ETH Zurich’s David Norris turns one pixel into a two-way light tool
“Fourier pixels” can emit and measure light, enabling bidirectional screens, holograms, and new optical communication building blocks.

David Norris, a professor at ETH Zurich's Optical Materials Engineering Laboratory, led a team that developed “Fourier pixels” able to both control and analyze light. The result could unlock two-way displays, holographic systems, optical communications, and quantum information processing.
ETH Zurich professor David Norris and his team say they have figured out how to make a single pixel do two jobs at once: emit light and measure light. Traditional pixels usually specialize. In a display, pixels primarily illuminate. In a camera sensor, the pixels primarily detect. The ETH Zurich researchers’ pitch is different. Their “Fourier pixels” can create and detect properties of optical fields, including amplitude, phase, and polarization, using one multifunction element.
The core enabling move is not “just add a sensor” or “just add a laser.” It is a specific measurement and control method built around Fourier analysis. According to the Nature article “Fourier pixels for bidirectional light control,” the team measures light wave interference patterns over a metallic surface. From those interference patterns, the researchers generate Fourier pixels. Put simply, the pixels represent spatial frequency of light rather than only brightness at a point. That distinction matters because it lets the system work with the structure of the wave itself, not just how bright it looks after the fact. As ETH Zurich put it in a press release, post-doc Sander Vonk said, “Thanks to the fact that the relevant surface profiles of the pixels can be determined using Fourier analysis, we can combine the control and analysis of amplitude, phase and polarisation on a single pixel.”
For executives, the practical question is always the same: does this change what systems can do, or just make them marginally better? This one is aimed squarely at changing the architecture of the interface between light and computation. If a “pixel” can both send and sense, then the display stops being a one-way projector and starts behaving more like a communication device. That is where the source’s near-term expectation becomes important. Norris expects to put Fourier pixels into a matrix to construct more sophisticated camera displays. That phrasing matters. Camera displays are not just for showing images. They are for collecting information from what the display is interacting with, which can enable bidirectional behavior.
Now zoom out to the bigger ecosystem the source names. The researchers say the work raises the possibility of two-way screens that take and present pictures, holographic displays, optical communication systems, and quantum information processing. None of those are small upgrades. Each one changes where value is created and who owns the stack. Two-way screens imply a tight loop between what is shown and what is sensed. Holographic displays imply a need to control the wavefront in more precise ways than typical intensity-only rendering. Optical communication systems imply new degrees of freedom to modulate signals, because controlling amplitude, phase, and polarization can enrich what information can carry. Quantum information processing implies relevance to optics at the level where wave properties matter for protocols.
The method also hints at why investors and operators should care about supply chain and manufacturing realities, even if the source does not spell them out. Measuring interference patterns over a metallic surface suggests a design philosophy tied to surface profiles. The “surface profiles of the pixels” can be determined using Fourier analysis, and the same surfaces can support both control and analysis. In other words, the system is not just “two devices glued together.” It is a coupled design, which typically has consequences for yield, repeatability, and cost. If you can truly standardize the pixel profile generation process, you can potentially reduce complexity. If you cannot, you may find that performance exists but manufacturability becomes the constraint. The source does not settle that question. But it does set up the core engineering bet: you can bake a wavefront controller and a wave property analyzer into the same pixel geometry.
There is also a governance and standards angle that quietly matters for decision-makers. Optical communication and imaging technologies often face verification hurdles. Regulators in many jurisdictions are less concerned with “how clever the physics is” and more with what the system does in the field, including performance reliability and safety. The source does not mention regulators. Still, when display hardware becomes bidirectional and sensing-capable, questions about data handling, consent, and system behavior often move from technical teams into legal and compliance. A “two-way screen” that “takes and presents pictures” raises the possibility of user data capture in contexts where people may not expect cameras or sensors. Even if the early deployments are narrow, executives building product roadmaps usually need to assume privacy and policy scrutiny will follow sensing capability.
Finally, look at the competitive implications. If Fourier pixels can “create and detect” amplitude, phase, and polarization on a single pixel, then the interface between optics and computation becomes both richer and tighter. That can compress steps that imaging pipelines currently separate into distinct components, like projecting, then sensing, then reconstructing. For leaders in display technology, AR and spatial computing, and optical networking, that means fewer handoffs between subsystems could translate into lower latency or better fidelity, depending on implementation. For quantum technology stakeholders, anything that helps control and read out optical field properties with compact elements is a potential accelerant, especially if it scales beyond lab prototypes.
The takeaway for peers is simple: Norris’s team is not just improving a pixel. They are changing what “a pixel” means in the architecture of light-based systems. If Fourier pixels move from a research demonstration into matrix-based camera displays as expected, boards and product teams will need to revisit assumptions about bidirectionality, sensing integration, and what claims they can safely make about wave control and measurement. That is the strategic stake. In a world where the bottleneck is often how devices talk back to the environment, one element that can both send and measure could become a foundation layer for the next wave of optical products.
This story's Key Insights and Take-aways are locked.
Create a free account to unlock Executive Actions for one credit.
Register to UnlockAlways free for Executives Club members. Join the Club
More in Technology

Meccha Chameleon hits 10 million sales in under a month, nearly 200,000 Steam concurrents
An indie hide-and-seek spinoff turns friendsloppish joy into massive Steam traction, and it offers a business lesson.

Prompt injection turns Slack and Copilot into entry points for stolen data
OWASP and real breaches show the same flaw: LLMs cannot reliably separate instructions from data.

Firmus locks 360MW in Batam, targets $30B in Nvidia off-take deals for Indonesia build
A $5.5B Australian AI infrastructure firm signs an eight-year Nvidia deal to launch its first data center and chase massive demand commitments.

