Computational Geometry for Infinite Hypnotic 3D Environments: Real-Time Rendering and Mathematical Structures

Full report

On this page

Introduction to Abstract Continuous Visualization

The generation of infinitely scalable, mesmerizing 3D environments in real time requires a fundamental departure from traditional explicit polygonal rendering, commonly known as rasterization. To achieve continuous abstract motion, persistent visual depth, and mathematically coherent seamless looping, procedural graphics rely entirely on implicit representations of space. In these systems, geometry is not defined by discrete vertices and triangles, but by continuous mathematical functions evaluated per-pixel via a ray marching architecture1. The viewer is immersed in a scalar field where surfaces are defined by the zero-level set of a Signed Distance Function (SDF).

The objective in this specific domain is to induce a hypnotic, visually infinite state. This requires mathematical structures capable of continuous transformation without topological cuts, self-similar detail, and deep spatial recursion. The synthesis of such environments draws heavily from complex dynamics, differential geometry, four-dimensional topology, and non-Euclidean spaces. Through the manipulation of the coordinate domain and the careful integration of continuous fractal iterations, environments can be created that subvert human spatial intuition. This report exhaustively details the mathematical frameworks, distance estimation techniques, spatial transformations, and implementation strategies required to produce optimal hypnotic abstract visual experiences.

Core Rendering Paradigm: Sphere Tracing and Distance Estimation

To render implicit mathematical structures, the standard approach is sphere tracing, an iterative ray marching technique. A Signed Distance Function, [Figure omitted from source export], returns the shortest perpendicular distance from any arbitrary point [Figure omitted from source export] to the nearest surface3. Positive values indicate the point is outside the object, negative values indicate the interior, and exactly zero defines the boundary surface3.

Because the function guarantees that no surface exists within a sphere of radius [Figure omitted from source export] centered at [Figure omitted from source export], a ray can safely step forward by this exact distance without intersecting or bypassing geometry2. The standard sphere tracing algorithm in GLSL is expressed as follows, utilizing an iterative loop to accumulate distance along a directional vector:

 

 

 

OpenGL Shading Language

float raymarch(vec3 rayOrigin, vec3 rayDirection) {     float distanceTraveled \= 0.0;     for(int i \= 0; i \< MAX\_STEPS; i++) {         vec3 currentPosition \= rayOrigin \+ rayDirection \* distanceTraveled;         float distanceToScene \= map(currentPosition);         if(distanceToScene \< EPSILON) return distanceTraveled;          if(distanceTraveled \> MAX\_DIST) break;          distanceTraveled \+= distanceToScene;      }     return \-1.0; }

The economics of this raymarching technique differ fundamentally from rasterization. Every pixel runs its own iterative loop, each iteration re-evaluating every primitive in the distance mapping function2. This makes the technique ideal for ambient, slow-moving focus environments where perfect curvature and organic blending are prioritized over raw geometric throughput.

Lipschitz Continuity and Step Size Reduction

For sphere tracing to converge reliably and avoid visual artifacts, the SDF must be Lipschitz continuous with a Lipschitz constant [Figure omitted from source export]. This mathematical constraint implies that the rate of change of the distance field cannot exceed the physical distance moved in space5. If a mathematical operation, such as a coordinate deformation or a chaotic fractal iteration, distorts space such that the gradient exceeds 1, rays will overstep the surface, causing severe visual artifacts, "holes" in the geometry, and structural breakdown2.

When mapping highly chaotic structures, such as deep fractals or strange attractors, finding a global Lipschitz bound is often impossible. The rate of change fluctuates wildly depending on the local region of space. A common mitigation strategy is step size reduction, multiplying the distance estimate by a safety factor (e.g., [Figure omitted from source export])8. While this prevents overstepping, it significantly increases the number of marching steps required, degrading real-time performance.

More advanced implementations utilize generalized Lipschitz tracing or Harnack bounds. These frameworks compute local directional Lipschitz upper bounds via polynomial approximations, allowing the algorithm to dynamically adjust step sizes to maximize performance while preventing overstepping5. By reasoning about harmonic functions and shifting the local field, the step size can grow in proportion to the unbounding radius without risking a surface collision, maintaining mathematical coherence even in heavily distorted spatial domains9.

Analytical Gradients via Dual Numbers and Jacobians

Complex recursive systems like the Mandelbox and Mandelbulb require evaluating the length of the gradient to formulate an accurate distance estimator. The distance estimate is generally defined as [Figure omitted from source export], where [Figure omitted from source export] is the escape length of the function10. Calculating the exact Jacobian matrix iteratively through numerical finite differences (sampling the space at multiple offset points) introduces precision errors and severe GPU performance penalties.

