Exercises

  1. One nice property of mesh-based shapes like triangle meshes and subdivision surfaces is that the shape’s vertices can be transformed into world space, so that it isn’t necessary to transform rays into object space before performing ray intersection tests. Interestingly enough, it is possible to do the same thing for ray–quadric intersections. The implicit forms of the quadrics in this chapter were all of the form
    upper A x squared plus upper B x y plus upper C x z plus upper D y squared plus upper E y z plus upper F z squared plus upper G equals 0 comma
    where some of the constants upper A ellipsis upper G were zero. More generally, we can define quadric surfaces by the equation
    upper A x squared plus upper B y squared plus upper C z squared plus 2 upper D x y plus 2 upper E y z plus 2 upper F x z plus 2 upper G x plus 2 upper H y plus 2 upper I z plus upper J equals 0 comma
    where most of the parameters upper A ellipsis upper J don’t directly correspond to the earlier upper A ellipsis upper G . In this form, the quadric can be represented by a 4 times 4 symmetric matrix  bold upper Q :
    Start 1 By 4 Matrix 1st Row 1st Column x 2nd Column y 3rd Column z 4th Column 1 EndMatrix Start 4 By 4 Matrix 1st Row 1st Column upper A 2nd Column upper D 3rd Column upper F 4th Column upper G 2nd Row 1st Column upper D 2nd Column upper B 3rd Column upper E 4th Column upper H 3rd Row 1st Column upper F 2nd Column upper E 3rd Column upper C 4th Column upper I 4th Row 1st Column upper G 2nd Column upper H 3rd Column upper I 4th Column upper J EndMatrix Start 4 By 1 Matrix 1st Row x 2nd Row y 3rd Row z 4th Row 1 EndMatrix equals normal p Superscript upper T Baseline bold upper Q normal p Subscript Baseline equals 0 period
    Given this representation, first show that the matrix bold upper Q prime representing a quadric transformed by the matrix bold upper M is
    bold upper Q prime equals left-parenthesis bold upper M Superscript upper T Baseline right-parenthesis Superscript negative 1 Baseline bold upper Q bold upper M Superscript negative bold 1 Baseline period
    To do so, show that for any point normal p Subscript where normal p Subscript Baseline Superscript upper T Baseline bold upper Q normal p Subscript Baseline equals 0 , if we apply a transformation bold upper M to normal p Subscript and compute normal p prime equals bold upper M normal p Subscript , we’d like to find bold upper Q prime so that left-parenthesis normal p prime right-parenthesis Superscript upper T Baseline bold upper Q prime normal p Superscript prime Baseline equals 0 . Next, substitute the ray equation into the earlier more general quadric equation to compute coefficients for the quadratic equation a t squared plus b t plus c equals 0 in terms of entries of the matrix bold upper Q to pass to the Quadratic() function. Now implement this approach in pbrt and use it instead of the original quadric intersection routines. Note that you will still need to transform the resulting world space hit points into object space to test against theta Subscript normal m normal a normal x , if it is not 2 pi , and so on. How does performance compare to the original scheme?
  2. Improve the object space bounding box routines for the quadrics to properly account for phi Subscript normal m normal a normal x Baseline less-than 3 pi slash 2 , and compute tighter bounding boxes when possible. How much does this improve performance when rendering scenes with partial quadric shapes?
  3. There is room to optimize the implementations of the various quadric primitives in pbrt in a number of ways. For example, for complete spheres some of the tests in the intersection routine related to partial spheres are unnecessary. Furthermore, some of the quadrics have calls to trigonometric functions that could be turned into simpler expressions using insight about the geometry of the particular primitives. Investigate ways to speed up these methods. How much does doing so improve the overall run time of pbrt when rendering scenes with quadrics?
  4. Currently pbrt recomputes the partial derivatives partial-differential normal p slash partial-differential u and partial-differential normal p slash partial-differential v for triangles every time they are needed, even though they are constant for each triangle. Precompute these vectors and analyze the speed/storage trade-off, especially for large triangle meshes. How do the depth complexity of the scene and the size of triangles in the image affect this trade-off?
  5. Implement a general polygon primitive that supports an arbitrary number of vertices and convex or concave polygons as a new Shape in pbrt. You can assume that a valid polygon has been provided and that all of the vertices of the polygon lie on the same plane, although you might want to issue a warning when this is not the case. An efficient technique for computing ray–polygon intersections is to find the plane equation for the polygon from its normal and a point on the plane. Then compute the intersection of the ray with that plane and project the intersection point and the polygon vertices to 2D. Then apply a 2D point-in-polygon test to determine if the point is inside the polygon. An easy way to do this is to effectively do a 2D ray-tracing computation, intersect the ray with each of the edge segments, and count how many it goes through. If it goes through an odd number of them, the point is inside the polygon and there is an intersection. See Figure 3.47 for an illustration of this idea. You may find it helpful to read the article by Haines (1994) that surveys a number of approaches for efficient point-in-polygon tests. Some of the techniques described there may be helpful for optimizing this test. Furthermore, Section 13.3.3 of Schneider and Eberly (2003) discusses strategies for getting all the corner cases right: for example, when the 2D ray is aligned precisely with an edge or passes through a vertex of the polygon.
