Get the full ebook with Bonus Features for a Better Reading Experience on ebookgate.

com

Foundations of 3D Computer Graphics 1st Edition

Steven J. Gortler

https://ebookgate.com/product/foundations-of-3d-computer-
graphics-1st-edition-steven-j-gortler/

OR CLICK HERE

DOWLOAD NOW

Download more ebook instantly today at https://ebookgate.com

Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

Mathematics for 3D Game Programming and Computer Graphics

Second Edition Eric Lengyel

https://ebookgate.com/product/mathematics-for-3d-game-programming-and-
computer-graphics-second-edition-eric-lengyel/

ebookgate.com

Mobile 3D graphics learning 3D graphics with the Java

micro edition Claus Ho■Fele

https://ebookgate.com/product/mobile-3d-graphics-learning-3d-graphics-
with-the-java-micro-edition-claus-ho%cc%88fele/

ebookgate.com

Fundamentals of Computer Graphics 2nd Edition Shirley P.

https://ebookgate.com/product/fundamentals-of-computer-graphics-2nd-
edition-shirley-p/

ebookgate.com

Essentials of Interactive Computer Graphics Concepts and

Implementation Kelvin Sung

https://ebookgate.com/product/essentials-of-interactive-computer-
graphics-concepts-and-implementation-kelvin-sung/

ebookgate.com
Computer Graphics with OpenGL 4th Edition Donald Hearn

https://ebookgate.com/product/computer-graphics-with-opengl-4th-
edition-donald-hearn/

ebookgate.com

Computer Graphics with OpenGL 3rd Edition Donald D. Hearn

https://ebookgate.com/product/computer-graphics-with-opengl-3rd-
edition-donald-d-hearn/

ebookgate.com

DirectX 9 Graphics The Definitive Guide to Direct 3D 1st

Edition Alan Thorn

https://ebookgate.com/product/directx-9-graphics-the-definitive-guide-
to-direct-3d-1st-edition-alan-thorn/

ebookgate.com

3D Math Primer for Graphics and Game Development 2nd

Edition Fletcher Dunn

https://ebookgate.com/product/3d-math-primer-for-graphics-and-game-
development-2nd-edition-fletcher-dunn/

ebookgate.com

3D Nanoelectronic Computer Architecture and Implementation

1st Edition D. Crawley

https://ebookgate.com/product/3d-nanoelectronic-computer-architecture-
and-implementation-1st-edition-d-crawley/

ebookgate.com
Foundations of 3D Computer Graphics

Steven J. Gortler

August 23, 2011

Foundations of 3D Computer Graphics 2
S.J. Gortler

MIT Press, 2012

To A.L.G., F.J.G, and O.S.G.
Foundations of 3D Computer Graphics ii
S.J. Gortler

MIT Press, 2012

Contents

Preface ix

I Getting Started 1

1 Introduction 3
1.1 OpenGL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 Linear 9
2.1 Geometric Data Types . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Vectors, Coordinate Vectors, and Bases . . . . . . . . . . . . . . . . 11
2.3 Linear Transformations and 3 by 3 Matrices . . . . . . . . . . . . . . 12
2.4 Extra Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.5 Rotations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.6 Scales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

3 Affine 21
3.1 Points and Frames . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.2 Affine transformations and Four by Four Matrices . . . . . . . . . . . 22
3.3 Applying Linear Transformations to Points . . . . . . . . . . . . . . 24
3.4 Translations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.5 Putting Them Together . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.6 Normals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

4 Respect 29

iii
CONTENTS CONTENTS

4.1 The Frame Is Important . . . . . . . . . . . . . . . . . . . . . . . . . 29

4.2 Multiple Transformations . . . . . . . . . . . . . . . . . . . . . . . . 31

5 Frames In Graphics 35
5.1 World, Object and Eye Frames . . . . . . . . . . . . . . . . . . . . . 35
5.2 Moving Things Around . . . . . . . . . . . . . . . . . . . . . . . . . 37
5.3 Scales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
5.4 Hierarchy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

