09/07/04© University of Wisconsin CS 559: Computer Graphics Prof Stephen Chenney Fall

09/07/04© University of Wisconsin Today Course overview and information Digital Images Homework 1 – due Sept 14 in class

09/07/04© University of Wisconsin What is Computer Graphics? Technically, it’s about the production, manipulation and display of images using computers Practically, it’s about movies, games, design, training, art, advertising, communication, …

09/07/04© University of Wisconsin This Course: Building Blocks Images and computers –Sampling, Color, Filters, … Drawing in 2D –Drawing lines and triangles, clipping, transformations Drawing in 3D –Viewing, transformations, lighting, real-time graphics Modeling in 3D –Describing volumes and surfaces, drawing them effectively Miscellaneous topics –Raytracing, animation, …

09/07/04© University of Wisconsin People Prof. Stephen Chenney –Room 6387 –Office Hours: TBD TA: Leo Chao –Room 1346 –Office Hour TBD Send all class to It gets the fastest response.

09/07/04© University of Wisconsin Web and The class web site is –It is updated very frequently –Lecture notes are put online before class, and updated after class –Additional resources and links are provided –Reading for future classes is listed is sent out on the class mailing list –This list sends to your official university account Check it regularly or set up forwarding –All notices are sent out on the mailing list, including things not mentioned in class –Only registered students get You read an archive at

09/07/04© University of Wisconsin CS Computer Account Everyone in the class needs a Computer Science department computer account –For completing and submitting projects If you don’t have one, go to the CSL on the 2 nd floor CS559 provides you with a Windows account –You will only have a Unix account if you are taking some other course

09/07/04© University of Wisconsin Textbooks and Readings Peter Shirley, “Fundamentals of Computer Graphics”, A.K. Peters, 2002 Woo et. al., "OpenGL Programming Guide", Third or Fourth Edition, Adison-Wesley, 1999 or 2003 –The definitive guide to OpenGL, and a reasonable description of general real-time 3D graphics –It doesn’t really matter which edition – the differences are not pertinent to this class Class readings –Online via the library’s reserve system – see class web page –A collection of papers, textbook chapters, and other documents –Some essential material not contained in the textbook

09/07/04© University of Wisconsin Projects There will be three projects for the course, spread evenly through the semester Project 1: Image editing Project 2: Running a maze Project 3: Building a virtual theme park You must submit all three in order to pass the course –This rule has never been waived

09/07/04© University of Wisconsin Homeworks There will be a homework every two weeks or so They are intended primarily to explore topics further and to prepare you for the exams They will be graded, but only the best five will count Some essential techniques will be presented only in homework –For example, an review of linear algebra

09/07/04© University of Wisconsin Grading (approximate) 45% Midterm and Final 45% Projects 10% Homework Everyone must write up their own homework For projects 1 and 3, you have the option of working in pairs

09/07/04© University of Wisconsin Lab Facilities Room B240 contains machines for use in this class They have adequate hardware (circa 2001) and the software to make it work Students in CS 559 have priority in the lab There are benefits to working in a lab with your classmates –For instance, the blackboards frequently display useful hints that someone else scrawled –But they also sometimes have incorrect information!!

09/07/04© University of Wisconsin Software Infrastructure FLTK will be the user interface toolkit –Provides windows, buttons, menus, etc –C++ class library, completely portable –We are currently at version 1.1.0rc6, available for free: OpenGL will be the 3D rendering toolkit –Provides an API for drawing objects specified in 3D –Included as part of Windows and in most Unix distributions Although getting hardware acceleration may take some doing Visual Studio.NET (Visual C++ 7.1) will be the programming environment for grading To be graded, your projects must compile under Visual C++ on the machines in room B240

09/07/04© University of Wisconsin C++ This is probably the first, and only, class in which you must complete large software projects in C++ without step by step instructions There is a great deal of freedom in the projects, which requires that you do your own software design If you are not comfortable in C++, you will have to take action There are tutorials intended to teach you C++ assuming you know Java: – –These are intended for the course CS 368, but just do the tutorials The transition to C++ was one of the biggest issues for past CS559 students –Visual Studio also causes issues – we will provide a transition tutorial

09/07/04© University of Wisconsin Adding the Class Everyone should be able to add at this point If not, with your name, ID and –If the system gave you a reason for denying registration, include it in your You have to wait a couple of days after adding before you can get a CS computer account

09/07/04© University of Wisconsin Admin Questions?