    Figure 3.47: A ray–polygon intersection test can be performed by finding the point where the ray intersects the polygon’s plane, projecting the hit point and polygon vertices onto an axis-aligned plane, and doing a 2D point-in-polygon test there.
  6. Constructive solid geometry (CSG) is a classic solid modeling technique, where complex shapes are built up by considering the union, intersection, and differences of more primitive shapes. For example, a sphere could be used to create pits in a cylinder if a shape was modeled as the difference of a cylinder and set of spheres that partially overlapped it. See Hoffmann (1989) for further information about CSG. Add support for CSG to pbrt and render images that demonstrate interesting shapes that can be rendered using CSG. You may want to read Roth (1982), which first described how ray tracing could be used to render models described by CSG, as well as Amanatides and Mitchell (1990), which discusses precision-related issues for CSG ray tracing.
  7. Procedurally described parametric surfaces: Write a Shape that takes a general mathematical expression of the form f left-parenthesis u comma v right-parenthesis right-arrow left-parenthesis x comma y comma z right-parenthesis that describes a parametric surface as a function of left-parenthesis u comma v right-parenthesis . Evaluate the given function at a grid of left-parenthesis u comma v right-parenthesis positions, and create a triangle mesh that approximates the given surface.
  8. Adaptive curve refinement: adjust the number of levels of recursive refinement used for intersection with Curve shapes based on the on-screen area that they cover. One approach is to take advantage of the RayDifferential class, which represents the image space area that a given ray represents. (However, currently, only Rays—not RayDifferentials—are passed to the Shape::Intersect() method, so you’d need to modify other parts of the system to make ray differentials available.) Alternatively, you could modify the Camera to provide information about the projected length of vectors between points in world space on the image plane and make the camera available during Curve creation.
  9. Almost all methods for subdivision surfaces are based on either refining a mesh of triangles or a mesh of quadrilaterals. If a rendering system only supports one type of mesh, meshes of the other type are typically tessellated to make faces of the expected type in a preprocessing step. However, doing this can introduce artifacts in the final subdivision surface. Read Stam and Loop’s paper on a hybrid subdivision scheme that supports meshes with both quadrilateral and triangular faces (Stam and Loop 2003), and implement their method. Demonstrate cases where the subdivision surface that your implementation creates does not have artifacts that are present in the output from pbrt’s current subdivision implementation.
  10. The smoothness of subdivision surfaces isn’t always desirable. Sometimes it is useful to be able to flag some edges of a subdivision control mesh as “creases” and apply different subdivision rules there to preserve a sharp edge. Extend the subdivision surface implementation in this chapter so that some edges can be denoted as creases, and use the boundary subdivision rules to compute the positions of vertices along those edges. Render images showing the difference this makes.
  11. Adaptive subdivision: a weakness of the subdivision surface implementation in Section 3.8 is that each face is always refined a fixed number of times: this may mean that some faces are underrefined, leading to visible faceting in the triangle mesh, and some faces are overrefined, leading to excessive memory use and rendering time. With adaptive subdivision, individual faces are no longer subdivided once a particular error threshold has been reached. An easy error threshold to implement computes the face normals of each face and its directly adjacent faces. If they are sufficiently close to each other (e.g., as tested via dot products), then the limit surface for that face will be reasonably flat and further refinement will likely make little difference to the final surface. Alternatively, you might want to approximate the area that a subdivided face covers on the image plane and continue subdividing until this area becomes sufficiently small. This approximation could be done using ray differentials; see Section 10.1.1 for an explanation of how to relate the ray differential to the screen space footprint. The trickiest part of this exercise is that some faces that don’t need subdivision due to the flatness test will still need to be subdivided in order to provide vertices so that neighboring faces that do need to subdivide can get their vertex one-rings. In particular, adjacent faces can differ by no more than one level of subdivision. You may find it useful to read recent papers by Patney et al. (2009) and Fisher et al. (2009) for discussion of how to avoid cracks in adaptively subdivided meshes.
  12. Ray-tracing point-sampled geometry: extending methods for rendering complex models represented as a collection of point samples (Levoy and Whitted 1985; Pfister et al. 2000; Rusinkiewicz and Levoy 2000), Schaufler and Jensen (2000) have described a method for intersecting rays with collections of oriented point samples in space. They probabilistically determined that an intersection has occurred when a ray approaches a sufficient local density of point samples and computes a surface normal with a weighted average of the nearby samples. Read their paper and extend pbrt to support a point-sampled geometry shape. Do any of pbrt’s basic interfaces need to be extended or generalized to support a shape like this?