6 Hello World 3D 45
6.1 Coordinates and Matrices . . . . . . . . . . . . . . . . . . . . . . . . 45
6.2 Drawing a Shape . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
6.3 The Vertex Shader . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
6.4 What Happens Next . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
6.5 Placing and Moving With Matrices . . . . . . . . . . . . . . . . . . . 53

II Rotations and Interpolation 57

7 Quaternions (a bit technical) 59

7.1 Interpolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
7.2 The Representation . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
7.3 Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
7.4 Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
7.5 Code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
7.6 Putting Back the Translations . . . . . . . . . . . . . . . . . . . . . . 68

8 Balls: Track and Arc 73

8.1 The Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
8.2 Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
8.3 Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

9 Smooth Interpolation 77
9.1 Cubic Bezier Functions . . . . . . . . . . . . . . . . . . . . . . . . . 78

Foundations of 3D Computer Graphics iv

S.J. Gortler

MIT Press, 2012

CONTENTS CONTENTS

9.2 Catmull-Rom Splines . . . . . . . . . . . . . . . . . . . . . . . . . . 80

9.3 Quaternion Splining . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
9.4 Other Splines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
9.5 Curves in Space ...... . . . . . . . . . . . . . . . . . . . . . . . . . . 83

III Cameras and Rasterization 87

10 Projection 89
10.1 Pinhole Camera . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
10.2 Basic Mathematical Model . . . . . . . . . . . . . . . . . . . . . . . 90
10.3 Variations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
10.4 Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97

11 Depth 101
11.1 Visibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
11.2 Basic Mathematical Model . . . . . . . . . . . . . . . . . . . . . . . 102
11.3 Near And Far . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
11.4 Code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

12 From Vertex to Pixel 109

12.1 Clipping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
12.2 Backface Culling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
12.3 Viewport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
12.4 Rasterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

13 Varying Variables (Tricky) 119

13.1 Motivating The Problem . . . . . . . . . . . . . . . . . . . . . . . . 119
13.2 Rational Linear Interpolation . . . . . . . . . . . . . . . . . . . . . . 120

IV Pixels and Such 125

14 Materials 127
14.1 Basic Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

Foundations of 3D Computer Graphics v

S.J. Gortler

MIT Press, 2012

CONTENTS CONTENTS

14.2 Diffuse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

14.3 Shiny . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
14.4 Anisotropy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134

15 Texture Mapping 137

15.1 Basic Texturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
15.2 Normal Mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
15.3 Environment Cube Maps . . . . . . . . . . . . . . . . . . . . . . . . 140
15.4 Projector Texture Mapping . . . . . . . . . . . . . . . . . . . . . . . 142
15.5 Multipass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145

16 Sampling 149
16.1 Two Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
16.2 The Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
16.3 The Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
16.4 Alpha . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

17 Reconstruction 161
17.1 Constant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
17.2 Bilinear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
17.3 Basis functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

18 Resampling 167
18.1 Ideal Resampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
18.2 Blow up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
18.3 Mip Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169

V Advanced Topics 173

19 Color 175
19.1 Simple Bio-Physical Model . . . . . . . . . . . . . . . . . . . . . . . 176
19.2 Mathematical Model . . . . . . . . . . . . . . . . . . . . . . . . . . 179
19.3 Color Matching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182

Foundations of 3D Computer Graphics vi

S.J. Gortler

MIT Press, 2012

CONTENTS CONTENTS

19.4 Bases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

19.5 Reflection Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . 186
19.6 Adaptation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
19.7 Non Linear Color . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188

20 What is Ray Tracing 197

20.1 Loop Ordering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
20.2 Intersection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
20.3 Secondary Rays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201

21 Light (technical) 205

21.1 Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
21.2 Reflection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
21.3 Light Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
21.4 Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
21.5 Integration Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . 221
21.6 More General Effects . . . . . . . . . . . . . . . . . . . . . . . . . . 222

22 Geometric Modeling: Basic Intro 225

22.1 Triangle Soup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
22.2 Meshes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
22.3 Implicit Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
22.4 Volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
22.5 Parametric Patches . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
22.6 Subdivision Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . 229

