The Phenomenology and Computation of Mesmeric 3D Moving Images: From Op Art to Neural Radiance Fields

Full report

On this page

1. Introduction to Visual Mesmerism and Dimensionality

The human visual system is a highly optimized biological inference engine, evolved over millions of years to parse complex spatial environments and extract meaningful three-dimensional structures from the two-dimensional projections cast upon the retina. When this neurological architecture is presented with stimuli that intentionally exploit its heuristics—whether through static optical illusions, stereoscopic disparity, procedural fractal generation, or continuous neural transformation—the result is a state of cognitive fascination often described as "mesmerism." In the context of visual computing, cognitive neuroscience, and art history, mesmerism refers to the profound capture of human attention, the induction of altered perceptual states such as trance or cognitive flow, and the temporary hijacking of the brain's predictive coding mechanisms.

The pursuit of synthesizing mesmerizing moving images in three dimensions represents a profound convergence of visual art, advanced mathematics, and neurobiology. Historically, this endeavor began with mechanical illusions and the simulation of depth on flat planes. Today, it has evolved into real-time computational architectures that generate infinitely complex fractal landscapes and neural volumetric fields using generative artificial intelligence. This report provides an exhaustive, multi-disciplinary analysis of the mechanisms underlying 3D visual mesmerism. It systematically examines the historical evolution of optical illusions, the neurobiological foundations of visual entrainment and predictive coding, the mathematical frameworks of procedural ray marching and generative 3D rendering, and the physiological impact of modern immersive light field display technologies.

2. Historical Foundations of 3D Illusion and Optical Manipulation

The endeavor to trick the human eye into perceiving three dimensions on a two-dimensional surface long predates the advent of digital computation. The mechanisms of optical mesmerism have historically relied on a deep, intuitive understanding of human perceptual vulnerabilities, exploiting the brain's tendency to construct depth from ambiguous cues.

2.1 Trompe-l'œil and the Origins of Spatial Illusion

The earliest systematic attempts to manipulate depth perception can be traced back to classical antiquity and the Renaissance through the technique of trompe-l'œil, a French term translating to "deceive the eye"1. Ancient Greek artists, such as Apollodorus and Zeuxis, were among the first to master visual deception, famously creating paintings so realistic that audiences mistook painted objects for physical reality2. However, it was the codification of linear perspective and the vanishing point by Filippo Brunelleschi in 1425 that provided artists with a mathematical framework for depth3.

Renaissance masters utilized these techniques to overwhelm the observer's spatial reasoning. Andrea Mantegna’s mastery of foreshortening, most notably in his depiction of the human form, achieved an incredibly lifelike composition that gave viewers the distinct impression of looking down into an actual physical space2. In the 17th century, Andrea Pozzo elevated this illusionism to an architectural scale. His fresco at the Sant'Ignazio Church in Rome utilizes extreme perspective to create the illusion of a towering dome opening into the heavens, complete with figures seemingly suspended in midair, despite the ceiling being entirely flat2. These early techniques established the foundational principle of all 3D imagery: depth is not an inherent property of the physical world, but a construct synthesized by the brain from geometric and shading cues.

2.2 The Discovery of Stereoscopy and Binocular Disparity

The transition from monocular perspective cues to true binocular depth representation occurred in the 1830s, fundamentally altering the trajectory of visual media. In 1838, the British physicist Sir Charles Wheatstone presented a seminal paper to the Royal Society detailing the first formal description of stereoscopic vision4. Wheatstone demonstrated that depth perception is primarily a phenomenon of binocular disparity—the brain's fusion of two slightly different perspective projections received independently by the left and right eyes5.

To prove his theory experimentally, Wheatstone invented the reflecting stereoscope in 18327. The device consisted of two mirrors placed at a 90-degree angle; an observer placed their eyes close to the apex, viewing two separate, laterally displaced drawings4. The brain fused these disparate images into a single, floating, three-dimensional object, entirely decoupling the natural visual environment from the observer5.

In 1849, the Scottish physicist Sir David Brewster refined this concept by introducing the lenticular stereoscope, replacing Wheatstone's mirrors with lenses and prisms to create a more compact, portable viewing device4. Combined with the 1839 advent of photography by Louis Jacques Mandé Daguerre and William Henry Fox Talbot, the stereoscope became a parlor phenomenon6. American author Oliver Wendell Holmes, who amassed a collection of thousands of views, coined the term "stereograph" and described the experience as a profound surprise where the mind "feels its way into the very depths of the picture"10. By engineering a simplified, affordable viewer in 1859, Holmes sparked a craze that laid the physiological groundwork for modern Virtual Reality10.

2.3 Optical Art and the Induction of Physiological Movement

While stereoscopy manipulated binocular disparity, the Optical Art (Op Art) movement of the mid-20th century sought to manipulate the very mechanics of retinal processing to create illusions of kinetic movement and depth on a flat surface1. Emerging from the aesthetic theories of Pointillism, the Divisionism of the Italian Futurists, and the multi-dimensional planes of Cubism, Op Art intentionally exploits the physiological limitations of the eye, creating sensory overload through the precise mathematical arrangement of high-contrast patterns1.

Victor Vasarely, widely considered the father of Op Art, demonstrated as early as the 1920s and 1930s (e.g., his 1938 work Zebra) that manipulating line and contrast alone could completely distort a two-dimensional surface into an apparent three-dimensional space1. By the 1960s, British artist Bridget Riley took this manipulation to its neurophysiological limits. In works like Movement in Squares (1961) and Cataract 3 (1967), Riley utilized repeating geometric patterns, alternating black-and-white checkers, and the strategic placement of lines to induce profound illusions of kinetic motion3.