A superior mathematical approach utilizes automatic differentiation via dual numbers11. A dual number extends the real numbers and takes the form [Figure omitted from source export], where [Figure omitted from source export] is a dual unit such that [Figure omitted from source export]. Evaluating a function with a dual number automatically carries the exact analytical derivative through the calculation. Using Taylor series expansion, it is established that [Figure omitted from source export] because the higher-order terms vanish13.

By tracking dual vectors inside a [Figure omitted from source export] matrix during the fractal iteration loop, the exact running scalar derivative or Jacobian norm is produced10. For example, the gradient magnitude [Figure omitted from source export] is calculated natively alongside the spatial position, yielding the final exact distance estimate [Figure omitted from source export]10. This guarantees a mathematically precise distance estimator without arbitrary finite difference step sizes, eliminating visual noise and providing immense stability for deep zooming10.

Coordinate Domain Manipulation and Spatial Folding

The foundation of infinite perception relies on folding and manipulating the coordinate domain before the distance function is evaluated. Domain manipulation is computationally inexpensive yet perceptually powerful, making it highly suitable for real-time environments targeting fluid, seamless looping.

Cartesian Modular Repetition

Infinite expanses are generated by taking the modulus of the spatial coordinates, essentially folding infinite Euclidean space into a single fundamental geometric domain14. When a ray enters the domain boundary, the mathematical operation teleports it back to the opposing side, yet the visual output remains continuous.

 

 

 

OpenGL Shading Language

vec3 domainRepetition(vec3 p, vec3 spacing) {     return mod(p \+ 0.5 \ spacing, spacing) \- 0.5 \ spacing; }

To ensure continuity at the cell boundaries and avoid visible cuts, the geometry inside the domain must be completely symmetric across all axes, or the space must be folded using mirrored repetition to prevent seams. By continuously translating the camera through this modular space, the viewer experiences infinite forward motion. When the camera translates by exactly one module size over a fixed duration, the animation loops flawlessly.

Polar and Log-Polar Mappings

To create the perception of falling down an infinite tunnel or zooming endlessly into a fractal structure, polar and log-polar coordinate transformations are utilized15. Converting a two-dimensional Cartesian slice of space into polar coordinates [Figure omitted from source export] wraps the plane into a radial configuration, turning straight parallel lines into radiating sunbursts or concentric rings.

A log-polar transform is defined in the complex plane as [Figure omitted from source export]15. This maps concentric circles in the Cartesian plane to vertical lines in the logarithmic plane, and radial rays to horizontal lines16. The perceptual strength of this mapping is unparalleled for hypnotic visualizers. If domain repetition is applied after a log-polar transform, a simple linear translation along the logarithmic axis corresponds to an exponential scaling (zooming) in the original Cartesian space16. By animating this linear translation phase, the geometry scales indefinitely, producing an optically flawless infinite zoom loop without the catastrophic floating-point precision breakdown that normally accompanies deep scaling.

Continuous Curves and Topological Surfaces

To generate a hypnotic environment, the base structures must eschew sharp polygons in favor of continuous, weaving, or mathematically paradoxical topologies. These forms must exhibit continuous transformation without visual cuts.

Spirals: Logarithmic and Archimedean

Spirals are fundamental to inward and outward flow perception. The visual effectiveness of a spiral depends heavily on its mathematical formulation. The Archimedean spiral, defined in polar coordinates as [Figure omitted from source export], maintains a constant distance between successive turns. When animated through rotation, it produces a mesmerizing, steady radial outward flow, ideal for flat, hypnotic disk environments.

Conversely, the logarithmic spiral, defined as [Figure omitted from source export], exhibits self-similarity; it appears identical at all scales16. Because the distance between turns increases exponentially, rotating a logarithmic spiral induces a visceral sensation of infinite forward tunneling or deep inward falling. When mapped into three dimensions as a helical structure along a log-polar extruded tube, it forms the perfect basis for an infinite tunnel perception.

Toroidal Paths, Helices, and Twisted Tori

A torus provides the simplest closed framework for a continuous looping camera path. By defining the coordinate space around the medial axis of a torus, a scene can loop seamlessly over a set temporal duration17. Helices provide infinite vertical or axial progression.

Twisted tori and complex knot topologies (such as torus knots or trefoil knots defined via parametric equations turned into tubular SDFs) create complex interwoven pathways19. An effective technique for geometric surprise involves sweeping an arbitrary 2D cross-section along a 3D toroidal path while applying a slow rotation to the cross-section. As the user watches, the walls of the tunnel twist endlessly around the camera, generating the sensation of spiraling without any actual physical displacement of the camera origin.