23 Animation: Not Even an Introduction 237

23.1 Interpolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
23.2 Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
23.3 Human Locomotion . . . . . . . . . . . . . . . . . . . . . . . . . . . 243

A Hello World 2D 245

A.1 APIs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245

Foundations of 3D Computer Graphics vii

S.J. Gortler

MIT Press, 2012

CONTENTS CONTENTS

A.2 Main Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246

A.3 Adding some Interaction . . . . . . . . . . . . . . . . . . . . . . . . 253
A.4 Adding a Texture . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
A.5 What’s Next . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258

B Affine Functions 261

B.1 2D Affine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
B.2 3D Affine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
B.3 Going Up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
B.4 Going Down . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
B.5 Going Sideways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263

Foundations of 3D Computer Graphics viii

S.J. Gortler

MIT Press, 2012

Preface

This book has developed out of an introductory computer graphics course I have been
teaching at Harvard since 1996. Over the years I have had the pleasure of teaching
many amazing students. During class, these students have asked many good questions.
In light of these questions, I often realized that some of my explanations in class were
a bit sloppy and that I didn’t fully understand the topic I had just tried to explain. This
would often lead me to rethink the material and change the way I taught it the next
time around. Many of these ideas have found their way into this book. Throughout
the course of the book, I cover mostly standard material, but with an emphasis on
understanding some of the more subtle concepts involved.
In this book, we will introduce the basic algorithmic technology needed to produce
3D computer graphics. We will cover the basic ideas of how 3D shapes are represented
and moved around algorithmically. We will cover how a camera can be algorithmically
modeled turning this 3D data into a 2D image made up of a discrete set of dots or
pixels on a screen. Later in the book, we will also cover some more advanced topics
discussing the basics of color and light representations. We will also briefly introduce
some more advanced topics on light simulation for producing photo-realistic images,
on various ways of dealing with geometric representations and on producing animated
computer graphics.
In this book, we include material that is both above and below the API-hood. Much
of the material (especially early on) is stuff you simply need to know in order to do
3D computer graphics. But we also spend time to explain what is going on inside of
OpenGL. This is necessary to be a highly competent computer graphics programmer.
But also, it is simply fascinating to learn the hows and whys of our amazing computer
graphics computational infrastructure.
We will not cover the hardware and compiler aspects of computer graphics in this
book. We will also not focus on 2D computer graphics or human computer interfaces.
These topics are all interesting in their own rights but are fairly distinct from the algo-
rithmic side of 3D computer graphics.
In this book, we structure our explanations around OpenGL, a real-time “rasterization-
based” rendering environment. We have done this (rather than, say, a “ray-tracing-
based” environment) since so much of computer graphics is done in this setting. Any-
one, for example, who works in 3D video games needs to master this material (and

ix
 CHAPTER 0. PREFACE

more). We have chosen the OpenGL API (with the GLSL shading language) in partic-
ular, since it can be run on a wide variety of computing platforms.
This book is intended for upper level CS/math/physics undergraduate students with
at least a year of programming under their belt and at least a rudimentary understanding
of linear algebra.

For the Professor

In the following paragraphs I will describe some of the subtle issues that require some
care to get right and are often not taught clearly. I hope that students will master these
topics from this book.
Appendix A: We try to get up and running as painlessly as possible with a first
OpenGL program. Since we will be using a modern version of OpenGL, there is
thankfully not much API left to teach anymore. All OpenGL needs to do is man-
age the shader programs, vertex buffer objects and textures. We do not teach any of the
deprecated elements of old OpenGL, as they are not used in modern computer graphics
programming.
Chapters 2-4: In computer graphics we need to think about points and vectors
from both a coordinate-free and coordinate-full approach. We need to use coordinates
to obtain a concrete representation ultimately resulting in our rendered images. But
it is often important to represent and transform our points with respect to different
coordinate systems. As such it is important to

• Distinguish between a geometric point, and the coordinates used to represent that
point with respect to some frame.

• Use a notation to keep explicitly keep track of the basis used for a set of coordi-
nates.