The visual mesmerism of Op Art was not purely aesthetic; it had direct, measurable physiological consequences. Riley's high-frequency geometric patterns often caused viewers to experience motion sickness or vertigo, a testament to the fact that artificial visual stimuli can trigger systemic physical reactions by overwhelming visual processing centers and creating sensory conflict1. The movement also expanded into physical three-dimensional space with kinetic sculptors like Jesús Rafael Soto, whose Penetrables—interactive installations consisting of hundreds of dangling plastic tubes—created hovering concrete forms that adapted dynamically to the viewer's physical movement1.

2.4 The Autostereogram Phenomenon: Magic Eye and Cyclopean Vision

The manipulation of binocular vision and repetitive Op Art patterning converged in the late 20th century with the development of the autostereogram, colloquially popularized as the "Magic Eye"12. The neurobiological foundation for this was laid in the 1960s by neuroscientist Bela Julesz, who utilized computer-generated random dot stereograms to study "cyclopean vision"12. Julesz proved that depth could be perceived purely from retinal disparity without any monocular depth cues, simply by shifting a section of dots horizontally between two side-by-side images12.

In the 1970s, Christopher Tyler advanced this research by developing a method to achieve stereoscopic depth within a single image, utilizing repeating vertical patterns with slight lateral shifts12. When a viewer diverges their eyes—looking "through" the image rather than focusing on the surface—the brain mistakenly fuses adjacent repeating patterns6. The slight shifts in the pattern are interpreted by the visual cortex as binocular disparities, rendering a hidden 3D shape that appears to float above or below the static visual static12.

The commercialization of this phenomenon began in 1991 when Tom Baccei, a manager at Pentica Systems, attempted to create a unique advertisement for an in-circuit emulator12. Working with a pantomimist and 3D photography enthusiast named Ron Labbe, and later partnering with 3D graphics artist Cheri Smith, Baccei began designing "gaze toys" using Single Image Random Dot Stereograms (SIRDS)12. Forming N.E. Thing Enterprises, they partnered with Tenyo Co., a Japanese magic products company, to release the first Magic Eye books12. The mesmerizing appeal of the autostereogram lay in the active, almost meditative participation required by the viewer; extracting the 3D signal from the chaotic noise required holding the oculomotor system in an unnatural state of decoupled vergence and accommodation, inducing a minor trance state12.

3. The Cognitive Neuroscience of Visual Entrainment and Trance

To understand why human observers find dynamic 3D visuals inherently "mesmerizing," it is necessary to examine the underlying neurobiology of visual perception. Mesmerism is not merely a poetic metaphor; it represents measurable, quantifiable shifts in cortical oscillations, predictive processing, and multisensory integration within the brain.

3.1 Predictive Coding, Active Inference, and Machine Architecture

The modern neuroscientific understanding of perception relies heavily on hierarchical predictive coding (hPC)14. According to this theory, the brain is not a passive receiver of sensory data; rather, it is a highly proactive inference engine14. Higher cortical areas continuously generate generative models (predictions) about the spatial environment and send these predictions down the cortical hierarchy to primary sensory areas14. Incoming sensory data is continuously compared against these top-down predictions, and only the "prediction errors"—the unpredicted variance or novelty—are propagated forward to update the internal models14.

Mesmerizing 3D visuals—such as infinite mathematical fractals, generative neural loops, or high-frequency Op Art—essentially hack this predictive architecture. Because these stimuli exhibit continuous novelty, extreme self-similarity, and non-Euclidean transformations, they prevent the visual cortex from ever settling on a stable predictive model17. The brain generates continuous, unresolved prediction errors, keeping the visual and attentional networks in a highly engaged, computationally active state14.

Recent models of dendritic predictive coding suggest that these error calculations occur locally within the apical and basal dendrites of balanced spiking neural networks, rather than in separate error units14. When a viewer navigates a seamlessly looping 3D animation, the visual cortex is flooded with bottom-up sensory data that mimics physical geometry but disobeys natural physical laws, such as endless zooming without reaching a finite scale14. This failure to resolve prediction errors leads to a state of profound cognitive fascination, characterized by the suspension of time perception and the induction of a flow state17. In Artificial Intelligence, architectures like Active Predictive Coding Networks (APCNs) explicitly attempt to model this biological process, using hypernetworks and reinforcement learning to dynamically generate recurrent neural networks that parse visual scenes into part-whole hierarchies, mirroring the saccadic sampling of the human retina15.

3.2 Brainwave Entrainment and Steady-State Visual Evoked Potentials

Prolonged exposure to rhythmic visual and auditory stimuli can induce a physiological phenomenon known as brainwave entrainment, wherein the brain's endogenous electrical oscillations synchronize to the frequency of the external stimulus19. Neural oscillations in different frequency bands dictate human cognitive states.

Rhythmic visual stimulation, particularly flickering stereoscopic imaging, elicits Steady-State Visual Evoked Potentials (SSVEPs) in the occipital and parietal cortices21. When exposed to a flickering 3D visual stimulus, neural populations synchronize their firing rates to the exact flicker frequency23. Research demonstrates that prolonged visual entrainment tailored to a subject's Individual Alpha Frequency (IAF) can induce long-lasting, spatially distributed modulations in resting-state alpha oscillations, increasing phase coherence that persists long after the stimulus is removed19.

Increases in pre-stimulus Theta and Alpha power have also been shown to correlate directly with the successful encoding of crossmodal associations and episodic memory25. The phase locking of theta rhythms in the hippocampus—a process known as phase precession—is critical for spatial navigation and mapping25. Therefore, traversing a 3D digital landscape inherently engages the brain's spatial processing networks, and doing so rhythmically can coax the brain into the Theta state associated with deep trance25. During the anticipation of these visual events, strong frontal stimulus-preceding negativity (SPN) is observed, reflecting intense cortical preparation and engagement27.