Möbius-like Surfaces

Möbius strips represent non-orientable manifolds characterized by a single continuous surface and a single boundary edge17. In a volumetric raymarching context, a Möbius strip can be constructed by applying a domain twist along a toroidal axis18.

When particles, organic textures, or secondary recursive geometries are projected onto this surface, their behavior becomes mathematically paradoxical. An element moving along the surface must travel twice the circumference of the loop before returning to its original state and orientation. Animating the flow of abstract geometry over a Möbius manifold induces deep geometric surprise and subverts human spatial expectations, forcing the viewer's eye to follow paths that defy standard Euclidean topology.

Harmonic Oscillations: Lissajous and Rose Curves

Lissajous curves are defined by parametric harmonic equations, mapping two oscillating wave motions against one another: [Figure omitted from source export] When expanded into three dimensions and utilized as the medial spine of a volumetric tubular SDF, Lissajous curves create highly complex, oscillating webs22. If the frequency ratios ([Figure omitted from source export]) are irrational or properly tuned, the curves weave through space without self-intersecting22.

The visual power of the Lissajous structure lies in the temporal phase shift [Figure omitted from source export]. Animating this phase causes the entire tubular lattice to continuously reshape itself, passing seamlessly through phases of extreme density and sudden, spacious openness24. Rose curves, defined by the polar equation [Figure omitted from source export], alongside epicycloids and hypocycloids, function similarly25. They introduce sharp, cusp-like inversions within the continuous flow, adding aggressive structural articulation to an otherwise smooth environment25.

Four-Dimensional Projections and the Hopf Fibration

One of the most potent mathematical tools for generating mesmerizing motion is the rigid rotation of higher-dimensional objects projected down to our three-dimensional visual space. Because the human brain cannot natively process four spatial dimensions, the resulting 3D shadows behave in ways that appear organically impossible.

Stereographic Projection and the Hypersphere

To visualize a four-dimensional object in 3D, stereographic projection maps the hypersphere [Figure omitted from source export] embedded in [Figure omitted from source export] down to [Figure omitted from source export]20. Given a 4D point [Figure omitted from source export] on the unit hypersphere, the projection to 3D Cartesian coordinates is defined as:

[Figure omitted from source export]

As the four-dimensional object undergoes a rigid rotation (computed using rotor matrices or unit quaternions), the projected 3D object undergoes non-linear, mesmerizing deformations26. Objects appear to expand infinitely from the origin, flow outward to the horizon, and seamlessly return from the opposite direction, creating a perception of impossible spatial continuity20. Four-dimensional rotations are generally "double rotations," composed of two simultaneous simple rotations around orthogonal planes, allowing for immensely complex but flawlessly smooth geometric morphing27.

The Hopf Fibration

The Hopf fibration is a specific mapping from the 3-sphere [Figure omitted from source export] to the ordinary 2-sphere [Figure omitted from source export] where every point on the 2-sphere corresponds to a distinct great circle (a fiber) in [Figure omitted from source export]20. Under stereographic projection into 3D space, these fibers map to a set of perfectly linked, nested tori known as Villarceau circles.

Animating a rotation in the 4D space causes these interlocked rings to flow endlessly through one another in a perfectly smooth, cyclic symmetry26. Because no two fibers ever intersect, the visual output is a dense, non-colliding weave of geometry that turns itself inside out perpetually. Implementing this via a GLSL shader requires minimal overhead—just a few matrix multiplications to rotate the 4D coordinate before applying the stereographic projection—making it highly efficient on the GPU yet perceptually profound26.

Recursive Geometries and 3D Fractals

Fractals natively fulfill the requirement for infinite self-similar detail and nested depth. However, because standard complex numbers only exist in two dimensions, producing 3D fractals in real-time requires specific algebraic extensions and spatial folding techniques.

The Mandelbulb

The Mandelbulb is constructed using a three-dimensional spherical analog of the complex multiplication used in the traditional 2D Mandelbrot set ([Figure omitted from source export])10. By converting Cartesian coordinates to spherical coordinates [Figure omitted from source export], raising the radius to a scalar power (usually 8), and multiplying the angles, the algorithm produces highly detailed, organic-looking bulbs29.

 

 

 

OpenGL Shading Language

float r \= length(z); float theta \= acos(z.z / r) \ Power; float phi \= atan(z.y, z.x) \ Power; z \= pow(r, Power) \ vec3(sin(theta)\cos(phi), sin(theta)\sin(phi), cos(theta)) \+ c; dr \= pow(r, Power \- 1.0) \ Power \* dr \+ 1.0;

The Mandelbulb is highly effective for visual recursion and nested depth, particularly when zooming endlessly into its crevices. However, because the inner loop relies on expensive trigonometric functions (acos, atan, sin, cos), it presents a high GPU cost29. Furthermore, its geometry becomes noisy and numerically unstable at high iterations, requiring careful tuning of the distance estimator and bounding volumes to prevent visual tearing.