• Distinguish in our notation between matrix equations that represent basis changes,
vs. matrix expressions that represent geometric transformations being applied to
points.

This ultimately leads to what we call the left-of rule, which allows us to interpret matrix
expressions and understand with respect to which basis a transformation is acting.
It is our hope that by mastering this explicit notational framework, a student can
easily figure out how to do complicated transformations. This is in contrast to the “try
lots of orderings and inverses of sequences of matrices until the program does the right
thing” approach. One loose inspiration for our approach is the manuscript [16] by Tony
DeRose.
Chapters 5 and 6: We describe an organized framework for dealing with frames in
computer graphics, and how this translates into simple 3D OpenGL code. In particular,
we derive useful ways to move objects around using a “mixed auxiliary frame”. This

Foundations of 3D Computer Graphics x

S.J. Gortler

MIT Press, 2012

CHAPTER 0. PREFACE

allows us to, say, rotate an object correctly about its own center, but in directions that
correspond naturally to those on the screen.
Chapters 7: This is a simple and straightforward description of the quaternion rep-
resentation for rotations. We also derive how to properly combine quaternions and
translation vectors to define a rigid-body transformation data type that multiplies and
inverts just like matrices.
Chapter 8 is a simple and straightforward description of the trackball and arcball
rotation interface. We also show why the trackball interface is mouse-path-invariant.
Chapter 9: We do a quick and dirty introduction to Bezier and Catmull-Rom splines.
Chapters 10-13: In these chapters we describe how camera projection is modeled
using 4-by-4 matrices. We also describe the fixed function operations in OpenGL.
We pay special attention to deriving the correct formulas for interpolating the varying
variables. Some of the background on affine functions and interpolation are relegated
to Appendix B. Many of the subtleties here are described nicely in essays by Jim
Blinn [5]. In these chapters, we do not cover details about clipping or rasterization
algorithms.
Chapter 14 and 15: We give some simple example shaders for diffuse, shiny, and
anisotropic materials. We also point to more advanced real-time rendering techniques
such as multipass rendering and shadow mapping. These sections are admittedly too
brief. Presumably students pursuing more aggressive rendering projects will need to
learn a lot more about the ever-evolving techniques (and tricks) used in modern real-
time rendering.
Chapter 16-18: We cover the basics of why filtering is needed to compute discrete
images, and how basis functions are used to reconstruct continuous ones. In particular,
we show how these two concepts need to be combined to properly do filtering during
texture mapping (a la Paul Heckbert’s MS thesis [28].) We do not delve into the details
of Fourier analysis here as we think it would pull us a bit too far off the main narrative
(and in the end we use box filters anyway).
Chapter 19: We describe the basic theory of human color perception. From a math-
ematical perspective, we attempt to be clear about the very definition of what a color
is, and why such things form a linear space. A related treatment to ours can be found
in Feynman’s Lectures [20]. For many of the technical issues of color computations,
we rely on the Color FAQ of Charles Poynton [58].
Chapter 20: For completeness, we briefly describe ray tracing computations. As
this is not the focus of the course, we only touch on the topic.
Chapter 21: As an introduction to advanced topics in photo-realistic rendering,
we do a careful introduction to the physical units for describing light as well as the
reflection, rendering, and measurement equations. One thing we pay close attention to,
for example, is motivating why reflection is measured in the particular units used. A
good reference on these basics and more is Eric Veach’s PhD Thesis [71].
Chapter 22: We outline some of the ways that surfaces are modeled and represented

Foundations of 3D Computer Graphics xi

S.J. Gortler

MIT Press, 2012

 CHAPTER 0. PREFACE

in computer graphics. This is a non-technical discussion that only just tries to give a
quick intro to this rich topic. We do go into enough detail to be able to implement
Catmull-Rom subdivision surfaces (assuming you have a mesh data structure handy),
as these are a quick and easy way to represent a broad family of surfaces.
Chapter 23: We outline some of the ways animation is done in computer graphics.
Again, this is a non-technical discussion that tries to give a quick intro to this rich topic.
One nice place to start for this material is Adrien Treuille’s course on its CMU website.
This book covers a bit more than what would be possible to do in a one-semester
course. It does not attempt to be an encyclopedic reference to all of the rich topics in
computer graphics theory and practice. Additionally, this book stays close to material
in wide use today. We do not cover the many current research results and ideas that
may one day become part of standard practice.

