9. Texturing, Sampling, and Image-Based Rendering

Introduction and Context

A geometric model describes the shape of an object, but geometry alone rarely produces a convincing image. Real surfaces contain colour variation, fine-scale relief, gloss, stains, scratches, woven patterns, and other visual properties that would be expensive or impractical to represent entirely with additional polygons. Texturing provides a way to attach such information to geometry. A texture may specify colour, opacity, surface normal, roughness, metallic response, displacement, or illumination-related data.

Texturing is closely connected with sampling. A continuous surface must be represented by discrete texture elements, and a continuous image must be reconstructed on a discrete pixel grid. Incorrect sampling produces aliasing: jagged edges, shimmering patterns, moiré effects, and unstable detail during animation. Filtering and mipmapping address these problems by controlling how texture information is evaluated at different image resolutions.

Image-based rendering extends the use of images beyond surface decoration. Environment maps represent distant surroundings, image-based lighting uses captured illumination to shade objects, and high-dynamic-range imaging preserves the large range of brightness found in real scenes. Together, these techniques support interactive graphics, film production, virtual reality, games, scientific visualization, product design, and digital twins.

The central engineering challenge is to preserve visual detail while maintaining predictable performance, memory use, physical plausibility, and temporal stability. This chapter develops the mathematical and practical foundations needed to make informed choices among texture representations, filtering methods, parameterization strategies, and image-based rendering pipelines.

Learning Objectives

After studying this chapter, a student should be able to:

1. Explain texture mapping, UV parameterization, texture coordinates, and common texture representations.

2. Analyse how sampling, reconstruction, filtering, and aliasing affect rendered images.

3. Select and justify bilinear, trilinear, anisotropic, and mipmap-based filtering methods.

4. Distinguish colour, normal, roughness, metallic, displacement, and ambient-occlusion maps.

5. Describe procedural texturing, environment mapping, image-based lighting, compositing, and HDR workflows.

6. Identify practical sources of texture distortion, seams, shimmering, memory waste, and colour errors.

7. Apply a structured texturing pipeline to a realistic graphics problem.

8. Evaluate professional, ethical, and technical considerations in the use of captured imagery and generated content.

From Surface Geometry to Surface Appearance

A texture is a function defined over a domain, commonly a two-dimensional coordinate space. For a surface point \(p\), texture evaluation can be written as

\[

T(u,v) = c,

\]

where \((u,v)\) are texture coordinates and \(c\) may be a colour, scalar value, vector, or complete material description. The renderer first determines the surface point visible at a pixel, obtains its texture coordinates, and samples one or more texture maps before evaluating the illumination model.

The most familiar map is the albedo or base-colour map. It describes the intrinsic reflected colour under a specified colour-management convention. It should not normally contain strong lighting, shadows, or highlights, because those effects belong to the illumination model. Other maps carry different physical or artistic properties:

• A normal map modifies the direction used for lighting without changing the actual silhouette.
• A roughness map controls the spread of specular reflection.
• A metallic map distinguishes conducting and non-conducting material behaviour in a physically based workflow.
• A height or displacement map represents geometric variation.
• An opacity map controls visibility or blending.
• An emissive map identifies areas that contribute self-emitted light.
• An ambient-occlusion map approximates local visibility and is often used cautiously as an artistic or auxiliary control.

Separating these properties improves authoring, reuse, and physically based rendering. It also prevents a common mistake: encoding all visual information into a single colour image, which makes material response incorrect when the lighting changes.

UV Parameterization

A UV parameterization maps a three-dimensional surface to a two-dimensional texture domain. For a point \(p(x,y,z)\) on a mesh, a parameterization assigns coordinates \((u,v)\). The mapping is rarely perfect because a curved or topologically complex surface cannot generally be flattened without distortion.