The Mandelbox and Spatial Folding

In contrast to the trigonometric scaling of the Mandelbulb, the Mandelbox relies on conditional geometric folding12. It alternates between a boxFold (which reflects points outside a cubic boundary back inwards) and a sphereFold (which inverts points inside a spherical boundary outwards), followed by a uniform spatial scale and translation12.

The result is a heavily architectural, biomechanical landscape of infinite nested cubes, intersecting spheres, and alien circuitry. The true visual power of the Mandelbox is unlocked through temporal phase changes: animating the scaling factor or the folding thresholds smoothly morphs the Mandelbox between rigid, grid-like structures and organic, flowing forms12.

Kaleidoscopic Iterated Function Systems (KIFS)

Kaleidoscopic IFS generate classic recursive fractals (such as Sierpinski tetrahedra, Menger sponges, and nested Platonic solids) by reflecting space across hyperplanes using the absolute value function: p \= abs(p)32.

Introducing arbitrary 3D rotation matrices between the folding and scaling steps perturbs the symmetry, transforming rigid mathematical forms into flowing, organic matrices33. These are exceptionally efficient on the GPU as they lack trigonometric functions in the iterative loop, relying solely on matrix multiplication and sign flipping. By flying the camera through a KIFS lattice while rotating the internal folding matrices, the viewer experiences a continuous transformation without cuts, traversing a landscape that constantly rebuilds itself.

Quaternion Julia Sets

A four-dimensional Julia set utilizes full quaternion algebra ([Figure omitted from source export]). The equation [Figure omitted from source export] is evaluated in 4D space, and a 3D slice is rendered by setting one dimension (typically the real axis or one imaginary axis) to a constant or zero34.

By slowly animating the fourth-dimensional component ([Figure omitted from source export]) of the constant [Figure omitted from source export], the 3D shape undergoes impossible topological shifts. Because the observer is sliding a 3D window through a 4D object, solid components separate, merge, spawn out of thin air, and dissolve organically35. Utilizing unbounding volumes and accurate analytical distance estimators ensures these complex forms render cleanly at high frame rates34.

Strange Attractors and Chaotic Dynamics

Strange attractors, such as the Lorenz, Rössler, and Rikitake systems, map the chaotic evolution of a continuous dynamical system over time37. Rather than being evaluated via exact SDFs—which is analytically impossible for most chaotic systems due to their non-Lipschitz nature—they are rendered either by marching through a pre-calculated 3D volumetric density grid (using compute shaders) or by simulating millions of particles that coalesce along the attractor's manifold39.

For abstract visual environments, strange attractors are best used as underlying vector fields that drive the movement of light, particles, or domain warping. By applying a smooth, slow rotation to the system parameters (e.g., modulating the Prandtl number in the Lorenz equations), the attractor physically reshapes. This produces a hypnotic inward and outward flow of visual information, resembling magnetic field lines or chaotic fluid dynamics, ensuring the scene never visibly repeats39.

Non-Euclidean and Hyperbolic Geometries

Generating a feeling of impossible spatial continuity is natively achieved by replacing standard Euclidean space with curved metric spaces.

Hyperbolic Tilings (Poincaré Disk and Ball)

In the hyperbolic plane ([Figure omitted from source export]) and space ([Figure omitted from source export]), the parallel postulate fails: parallel lines diverge, and the volume of space expands exponentially rather than polynomially41. Visualized through the Poincaré ball model, infinite hyperbolic space is mathematically compressed into a finite Euclidean sphere42.

As the virtual camera translates, objects shrink logarithmically as they approach the boundary. Using distance estimators derived from the Minkowski hyperboloid model ([Figure omitted from source export])41, ray marching through a hyperbolic honeycomb produces an endlessly recursive, non-repeating yet mathematically perfect grid. To the viewer, it feels like flying through an infinite crystal lattice where the edges of the universe are always visible but unreachable, creating a profound sense of depth and geometric recursion.

Thurston Geometries

Thurston's geometrization theorem categorizes eight homogeneous 3-manifolds that form the building blocks of all possible 3D spaces. Geometries like Nil, Sol, and the universal cover of [Figure omitted from source export] exhibit severe anisotropic curvature44. Ray marching in these spaces cannot use straight Euclidean rays. Instead, rays must trace the actual geodesics of the curved metric tensor45.