3.3 Multisensory Integration: Binaural Beats and Digital Therapeutics

The mesmerizing effect of 3D visuals is exponentially magnified when integrated with auditory entrainment, specifically through binaural beats and isochronic tones20. A binaural beat is a psychoacoustic illusion generated when two pure tones of slightly different frequencies (typically below 1000 Hz) are presented independently to each ear via headphones20. Because the sound waves never physically meet in the air, the brain processes this discrepancy in the superior olivary complex of the brainstem, perceiving a phantom beat oscillating at the exact difference between the two frequencies20. For instance, a 400 Hz tone in the left ear and a 410 Hz tone in the right ear produces a 10 Hz binaural beat, targeting the Alpha band20. The maximum frequency difference for this illusion to hold is approximately 30 Hz; beyond this, the tones are perceived separately20.

Advanced generators, such as those developed by myNoise, utilize up to ten carrier frequencies simultaneously, producing amplitude-modulated beating patterns within each ear canal that create layered, evolving rhythmic structures28. Utilizing warm carrier frequencies around 140 Hz at near-subliminal volumes can create a "vacuum effect" that masks external distractions, pulling the viewer's focus entirely into the visual field28. Because entrainment relies on the thalamic relay to propagate the rhythmic stimulus to the broader cortex—bypassing the olivary bodies where the beats are initially heard—the integration of 3D spatial audio and pure sine wave carriers is highly effective at inducing trance-like states28.

This neurobiological pathway has given rise to a new clinical paradigm of "digital therapeutics"32. By combining immersive Virtual Reality (VR) 3D visuals with precise audio-visual entrainment (AVE), developers can create simulated environments that suppress the perception of chronic pain, anxiety, and isolation32. In these multimodal trance states, the visual cortex processes the expansive 3D graphics while the auditory cortex synchronizes to the binaural frequencies, creating a profound dissociative effect that guides the user into deep meditation or cognitive flow26. Clinical trials have demonstrated that 40 Hz gamma binaural beats can significantly improve working memory and cognitive flexibility, while 10 Hz alpha beats can reduce pre-operative anxiety and lower blood pressure30.

4. Algorithmic Mesmerism: The Mathematics of 3D Fractals and Ray Marching

The computational generation of mesmerizing 3D images relies on pushing far beyond the boundaries of traditional polygon-based rasterization. By utilizing continuous mathematical spaces and implicit surfaces, rendering algorithms can produce infinitely detailed, self-similar topologies that defy conventional Euclidean geometry.

4.1 The Geometry of the Infinite: Mandelbulbs and Mandelboxes

The core of 3D fractal art lies in the conversion of 2D complex plane fractals, such as the classic Mandelbrot set ([Figure omitted from source export]), into three-dimensional Cartesian space35. The Mandelbulb, mathematically derived by Daniel White and Paul Nylander, represents the three-dimensional analogue of the Mandelbrot set36. Because there is no perfect 3D mathematical equivalent of complex numbers, the algorithm operates by converting 3D Cartesian coordinates [Figure omitted from source export] into spherical coordinates [Figure omitted from source export], raising the angles and radius to a specific power (typically [Figure omitted from source export]), and converting back to Cartesian coordinates35.

The iterative mathematical transformation is defined as:

[Figure omitted from source export]

[Figure omitted from source export]

When this equation is iterated across space, points that do not escape to infinity (remaining within a defined bailout radius) define the physical structure of the Mandelbulb35. Other geometries, such as the Mandelbox, rely on complex spatial folding operations—reflecting points across spherical or cubical planes of symmetry—rather than trigonometric exponentiation36. The result is a mathematically pure landscape characterized by intricate, organic-looking details that resolve infinitely as the observer zooms closer, creating a visual feedback loop that is inherently hypnotic18.

4.2 Signed Distance Functions (SDFs) and Space Folding

Traditional computer graphics rely on rasterizing fixed triangles. However, 3D fractals and many hypnotic generative shapes are represented implicitly using Signed Distance Functions (SDFs)40. An SDF is a mathematical formula that takes a 3D coordinate as an input and returns the shortest orthogonal distance to the nearest surface of an object41. The returned value is signed: positive if the point is located outside the object, negative if inside, and exactly zero on the surface boundary40.

The computational power of SDFs lies in their composability. Because they represent shapes purely as continuous mathematical distance fields, they can be combined using Boolean operations (addition, subtraction, intersection) with minimal computational cost44. More importantly for the creation of mesmerism, SDFs allow for advanced spatial manipulations like "smooth minimums" (which seamlessly blend shapes together like liquid droplets) and "domain repetition" using modulo arithmetic40. Domain repetition replicates a shape infinitely across coordinate space for almost zero extra computational cost, making it ideal for rendering endless 3D fractals and repeating architectural illusions40.

4.3 The Ray Marching (Sphere Tracing) Algorithm

Because SDFs do not possess physical vertices, they cannot be rendered via traditional ray tracing algebraic intersections40. Instead, they are rendered using a highly optimized algorithm known as Ray Marching, or Sphere Tracing, typically executed within custom fragment shaders using GLSL (OpenGL Shading Language)39.

In Ray Marching, a virtual ray is cast from the camera through each pixel on the screen into the 3D scene35. At the ray's current position, the SDF is evaluated to find the distance [Figure omitted from source export] to the nearest surface40. The algorithm then "marches" the ray forward along its normalized direction vector by exactly [Figure omitted from source export] units40. Since the distance estimator guarantees that no object lies within a sphere of radius [Figure omitted from source export] centered at the current point, it is perfectly safe to step forward without overshooting the geometry35. This process iterates until the distance drops below a microscopic threshold (a "hit", often denoted as EPS\_H) or exceeds the bounds of the scene (FAR\_T)35.

For highly complex fractals like the Mandelbulb, the distance estimator (DE) is approximated by comparing the escape-time length ([Figure omitted from source export]) to its spatial derivative ([Figure omitted from source export]):

[Figure omitted from source export]

Where [Figure omitted from source export] is updated iteratively alongside the spatial coordinates to maintain a safe lower bound for the ray step35.

