DeepSite - 🧬 RemixMaster computer graphics concepts through interactive learning
Click to reveal details
A collection of vertices, edges and faces that defines the shape of a 3D object. Most common in real-time applications like games.
Elements: Vertices, Edges, Polygons (usually triangles or quads)
Optimization: Level of Detail (LOD), Normal maps
Click to reveal details
Non-Uniform Rational B-Splines are mathematical representations for modeling curves and surfaces in 3D space, offering precise control.
Advantages: Smooth surfaces, Precise control, Scalable
Uses: Automotive design, Industrial modeling
Click to reveal details
A method of representing smooth surfaces by specifying a coarse control mesh and applying subdivision rules to create a smoother surface.
Types: Catmull-Clark, Loop, Doo-Sabin
Used in: Character modeling, Organic shapes
Click to reveal details
The sequence of steps that transforms 3D models into 2D images on the screen. Includes vertex processing, rasterization, and fragment processing.
Stages: Vertex Shader, Primitive Assembly, Rasterization, Fragment Shader
Optimization: Culling, Instancing, Batching
Click to reveal details
Mathematical models that determine how light interacts with surfaces. Common models include Lambertian, Phong, Blinn-Phong, and PBR.
Components: Diffuse, Specular, Ambient, Emissive
Advanced: Physically Based Rendering (PBR), Subsurface Scattering
Click to reveal details
The process of applying 2D images to 3D surfaces to add detail without increasing geometric complexity.
Map Types: Diffuse, Normal, Specular, Displacement
Techniques: UV Mapping, Procedural Texturing, PBR Materials
Click to reveal details
Neural Radiance Fields represent scenes as continuous 5D functions that output color and density given a 3D location and 2D viewing direction.
Input: 3D position + 2D viewing direction
Output: RGB color + volume density
Click to reveal details
Instant Neural Graphics Primitives use hash encoding and small MLPs to achieve real-time NeRF rendering with significantly faster training.
Key Features: Hash grids, Multi-resolution, Real-time
Performance: Seconds to minutes training time
Click to reveal details
Extensions of NeRFs that can model dynamic scenes with moving objects or changing viewpoints over time.
Approaches: Time-conditioned, Deformation fields, Latent codes
Applications: Free-viewpoint video, VR/AR
Click to reveal details
Point-based representation where each point is an anisotropic 3D Gaussian with position, covariance, color and opacity attributes.
Properties: Position, Covariance, Color (SH), Opacity
Rendering: Differentiable, Tile-based rasterization
Click to reveal details
The rendering process for Gaussian splats involves projecting 3D Gaussians to 2D, sorting them, and alpha-blending them back-to-front.
Steps: Projection, Tile assignment, Sorting, Blending
Optimization: CUDA acceleration, Level of Detail
Click to reveal details
Comparison between explicit (splatting) and implicit (NeRF) scene representations for novel view synthesis.
Splatting: Explicit, Fast rendering, Large scenes
NeRF: Implicit, High quality, View-dependent effects
Click to reveal details
The most basic form of ray tracing that shoots rays from the camera through each pixel to find the nearest intersection with objects.
Characteristics: Primary rays only, No reflections/refractions
Usage: Early 3D games, Simple ray tracers
Click to reveal details
Extension of ray casting that recursively traces secondary rays for reflections, refractions, and shadows to achieve more realistic images.
Ray Types: Primary, Shadow, Reflection, Refraction
Optimization: Ray trees, Russian roulette
Click to reveal details
A Monte Carlo method for solving the rendering equation by randomly sampling light paths to simulate global illumination.
Features: Unbiased, Physically accurate, Slow convergence
Optimizations: Importance sampling, Next event estimation
Click to reveal details
Techniques for displaying 2D projections of 3D volumetric data sets, typically used for medical imaging and scientific visualization.
Methods: Ray casting, Texture slicing, Shear-warp
Data Types: CT scans, MRI, Scientific simulations
Click to reveal details
Mappings from scalar values in volume data to optical properties like color and opacity, crucial for meaningful visualization.
Types: 1D, 2D, Multi-dimensional
Design: Histogram-based, Interactive widgets
Click to reveal details
The process of extracting surfaces of constant value (isosurfaces) from volumetric data, typically using Marching Cubes algorithm.
Algorithms: Marching Cubes, Marching Tetrahedra, Dual Contouring
Applications: Medical imaging, CFD visualization
Click to reveal details
The process of converting vector graphics (points, lines, triangles) into a raster image (pixels) for display on a screen.
Pipeline: Vertex processing, Primitive assembly, Fragment processing
Optimizations: Early Z, Hierarchical Z, Tile-based
Click to reveal details
A depth buffering technique that stores the depth of each pixel and only draws fragments that are closer than what's already stored.
Advantages: Simple, Handles arbitrary geometry
Variants: W-buffer, Reverse Z, Floating-point Z
Click to reveal details
A rendering technique that separates geometry processing from lighting calculations to handle many lights efficiently.
Steps: G-buffer creation, Light accumulation, Post-processing
Tradeoffs: Memory intensive, No MSAA, Many lights
Made with DeepSite - 🧬 Remix