AR, VR, MR, XR - Defined

The cleanest way to hold VR, AR and MR in your head is not as three separate boxes but as points along a single line. In 1994 the researchers Paul Milgram and Fumio Kishino drew that line and called it the reality-virtuality continuum. At its left end sits the completely real environment - the room you are in right now, unmediated. At its right end sits a completely virtual environment - a world made entirely of computed pixels, with nothing of your real surroundings in it. Everything interesting lives in between, in the broad band they named mixed reality: any situation where real and computed content are combined in a single view. This one picture does more to prevent confusion than any list of definitions, because it shows the family as a spectrum of how much of what you see is real versus computed, rather than as rival products fighting for the same slot.

Reading the continuum from left to right, you pass through recognisable stations. Just in from the real end, computed content is thin and floats loosely over the real world - a label, an arrow, a dimension hovering in your view; the real world dominates. Move rightward and the computed content grows richer and more tightly bound to reality until it appears to genuinely sit in your space, aware of your walls and furniture. Keep going and the balance tips until the real world all but disappears and you are effectively inside a computed place. At the far right the real world is gone entirely. The point is that these are not hard walls but a gradient, and a given device or app occupies a region on the line rather than a single dot - the same modern headset can push you far right into a fully virtual walkthrough one moment and pull you back toward the middle to show computed furniture in your real room the next.

For a designer this continuum is the mental model to keep. When someone demonstrates an immersive tool, your first question is simply: where on this line does it sit - how much of what I am seeing is my real world, and how much is computed? That single question cuts through most of the terminology fog, and it maps directly onto design use. The far-virtual end is for experiencing an unbuilt space as if you were standing in it; the middle is for placing your design into the real world you are already standing in. Hold the line, and the labels hang off it in their proper places rather than blurring into one another.

Virtual reality is the far-right end of the continuum: a fully computed world that replaces your view of the real one. You put on an opaque headset, and the real room vanishes - two small screens, one per eye, fill your entire field of view with a stereoscopic 3D scene that updates instantly as you turn your head and move, so the computed world holds still around you as though it were really there. That convincing sense of being somewhere else is called presence, and it is VR's superpower: your body stops treating the scene as a picture and starts treating it as a place, bringing your real spatial intuition - your read of height, distance, enclosure and scale - to bear on something that does not yet exist.

The precise, distinguishing fact about VR is occlusion of the real world. You cannot see your real hands, your real desk or the real people around you unless the system deliberately draws computed stand-ins for them. This is exactly what makes VR powerful and exactly what makes it demanding. Powerful, because nothing of the messy real room distracts from the designed space - you are wholly inside the corridor, the atrium, the bedroom, at true one-to-one scale. Demanding, because you are sealed off: you can trip over real furniture, you cannot glance at your notes, and being cut off from the room is part of why headsets feel isolating and tiring, and why some users feel motion-sick when the computed motion and their inner ear disagree.

The canonical design example is the VR design review. A model of an unbuilt house is loaded into a VR headset, and the client - who has never been able to read the plans - simply stands in the living room, looks up at the real height of the double-height void, walks to the window and sees the actual view, and understands the space in seconds in a way no drawing or render achieved in months. The architect uses the same walkthrough to catch a corridor that feels tighter than the dimension suggested, or a sightline that does not work. Note the honest boundary even here: the client is understanding and judging the space, not measuring it. The felt tightness is a real design signal; the exact clear width is not read off the headset but confirmed from the verified drawings.

Cross to the other side of the continuum and the defining move reverses: the real world stays fully visible and computed content is added to it. But two quite different things get called by these names, and keeping them apart matters. Augmented reality, strictly, means computed content laid over your view of the real world - information, images or 3D objects that appear in your sight but do not necessarily understand or interact with the real space. The classic, everyday example is holding up a phone or tablet: the camera shows the real room on screen and the app draws a computed sofa or a building model into the picture. The real world is kept and something is added on top; that is augmentation.

Mixed reality is the richer, more demanding case, and the crucial extra ingredient is that the computed content sits in and interacts with reality. An MR system has scanned and understood the real room - it knows where the floor, walls and furniture are - so a computed object can rest on your actual table, be hidden correctly when your real sofa passes in front of it (proper occlusion), and stay locked to the same real spot as you walk around it. The computed and the real share one coherent space and respect each other's presence. Modern passthrough headsets deliver this: outward cameras show you the real room in real time and the headset composites anchored computed content into it convincingly. The everyday test that separates MR from plain AR is anchoring and occlusion - does the computed thing stay put in the real world and get correctly hidden by real objects, or does it merely float over the image?

The design examples divide cleanly. AR is the tablet you hold up on site to see a proposed extension roughly in place, or the app that drops a to-scale model onto your meeting table for a client to walk around. MR is the passthrough headset that anchors the full model onto the real construction site so the team sees exactly where the ductwork will run against the real structure, or that lets you place computed furniture in your real, empty room and walk around it as if it were installed. Both keep you in the real world - safer, less isolating, better for collaboration than sealed VR - and both, again, are for seeing and communicating; the precise position where that duct actually goes is set out by the surveyor from the verified model, never trusted off the overlay.

XR - extended reality - is simply the umbrella term for the whole family: VR, AR, MR and everything in between along the continuum. It is a useful word precisely because it refuses to commit to a point on the line; when you mean immersive and spatial in general, without specifying how much of the view is real, XR is the honest, non-committal label. Some people expand the X as extended; some treat it as a variable standing in for any of V, A or M reality. Either way it is a category, not a device - there is no such thing as an XR headset that is somehow distinct from VR and AR; XR contains them.

Here is why the sloppiness is not harmless. These words are marketing battlegrounds, and companies stretch them to flatter their products - a plain phone-AR app gets sold as mixed reality, a passthrough headset gets called a spatial computer to sound more advanced, and everything immersive gets swept up as the metaverse. When the terms blur, you cannot tell what is actually being offered, and you make bad decisions: you might budget for a sealed VR review when the client actually needs to stay in their real room, or expect a computed object to stay anchored on site when the tool only floats it loosely over a camera feed. Precise language is precise thinking - if you can say exactly where on the continuum a proposal sits, you know what it will and will not do for the design task in front of you.

So keep the discipline. VR replaces the world; AR overlays the world; MR anchors computed content in the world so it interacts with reality; XR is the umbrella over all of it. When you read a claim or watch a demo, translate the marketing back into the continuum: how much is real, how much is computed, and does the computed content merely float or genuinely sit in the space? That translation is a skill you will use in every remaining lesson, because the rest of this course is organised around exactly these distinctions - VR for design in Module 3, AR and MR in Module 4 - and each earns its place on different design tasks. Get the family straight now and everything after it is clearer.