A practical UV workflow divides the surface into charts or islands, cuts the mesh along selected seams, unfolds the charts, and packs them into the texture rectangle. The quality of a parameterization depends on several factors:

• Stretching changes the apparent scale of texture features.
• Angular distortion changes the shape of patterns.
• Seams may become visible if neighbouring texels do not agree.
• Packing efficiency determines how much of the texture memory is useful.
• Texel density determines the number of texture samples allocated per unit of surface area.

Cylindrical, spherical, planar, box, and projection-based mappings are useful for simple objects. Complex assets generally require manually controlled seams or automated parametrization algorithms. A good unwrap places seams where they are visually inconspicuous, follows natural construction boundaries, and allocates resolution according to visual importance.

Texture coordinates may extend beyond the nominal interval [0,1]. Depending on the addressing mode, such coordinates may repeat, clamp to an edge, mirror, or use a border colour. Repetition is effective for tiles and fabrics, while clamping is appropriate for a single image intended to cover a surface only once.

Tangent Space and Normal Mapping

A tangent-space normal map stores a normal relative to a local basis consisting of tangent, bitangent, and geometric normal vectors. At a surface point, the tangent-space normal is transformed into world or view space before the lighting calculation. This permits high-frequency detail to be represented without increasing polygon count.

The tangent basis must be generated consistently between the modelling, baking, and rendering tools. Differences in handedness, vertex splitting, UV orientation, or normalisation conventions can create seams and shading errors. Normal maps are directional data, not ordinary colour data; they must be interpreted in a suitable linear/vector representation and not subjected to inappropriate colour transformations.

Normal mapping cannot change the actual silhouette or cast accurate self-shadowing for large features. Displacement mapping changes the surface position and can affect silhouettes, but it requires suitable tessellation or a geometry-processing stage. Parallax and relief techniques provide an intermediate solution by changing the apparent intersection of view rays with a height field, although they are view-dependent approximations.

Sampling: From Continuous Surfaces to Pixels

Rendering involves several sampling operations. The camera samples a projected scene onto pixels; rasterization samples geometric coverage; texture lookup samples a texture function; and image display samples or resamples the final image. The sampling theorem states, in simplified form, that a band-limited signal can be reconstructed from samples when the sampling frequency exceeds twice its highest frequency. Real textures are not perfectly band-limited, and projected texture frequency varies over the surface, so aliasing is a persistent problem.

Consider a checkerboard viewed at a distance. Many checker cells may project into one pixel. Selecting one texel gives an unstable answer: a slight camera movement may switch the chosen texel, producing flicker. The desired pixel value is closer to the average of the texture over the pixel footprint. This observation motivates prefiltering and filtered texture lookup.

The projected footprint is the region of texture space corresponding to a screen pixel. It may be approximately square, rectangular, or highly elongated. Derivatives of texture coordinates with respect to screen coordinates, often represented by \(\partial u/\partial x\), \(\partial v/\partial x\), \(\partial u/\partial y\), and \(\partial v/\partial y\), help the renderer estimate this footprint.

Filtering Methods

Nearest-neighbour filtering selects the closest texel. It is inexpensive and preserves hard pixel boundaries, making it useful for pixel art, masks, or diagnostic views. For ordinary photographic textures it causes blockiness and temporal instability.

Bilinear filtering computes a weighted interpolation of the four nearest texels. It produces smooth magnification but does not solve minification aliasing. It may also blur detail when a large area of texture is compressed into a small number of pixels.

Mipmap filtering stores a sequence of progressively smaller images. Each level is commonly produced by reducing the previous level, often by a box or higher-quality filter. The renderer selects a level whose texel scale approximately matches the projected footprint and then performs bilinear filtering within that level. This is called trilinear filtering when interpolation also occurs between adjacent mip levels. Mipmaps improve cache behaviour, reduce aliasing, and provide predictable lookup cost, at the expense of additional memory, usually about one-third of the base level for a complete power-of-two pyramid.