The Nil geometry acts as a twisted product of [Figure omitted from source export] and [Figure omitted from source export]. Its geodesics spiral identically like coiled springs, creating mesmerizing, twisted vortex effects regardless of where the camera looks44. The Sol geometry contains planes isometric to the hyperbolic plane, but stretching in opposing directions. Visualizing Sol produces aggressive stretching and compression depending on the viewing angle, highly effective for disorienting, hypnotic motion46. Adapting a real-time engine to shoot geodesics requires solving systems of ordinary differential equations (ODEs) per ray, presenting a massive GPU load, but delivering visual topologies completely alien to human intuition45.

Temporal Dynamics, Morphing, and Loopability

To create interactive experiences where users deliberately opt in to view continuous abstract motion, the geometry must shift and flow without jarring cuts.

Exponential and Polynomial Smooth Blending

Transitioning between distinct geometric families (e.g., from a fractal lattice to a toroidal flow) is flawlessly achieved via smoothed minimum functions (smoothmin). Popularized in computational geometry by Iñigo Quílez, a polynomial smoothmin blends two distinct SDFs ([Figure omitted from source export]) seamlessly into a unified implicit surface38:

 

 

 

OpenGL Shading Language

float smin(float a, float b, float k) {     float h \= clamp(0.5 \+ 0.5 \ (b \- a) / k, 0.0, 1.0);     return mix(b, a, h) \- k \ h \* (1.0 \- h); }

By animating the interpolation parameter [Figure omitted from source export] or the smoothness threshold [Figure omitted from source export], disparate geometries merge like liquid metal or biological tissue. This operation guarantees [Figure omitted from source export] derivative continuity, meaning that the surface normals (and therefore lighting, ambient occlusion, and reflections) never pop or fracture38.

Phase Angles and Seamless Looping

Perfectly looping animations require all mathematically driven variables to return precisely to their initial state. Phase angles ([Figure omitted from source export]) passed into trigonometric functions drive translations, rotations, and folding thresholds. In modular spaces, translating the camera by exactly one module size [Figure omitted from source export] over duration [Figure omitted from source export] results in a perfect structural loop, tricking the brain into perceiving infinite forward motion through an endless corridor14. Similarly, continuous rotation of the stereographic projection pole in 4D creates a perpetual morph that flawlessly wraps back to its starting state, guaranteeing fixed-duration loopability.

Implementation Hurdles and Numerical Stability

Real-time implicit rendering pushes modern hardware to its architectural limits, presenting several unique computational challenges that must be mitigated.

Normal Estimation and High-Frequency Aliasing

Because implicit SDFs do not supply explicit surface normals, normals must be estimated via the spatial gradient: [Figure omitted from source export]. The standard numerical approach evaluates the SDF at four tetrahedrally offset points2. However, for high-frequency recursive structures like the Mandelbox or KIFS, the finite difference [Figure omitted from source export] must be dynamically scaled relative to the distance from the camera.

If [Figure omitted from source export] is fixed too large, normals are smoothed out, wiping away fractal detail and causing lighting to look plastic. If [Figure omitted from source export] is too small, floating-point precision loss introduces severe visual noise and black rendering artifacts11. Furthermore, highly repetitive geometries (like infinite grids or nested fractals) generate extreme Moiré patterns at a distance. Mitigating this requires abandoning the distance field approach for distant geometry and faking the structural detail with a smoothly bump-mapped function or analytic analytical filtering51.

Floating-Point Limits and Camera Clipping

In Euclidean ray marching, objects far from the origin suffer from standard IEEE 754 32-bit floating-point (FP32) degradation. As values grow large, the mantissa loses precision, causing vertices to jitter and space to structurally fracture. Domain repetition completely eliminates this issue. By keeping the camera logically anchored near the origin and wrapping the spatial coordinates backwards around the camera, infinite distance can be simulated with absolute precision14.

Mobile GPU Suitability: FP16 vs FP32 Constraints

Mobile GPUs heavily rely on 16-bit floating-point (FP16) precision to maximize memory bandwidth and power efficiency in tile-based rendering architectures52. Ray marching fractals in FP16 results in catastrophic failure due to numerical cancellation during the iterative folding phases; the tiny spatial steps required near fractal surfaces round to zero, causing the ray to halt prematurely.

Rendering these scenes on mobile requires strict data-type segmentation. The developer must maintain mathematical coordinates and distance evaluation strictly in FP32, while safely downcasting the resultant surface normals, ambient occlusion, lighting calculations, and color aggregation to FP16. Recent architectural studies in GPU rendering demonstrate the need to preemptively cull alpha and distance checks before deep mathematical logic engages on mobile tensor units, ensuring that only rays intersecting geometry trigger heavy FP32 computations52.