Ray marching inherently facilitates complex visual phenomena essential for hypnotic imagery. Soft shadows can be generated by tracking how close the ray passes to objects without hitting them, while volumetric lighting and ambient occlusion emerge naturally from the step data36. Furthermore, surface normals required for complex BRDF (Bidirectional Reflectance Distribution Function) shading and Albedo calculations can be easily estimated by sampling the SDF at small offsets using a tetrahedral sampling pattern, computing the gradient to determine how light scatters at a single point35.

4.4 The Cinematography of Mathematics: Julius Horsthuis

The aesthetic translation of these raw mathematical algorithms into mesmerizing cinematic art is epitomized by digital artists like Julius Horsthuis. Utilizing free software like Mandelbulb 3D (MB3D), Horsthuis treats fractal environments not as traditional design canvases, but as pre-existing alien landscapes awaiting discovery18.

The process of exploring an SDF-driven fractal space is akin to wildlife photography in a mathematical realm51. By adjusting the variables of the non-linear equations, the artist navigates an unstructured, multidimensional topology51. Horsthuis emphasizes grounding these highly abstract, potentially chaotic visual spaces by applying familiar cinematic techniques: simulated depth-of-field, lens distortion, strategic camera dollies, and hyper-realistic lighting schemas51.

Crucially, he employs heavy color correction and post-processing in Adobe After Effects (often utilizing tools like the Magic Bullet Looks Suite) to provide the human brain with recognizable psychological anchors51. By borrowing color palettes from films like Terminator 2 or Prometheus—favoring deep, muted blues—he colors abstract fractal spires to resemble realistic rock faces illuminated by a low sun51. This deliberate juxtaposition of the alien (infinite mathematical iteration) and the familiar (photorealistic lighting and color) acts as a powerful hypnotic hook, triggering the brain's spatial mapping networks while suspending its logical predictions18. His immersive exhibitions, such as Foreign Nature at the Nxt Museum in Amsterdam and Geometric Properties at ARTECHOUSE, require massive computational overhead, with 360-degree VR renders topping 3GB for a mere 10 minutes of 4000x2000 resolution footage53.

5. The Neural Turn: Generative 3D and Machine Hallucinations

While procedural fractals rely on explicit, hand-crafted mathematical formulas, the current frontier of 3D visual mesmerism utilizes deep learning to synthesize novel spatial environments from massive visual datasets. This represents a fundamental shift from mathematical complexity to semantic, data-driven complexity.

5.1 Neural Radiance Fields (NeRF) and Volumetric Rendering

Introduced in 2020, Neural Radiance Fields (NeRFs) revolutionized the field of novel-view synthesis57. A NeRF implicitly represents a continuous 3D scene by optimizing a Multilayer Perceptron (MLP) neural network58. The network maps a 5D spatial coordinate (3D position plus 2D viewing direction) to a volume density and view-dependent emitted radiance (color)59. Using traditional volume rendering techniques, NeRFs cast rays through this neural field, accumulating color and density to synthesize highly photorealistic images from unobserved angles58. While NeRFs produce mesmerizingly accurate spatial reconstructions with perfect specular reflections and view-dependent effects, they are extraordinarily computationally intensive, requiring significant time for both training and querying the MLP for every pixel, which historically hindered real-time applications61.

5.2 3D Gaussian Splatting (3DGS) and Dynamic Loops

To overcome the severe latency bottlenecks of NeRFs, the spatial computing paradigm has rapidly shifted toward 3D Gaussian Splatting (3DGS)61. Rather than using an implicit neural network, 3DGS represents a scene using a vast, unstructured cloud of explicit 3D Gaussian primitives61. Initialized from sparse point clouds derived from Structure-from-Motion (SfM) camera calibration, these millions of Gaussians are optimized for position, covariance (scaling and rotation), opacity, and color62. Crucially, color is represented via Spherical Harmonics, allowing the scene to accurately replicate complex view-dependent reflections62.

By interleaved optimization with adaptive density control (adding or pruning Gaussians as needed) and employing highly optimized tile-based rasterization with alpha-blending, 3DGS achieves state-of-the-art visual fidelity at real-time rendering speeds—often exceeding 135 fps at 1080p resolution62. This real-time capability is crucial for mesmerism, as any latency or frame drop immediately breaks the illusion of presence and disrupts cognitive flow61. Advancements like LightSplat synthesize local sparse features with a dual-thread backend that progressively constructs dense Gaussian submaps, enabling online loop closure and pose graph optimization for robust real-time robotic SLAM61.

Furthermore, 3DGS has been adapted for dynamic, moving scenes65. Advanced frameworks like LoopGaussian generate seamless 3D cinemagraphs by extracting Eulerian motion fields from static scenes66. By applying bidirectional animation techniques to the 3D Gaussian points moving along these motion fields, developers can create spatially contiguous, infinitely looping 3D environments that evoke a serene, hypnotic dynamic realism unattainable by static rendering alone66. Other implementations, such as GPS-Gaussian, utilize parameter maps to regress 3D properties directly for instant, real-time novel view synthesis of human subjects without per-subject fine-tuning67.

5.3 Latent Space Walk and AI Hallucinations: Refik Anadol

The intersection of AI and hypnotic visual art reaches its zenith in the exploration of latent spaces within Generative Adversarial Networks (GANs), particularly using architectures like StyleGAN2-ADA68. When trained on massive image archives, GANs learn a continuous "latent space"—a high-dimensional abstract coordinate grid where every mathematical point can be decoded into a unique, photorealistic image71.

Media artist Refik Anadol utilizes these algorithms to create macroscopic, data-driven spatial paintings. For his critically acclaimed installation Unsupervised at the Museum of Modern Art (MoMA), Anadol trained a unique AI model on 178,000 images from MoMA's archive, representing over 200 years of modern art history71. By executing smooth, continuous paths (latent space walks) through this multi-dimensional coordinate space, the AI generates fluid, ever-morphing "machine hallucinations" that mathematically interpolate between distinct concepts of art68.

