How a 3D viewer works in a browser
A web based 3D viewer lets a reader turn, tilt and zoom a digital model of an object on a page, without installing anything. The browser downloads a model file describing the object’s shape and surface, then draws it on the reader’s screen using the graphics hardware of the device. When the reader drags a finger or mouse, the browser redraws the model from the new angle.
Most viewers are built with ready-made code libraries or embeddable components rather than from scratch. That keeps the work manageable, though the underlying demands remain: the model must be small enough to download, and the device must be capable enough to draw it smoothly.
Where the models come from
A viewable object is either scanned from a real item or built by hand. The two usual capture approaches are photogrammetry, which reconstructs shape from many overlapping photographs, and structured light or laser scanning, which measures the surface directly. Both are explored in the discussion of 3D scanning of collection objects.
Raw results are generally too heavy for the web. Models are usually simplified, which reduces the number of surface polygons, and their surface images are shrunk, so that a page remains usable. The result is a tradeoff between detail and speed. Details like fine engraving or textile weave may survive only partly, so check that the simplified model still shows what matters about the object.
Models also need care in presentation. Lighting changes how an object looks, and unrealistic or flat lighting can mislead. Colors may differ from the real object, so state clearly that this is a digital representation and note where it differs. These digital copies are best understood as digital surrogates, helpful for study but not the same as handling the original.
Where 3D helps
A rotating view is most useful when the shape itself is the point.
- Objects with several important sides. A vessel with decoration all around, a tool whose handle and blade both matter, or a sculpture designed to be seen from many angles.
- Showing the underside or inside. Marks on the base of a ceramic, for example, are rarely visible in a gallery case.
- Comparing form and wear. Close inspection of a worn edge or a repair is easier when the reader controls the view.
- Fragile or remote items. People can examine something that could never be handled or traveled.
- Learning and study. Students and specialist users can look at details and measure proportions, if the model is accurate enough for that.
In these cases the reader gains something that a single photograph cannot give.
Where it distracts
A spinning model is not automatically better than a good photograph.
- Flat or simple objects. A letter, a coin seen mainly from the front, or a painting gains little from rotation. A sharp, well-lit image with zoom does the job better.
- Poor models. A rough, blurry or misshapen model can make a fine object look cheap. If the quality is not good enough, use photography.
- Slow pages. Heavy models can slow a page badly, especially on older phones or poor connections. If your readers may face those limits, offer the model as an optional step, not the page’s main content.
- Gimmick effect. When 3D is added to every object in a show, it becomes background noise, and attention goes to the controls rather than the object.
- Missing story. A model with no explanation is just a shape. The reader still needs to know what to look for.
Good object photography remains the foundation. A strong still image should be the default view, with the interactive version as an addition.
Design the controls and the page
Make it obvious that the model can be moved. A short hint such as “Drag to rotate” and a visible control for resetting the view help first-time users. Test on touch screens, where dragging a model can clash with scrolling the page; provide a clear way to move past the viewer without being trapped in it. Controls should be large enough to tap comfortably, in line with advice on touch targets.
Offer a few preset views, such as front, back and base, so that readers who do not want to rotate can still see the main angles. Add short text notes pointing to features worth noticing, tied to those views.
Avoid automatic spinning that cannot be stopped. Continuous motion can be tiring or disorienting for some people, and it makes it hard to read nearby text.
Access and alternatives
Interactive 3D is mostly visual and mouse-driven, which creates real barriers. Provide:
- A written description of the object’s shape, material and key features, which also helps people using screen readers.
- Keyboard-operable controls where practical.
- Still images from several angles, with captions.
- Sufficient contrast for any overlaid labels, bearing in mind that low vision readers may struggle with small or pale text.
Without these, a person who cannot use the viewer receives nothing, so treat the alternatives as part of the feature, not an extra.
Rights, credit and care
Consider who owns the model. A scan of an old object may have different rights from the object itself, and creators of scans may expect credit. Some institutions release models openly, while others restrict them. Decide the terms before publishing and state them near the viewer. For items with cultural sensitivities, discuss with source communities whether showing and sharing a detailed 3D copy is appropriate.
Finally, think about long-term care. 3D formats and viewing code change faster than most other web content. Keep the original models and the simplified versions, record which tools produced them, and keep still images as a fallback, so the object can still be seen if the viewer stops working.