Anisotropic filtering improves quality when the projected footprint is elongated, as on a road receding into the distance or a wall viewed at a shallow angle. It takes multiple samples along the major axis of the footprint. Higher anisotropy generally improves oblique detail but increases bandwidth and computation.

More sophisticated methods include summed-area tables, elliptical weighted average filters, stochastic filtering, and texture-space supersampling. These can provide better reconstruction for difficult footprints but may be less suitable for real-time pipelines because of memory, bandwidth, or implementation cost.

Aliasing and Temporal Stability

Aliasing appears as jagged silhouettes, crawling textures, moiré patterns, and flickering highlights. It can arise from geometry, texture detail, shader functions, shadow maps, or post-processing. Increasing resolution helps but does not remove the underlying problem. Anti-aliasing methods such as multisample anti-aliasing, temporal anti-aliasing, supersampling, and reconstruction filters address different portions of the rendering pipeline.

Temporal rendering introduces an additional trade-off. A filter that accumulates information across frames can reduce noise and aliasing, but motion vectors, disocclusions, changing materials, and camera cuts may cause ghosting. Texture content should therefore be designed not only for one frame but also for stable behaviour during movement.

Procedural Textures and Material Systems

A procedural texture is generated from mathematical functions, noise fields, simulation data, or rules rather than stored entirely as an image. Common building blocks include gradients, cellular patterns, turbulence, fractal noise, Voronoi distance fields, and coordinate transformations. Procedural textures can be resolution-independent, tileable, compact, and parameterically editable.

For example, a procedural wood material may combine directional coordinates, low-frequency colour variation, ring functions, and high-frequency noise. A terrain material may blend rock, soil, and vegetation according to altitude, slope, moisture, and noise. Procedural approaches are particularly valuable when large worlds require repetition-free detail or when artists must generate many controlled variants.

Their disadvantages include computational cost, difficulty of reproducing exact visual results, and possible scale inconsistency. A procedural function that looks convincing on a close-up object may appear noisy when applied to a large environment. Production systems often combine procedural masks with artist-painted textures and baked maps. This hybrid method provides both control and variation.

Material systems increasingly use physically based parameters rather than ad hoc colour adjustments. However, physical terminology does not guarantee physical correctness. Values must remain within plausible ranges, units and colour spaces must be consistent, and the chosen shading model must correspond to the intended material.

Environment Mapping and Image-Based Lighting

An environment map represents the radiance arriving from directions around a point, usually assuming that the environment is sufficiently distant. A common representation is the latitude-longitude, or equirectangular, image. Cube maps divide the sphere into six square faces and are efficient for hardware lookup. Other projections may be used for acquisition or storage.

In reflection environment mapping, a reflection vector selects a direction in the environment map. This produces convincing distant reflections on shiny objects but is an approximation: it usually assumes the reflected environment is infinitely far away and does not account for nearby objects, parallax, or self-reflection. For local environments, reflection probes or multiple spatially placed maps can reduce these errors.

Image-based lighting (IBL) uses an environment image as an illumination source. A high-dynamic-range environment can be integrated over the hemisphere to estimate diffuse irradiance and can be prefiltered for different roughness values to support specular reflection. In a physically based rendering pipeline, the environment may be decomposed into:

1. A diffuse irradiance representation for broad, low-frequency illumination.

2. A prefiltered environment for roughness-dependent specular reflection.

3. A two-dimensional lookup table or analytical approximation for view- and roughness-related terms.

The resulting lighting is efficient because expensive integrals are precomputed. Nevertheless, an environment map alone does not reproduce contact shadows or the precise illumination of nearby objects. It is commonly combined with local lights, shadowing, ambient-occlusion approximations, or ray tracing.

High-Dynamic-Range Imaging and Compositing

Ordinary display images often use a limited range of encoded values, while real scenes may contain illumination differences of many orders of magnitude. High-dynamic-range (HDR) imaging stores a broader range of radiance values, frequently using floating-point formats. HDR is essential for preserving bright light sources, windows, reflections, and subtle shadow detail during rendering and compositing.