The visual output is highly mesmerizing because it presents the human eye with recognizable textures, brushstrokes, and architectural shapes that continuously dissolve and evolve before the brain can semantically categorize them72. Anadol enhances this mesmeric state by turning the digital installation into a living system; sensors capture the museum's ambient light, acoustics, weather, and crowd movement to dynamically alter the trajectory of the latent walk in real time72. The installation is further cemented in the digital realm by encoding the generated artifacts on the blockchain as minted mementos72. This positions AI not merely as a generative tool, but as a dynamic engine for perceptual exploration that fuses massive data abstraction with sensory immersion74.

6. Hardware Interfaces and the Physiology of Immersion

The psychological impact and mesmeric potential of 3D moving images are inextricably linked to the hardware utilized to display them. As displays bridge the gap between flat 2D screens and spatial reality, they must contend directly with the physiological limitations and evolutionary design of human binocular vision.

6.1 Virtual Reality, Vection, and Cybersickness

Head-Mounted Displays (HMDs) provide absolute sensory enclosure, placing the user directly within the 3D generated environment. However, this profound immersion frequently induces cybersickness (Virtual Reality Sickness)64.

Cybersickness is primarily explained by the sensory conflict theory76. In a virtual environment, dynamic optic flow creates a powerful illusion of self-motion, known as vection76. The visual cortex informs the brain that the body is accelerating or rotating, yet the vestibular system (the inner ear) and proprioceptive networks report that the body is entirely stationary76. This sensory mismatch triggers a maladaptive autonomic response, leading to disorientation, oculomotor strain, nausea, and vertigo76.

The severity of these symptoms is heavily monitored using standardized metrics like the Virtual Reality Sickness Questionnaire (VRSQ) and the Motion Sickness Assessment Questionnaire (MSAQ), which track gastrointestinal, central, and peripheral distress80. Mitigation requires meticulous content design: maintaining a strict minimum refresh rate of 90Hz, minimizing system latency (motion-to-photon lag), utilizing gradual acceleration curves, and restricting field-of-view during artificial locomotion64. Furthermore, there is a minor but critical risk of Photosensitive Epilepsy (PSE) in virtual environments; highly evocative stimuli featuring high-contrast, bright flashes—particularly red flickers at a frequency of 15-25 Hz—can provoke seizures in susceptible individuals, though the overall incidence of PSE remains exceptionally low at 0.002%81.

6.2 The Vergence-Accommodation Conflict (VAC)

Even for stationary VR users free from vection, a profound physiological barrier to viewing comfort exists: the Vergence-Accommodation Conflict (VAC)83.

In natural viewing, the human oculomotor system tightly couples two mechanisms:

1. Vergence: The simultaneous rotation of the eyes inward (convergence) or outward (divergence) to place the object of interest on the fovea of both retinas, resolving binocular disparity83.

2. Accommodation: The ciliary muscles adjusting the shape of the crystalline lens to bring the object into sharp focus on the retina83.

In conventional stereoscopic displays and HMDs, this natural neural coupling is broken. The eyes must converge to varying virtual depths defined by the parallax of the stereoscopic images, but they must simultaneously remain accommodated to the fixed physical distance of the display screen to maintain a sharp image85. This unnatural uncoupling degrades depth perception, reduces stereoacuity, increases the time required for binocular fusion, and induces significant visual fatigue83. Vision science identifies a "zone of comfort" for stereoscopic viewing, typically spanning [Figure omitted from source export] diopters around the focal plane; outside this zone, visual effort spikes dramatically85. Research indicates that rapid, phasic changes in the vergence-accommodation conflict—such as objects quickly flying toward the camera in a 3D animation—cause the most severe viewer discomfort89.

6.3 Volumetric and Light Field Displays

To circumvent the physiological drawbacks of head-mounted VR, the display industry has pushed toward "glass-free" holographic solutions, most notably Light Field Displays and Volumetric Displays, championed by companies like Looking Glass Factory90.

Rather than sending a single stereoscopic pair to a headset, a Light Field Display acts as a spatial window. Looking Glass technology projects up to 100 discrete views of a 3D scene simultaneously over a viewing cone of roughly 53 to 60 degrees90. This is achieved using an optical lens array overlaid on an ultra-high-resolution 2D display (rendering up to 8K resolution, routing nearly 100 million subpixel points of light)92. The display sub-divides its pixels such that as an observer moves their head laterally, their left and right eyes continuously transition between adjacent, perspective-correct views90.

This provides two critical optical cues missing from traditional 2D displays: binocular stereopsis and true optical motion parallax90. Objects rendered at the "Zero-Parallax Plane" (the physical depth of the screen) appear sharpest and remain stationary in pixel-space, while objects in the virtual foreground and background undergo appropriate lateral shifting as the viewer moves90. Because light field displays emit physical directional light rays that reconstruct the wave front of a 3D scene, they significantly mitigate the Vergence-Accommodation Conflict, allowing for comfortable, prolonged, group-viewable mesmerism without headgear91.

 

Looking Glass has also introduced Hololuminescent Displays (HLDs), which embed a fixed 3D holographic volume within a high-resolution 2D display stack, allowing standard video production pipelines to create immersive spatial signage without complex 3D rendering96.

6.4 The Architecture of Illusion: Immersive Museum Installations

The culmination of these visual, optical, and computational technologies is currently manifesting in large-scale, interactive art installations, such as the WNDR Museum in Chicago and the international Museum of Illusions97. These spaces translate the hypnotic mechanics of 3D algorithms into macro-scale physical realities, replacing traditional gallery formats with hands-on, perception-altering environments100.

