csci5607/assignment-1a/ASSIGNMENT.md
2023-02-02 18:55:56 -06:00

8.6 KiB

Assignment 1a: Getting Started with Ray Casting

In this assignment, you are asked to begin writing a basic ray casting program.

This program should:

  1. read a simple scene description from a file that is provided as a command line argument
  2. define an array of pixels that will hold a computer-generated image of the scene
  3. use ray casting to determine the appropriate color to store in each pixel of the output image
  4. write the rendered image to an output file in ascii PPM format

As with assignment 0, the output file name should match the input file name except for the suffix. You can also earn extra credit by enabling your program to render a parallel projection and/or to render not only spheres but also cylinders.

You are asked to run your program on multiple different input files, featuring a diverse range of different input parameter settings. One of the objectives of this assignment is to help you to develop an intuitive understanding of how each of the various input settings affects the appearance of the rendered scene. Please be sure to specifically consider these three questions:

  • How does the direction of the 'up' vector affect the rendered view of a scene? You should specifically consider tipping the 'up' vector from side to side and forward to back. You will want to define scenes with multiple objects, asymmetrically arranged, to explore this question.
  • How do changes in the field of view settings affect the contents of your rendered images?
  • How do various modifications of the viewing parameters affect the amount of perspective distortion apparent in your image?

In subsequent assignments you will be asked to extend the code you write for this assignment to incorporate additional effects like illumination, shadows, specular reflections and transparency, with these latter effects being achieved by recursive ray tracing. Be sure that your code is structured with these extensions in mind. In particular, you are advised to write simple, modular code following the example described in class.

Please be sure to check the grading criteria to ensure that all requirements are understood.

Detailed instructions

  1. Your program should read the following information from an input scene description file (the syntax is shown below each item; entries in italics are variables, entries in bold are keywords):

    • The view origin, also variously referred to as the 'eye position', 'camera position' or 'center of projection' (a 3D point in space)

      eye   eyex eyey eyez
      
    • The viewing direction (a 3D vector)

      viewdir   vdirx  vdiry  vdirz
      
    • The 'up' direction (a 3D vector)

      updir   upx  upy  upz
      
    • The horizontal field of view (in degrees, please)

      hfov   fovh
      
    • The size of the output image (in pixel units)

      imsize   width  height
      
    • The 'background' color (using r, g, b components defined on a scale from 0-1)

      bkgcolor   r  g  b
      
    • A 'material' color (in terms of r, g, b components defined on a scale from 0-1). The material color should be treated as a state variable, meaning that all subsequently-defined objects should use the immediately-preceding material color

      mtlcolor   r  g  b
      
    • A sequence of one or more objects. For this assignment, your program is only required to be able to handle spheres. In subsequent assignments you will be asked to extend your code to handle triangles, so you will want to write your code with this future extension in mind. A sphere should be defined by the coordinates of its center point (a 3D point in space) and radius.

      sphere   cx  cy  cz  r
      
    • If you would like to attempt the extra credit portion of this assignment, please use this format to define a parallel projection

      projection  parallel
      
    • If you would like to attempt the extra credit portion of this assignment, please use this format to define a finite cylinder (without endcaps) that is centered at the point (cx, cy, cz) and oriented in the direction (dirx, diry, dirz).

      cylinder   cx  cy  cz  dirx  diry  dirz  radius  length
      
  2. Define an array of sufficient size to store the color values of your image.

  3. Using knowledge of the view origin, viewing direction, up direction, horizontal field of view and desired image aspect ratio, define an appropriate “viewing window” in world coordinate space, and a 1-1 mapping between points within this viewing window and pixel locations in your output image.

  4. For each pixel in the output image:

    • Define the equation of the ray that begins at the view origin and passes through the corresponding 3D point in the viewing window
    • Then, for each object in the scene:
      • Determine whether the current viewing ray intersects that object, and if so at what point or points
      • If there are one or more ray/object intersection points that are 'in front of' the view origin with respect to the positive viewing direction, determine which of them is closest to the view origin, remembering that ray/object intersection points that are 'behind' the view origin, with respect to the viewing direction, should be ignored
      • Determine the color of the object at the closest forward ray/object intersection point, if there is one, and return that value to be stored at the appropriate location in your image array
      • If no ray/object intersection is found, return the background color
  5. After all of the rays have been cast and a color has been determined for each pixel in the output image, write the final image to an output file, using the ascii PPM format.

You are strongly encouraged to begin with a very simple scene description, consisting for example of a single sphere located directly in front of the eye in the direction of view. You are also advised to begin by using a very small image size, to expedite the debugging process.

To explore the effects of various scene parameters, you are advised to start by varying the values of just one viewing parameter at a time, such as: the eye position, the viewing direction, the direction of the 'up' vector, the vertical field of view, the image aspect ratio, etc. Once you understand the effect of each parameter in isolation, you should experiment with combinations of changes, such as: co-varying the field of view and the proximity of the eye to the objects in your scene. Your ability to appreciate the effects of different view settings will be improved if you use a relatively complex but intuitively organized scene, with some asymmetric features and not too much empty space. One example is: spheres centered at each of the 8 corners of a cube, using different colors for each corner, or cylinders arranged along each of the 12 edges of a cube.

What you should turn in

  • All of your source code, clearly commented, plus a readme file with compiling instructions or a Makefile or CMake file that the TA can use to compile your code.
  • One "showcase" image produced by your program, that I can share with the rest of the class. To save time for the TA, we would be grateful if you could provide this image in a format that is supported for direct display in Google Slides or Powerpoint.
  • A 1-3 page writeup (including pictures) in which you discuss your observations on how the key viewing parameters affect the appearance of the rendered scene. Please incorporate sufficient images produced by your ray casting program to illustrate and explain your findings. In your writeup, be sure to specifically address each of these three points:
    • How does the apparent rotation of the scene with respect to the viewpoint change with changes in the direction of the 'up' vector?
    • How do changes in the field of view settings affect the appearance of the scene in your rendered image?
    • How can the viewing parameters (e.g. the camera location, field of view settings, …) be adjusted to achieve a less exaggerated vs more exaggerated amount of apparent perspective distortion in your image?
  • Please submit all of the above as a single zip file, containing appropriate subfolders, using the naming convention hw1a.firstname.lastname.zip. Please do not use tar or rar or any other alternative compression mode that could complicate the grading process for the TA.

Two sample input files with their corresponding output images are provided in the Files directory of this Canvas website. Please feel free to use these examples to partially check the functionality of your program. We will be testing your code on additional input files.