A typical HDR workflow begins with multiple exposures or a physically based render. Exposure alignment and response calibration may be followed by merging into a radiance map. In computer graphics, an HDR image may be used as an environment light, a render target, or a compositing layer. Before display, tone mapping compresses the scene-referred range into the display range. Tone mapping is not merely brightness adjustment; it affects contrast, colour appearance, highlight roll-off, and artistic interpretation.

Colour management is critical. Linear-light values are appropriate for physically meaningful blending and lighting calculations, while display-referred encodings such as sRGB are designed for visual presentation. Adding or interpolating gamma-encoded values directly can produce incorrect results. Alpha compositing should generally be performed in linear light, with careful attention to whether alpha is straight or premultiplied. Premultiplied alpha is often safer for filtering and repeated compositing because colour values are already weighted by coverage.

The basic “over” operation for premultiplied foreground colour \(C_f\), background colour \(C_b\), and opacity \(\alpha_f\) is

\[

C_{out}=C_f+C_b(1-\alpha_f),

\]

with output opacity

\[

\alpha_{out}=\alpha_f+\alpha_b(1-\alpha_f).

\]

Incorrect alpha conventions cause dark or bright fringes around cut-out objects. HDR compositing also requires matching exposure, white balance, camera response, lens effects, and noise characteristics between layers.

A Production Workflow for Textured Assets

A reliable asset workflow is more important than any individual filter or map. A typical sequence is as follows.

First, the mesh is inspected for topology, scale, normals, and material regions. UV charts are then created with seams chosen according to visibility and deformation. Texel density is established so that assets in the same scene maintain a consistent apparent resolution.

Next, high-resolution geometric detail may be baked into normal, curvature, or ambient-occlusion maps for a lower-resolution real-time mesh. Texture maps are authored or generated in a linear-aware, physically based workflow. Channel packing may place several grayscale maps into different channels to reduce memory and texture fetches, provided that compression and filtering requirements are compatible.

The asset is then validated under several lighting conditions rather than under a single studio light. Validation should include close-up views, grazing angles, distance views, motion, and different exposure levels. Mip transitions, seams, normal orientation, texture colour space, and compression artefacts should be inspected.

For deployment, texture resolution and format are selected according to platform constraints. Streaming systems may load mip levels progressively. Virtual texturing divides large surfaces into pages and loads only visible portions, reducing the need to keep complete giant textures in memory. The cost is increased management complexity, page-table overhead, and possible visible pop-in if streaming cannot keep pace with camera motion.

Applied Case: Virtual Heritage Reconstruction

Background

Consider an illustrative university project that reconstructs a historic stone courtyard for an interactive virtual-reality application. The project combines photogrammetric geometry, photographs of masonry, scanned wood, and an HDR capture of the courtyard at different times of day.

Situation and Problem

The initial prototype looks detailed in still images but performs poorly in a headset. Stone textures shimmer when the user walks, mortar seams appear distorted around corners, and the courtyard looks unnaturally flat under changing illumination. Large photographic textures consume excessive memory, while direct use of photographs embeds shadows and makes relighting inconsistent.

Analysis

The problems have different causes. Shimmering results from undersampled high-frequency texture detail and inadequate minification filtering. Distortion is caused by poorly placed UV seams and uneven texel density. Flat appearance indicates insufficient normal and roughness variation. Photographic shadows in the albedo maps conflict with the real-time lighting model. Finally, the HDR capture is being used as a visible background but not as a properly filtered illumination source.

Approach and Solution

The team first creates a texel-density standard and re-unwraps major architectural surfaces, placing seams along corners and construction joints. Photographs are processed to remove strong cast shadows where possible, while separate maps are produced for albedo, roughness, and normal detail. High-resolution scans are baked onto optimized meshes.