Exhibits like the Infinity Mirror Room, pioneered by artists like Yayoi Kusama in installations such as Love Is Calling, utilize recursive reflections and signature polka-dotted forms to simulate the infinite coordinate space of a fractal algorithm within a physical room, profoundly disorienting the participant's spatial boundaries and simulating a galaxy-like suspension97. Similarly, the WNDR Light Floor utilizes thousands of optical sensors (rather than basic motion sensors) to allow participants to dynamically manipulate generative particle systems through physical motion; the graphics lie dormant until movement magnetizes the animation, creating rushing trails of color100.

Other exhibits, such as the SENSE installation, utilize embedded flexible screens to create a dialogue between physical intention and digital response, while Sound-Responsive Rooms visualize voice and footsteps as dancing waves of color97. By combining trompe-l'œil wall distortions, augmented reality overlays, and high-frequency light tunnels, these physical venues act as architectural-scale stereoscopes97. They synthesize historical optical trickery with real-time neural computation, transforming passive observation into an active, embodied trance.

7. Conclusions

The phenomenon of mesmerizing 3D moving images is the result of a precise, historically evolving intersection between computational algorithms, optical engineering, and the neurobiology of human perception. Tracing its lineage from the static deception of Renaissance trompe-l'œil and Wheatstone’s original stereoscope to the physiological manipulations of Op Art and the Magic Eye, the goal has consistently been to bypass the brain's logical processing and directly stimulate the lower-level sensory cortex.

Today, this manipulation is achieved with unprecedented mathematical supremacy through Ray Marching algorithms, Signed Distance Functions, and 3D Gaussian Splatting, which can render infinite fractal complexity and dynamic volumetric motion in real time. When these generative structures are mapped onto advanced Light Field Displays or VR systems, and paired with the exact audio-visual entrainment of binaural beats and isochronic tones, they do far more than entertain. They engage the brain's hierarchical predictive coding networks in a continuous loop of prediction errors, phase-locking neural oscillations into deep Theta and Alpha trance states. Ultimately, the future of 3D visual mesmerism lies not just in aesthetic appreciation, but in its profound capability as a cognitive interface—a technology with rapidly expanding applications in digital therapeutics, psychological entrainment, memory modulation, and the seamless integration of artificial spatial computing into human reality.

Works cited

1. Optical Illusion Art That Marked the 20th Century \- Ideelart, https://ideelart.com/blogs/magazine/optical-illusion-art-that-marked-the-20th-century

2. From Trompe-l'œil to Digital Magic: The Evolution of Art Illusions, https://townquaystudios.co.uk/from-trompe-loeil-to-digital-magic-the-evolution-of-art-illusions/

3. 5 Optical Illusions in Some of History's Most Iconic Masterpieces, https://www.riseart.com/article/2684/5-optical-illusions-in-some-of-history-s-most-iconic-masterpieces

4. Stereoscopy: the birth of 3D technology \- Google Arts & Culture, https://artsandculture.google.com/story/stereoscopy-the-birth-of-3d-technology-the-royal-society/pwWRTNS-hqDN5g?hl=en

5. Wheatstone-1838-CPV.pdf, https://vis.cs.brown.edu/docs/pdf/Wheatstone-1838-CPV.pdf

6. History, http://acorn.stanford.edu/psych221/projects/2003/jgin/history.htm

7. A short history of Stereoscopy \- Watts Gallery, https://www.wattsgallery.org.uk/explore-victorian-virtual-reality/a-short-history-of-stereoscopy

8. Wheatstone and the origins of moving stereoscopic images \- PubMed, https://pubmed.ncbi.nlm.nih.gov/23362669/

9. History of photography \- Stereoscopic, Daguerreotype, Calotype, https://www.britannica.com/technology/photography/Development-of-stereoscopic-photography

10. Stereoscopes \- Scalar, https://scalar.chapman.edu/scalar/this-land-is-your-land/stereoscopes

11. Explore the mesmerizing world of optical illusion art | RMCAD, https://www.rmcad.edu/blog/beyond-reality-exploring-the-mesmerizing-world-of-optical-illusion-art/

12. The Hidden History of Magic Eye, the Optical Illusion That Briefly, https://eyeondesign.aiga.org/the-hidden-history-of-magic-eye-the-optical-illusion-that-briefly-took-over-the-world/

13. How 1990s Magic Eye 3D Images Were Made \- YouTube, https://www.youtube.com/watch?v=uvXY99HysrU

14. A theory of cortical computation with spiking neurons \- arXiv, https://arxiv.org/pdf/2205.05303

15. Active Predictive Coding Networks \- arXiv, https://arxiv.org/pdf/2201.08813

16. Alpha oscillations and traveling waves: Signatures of predictive, https://pubmed.ncbi.nlm.nih.gov/31581198/

17. \#1004: ANANDALA is an Awe-Inspiring, Generative Art of Embodied, https://voicesofvr.com/1004-anandala-is-awe-inspiring-generative-art-of-embodied-ai-that-cultivates-perpetual-novelty/

18. 3D Fractal Animations | gasathj, https://gasathj.com/en/article34-3D-Fractal-Animations

19. Prolonged Visual Entrainment Induces Long-Lasting Alpha-Band, https://pubmed.ncbi.nlm.nih.gov/42144890/

20. Binaural beats to entrain the brain? A systematic review of the ... \- PMC, https://pmc.ncbi.nlm.nih.gov/articles/PMC10198548/

21. Exploring Attention in Depth: Event-Related and Steady-State Visual, https://pmc.ncbi.nlm.nih.gov/articles/PMC12015859/

22. Generating Visual Flickers for Eliciting Robust Steady-State Visual, https://sccn.ucsd.edu/\~yijun/pdfs/PONE14.pdf

23. Steady-state visual evoked potentials: distributed local sources and, https://scispace.com/pdf/steady-state-visual-evoked-potentials-distributed-local-1rqrztjqmg.pdf

24. Exploring neural entrainment and synchrony in response to, https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0332310