Thanks to...

During the development of this book over the last year, I received lots of helpful input
from Hamilton Chong, Guillermo Diez-Canas and Dylan Thurston.
Helpful advice also came from Julie Dorsey, Hugues Hoppe, Zoe Wood, Yuanchen
Zhu and Todd Zickler.
Other comments and assistance came from Gilbert Bernstein, Fredo Durand, Ladislav
Kavan, Michael Kazhdan and Peter-Pike Sloan.
During the development of the course, I was greatly assisted by Xiangfeng Gu,
Danil Kirsanov, Chris Mihelich, Pedro Sander and Abraham Stone.
Finally, over the years, the course has had many talented students acting as teach-
ing assistants. They have of course contributed in many ways to the evolution of this
material. These have included Brad Andalman, Keith Baldwin, Forrester Cole, Ashley
Eden, David Holland, Brad Kittenbrink, Sanjay Mavinkurve, Jay Moorthi, Doug Nac-
hand, Brian Osserman, Ed Park, David Ryu, Razvan Surdulescu and Geetika Tewari.

Foundations of 3D Computer Graphics xii

S.J. Gortler

MIT Press, 2012

 Part I

Getting Started

1
Chapter 1

Introduction

Computer graphics is an amazing technology success story. The basic ideas, represen-
tations, algorithms and hardware approaches were forged in the 1960’s and 1970’s and
developed to maturity in the following two decades. By the mid 90’s computer graph-
ics techniques were reasonably mature, but their impact was still somewhat restricted
to “high end” applications such as scientific visualization on super-computers, and ex-
pensive flight simulators. Its hard to believe now, but many undergraduate CS majors
had no idea what 3D computer graphics even was!
The previous decade has finally seen the mass commodification of graphics. Every
modern PC is capable of producing high quality computer generated images, mostly
in the form of video games and virtual-life environments. The entire animation indus-
try has been transformed from its high end (e.g., Pixar films) down to daily children’s
television. For live-action movies, the field of special effects has been completely rev-
olutionized; viewers nowadays don’t flinch when they see the most amazing computer
generated special effects, it is simply expected.

1.1 OpenGL

OpenGL began as an API that implemented a very specific sequence of operations used
for doing 3D computer graphics. As the underlying hardware became cheaper, more
and more flexibility was put into the hardware and exposed to the user by the OpenGL
API. Over time, it has become feasible to fully control certain portions of the graphics
computation. The user does this by writing small special purpose programs, called
shaders, that are passed to and complied by the API. In OpenGL these shaders are
written in a C-like special purpose language called GLSL. As we will describe below,
the two main programmable portions are controlled by a vertex shader and a fragment
shader. These specific sections of the computation have been made programmable both
because there is great utility in giving the user flexibility in them, but also because this

3
 CHAPTER 1. INTRODUCTION

Figure 1.1: A cat head is described by a soup of triangles. Some of the vertices are
highlighted with black dots. From [64], c Eurographics and Blackwell Publishing
Ltd.

computation can be done using single instruction multiple data (SIMD) parallelism.
Data stored with each geometric vertex can be processed independently from the other
vertices. Likewise, computation to determine the color of one screen pixel can be
computed completely independently from the other pixels.
In a current OpenGL program, much (but not all) of the actual 3D graphics is done
by the shaders that you write, and are no longer really part of the OpenGL API itself.
In this sense, OpenGL is more about organizing your data and your shaders, and less
about 3D computer graphics. In the rest of this section, we will give an overview of
the main processing steps done by OpenGL. But we will also give some high level
descriptions of how the various shaders are typically used in these steps to implement
3D computer graphics.
In OpenGL we represent our geometry as a collection of triangles. On the one
hand, triangles are simple enough to be processed very efficiently by OpenGL, while
on the other hand, using collections of many triangles, we can approximate surfaces
with complicated shapes (see Figure 1.1). If our computer graphics program uses a
more abstract geometric representation, it must first be turned into a triangle collection
before OpenGL can draw the geometry.
Briefly stated, the computation in OpenGL determines the screen position for each
vertex of each of the triangles, figures out which screen dots, called pixels, lie within
each triangle, and then performs some computation to determine the desired color of
that pixel. We now walk through these steps in a bit more detail.
Each triangle is made up of 3 vertices. We associate some numerical data with
each vertex. Each such data item is called an attribute. At the very least, we need
to specify the location of the vertex (using 2 numbers for 2D geometry or 3 numbers
for 3D geometry). We can use other attributes to associate other kinds of data with our