Complete mipmap chains are generated with appropriate alpha handling. Anisotropic filtering is enabled for oblique masonry surfaces, and distant decorative detail is reduced through controlled filtering rather than simply increasing texture resolution. Repeating stone patterns are varied with procedural masks to avoid obvious tiling.

The HDR environment is converted into a cube-map representation. Diffuse irradiance and roughness-dependent specular prefilters are generated offline. Local shadowing and a small number of dynamic lights are retained to provide contact and directional cues. Tone mapping and colour management are standardized across the headset and authoring tools.

Outcome

The revised scene uses less memory, remains visually stable during movement, and supports consistent relighting. The geometry is simpler than the original scan, but the combination of normal detail, roughness variation, filtered textures, and image-based lighting produces a more convincing result. Performance testing is conducted at the target headset frame rate rather than only on a desktop monitor.

Learning Points

The case illustrates that texture quality is a pipeline problem. More pixels do not automatically solve aliasing, distortion, or incorrect material decomposition. UV design, filtering, colour management, lighting representation, and platform constraints must be considered together.

Technical Risks, Limitations, and Common Errors

A frequent error is confusing albedo with a lit photograph. Baked shadows and highlights restrict reuse and can produce double lighting. Another is treating normal, roughness, and metallic maps as sRGB colour images. Data maps require appropriate interpretation and often linear storage.

Poor UV packing wastes memory and may cause bleeding between islands during mip generation. Padding or dilation around islands is necessary because filtering samples neighbouring texels. Insufficient padding creates coloured halos at seams.

Texture resolution is not a substitute for correct sampling. A high-resolution image can still alias when viewed at a distance. Conversely, excessive resolution increases download time, memory usage, and energy consumption without improving the final image.

Environment maps have geometric limitations. They cannot automatically represent nearby occluders, correct parallax, or dynamic objects. Probe placement, blending, and update frequency must be selected according to scene scale and motion.

HDR pipelines can fail through clipped highlights, inconsistent exposure, incorrect gamma operations, or incompatible transfer functions. Automated asset conversion should therefore record colour space, channel semantics, units, compression format, and version information. Reproducibility is especially important in collaborative production.

Professional and Ethical Considerations

Texture libraries and HDR photographs may be protected by copyright, licence restrictions, privacy law, or cultural heritage requirements. A production team must document acquisition rights, permitted uses, attribution obligations, and whether captured environments contain identifiable people or private property. Digitizing heritage sites also raises questions of consent, ownership, representation, and long-term preservation.

Procedural and generative tools can accelerate material creation, but their outputs should be evaluated for hidden repetition, inappropriate training provenance, and misleading realism. In scientific or engineering visualization, texture should not imply measurements that were not actually observed. Decorative enhancement must remain distinguishable from data-bearing imagery.

From a professional perspective, asset metadata is part of quality assurance. Teams should retain source imagery, calibration information, map meanings, colour profiles, and transformation history. This supports debugging, auditing, and future reuse.

Current Developments

Modern graphics systems increasingly combine rasterization with ray tracing. Texture maps continue to provide efficient material data, while ray-traced visibility improves reflections, shadows, and indirect illumination. Neural texture compression and learned super-resolution seek to reduce storage and bandwidth, although they introduce model-dependence and validation challenges.

Virtual texturing, sparse image representations, and streaming-first pipelines are becoming important for expansive worlds. Material authoring is also moving toward node-based procedural systems that can produce variants from semantic parameters such as age, moisture, damage, and manufacturing process. Machine-learning methods can assist in image relighting, texture synthesis, inverse rendering, and material estimation from photographs. Their outputs remain dependent on training data, camera calibration, and the quality of physical assumptions.

Display technology is also evolving toward wider colour gamuts, higher peak brightness, and improved HDR. These developments make colour management more important, not less, because an asset that looks correct on one display may clip or appear overly dark on another.