25. Increases in pre-stimulus theta and alpha oscillations precede, https://pmc.ncbi.nlm.nih.gov/articles/PMC10991485/

26. What is brainwave entrainment & audio visual entrainment \- Roxiva, https://roxiva.com/what-is-brainwave-entrainment-audio-visual-entrainment/

27. Dissociating nociceptive modulation by the duration of pain, https://www.researchgate.net/publication/23448236\_Dissociating\_nociceptive\_modulation\_by\_the\_duration\_of\_pain\_anticipation\_from\_unpredictability\_in\_the\_timing\_of\_pain

28. The Ultimate Binaural Beat Generator — Online & Free \- myNoise, https://mynoise.net/NoiseMachines/binauralBrainwaveGenerator.php

29. Entraining Tones and Binaural Beats \- Mind Alive, https://mindalive.com/pages/entraining-tones-and-binaural-beats

30. The Science Behind Brainwave Entrainment How Multisensory, https://www.somatic-learning.com/blog/what-the-science-says-brainwave-entrainment-and-multisensory-stimulation

31. Entrain: Binaural Beats \- Apps on Google Play, https://play.google.com/store/apps/details?id=com.entrain.beats

32. VR Based Digital Therapeutics \- Mynd Immersive, https://www.myndimmersive.com/vr-based-digital-therapeutics/

33. Telehealth virtual reality intervention reduces chronic pain in a, https://pmc.ncbi.nlm.nih.gov/articles/PMC11976909/

34. Super Focus: Flow State Music \- Binaural Alpha Brainwaves 3D Audio, https://www.youtube.com/watch?v=CO1TDz-ycCI

35. Rendering the Mandelbulb \- Nikos Papadopoulos, https://www.4rknova.com/blog/2025/09/01/mandelbulb

36. 3D Fractals, https://users.csc.calpoly.edu/\~zwood/teaching/csc471/final13/kdanie03/

37. Rendering of a Mandelbulb Fractal in Real Time, https://celarek.at/wp/wp-content/uploads/2014/05/realTimeFractalsReport.pdf

38. Distance Estimated 3D Fractals (V): The Mandelbulb & Different DE, http://blog.hvidtfeldts.net/index.php/2011/09/distance-estimated-3d-fractals-v-the-mandelbulb-different-de-approximations/

39. CUDA Ray Marching, https://users.csc.calpoly.edu/\~zwood/teaching/csc473/final13/kschapan/

40. Ray Marching — carl-vbn.dev, https://carl-vbn.dev/projects/raymarching/

41. What is a Signed Distance Function? \- DIGITAL SALMON, https://www.digitalsalmon.co.uk/blog/signed-distance-functions

42. Signed distance function \- Wikipedia, https://en.wikipedia.org/wiki/Signed\_distance\_function

43. Morphing Geometric Shapes with SDF in GLSL Fragment Shaders, https://medium.com/@den4icccccc/morphing-geometric-shapes-with-sdf-in-glsl-fragment-shaders-and-visualization-in-jetpack-compose-48fd8d403e24

44. An Introduction to Ray Marching | Michael Crum | Portfolio, https://michael-crum.com/raymarching/

45. A Recursive Algorithm to Render Signed Distance Fields, https://pointersgonewild.com/2026-03-06-a-recursive-algorithm-to-render-signed-distance-fields/

46. Ray marching \- Wikipedia, https://en.wikipedia.org/wiki/Ray\_marching

47. Ray Marching \- The Graphics Codex, https://graphicscodex.com/app/\_rn\_rayMrch.xml.html

48. Ray Marching \- Evelyn's Portfolio, https://www.evelynsalie.com/ray\_marching/ray\_marching.html

49. Distance Estimated 3D Fractals (Part I) \- Syntopia, http://blog.hvidtfeldts.net/index.php/2011/06/distance-estimated-3d-fractals-part-i/

50. Mandelbulb | Syntopia, http://blog.hvidtfeldts.net/index.php/category/mandelbulb/

51. Julius Horsthuis: The Beauty of Fractal Art \- 80 Level, https://80.lv/articles/julius-horsthuis-the-beauty-of-fractal-art

52. Julius Horsthuis Talking About His Immersive Style As a Fractal Artist, https://www.psytshirt.com/blog/psychedelic-shirts-trance-festival-clothing-Julius-Horsthuis-interview-fractal-art.html

53. Fractal Immersive Experience by Julius Horsthuis \- YouTube, https://www.youtube.com/shorts/QhJ8QMcA\_nw

54. An Interview With Julius Horsthuis \- DomaChroma, https://domachroma.com/artists/an-interview-with-julius-horsthuis/

55. 'Foreign Nature' Takes You on a Geometry Bending VR Trip, https://roadtovr.com/foreign-nature-oculus-rift-dk2-virtual-reality-fractal-download-julius-horsthuis/

56. Fractal Art with Artist Julius Horsthuis \- Artechouse, https://www.artechouse.com/fractal-art-with-artist-julius-horsthuis/

57. Neural radiance field | IEEE Technology Navigator, https://technav.ieee.org/topic/neural-radiance-field/

58. 3D Rendering Using Neural Radiance Fields \- ijisae.org, https://www.ijisae.org/index.php/IJISAE/article/view/6914

59. (PDF) An overview of Neural Radiance Fields \- ResearchGate, https://www.researchgate.net/publication/382373557\_An\_overview\_of\_Neural\_Radiance\_Fields

60. Application of Neural Radiance Fields (NeRFs) for 3D Model ... \- MDPI, https://www.mdpi.com/2076-3417/14/5/1825

61. Real-Time High-Fidelity 3D Gaussian SLAM with Loop Closure \- arXiv, https://arxiv.org/html/2609.07274v1

62. 3D Gaussian Splatting for Real-Time Radiance Field Rendering, https://arxiv.org/abs/2308.04079

63. 3D Gaussian Splatting for Real-Time Radiance Field Rendering, https://arxiv.org/pdf/2308.04079

