Introduction

Ron Mertens of OLED-Info, MicroLED-Info, the MicroLED Association, and one of the organizers of MicroLED & AR/VR Connect asked for my input on “Why is there no ‘simple’ solution, where we place a transparent display in front of the user [for AR]?” Or more simply, why aren’t ‘transparent’ OLEDs/MicroLEDs being used in optical-see-through (OST) AR headsets?

I ended up writing a rather long response to Ron, and he wanted to publish it on his OLED-Info Pro (paid monthly or yearly subscription (link: https://www.oled-info.com/karl-guttag-transparent-ar-displays-article). I agreed on the condition that I put a summary on this blog. Ron has also extended a discount for my readers who are interested in a yearly subscription.

“Transparent” Microdisplays for AR

Basic Issues

When evaluating OST AR solutions, we need to consider the following:

1. Effect on the see-through image
   • Darkening
   • Distortion and uniformity (including diffraction) of the real world
   • Artifacts such as diffraction light capture (so-called rainbow capture)
2. Effect on the virtual image
   • Uniformity in brightness and color, plus image contrast
   • Efficiency (power-in in Watts to light to the eye in nits)
3. Forward Projection (“Eye Glow”) and other adverse social effects
4. Eye Box size (how much the eye can move and still see an image)
5. Cost, weight, thickness, and overall size

I often say that the most important measure of an AR display is how it behaves when the display is off (#1 above). As will be discussed, this is not possible with a so-called “transparent” display that is near to the eye.

Summary of Article on OLED-Info

While it is possible to make a large OLED-, LCD-, or LED-based display that is mostly transparent, the same cannot be said for transparent near-to-eye displays. The fundamental differences are that the ratio of pixel pitch to pixel size is much greater with large displays, and, often overlooked, near-to-eye displays are so close to the eye that optics are required between the pixels and the eye to enable the eye to focus. Or more simply, the pixels (plus their optics) have an order-of-magnitude larger impact when they are so close to the eye.

Fraunhofer, LusoVu, and Newsight Reality have all presented papers, published patents/applications, and had booths at conferences on near-eye “transparent” displays, which I used to study the various approaches.

With optical AR, you are always balancing the effect on the real world (focus, darkness, artifacts, etc.) versus the effect on the virtual image. Near-to-eye transparent microdisplays have major issues that are unlikely to be solved in the foreseeable future. The major issues include:

1. Moving the focus of the display from about 2.5cm (typical display distance) to at least 25cm, and preferably about 2m, without damaging the focus and otherwise distorting the view of the real world.
2. The pixels and their associated circuitry, when placed so close to the eye, cause serious problems with viewing the real world. Not only will there be darkening, but there will be both dark spots in the user’s vision and distortion due to diffraction. What’s more, the diffraction will cause a wobbling effect of the real world as the eye moves relative to the display.
3. To make the pixels small enough to allow for some level of transparency, they will be built on semiconductor substrates. To allow for transparency, the pixels will be spaced out, resulting in larger devices. There are also extra significant process steps to make the devices transparent. All these factors combined to make an expensive component compared to a non-transparent display of the same resolution.

At first, though, you might assume they use an array of sparsely spaced individual pixels. I don’t see how it is possible to put optics at the individual-pixel level to solve the focus problem, and nobody, as far as I know, has even proposed an approach. Fraunhofer, LusoVu, and Newsight all use clusters of pixels on the order of 64 by 64 to 128 by 128 pixels, with optics over each group and spacing between clusters. The optics not only change the focus, but they also combine the image from multiple clusters to try and form a single larger image. Unfortunately, these clusters with their optics end up being on the order of 1 to 2 millimeters wide, which is significant when placed close to the eye.

Another issue with the “cluster” optics approach is that the image can only be seen as a single image if the eye is in exactly the right location in 3-D space. There is a very small eye box where the image can be seen as a single image rather than segmented (see slide from LusoVu, right, and set of stills via a video about Fraunhofer’s semi-transparent display below)

While Fraunhofer and Newsight use “conventional” micro-lens optics, LusoVu uses Holographic Optics Elements (HOEs). But whether using micro-lenses or HOEs, I don’t see how that can get around the fundamental problems of light blockage/disturbance caused by the pixel clusters with their optics. And this is before considering all the other issues of getting a good-looking virtual image to the eye.

In the full article on OLED-Info Pro (link: https://www.oled-info.com/karl-guttag-transparent-ar-displays-article), I go into more detail on the various companies’ approaches and show more references.

I will be at MicroLED and AR/VR Connect in Eindhoven, Sept. 15-17, 2026 (repeating from last time)

This is my second year attending and presenting at MicroLED and AR/VR Connect in Eindhoven, the Netherlands. I very much enjoyed this conference last year, particularly for the ability to meet people at the several networking sessions. Last year at the conference, I met my Display Skeptics YouTube partner Radu Reit.

On Day-minus-1 of the conference (September 15th) at 10 AM, I’m presenting a 1-hour Master Class on “The Latest Trends in Optical See-Through (OST) AR Hardware and Optics Technology.” The first day, September 15, is Master Classes and Tours of some Local Labs. The conference and exhibition proper are on September 16-17th.

Radu Reit is presenting as a Master Class on “Tearing down smart-glasses to understand the displays & optics driving cost, performance & user experience,” right before me (but in a different room) at 9:00 AM

This year, Radu and I will be hosting a great panel. We will discuss MicroLEDs, LCOS, Lasers, and Micro-OLED for use in AR Light Engines. The Panelist will include Bernard Kress, from Google and SPIE Fellow and past President (and formerly with Microsoft’s HoloLens), Berthold Hahn, from Meta working on MicroLEDs, and Barry Silverstein, from the Center for Extended Reality at the University of Rochester (and until recently Meta), who was a proponent for laser displays.

If you are going to MicroLED and AR/VR Connect this year, hopefully we can meet. Send an email to meet@kgontech.com.

MicroLED and AR/VR connect Discount code

The Conference is offering my readers a €150 discount if they use the discount code KarlARVR. This discount code is valid for both the “Virtual Pass” and “Hybrid Pass” for virtual and/or physical conferences, which can be attended remotely (this blog receives remuneration for the use of this code).