Best Practices for Engineering-Quality Results

1. Define the material model and colour-management conventions before authoring assets.

2. Use UV seams strategically and maintain consistent texel density.

3. Separate surface properties into semantically meaningful maps.

4. Generate mipmaps with adequate island padding and inspect them in motion.

5. Use anisotropic filtering for elongated projected footprints, not merely higher base resolution.

6. Keep lighting out of albedo maps unless the deliberate artistic effect is documented.

7. Validate assets under varied exposure, distance, viewing angle, and illumination.

8. Treat normal, roughness, metallic, and mask data according to their channel semantics.

9. Use HDR and linear-light operations for lighting and compositing calculations.

10. Profile memory, bandwidth, shader cost, and streaming behaviour on the target platform.

11. Record licences, source data, colour profiles, map conventions, and processing history.

12. Combine procedural variation with artist or domain-expert review rather than relying on automation alone.

Chapter Summary

Texturing extends geometric models with colour and material information, while UV parameterization provides the correspondence between surface points and texture space. Because rendering samples continuous appearance on a discrete pixel grid, texture lookup must account for projected footprints and aliasing. Nearest, bilinear, trilinear, mipmap, and anisotropic filtering represent progressively more suitable strategies for different sampling conditions.

Normal, roughness, metallic, displacement, opacity, and emissive maps encode distinct surface behaviours. Procedural textures provide compact, editable, and scalable variation, while image-based methods capture complex environmental appearance. Environment mapping approximates distant reflections, and image-based lighting uses irradiance and prefiltered radiance to illuminate physically based materials. HDR and linear-light compositing preserve brightness information and prevent colour and alpha errors. In production, successful results depend on coordinated UV design, map authoring, filtering, colour management, asset validation, performance profiling, and responsible use of captured or generated imagery.

Key Terms

Texture: A sampled or procedurally generated function that defines surface appearance or material data.

Texel: A texture element corresponding to a discrete sample in a texture image.

UV Parameterization: A mapping from a three-dimensional surface to two-dimensional texture coordinates.

Texture Chart: A connected region of a mesh unfolded into texture space.

Texel Density: The number of texture samples allocated per unit of surface area.

Albedo: The intrinsic base colour or diffuse reflectance component of a material representation.

Normal Map: A map that stores modified surface-normal directions for lighting calculations.

Tangent Space: A local coordinate system defined by tangent, bitangent, and surface-normal directions.

Mipmapping: Storage of progressively reduced texture representations for efficient minification filtering.

Aliasing: Visual error caused by insufficient sampling of spatial or temporal detail.

Bilinear Filtering: Interpolation among the four nearest texels.

Anisotropic Filtering: Filtering that uses multiple samples to address elongated projected texture footprints.

Procedural Texture: Texture data generated from functions, rules, noise, or simulations.

Environment Map: An image representation of radiance surrounding a viewpoint or scene region.

Image-Based Lighting: Illumination derived from captured or rendered environment radiance.

HDR: High-dynamic-range representation capable of storing a broad range of scene brightness.

Tone Mapping: Compression of scene-referred luminance into a displayable range.

Premultiplied Alpha: An alpha representation in which colour values have already been multiplied by opacity.

Virtual Texturing: A technique that streams only visible texture pages rather than storing all texture data in active memory.

Colour Management: The controlled interpretation and conversion of colour values across devices and processing stages.

Review Questions

Short Answer Questions

1. What is the purpose of UV parameterization in texture mapping?

2. Why can a normal map add visual detail without increasing the mesh silhouette?

3. What problem does a mipmap hierarchy address?

4. Distinguish an albedo map from an emissive map.

5. Why should HDR compositing generally be performed in linear light?

Long Answer Questions

1. Explain the causes of texture aliasing and compare bilinear, trilinear, and anisotropic filtering.

2. Describe a complete UV and material-authoring workflow for a real-time asset.

3. Explain how image-based lighting uses diffuse irradiance and roughness-dependent specular prefiltering.

4. Discuss the advantages and limitations of procedural textures in large interactive environments.

5. Explain HDR image formation, tone mapping, colour management, and premultiplied-alpha compositing.

Analytical / Application Questions

1. A road texture appears sharp near the camera but shimmers and becomes unreadable in the distance. Diagnose the problem and recommend a filtering and mipmapping strategy.