Structural Comparison Matrix

The following matrix evaluates the core mathematical structures based on specific rendering metrics required for continuous, hypnotic abstract experiences.

 

Applied Scene Concepts

Synthesizing the mathematical research above, the following ten concrete scene concepts are designed for real-time, interactive, mesmerizing graphical systems.

1. The Hopf Vortex

Structure: The Hopf Fibration projected stereographically to 3D.Execution: A series of interlaced Villarceau circles constructed as smooth tubular SDFs. By animating a slow rotation of the 4D space using a dual-quaternion matrix, the nested tori continuously turn inside out, expanding from the center and flowing outward infinitely. The viewer remains stationary, immersed in an impossible spatial flow. Polynomial smoothmin blending causes the independent fibers to melt into each other dynamically as they pass in close proximity, creating a liquid geometry effect.

2. Log-Polar Biomechanical Tunnel

Structure: A Log-Polar coordinate mapping applied to a Kaleidoscopic IFS.Execution: A rigid, sharply folded KIFS structure (such as a recursive Menger sponge) is subjected to a log-polar coordinate transformation. The result is a fractal that has been mathematically wrapped into an infinite cylindrical tunnel. Panning the virtual camera linearly along the Z-axis translates to a mesmerizing, exponentially accelerating zoom through a cavern of infinitely repeating, biomechanical fractal pillars.

3. Breathing Mandelbox Matrix

Structure: A Mandelbox with oscillating folding thresholds. Execution: A highly iterated Mandelbox rendered with exact dual-number analytical gradients to maintain crisp ambient occlusion and lighting11. The sphereFold radii and boxFold limits are bound to a low-frequency sine wave. As the thresholds shift over time, the entire landscape—resembling a vast, greebled alien city—unfolds and collapses into itself organically. It transitions seamlessly between rigid right angles and flowing, curved, metallic dunes without visual cuts.

4. Poincaré Tessellation Swarm

Structure: Hyperbolic tiling rendered within the Poincaré ball model. Execution: A cellular lattice structure rendered using the Minkowski metric tensor to simulate [Figure omitted from source export] hyperbolic space41. The geometry shrinks infinitely as it approaches the spherical boundary of the scene. The camera travels along a hyperbolic geodesic path. To the viewer, it feels like flying through an infinite crystal lattice where the edges of the universe are always visible but unreachable, creating a profound sense of nested depth and recursion.

5. Lissajous Phase Weave

Structure: Parametric Lissajous curves wrapped in a volumetric SDF tube.Execution: Multiple overlapping tubular structures driven by incommensurable Lissajous frequencies. The temporal phase variable [Figure omitted from source export] is continuously modulated. The tubes weave through one another without colliding, creating a dense, mesmerizing knot of metallic cables that slowly tighten into a dense sphere and then seamlessly unravel into a sparse, room-sized web, providing constant geometric surprise.

6. The Sol Geometry Corridor

Structure: Thurston's Sol geometry. Execution: A regular Cartesian lattice of glowing pillars mapped into a Sol manifold. Because light rays (geodesics) in Sol geometry stretch in one axis and compress in another46, as the camera rotates, the straight pillars severely deform, bending into hyperbolic arcs. Forward movement feels like swimming through a dense optical illusion where distances are deceptive, inducing a hallucinatory spatial vertigo ideal for an opt-in hypnotic visualizer.

7. Smoothmin Metamorphosis

Structure: Domain-repeated modular geometry with exponential smoothmin blending. Execution: An infinite Cartesian grid of spheres and an offset grid of tori occupy the same space via the modulus operator. Using an exponential smoothmin function, the two structures are blended. A slow spatial sine wave sweeps through the domain, shifting the balance factor from spheres to tori. The resulting visuals resemble ferrofluid or organic cell division, with droplets reaching out and seamlessly fusing with neighboring shapes in a continuous, flowing dance38.

8. Twisted Toroidal Ribbon

Structure: A twisting coordinate domain built around a torus medial axis. Execution: A single, wide, flattened box SDF is placed in a spatial domain that has been mathematically bent into a torus and twisted [Figure omitted from source export] times (creating a Möbius-like non-orientable surface)17. The camera is locked onto the surface of the ribbon, looking strictly forward. As the geometry is rotated along its medial axis, the environment twists endlessly around the viewer, creating a perpetual roller-coaster effect with no physical start or end point.