09/07/04© University of Wisconsin Images An image for our purposes is an abstract concept –It has no existence until it is displayed An image is intended to describe the light that arrives at your eyes when you view it –You can be even more abstract: image describes what you should think when you see it Different display devices convey the image content in different ways –e.g. Print and computer monitors use two fundamentally different approaches –The same image may look different on different monitors Who cares?

09/07/04© University of Wisconsin Image Formats We are familiar with many forms of image: –Photographs –Paintings –Sketches –Television (NTSC, PAL-SECAM) –Digital formats (JPEG, PNG, GIF, etc. etc. etc. etc.) Each form has its own way of obtaining and storing the information content We are primarily concerned with digital formats, but we will also discuss NTSC

09/07/04© University of Wisconsin Digital Images Many formats (100s) exist for storing images on a computer There are some conflicting goals: –The storage cost should be minimized –The amount of information stored should be maximized The size of something and the amount of information is contains are not the same thing –Original information versus perceptual equivalence –Tracking ownership may be important Most formats you are familiar with are raster images

09/07/04© University of Wisconsin Raster Images A raster is a regular grid of pixels (picture elements) Raster image formats store the color at each pixel, and maybe some other information –Easiest is to use a simple array of pixel values –Some formats store the pixel information in very different ways –e.g. a 5x3, floating point, grayscale image

09/07/04© University of Wisconsin Vector Images Vector formats offer an alternative way to store images The most common use of vector formats are in fonts – images of characters (Postscript, TrueType) Store images as collections of geometric primitives –E.g. Lines, polygons, circles, curves, … It is possible to go from a vector image to a raster image –We’ll learn how It is very hard to go the other way

09/07/04© University of Wisconsin Trade-Offs Which format, raster or vector, is easier to: –Display? –Resize (scale bigger or smaller)? –Rotate? –Crop (cut bits off at the edges)?

09/07/04© University of Wisconsin Obtaining Digital Images What are some methods for obtaining a digital image? –Digital camera –Scanning another image –Other forms of scanning (eg medical) –Editing existing digital images –Paint or drawing programs –Created from abstract data (e.g. math function plot) –Rendered from a scene description –…

09/07/04© University of Wisconsin Ideal Images The information stored in images is often continuous in nature For example, consider the ideal photograph: –It captures the intensity of light at a particular set of points coming from a particular set of directions (it’s called irradiance) –The intensity of light arriving at the camera can be any positive real number, and it mostly varies smoothly over space –Where do you see spatial discontinuities in a photograph? Film Focal point

09/07/04© University of Wisconsin Digital Images Computers work with discrete pieces of information How do we digitize a continuous image? –Break the continuous space into small areas, pixels –Use a single value for each pixel - the pixel value (no color, yet) –No longer continuous in space or intensity This process is fraught with danger, as we shall see Continuous Discrete Pixels: Picture Elements

09/07/04© University of Wisconsin Discretization Issues Can only store a finite number of pixels –Choose your target physical image size, choose your resolution (pixels per inch, or dots per inch, dpi), determine width/height necessary –Storage space goes up with square of resolution 600dpi has 4× more pixels than 300dpi Can only store a finite range of intensity values –Typically referred to as depth - number of bits per pixel Directly related to the number of colors available and typically little choice Most common depth is 8, but also sometimes see 16 for grey –Also concerned with the minimum and maximum intensity – dynamic range What is enough resolution and enough depth?

09/07/04© University of Wisconsin Perceptual Issues Spatially, humans can discriminate about ½ a minute of arc –At fovea, so only in center of view, 20/20 vision –At 0.5m, about 0.1mm (“Dot pitch” of monitors) –Sometimes limits the required number of pixels Humans can discriminate about 8 bits of intensity –“Just Noticeable Difference” experiments –Limits the required depth for typical dynamic ranges –Actually, it’s 9-10 bits, but 8 is far more convenient BUT, when manipulating images much higher resolution may be required

09/07/04© University of Wisconsin Next Time Color