Foundations of 3D Computer Graphics 4

S.J. Gortler

MIT Press, 2012

CHAPTER 1. INTRODUCTION

Uniform Variables

Vertex Shader
Attributes gl_Position
Varying variables
Vertex buffer Assembler

Figure 1.2: Vertices are stored in a vertex buffer. When a draw call is issued each
of the vertices passes through the vertex shader. On input to the vertex shader, each
vertex (black) has associated attributes. On output, each vertex (cyan) has a value for
gl Position and for its varying variables.

vertices that we will use to determine their ultimate appearances. For example, we may
associate a color (using 3 numbers representing amounts of red, green and blue) with
each vertex. Other attributes might be used to represent relevant material properties
describing, say, how shiny the surface at the vertex is.
Transmitting the vertex data from the CPU to the graphics hardware (the GPU)
is an expensive process, so it is typically done as infrequently as possible. There are
specific API calls to transfer vertex data over to OpenGL which stores this data in a
vertex buffer
Once the vertex data has been given to OpenGL, at any subsequent time we can
send a draw call to OpenGL. This commands OpenGL to walk down the appropriate
vertex buffers and draw each vertex triplet as a triangle.
Once the OpenGL draw call has been issued, each vertex (i.e., all of its attributes)
gets processed independently by your vertex shader (See Figure 1.2). Besides the
attribute data, the shader also has access to things called uniform variables. These are
variables that are set by your program, but you can only set them in between OpenGL
draw calls, and not per vertex.
The vertex shader is your own program, and you can put whatever you want in
it. The most typical use of the vertex shader is to determine the final position of the
vertices on the screen. For example, a vertex can have its own abstract 3D position
stored as an attribute. Meanwhile, a uniform variable can be used to describe a virtual
camera that maps abstract 3D coordinates to the actual 2D screen. We will cover the
details of this kind of computation in Chapters 2- 6 and Chapter 10.
Once the vertex shader has computed the final position of the vertex on the screen, it
assigns this value to the the reserved output variable called gl Position. The x and
y coordinates of this variable are interpreted as positions within the drawing window.
The lower left corner of the window has coordinates (−1, −1), and the upper right
corner has coordinates (1, 1). Coordinates outside of this square represent locations

Foundations of 3D Computer Graphics 5

S.J. Gortler

MIT Press, 2012

 CHAPTER 1. INTRODUCTION

gl_Position Varying variables

Varying variables
Assembler

Rasterizer

Figure 1.3: The data in gl Position is used to place the three vertices of the triangle
on a virtual screen. The rasterizer figures out which pixels (orange) are inside the
triangle and interpolates the varying variables from the vertices to each of these pixels.

outside of the drawing area.