  13. Deformation motion blur: the TransformedPrimitive in Section 4.1.2 of Chapter 4 supports animated shapes via transformations of primitives that vary over time. However, this type of animation isn’t general enough to represent a triangle mesh where each vertex has a position given at the start time and another one at the end time. (For example, this type of animation description can be used to describe a running character model where different parts of the body are moving in different ways.) Implement a more general Triangle shape that supports specifying vertex positions at the start and end of frame and interpolates between them based on the ray time passed to the intersection methods. Be sure to update the bounding routines appropriately. Triangle meshes with very large amounts of motion may exhibit poor performance due to triangles sweeping out very large bounding boxes and thus many ray–triangle intersections being performed that don’t hit the triangle. Can you come up with approaches that could be used to reduce the impact of this problem?
  14. Implicit functions: just as implicit definitions of the quadric shapes are a useful starting point for deriving ray-intersection algorithms, more complex implicit functions can also be used to define interesting shapes. In particular, difficult-to-model organic shapes, water drops, and so on can be well represented by implicit surfaces. Blinn (1982a) introduced the idea of directly rendering implicit surfaces, and Wyvill and Wyvill (1989) gave a basis function for implicit surfaces with a number of advantages compared to Blinn’s. Implement a method for finding ray intersections with general implicit surfaces and add it to pbrt. You may wish to read papers by Kalra and Barr (1989) and Hart (1996) for methods for ray tracing them. Mitchell’s algorithm for robust ray intersections with implicit surfaces using interval arithmetic gives another effective method for finding these intersections (Mitchell 1990), and more recently Knoll et al. (2009) described refinements to this idea. You may find an approach along these lines easier to implement than the others. See Moore’s book on interval arithmetic as needed for reference (Moore 1966).
  15. L-systems: a very successful technique for procedurally modeling plants was introduced to graphics by Alvy Ray Smith (1984), who applied Lindenmayer systems (L-systems) to model branching plant structures. Prusinkiewicz and collaborators have generalized this approach to encompass a much wider variety of types of plants and effects that determine their appearance (Prusinkiewicz 1986; Prusinkiewicz et al. 1994; Deussen et al. 1998; Prusinkiewicz et al. 2001). L-systems describe the branching structure of these types of shapes via a grammar. The grammar can be evaluated to form expressions that describe a topological representation of a plant, which can then be translated into a geometric representation. Add an L-system primitive to pbrt that takes a grammar as input and evaluates it to create the shape it describes.
  16. Given an arbitrary point left-parenthesis x comma y comma z right-parenthesis , what bound on the error from applying a scale transformation of left-parenthesis 2 comma 1 comma 4 right-parenthesis is given by Equation (3.16)? How much error is actually introduced?
  17. The quadric shapes all use the EFloat class for their intersection tests in order to be able to bound the error in the computed t value so that intersections behind the ray origin aren’t incorrectly reported as actual intersections. First, measure the performance difference when using regular Floats for one or more quadrics when rendering a scene that includes those shapes. Next, manually derive conservative error bounds for t values computed by those shapes as was done for triangles in Section 3.9.6. Implement your method. You may find it useful to use the EFloat class to empirically test your derivation’s correctness. Measure the performance difference with your implementation.
  18. One detail thwarts the watertightness of the current Triangle shape implementation: the translation and shearing of triangle vertices introduces round-off error, which must be accounted for in the extent of triangles’ bounding boxes; see Section 3.3 of Woop et al. (2013) for discussion (and a solution). Modify pbrt to incorporate a solution to this shortcoming. Can you find scenes where small image errors are eliminated thanks to your fix?
  19. Rendering in camera space: because floating-point arithmetic provides more precision near the origin than farther away from it, transforming the scene to a coordinate system with the camera at the origin can reduce the impact of error due to insufficient floating-point precision. (For example, consider the difference between rendering a scene in world space with the camera at left-parenthesis 100000 comma 100000 comma 100000 right-parenthesis , looking at a unit sphere two units away versus a scene with the camera at the origin, also looking at a sphere two units away: many more bits of precision are available to represent intersection points in the latter case.) Modify pbrt to primarily perform rendering computations in camera space, rather than world space, as it does currently. You’ll need to modify the Camera implementations to return camera-space rays and modify shapes to transform incoming rays from camera space to object space. You’ll want to transform the vertices of TriangleMeshes all the way to camera space. Measure the performance of your implementation versus an unmodified version of pbrt and render a variety of scenes with both systems. (In particular, make sure you test some scenes that have the camera far from the world space origin.) What sort of image differences do you see?