64. What is Virtual Reality Sickness (Cyber Sickness) \- Unity, https://unity.com/glossary/cyber-sickness

65. 3D Gaussian Splatting for Efficient Retrospective Dynamic Scene, https://arxiv.org/abs/2605.12437

66. \[2404.08966\] LoopGaussian: Creating 3D Cinemagraph with Multi, https://arxiv.org/abs/2404.08966

67. Generalizable Pixel-wise 3D Gaussian Splatting for Real-time, https://arxiv.org/abs/2312.02155

68. Unsupervised — Machine Hallucinations — MoMA \- Refik Anadol, https://refikanadol.com/works/unsupervised/

69. Adaptive 3D Augmentation in StyleGAN2-ADA for High-Fidelity Lung, https://www.mdpi.com/1424-8220/25/24/7404

70. ClaudioCarvalhoo/stylegan2-ada-pytorch-latent-space-exploration, https://github.com/ClaudioCarvalhoo/stylegan2-ada-pytorch-latent-space-exploration

71. Modern Dreams: Refik Anadol's Unsupervised \- Feral File, https://feralfile.substack.com/p/modern-dreams-refik-anadols-unsupervised

72. Refik Anadol: Unsupervised \- MoMA, https://www.moma.org/calendar/exhibitions/5535

73. Unsupervised by Refik Anadol at the Museum of Modern Art in New, https://www.youtube.com/watch?v=S9J96Pq\_rvg

74. Unsupervised — Burned — Data Universe — MoMA — 3D (2023), https://feralfile.com/exhibitions/series/unsupervised-burned-data-universe-moma-3d-bxu

75. Virtual Reality Sickness: A Review of Causes and Measurements, https://www.tandfonline.com/doi/full/10.1080/10447318.2020.1778351

76. Cybersickness: a Multisensory Integration Perspective | Request PDF, https://www.researchgate.net/publication/322400282\_Cybersickness\_a\_Multisensory\_Integration\_Perspective

77. Cybersickness and Its Severity Arising from Virtual Reality Content, https://pmc.ncbi.nlm.nih.gov/articles/PMC8963115/

78. Unexpected Vection Exacerbates Cybersickness During HMD, https://www.frontiersin.org/journals/virtual-reality/articles/10.3389/frvir.2022.860919/full

79. Review on cybersickness in applications and visual displays, https://search.proquest.com/openview/ad07083095df5849da6f684862916bf8/1?pq-origsite=gscholar\&cbl=26864

80. Focusing on cybersickness: pervasiveness, latent trajectories ... \- PMC, https://pmc.ncbi.nlm.nih.gov/articles/PMC8886867/

81. Concern of Photosensitive Seizures Evoked by 3D Video Displays, https://www.dovepress.com/concern-of-photosensitive-seizures-evoked-by-3d-video-displays-or-virt-peer-reviewed-fulltext-article-EB

82. Concern of photosensitive seizures evoked by 3D video displays or, https://profiles.wustl.edu/en/publications/concern-of-photosensitive-seizures-evoked-by-3d-video-displays-or/

83. Vergence–accommodation conflicts hinder visual performance and, https://arvojournals.org/arvo/content\_public/journal/jov/932853/jov-8-3-33.pdf

84. Evaluating the Vergence-Accommodation Conflict in Gaze-Based, https://arxiv.org/abs/2607.27369

85. Evaluating the Vergence–Accommodation Conflict in Gaze-Based, https://arxiv.org/html/2607.27369

86. Causes of discomfort in stereoscopic content: a review \- arXiv, https://arxiv.org/pdf/1703.04574

87. Evaluating the Vergence-Accommodation Conflict in Gaze-Based, https://arxiv.org/pdf/2607.27369

88. Vergence-accommodation conflicts hinder visual performance and, https://pubmed.ncbi.nlm.nih.gov/18484839/

89. The rate of change of vergence–accommodation conflict affects, https://escholarship.org/uc/item/17c9j1rp

90. How does Looking Glass Work? | Looking Glass | Light Field Displays, https://lfdocs.lookingglassfactory.com/keyconcepts/how-it-works

91. Light Field Displays from Looking Glass: Digital's New Dimension, https://www.youtube.com/watch?v=CeG2zd8PcZo

92. 16" Light Field Display \- Looking Glass Factory, https://checkout.lookingglassfactory.com/products/16-lightfield-display

93. Looking Glass Go, https://checkout.lookingglassfactory.com/products/looking-glass-go

94. \#1100: Looking Glass Factory's Holographic Displays feel like, https://voicesofvr.com/1100-looking-glass-factorys-holographic-displays-feel-like-magical-portals-demystifying-their-lightfield-display-technology/

95. Looking Glass Holographic Light Field Displays: Experience Group, https://lookingglassfactory.com/displays-overview

96. Our technology \- Looking Glass Factory, https://lookingglassfactory.com/how-it-works

97. WNDR Museum Chicago Tour – Book Your Immersive Art Experience, https://www.asapholidays.com/blog/wndr-museum-chicago/

98. Museum of Illusions Chicago Millennium Park Downtown, https://www.evisitorguide.com/chicago/brochures/museum-of-illusions-chicago-millennium-park-downtown.php?evgreturn=/chicago/guide/museums/chicago-museums.php

99. Exhibits | Museum of Illusions Chicago, https://moichicago.com/exhibits/

100. This Immersive Illinois Museum Is One Of The Most Interesting, https://everafterinthewoods.com/this-immersive-illinois-museum-is-one-of-the-most-interesting-places-in-the-state/

101. WNDR Light Floor | WNDR Museum Chicago, https://www.wndrmuseum.com/installations/chicago/wndr-light-floor

102. SENSE | WNDR Museum Chicago, https://www.wndrmuseum.com/installations/chicago/sense-chicago

103. Museum of Illusions Chicago: Experience Mind-Bending Exhibits in, https://www.gochicago.com/tourist-attraction/museum-of-illusions-chicago/