2. A scanned building asset has visible seams, inconsistent scale, and dark halos around transparent windows. Identify likely causes and propose a corrective pipeline.

3. An interactive museum application uses an HDR courtyard image for reflections but objects appear flat and fail to receive contact shadows. Analyse the limitations of the approach and design a hybrid lighting solution.

22. Multiple Choice Questions with Answers

Question 1

A texture contains many repeated high-frequency details that project into a single pixel during minification. Which method most directly reduces the resulting aliasing?

A. Increasing the mesh polygon count only

B. Using a mipmap hierarchy with filtered lookup

C. Replacing bilinear filtering with nearest-neighbour filtering

D. Encoding the texture in a wider colour gamut

Correct Answer: B. Using a mipmap hierarchy with filtered lookup

Question 2

Why are UV charts divided by seams during parameterization?

A. To increase the physical polygon count

B. To flatten surface regions with manageable distortion

C. To convert normal maps into height maps

D. To eliminate the need for texture filtering

Correct Answer: B. To flatten surface regions with manageable distortion

Question 3

Which map primarily changes the direction used in lighting calculations without changing an object's actual silhouette?

A. Normal map

B. Opacity map

C. Emissive map

D. Albedo map

Correct Answer: A. Normal map

Question 4

A wall is viewed at a very shallow angle, and its texture appears blurry despite mipmapping. Which technique is most appropriate?

A. Anisotropic filtering

B. Nearest-neighbour filtering

C. Reducing the texture resolution

D. Disabling texture derivatives

Correct Answer: A. Anisotropic filtering

Question 5

What is a major limitation of a conventional distant environment map used for reflections?

A. It cannot store colour values

B. It assumes all reflected content is effectively distant and lacks local parallax

C. It can represent only diffuse materials

D. It prevents the use of HDR values

Correct Answer: B. It assumes all reflected content is effectively distant and lacks local parallax

Question 6

Which statement best describes the role of tone mapping?

A. It converts polygons into texture charts

B. It compresses scene-referred brightness into a displayable range

C. It generates tangent vectors from UV coordinates

D. It removes all illumination from an albedo map

Correct Answer: B. It compresses scene-referred brightness into a displayable range

Question 7

Why should roughness and normal maps not automatically be treated as ordinary sRGB colour images?

A. They contain data with semantic and numerical meanings rather than display colour

B. They are always larger than albedo maps

C. They cannot be filtered

D. They are used only for offline rendering

Correct Answer: A. They contain data with semantic and numerical meanings rather than display colour

Question 8

In the virtual heritage case, visible texture shimmer during camera movement was primarily addressed by which combination?

A. Higher polygon count and painted shadows

B. Mipmaps, suitable filtering, and controlled high-frequency detail

C. Removing all normal maps and using nearest filtering

D. Using only an emissive texture

Correct Answer: B. Mipmaps, suitable filtering, and controlled high-frequency detail

Question 9

What is a principal advantage of procedural texturing?

A. It guarantees physically correct illumination

B. It can provide compact, editable, and potentially resolution-independent variation

C. It eliminates the need for material validation

D. It always renders faster than a bitmap texture

Correct Answer: B. It can provide compact, editable, and potentially resolution-independent variation

Question 10

Which compositing practice helps avoid dark fringes around filtered transparent objects?

A. Compositing gamma-encoded values directly

B. Using incompatible alpha conventions in different layers

C. Using premultiplied alpha with consistent linear-light processing

D. Discarding opacity during mipmap generation

Correct Answer: C. Using premultiplied alpha with consistent linear-light processing