9. Volumetric Lorenz Flow

Structure: A Lorenz Strange Attractor rendered via compute shader volumetric marching. Execution: A dense field of glowing, semi-transparent participating media mapped to the continuous vector flow of a Lorenz attractor38. Raymarching through the density volume reveals two massive, swirling orbital wings. The core mathematical parameters of the attractor are slowly interpolated over time, causing the chaotic butterfly wings to flutter, shift, and redraw their internal magnetic lines in real time, guaranteeing an ever-changing inward and outward flow.

10. Quaternion Julia Deep Dive

Structure: A three-dimensional slice of a four-dimensional Quaternion Julia Set. Execution: The complex equation [Figure omitted from source export] is evaluated using full 4D quaternions35. The viewer observes a 3D cross-section of the 4D fractal manifold. By slowly animating the fourth-dimensional component ([Figure omitted from source export]) of the constant [Figure omitted from source export] or shifting the slicing hyperplane, the 3D shape undergoes impossible topological shifts. Utilizing unbounding volumes and accurate analytical distance estimators ensures these complex forms render cleanly, showcasing pieces of the fractal spawning out of thin air, merging, and dissolving in perfect mathematical harmony34.

Works cited

1. Numerical Methods for Ray Tracing Implicitly Defined Surfaces⇤, https://kev.town/raymarching-toolkit/media/implicit.pdf

2. Raymarching and Signed Distance Functions: Procedural 3D From, https://anirone.com/blog/general/raymarching-signed-distance-functions

3. Signed distance function \- Grokipedia, https://grokipedia.com/page/Signed\_distance\_function

4. How to Create Interactive, Droplet-like Metaballs with Three.js and, https://tympanus.net/codrops/2025/06/09/how-to-create-interactive-droplet-like-metaballs-with-three-js-and-glsl/

5. (PDF) Generalized Lipschitz Tracing of Implicit Surfaces, https://www.researchgate.net/publication/384259016\_Generalized\_Lipschitz\_Tracing\_of\_Implicit\_Surfaces

6. (PDF) GPU-Based Sphere Tracing for Radial Basis Function Implicits, https://www.researchgate.net/publication/263652385\_GPU-based\_Sphere\_Tracing\_for\_Radial\_Basis\_Function\_Implicits

7. Enhanced Sphere Tracing | Request PDF \- ResearchGate, https://www.researchgate.net/publication/265020476\_Enhanced\_Sphere\_Tracing

8. Non-linear sphere tracing for rendering deformed signed distance, https://www.researchgate.net/publication/337129886\_Non-linear\_sphere\_tracing\_for\_rendering\_deformed\_signed\_distance\_fields

9. Ray Tracing Harmonic Functions \- Mark Gillespie & Denise Yang, https://www.scribd.com/document/815158988/Ray-Tracing-Harmonic-Functions-Mark-Gillespie-Denise-Yang-3640

10. Distance Estimation | Syntopia, http://blog.hvidtfeldts.net/index.php/category/distance-estimation/

11. Fractals | Syntopia | Page 2, http://blog.hvidtfeldts.net/index.php/category/fractals/page/2/

12. Distance Estimated 3D Fractals (VI): The Mandelbox \- Syntopia, http://blog.hvidtfeldts.net/index.php/2011/11/distance-estimated-3d-fractals-vi-the-mandelbox/

13. Generative Art, 3D Fractals, Creative Computing | Page 3 \- Syntopia, http://blog.hvidtfeldts.net/index.php/page/3/

14. Is there a method to do ray marching style modulus repeat with, https://computergraphics.stackexchange.com/questions/4094/is-there-a-method-to-do-ray-marching-style-modulus-repeat-with-raytracing

15. Archaeology in the Digital Era \- CS \- Huji, https://www.cs.huji.ac.il/\~werman/Papers/file516092.pdf

16. Fractals/Computer graphic techniques/2D/exp \- Wikibooks, https://en.wikibooks.org/wiki/Fractals/Computer\_graphic\_techniques/2D/exp

17. Articles & Tutorials | mysimulator.uk, https://www.mysimulator.uk/articles/

18. module-ParametricFunctions – three.js docs, https://threejs.org/docs/pages/module-ParametricFunctions.html

19. Articulating Space: \- UCSB MAT, https://www.mat.ucsb.edu/Dissertations/Pablo\_Colapinto\_PhD\_Dissertation.pdf

20. (PDF) Meshview: Visualizing the Fourth Dimension \- ResearchGate, https://www.researchgate.net/publication/2849675\_Meshview\_Visualizing\_the\_Fourth\_Dimension

21. (PDF) Line Drawings from 3D Models \- ResearchGate, https://www.researchgate.net/publication/334603188\_Line\_Drawings\_from\_3D\_Models

