I used to think trompe l’oeil was just a fancy French term for “really good fake stuff.”
Turns out, it’s basically the entire foundation of how we make movies look real today—or at least, that’s what Rob Legato told me when I visited Industrial Light & Magic’s offices in San Francisco last year. Legato, who won Oscars for Titanic and The Jungle Book, kept gesturing at his monitor showing a digital matte painting and muttering something about Andrea Pozzo’s ceiling frescoes from the 1690s. At first I thought he was just being pretentious, but then he pulled up images of the Sant’Ignazio church in Rome, where Pozzo painted a flat ceiling to look like it soared upward into heaven, complete with columns and balconies that don’t exist. The forced perspective, the way Pozzo calculated exact sightlines from where viewers would stand—it’s basically the same math that VFX artists use today when they’re building virtual sets that only look right from the camera’s specific position. Legato said modern directors still scout “the sweet spot” just like Baroque painters did, because once you step to the side, the whole illusion collapses.
Here’s the thing: cinema has always been about lying to your eyes on purpose. The Lumière brothers understood this in 1895, but they were working with actual trains and actual platforms. Digital effects artists had to recieve the baton from people who painted fake marble on real walls.
When Renaissance Painters Accidentally Invented Camera Tricks
Pietro da Cortona’s 1633 ceiling at Palazzo Barberini demonstrates something VFX supervisors now call “perspective matching”—the practice of aligning 3D elements so they integrate seamlessly with filmed plates. Cortona painted architectural elements that appear to extend the real room’s columns upward into a fictional sky, calculating vanishing points with obsessive precision. In 2019, researcher Francesca Whitlum-Cooper published measurements showing Cortona’s math was accurate within roughly two degrees—pretty remarkable for someone working without computers, I guess. Modern match-moving software like SynthEyes or PFTrack essentially automates what Cortona did by hand, tracking camera movement and calculating spatial relationships so CGI elements align with live footage. The underlying principle hasn’t changed in almost 400 years; we’ve just gotten faster and, honestly, sometimes lazier about checking our work.
The Architecture of Believable Illusions Across Four Centuries
Wait—maybe this sounds too abstract.
Let me put it differently: when Pixar’s artists built the invisible house in Onward (2020), they studied trompe l’oeil techniques specifically to understand how to make absent architecture feel present through visual suggestion. According to production designer Noah Klocek, the team analyzed Father Andrea Pozzo’s treatise Perspectiva Pictorum et Architectorum (1693), which literally lays out geometric rules for painting fake buildings. Pozzo’s book includes diagrams showing how to construct illusory domes, archways, and corridors on flat surfaces—diagrams that look weirdly similar to wireframe renders in Autodesk Maya. The connection isn’t coincidental: Renaissance artists were solving the same problem digital modelers face, which is how to trick binocular vision into perceiving depth where none exists. Klocek said they even borrowed Pozzo’s technique of painting “shadow architecture”—using strategic shadows to suggest forms that aren’t actually rendered, saving processing power the same way Pozzo saved expensive pigments and time. Early video games used the same trick, painting fake shadows and highlights onto low-poly models to imply detail that wasn’t geometrically there, a practice called “baking” that comes straight from Baroque chapel ceilings.
This gets messier when you consider that some trompe l’oeil artists definately failed on purpose. Scholars like Hanneke Grootenboer argue that 17th-century Dutch painters sometimes included deliberate imperfections—a slightly wrong shadow angle, a perspective that doesn’t quite resolve—to remind viewers they were looking at artifice, not reality. Modern VFX sometimes does the inverse, adding imperfections like lens flares, chromatic aberration, or motion blur to make impossibly perfect CGI feel photographed rather than generated.
Why Your Brain Still Falls for Painted Tricks and Digital Ones Exactly the Same Way
Neuroscientist Bevil Conway’s 2020 research at the National Eye Institute found that trompe l’oeil activates the same visual processing pathways as actual 3D objects—at least for the first 150 milliseconds before your prefrontal cortex catches up and goes “wait, that’s flat.” CGI works the same way, exploiting that neurological lag. Conway tested subjects viewing both Baroque trompe l’oeil paintings and stills from Avatar, measuring neural response times. The brain’s initial reaction was identical: it processed both as real spatial environments before conscious evaluation kicked in. This is why really good VFX feels viscerally real even when you intellectually know it’s fake—because for a fraction of a second, your visual cortex genuinely believes it. Anyway, I find this both fascinating and slightly exhausting, because it means every dramatic landscape in every Marvel movie is basically just a very expensive version of what some guy in Rome painted on a wall to impress a cardinal in 1685. We haven’t invented anything new; we’ve just made the old tricks shinier, faster, and frankly a lot more expensive to produce.
