brightness

Achieving such high brightness with LED LCoS projector is very challenging, and the brightest available projector today delivers 2–3 lm, a factor 10x of the targeted 20–30 lm required for 10,000 nits. Increasing the brightness further is challenging, as LEDs are a mature technology, meaning that significant increases in brightness are unlikely. The situation for microLED projector is similar, with a brightness of just ~1.7 lm shown to date.

Left: LEDs for illuminating an LCoS have a typical size of 1x1 mm, and their brightness is limited by their size. Right: Quantum Light Chip powered by three sets of RGB laser diodes (only red circuit shown). The laser diodes are very small and additional diodes can be added without changing the emission area.

This is why VitreaLab has replaced the LEDs with the Quantum Light Chip, a laser light-emitting device. Unlike LEDs, which are “area emitters,” laser diodes have a tiny emission point. This means that the light of several laser diodes can be combined to achieve very high brightness. While the brightness of an LED is fundamentally limited by its area, this is not the case for the Quantum Light Chip, which is limited only by heat-dissipation and power budget. Note that the size of the LED is fixed by the etendue of the projector, therefore using a larger LED will not work.

Integrating the Quantum Light Chip into an LCoS projector requires a working despeckling technique, but apart from this, the projector remains virtually unchanged compared to the LED counterpart.

Required Optical Flux A typical LCoS projector has an optical transmission from the light source to the exit pupil of ~50% due to multiple surfaces in the beam path, LCoS reflectivity, and polarizing optics transmission. Achieving 30 lm at the exit pupil, therefore, requires 60 lm to be emitted by the light source, in this case, the Quantum Light Chip. But this is only the average luminous flux. In practice, LCoS in AR will use field-sequential color to achieve color images, meaning that red, green, and blue illumination are cycled, and each color is ON only 33% of the time. The only other option would be to use an X-Cube with three LCoS, which is expensive to mass-produce and increases the projector size. Therefore the necessary peak power delivered by the Quantum Light Chip is 180 lm. Finally, given that the Quantum Light Chip has an optical efficiency of ~70%, the required peak flux from the laser diodes is 260 lm, and the average flux is 86 lm.

Laser diodes and efficiency Generating 260 lm of white light from red, green, and blue laser diodes require an optical power of 210 mW for blue, 450 mW for green, and 590 mW for red (at 445 nm, 515 nm, and 638 nm central wavelengths). Unfortunately, the current generation of laser diodes is not very efficient. For example, green diodes have a wall-plug efficiency below 10%, and the overall luminous efficiency of the red, green, and blue laser diodes is only ~33 lm/W for white light. This compares very unfavorable to LEDs (>200 lm/W) which have seen a tremendous increase in efficiency due to their widespread use over the past two decades. This low efficiency will hopefully improve soon!

But what lasers lack in wall-plug efficiency, they make up with their optical efficiency due to their tiny etendue. A laser LCoS projector based on the Quantum Light Chip can achieve 12 lm/W — the highest efficiency of all projectors presented to date!

Power Budget and Local Illumination Technique (LIT)

To deliver the full 10,000 nits to the viewer’s eye, the Quantum Light Chip has to generate an average optical flux of 86 lm, which will require 2.6W of electrical power per eye. This power is significantly more than the target of 1W! Since we do not want to wait for laser diode efficiencies to increase, another solution has to be found.

Left: In many cases, only a small part of the display area is used, showing the user some navigation information. Using Local Illumination Technique (LIT) most of the display can be switched off. Middle: Illustration of a Quantum Light Chip having three illumination zones. Right: If fewer pixels are active the display power consumption can be reduced using Local Illumination Technique. This is not possible for normal LED illumination.

An easy way forward is to use the low image content level typical of AR applications to reduce power consumption. In most scenarios, only 10–30% of the pixels will be ON, suggesting that there is potential to lower consumption by turning OFF unused pixels. LCoS is a reflective display technology that requires illumination by an external light source, which is the dominant power consumer as it generates all the light. Therefore, switching OFF certain pixels means turning OFF the illumination of those pixels.

This is easily possible with the Quantum Light Chip: the illumination can be segmented into different zones that allow for Local Illumination Technique (LIT) — the local dimming for AR glasses. As can be seen below, this technique leads to a large reduction in energy consumption.

In use cases where just 10% of the pixels are on, a ~70% power reduction can be achieved, bringing us safely below the 1W power target even when using the full brightness of 10,000 nits!

If you want to get in touch and learn more about our journey, contact us at office@vitrealab.com.

--

--