22. Shotcut User Guide Table of Contents, https://www.shotcut.org/stockmedia/Shotcut%20User%20Guide.pdf

23. Curves \- GeoGebra Tutorial \- Malin Christersson's, http://www.malinc.se/math/geogebra/curvesen.php

24. Computer animation of swaying trees based on physical simulation, https://www.researchgate.net/publication/220251896\_Computer\_animation\_of\_swaying\_trees\_based\_on\_physical\_simulation

25. AKTU Catalogue 2026 Aug \- S K Kataria & Sons, https://skkatariaandsons.com/Catalogue/Aktu\_Catalogue\_pdf.pdf

26. MathBox² \- Acko.net, https://acko.net/blog/mathbox2/

27. GL4D: A GPU-based architecture for interactive 4D visualization, https://www.researchgate.net/publication/38015495\_GL4D\_A\_GPU-based\_architecture\_for\_interactive\_4D\_visualization

28. Efficient Methods for Learning Visual Multi-object 6D Pose, https://bonndoc.ulb.uni-bonn.de/xmlui/bitstream/handle/20.500.11811/13878/8768.pdf?sequence=2\&isAllowed=y

29. 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/

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

31. Distance Estimated 3D Fractals (Part VIII): Epilogue \- Syntopia, http://blog.hvidtfeldts.net/index.php/2012/05/distance-estimated-3d-fractals-part-viii-epilogue/

32. skills/skills/shader-dev/reference/fractal-rendering.md at main, https://github.com/MiniMax-AI/skills/blob/main/skills/shader-dev/reference/fractal-rendering.md

33. Folding Space II: Kaleidoscopic Fractals \- Syntopia, http://blog.hvidtfeldts.net/index.php/2010/06/folding-space-ii-kaleidoscopic-fractals/

34. 4D Quaternion Julia Set Ray Tracer, http://2008.sub.blue/blog/2009/9/20/quaternion\_julia.html

35. Building a Fractal Explorer in Flutter with Fragment Shaders \- Medium, https://medium.com/@d3xvn/building-a-fractal-explorer-in-flutter-with-fragment-shaders-1031a51f756b

36. REAL-TIME RENDERING OF COMPLEX FRACTALS, https://www.visgraf.impa.br/Data/RefBib/PS\_PDF/rt-gems-2/rt-gems-2.pdf

37. Python Programming in OpenGL \- University of Illinois, https://new.math.uiuc.edu/math198/public198/ipython/stanblank/PyOpenGL.pdf

38. Fractals and Mathematical GLSL \- Daniel Gray, https://www.danbgray.com/fractals-and-mathematical-glsl

39. morphogenesis-resources/README.md at main \- GitHub, https://github.com/jasonwebb/morphogenesis-resources/blob/main/README.md

40. XRT Renderer, http://xrt.wikidot.com/

41. Adapting Game Engines to Curved Spaces, https://d-nb.info/1248461312/34

42. Modelling brain connectomes networks: Solv is a worthy ... \- arXiv, https://arxiv.org/pdf/2407.16077?

43. Catchup results for Computer Vision and Pattern Recognition ... \- arXiv, https://arxiv.org/catchup/cs.CV/2026-05-20?abs=True\&page=1

44. Visualization of Nil, Sol, and ^ SL2(R) Geometries \- Visgraf Lab, https://www.visgraf.impa.br/Data/RefBib/PS\_PDF/cag2020a/Visualization\_of\_Nil\_Sol\_SL2\_CAG.pdf

45. (PDF) Geodesics and geodesic spheres in SL(2,ℝ) ˜ geometry, https://www.researchgate.net/publication/264917439\_Geodesics\_and\_geodesic\_spheres\_in\_SL2R\_geometry

46. Non-Euclidean Virtual Reality IV: Sol \- arXiv, https://arxiv.org/html/2002.00369v1

47. \[2010.15801\] Ray-marching Thurston geometries \- arXiv, https://arxiv.org/abs/2010.15801

48. (PDF) Adapting Game Engines to Curved Spaces \- ResearchGate, https://www.researchgate.net/publication/355181831\_Adapting\_Game\_Engines\_to\_Curved\_Spaces

49. Gallery — Kuro, https://kuro.page/gallery.html

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

51. GLSLで超リアルなエフェクトを実装してみよう！実例10選 \- Workship, https://goworkship.com/magazine/glsl-shader-effect/

52. GSArch: Breaking Memory Barriers in 3D Gaussian Splatting, https://www.researchgate.net/publication/390592362\_GSArch\_Breaking\_Memory\_Barriers\_in\_3D\_Gaussian\_Splatting\_Training\_via\_Architectural\_Support

53. mwalczyk/hopf: 🧣 A graphical program for exploring the Hopf fibration., https://github.com/mwalczyk/hopf