Tomas Sluka | Creal: Why do reflective waveguides waste 99% of their light, and can holographic combiners save AR?
00:17:00 - 00:19:38
Other snippets from this talk
Summary of the clip:
Why do reflective waveguides waste 99% of their light, and can holographic combiners save AR?
The commercial AR display market has largely standardized on surface relief and holographic waveguides because they allow for generous eye-box tolerances through pupil replication. However, this optical architecture forces a severe design trade-off, sacrificing light efficiency, polarization integrity, and physical volume. Traditional waveguides output flat images that cannot preserve depth information, and they operate with extremely low optical efficiency, often wasting over 99% of the injected light before it reaches the eye.
Holographic combiners offer a highly efficient alternative by reflecting light directly from a specialized film laminated onto a standard prescription lens. Historically, the fatal flaw of holographic combiners was their prohibitively small eye-box, which made the virtual image disappear with even slight glasses movement. To bypass this limitation, a multi-pupil approach must be employed, projecting multiple small exit pupils to synthesize a large, robust overall eye-box.
By integrating a sequential light field projector with a multi-pupil holographic combiner, AR displays can achieve over 100 times the light efficiency of waveguide-based systems. This configuration delivers highly transparent, prescription-compatible lenses with no outward light leakage or distracting rainbow artifacts, proving that high-efficiency volumetric AR can fit within a standard glasses form factor.
In this short video, you can learn:
* The critical optical trade-offs between mainstream waveguides and highly efficient holographic combiners.
* How pupil replication in waveguides destroys light efficiency and locks the virtual image to a flat focal plane.
* The multi-pupil array technique used to expand the historically restrictive eye-box of holographic optical elements.
📋 **Clip Abstract** This clip compares the optical performance of mainstream waveguides against advanced holographic combiners. It explains how Creal’s multi-pupil design overcomes historical eye-box limitations to achieve 100x better light efficiency and integrated vision correction.
🔗 Link in comments 👇
#HolographicCombiners, #MultiPupilOptics, #LightFieldProjection, #WaveguideOptics, #NearEyeDisplays, #SpatialComputing
This is a highlight of the presentation:
More Highlights from the same talk.
03:42 - 05:29
Why does stabilizing an AR image require spare pixels, high frame rate, short persistence and high peak brightness?
Why does stabilizing an AR image require spare pixels, high frame rate, short persistence and high peak brightness?
Moving the image to counteract head movement stabilizes it, so part of the problem is solved. But a new one appears: the small display becomes too small because spare pixels must surround the image, giving it somewhere to move. Even for just an arrow or a letter, the image must move on the display for stabilization. Increasing field of view can improve the limited-size problem, yielding enough pixels to move the image and keep it stable.
The first fix means quite high frame rate is needed. Not only frame rate: high frame rate, short image persistence and high peak brightness in notoriously low efficiency optics. If high frame rate is available, the image will stabilize. For spare pixel area and field of view, just put more pixels, basically creating more, smaller and brighter pixels. Many talks at this conference indicate this is a critical problem.
The chain is clear: image movement for stabilization creates a spare-pixel area around the image. That area forces either a larger field of view or more pixels. More pixels then mean smaller brighter pixels. High frame rate, short image persistence and high peak brightness must all coexist with low efficiency optics. The result is a critical problem, as many talks at the conference indicate.
In this short video, you can learn:
* Moving the image to counteract head movement stabilizes it but requires spare pixels around the image on the display.
* Increasing field of view supplies enough pixels for image movement, yet demands high frame rate, short image persistence and high peak brightness in low efficiency optics.
* The remaining approach is more, smaller and brighter pixels, which the conference talks identify as a critical problem.
📋 **Clip Abstract** Moving the image to counteract head movement stabilizes the displayed image but makes a small display too small because spare pixels must surround the image, even for an arrow or letter. The response is higher field of view and more pixels, but this demands high frame rate, short image persistence, high peak brightness in notoriously low efficiency optics, and more smaller, brighter pixels.
About the speaker:
* Speaker: Tomas Sluka
* Company: CREAL
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#SparePixelArea, #ShortImagePersistence, #HighPeakBrightness, #LowEfficiencyOptics, #AROptics, #DisplayArchitecture
08:29 - 10:49
How can a fast FLCOS display create a 3K image from a 1K footprint?
How can a fast FLCOS display create a 3K image from a 1K footprint?
The design work started from the field-of-view direction: the display is fast enough to create multiple sections of the field of view simultaneously. It can show up to seven colored sections without flicker, and technically nine. A near-term AR use case uses the simple mode most of the time, but stabilized: the image hangs in the air, not shaking. Notifications can appear left and right within a 70-degree field of view, described as 25, 25, 25.
When you are plugged into power and compute, the same device can fill the full field of view, either short term or longer. The approach creates a world-locked, body-locked experience: the image hangs in the air rather than shaking. It can bring a wide field of view from around 30 times smaller footprint than brute force would require. That means a 3K display from a 1K footprint.
For FLCOS, which has smaller problems with making the pixel smaller because the limit is kind of 1.5 micron—at least this was tested—the approach can do 2.8 micron. The reason is simpler electronics, a thinner LC layer, and smaller cross-talk between pixels running at high speed. So the display can extend that sound barrier. We did not cross the sound barrier; we changed the environment.
In this short video, you can learn:
* The FLCOS display can create up to seven colored field-of-view sections without flicker, and technically nine.
* A 70-degree field of view can show notifications left and right, described as 25, 25, 25, while the same device fills the full field of view when plugged into power and compute.
* The approach delivers a 3K display from a 1K footprint, around 30 times smaller than brute force, using 2.8-micron pixels versus a tested 1.5-micron limit.
📋 **Clip Abstract** The clip explains using a fast FLCOS display to create multiple field-of-view sections simultaneously, up to seven colored without flicker and technically nine. It describes a stabilized world-locked AR experience, a 70-degree field of view with 25,25,25 notifications, a 3K display from a 1K footprint, and 2.8-micron pixels enabled by simpler electronics, a thinner LC layer, and smaller cross-talk.
About the speaker:
* Speaker: Tomas Sluka
* Company: CREAL
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#FLCOS, #PixelCrossTalk, #ThinnerLCLayer, #3KFrom1KFootprint, #ARDisplays, #NearEyeDisplays
06:33 - 08:15
How does an 8 kHz SLCoS modulator plus MEMS stabilize AR images with essentially no compute overhead?
How does an 8 kHz SLCoS modulator plus MEMS stabilize AR images with essentially no compute overhead?
The approach is a light-source modulator system: a laser with an SLCoS high frame rate liquid crystal modulator running at eight kilohertz. It can be made or worked on a 2.8 micron pixel pitch version and is combined with a MEMS. These components were already developed and used for a light field system, which works beautifully for vision care and medical application. They are unique because of the protocol needed to drive them.
First, it can do what other displays do: create an image in 25 degrees field of view. At 1K by 1K, that gives around 40 PPD, matching most of the resolution of waveguides or the NTF of the waveguides. On top of that, it can move the image a bit left, a bit right. There is an alert or indication. You can use it to stabilize against movement with practically any compute overhead.
If we forget about color corrections, no compute is required at all; it can be analog in fact. That is how the image then looks. The components first create the image, then shift it left or right for stabilization with practically any compute overhead. They are already developed and used for a light field system.
In this short video, you can learn:
* An eight kilohertz liquid crystal modulator with 2.8 micron pixel pitch is combined with MEMS for this light-source modulator system.
* A 1K by 1K image in a 25 degree field of view yields around 40 PPD, matching most waveguide resolution or NTF.
* The same system can shift the image left or right to stabilize against movement with practically any compute overhead, and without color corrections no compute is required.
📋 **Clip Abstract** The clip describes a light-source modulator system: an eight kilohertz liquid crystal modulator at 2.8 micron pixel pitch, combined with MEMS and already used for a light field system in vision care and medical applications. It creates a 1K by 1K image in a 25 degree field of view at around 40 PPD, then shifts the image left or right to stabilize against movement with practically any compute overhead, and without color corrections no compute is required.
About the speaker:
* Speaker: Tomas Sluka
* Company: CREAL
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#SLCoS, #MEMS, #LiquidCrystalModulator, #ZeroComputeStabilization, #ARDisplays, #NearEyeDisplays
Other talks by Tomas Sluka
- Vision Care at the core of AR · MicroLED Connect + AR/VR Connect
- Advances in light field display · AR, VR, and MR Vision Systems: Innovations, Promising Start-Ups, Future Roadmap
More highlights from Tomas Sluka
- ▶ Why does stabilizing an AR image require spare pixels, high frame rate, short persistence and high peak brightness? · MicroLED Connect 2026
- ▶ How does an 8 kHz SLCoS modulator plus MEMS stabilize AR images with essentially no compute overhead? · MicroLED Connect 2026
- ▶ How can a fast FLCOS display create a 3K image from a 1K footprint? · MicroLED Connect 2026
- ▶ What optical trade-offs occur when using a 0.8mm mirror and coherent lasers in waveguide-based AR displays? · MicroLED Connect 2026
- ▶ Why is the optical "flatness" of Apple Vision Pro and HoloLens making users physically sick? · AR/VR Connect 2025
- ▶ Can holographic combiners completely replace waveguides to deliver a 20x jump in optical efficiency? · AR/VR Connect 2025
- ▶ Why are microLEDs hitting a physics wall, and how can spatial multiplexing bypass it? · AR/VR Connect 2025
- ▶ Why are modern AR glasses chemically locked out of showing objects within arm's reach? · AR, VR, and MR Vision Systems 2023: Innovations, Promising Start-Ups, Future Roadmap
- ▶ Can we trick the human eye's natural focus lens using only 32 spatial viewpoints? · AR, VR, and MR Vision Systems 2023: Innovations, Promising Start-Ups, Future Roadmap