The vertex shader can also output other variables that will be used by the fragment
shader to determine the final color of each pixel covered by the triangle. These outputs
are called varying variables since as we will soon explain, their values can vary as we
look at the different pixels within a triangle.
Once processed, these vertices along with their varying variables are collected by
the triangle assembler, and grouped together in triplets.
OpenGL’s next job is to draw each triangle on the screen (see Figure 1.3). This step
is called rasterization. For each triangle, it uses the three vertex positions to place the
triangle on the screen. It then computes which of the pixels on the screen are inside
of this triangle. For each such pixel, the rasterizer computes an interpolated value for
each of the varying variables. This means that the value for each varying variable is set
by blending the three values associated with the triangle’s vertices. The blending ratios
used are related to the pixel’s distance to each of the three vertices. We will cover the
exact method of blending in Chapter 13. Because rasterization is such a specialized
and highly optimized operation, this step has not been made programmable.
Finally, for each pixel, this interpolated data is passed through a fragment shader
(see Figure 1.3). A fragment shader is another program that you write in the GLSL
language and hand off to OpenGL. The job of the fragment shader is to determine the
drawn color of the pixel based on the information passed to it as varying and uniform
variables. This final color computed by the fragment shader is placed in a part of GPU
memory called a framebuffer. The data in the framebuffer is then sent to the display,
where it is drawn on the screen.
In 3D graphics, we typically determine a pixel’s color by computing a few equa-
tions that simulate the way that light reflects off of some material’s surface. This calcu-
lation may use data stored in varying variables that represent the material and geometric
properties of the material at that pixel. It may also use data stored in uniform variables

Foundations of 3D Computer Graphics 6

S.J. Gortler

MIT Press, 2012

CHAPTER 1. INTRODUCTION

Uniform Variables

Fragment Shader
Varying variables Screen color

Frame Buffer

Figure 1.4: Each pixel is passed through the fragment shader which computes the final
color of the pixel (pink). The pixel is then placed in the framebuffer for display.

Figure 1.5: By changing our fragment shader we can simulate light reflecting off of
different kinds of materials.

that represent the position and color of the light sources in the scene. By changing the
program in the fragment shader, we can simulate light bouncing off of different types
of materials, this can create a variety of appearances for some fixed geometry, as shown
in Figure 1.5. We discuss this process in greater detail in Chapter 14.
As part of this color computation, we can also instruct fragment shader to fetch
color data from an auxiliary stored image. Such an image is called a texture and is
pointed to by a uniform variable. Meanwhile, varying variables called texture coordi-
nates tell the fragment shader where to select the appropriate pixels from the texture.
Using this process called texture mapping, one can simulate the “gluing” of a some
part of a texture image onto each triangle. This process can be used to give high visual
complexity to a simple geometric object defined by only a small number of triangles.
See Figure 1.6 for such an example. This is discussed further in Chapter 15.
When colors are drawn to the framebuffer, there is a process called merging which
determines how the “new” color that has just been output from the fragment shader
is mixed in with the “old” color that may already exist in the framebuffer. When z-
buffering is enabled a test is applied to see whether the geometric point just processed
by the fragment shader is closer to, or farther from the viewer, than the point that was
used to set the existing color in the framebuffer. The framebuffer is then updated only if

Foundations of 3D Computer Graphics 7

S.J. Gortler

MIT Press, 2012

 CHAPTER 1. INTRODUCTION

Figure 1.6: Texture mapping. Left: A simple geometric object described by a small
number of triangles. Middle: An auxiliary image called a texture. Right: Parts of the
texture are glued onto each triangle giving a more complicated appearance. From [65],
c ACM.

the new point is closer. Z-buffering is very useful in creating images of 3D scenes. We
discuss z-buffering in Chapter 11. In addition, OpenGL can also be instructed to blend
the old and new colors together using various ratios. This can be used for example
to model transparent objects. This process is called alpha blending and is discussed
further in Section 16.4. Because this merging step involves reading and writing to
shared memory (the framebuffer), this step has not been made programmable, but is
instead controlled by various API calls.
In Appendix A we walk through an actual code fragment that implements a simple
OpenGL program that performs some simple 2D drawing with texture mapping. The
goal there is not to learn 3D graphics, but to get an understanding of the API itself and
the processing steps used in OpenGL. You will need to go through this Appendix in
detail at some point before you get to Chapter 6.

Exercises
Ex. 1 — To get a feeling for computer graphics of the 1980’s, watch the movie Tron.

Ex. 2 — Play the video game Battlezone.

Foundations of 3D Computer Graphics 8

S.J. Gortler

MIT Press, 2012

Chapter 2

Linear

Our first task in learning 3D computer graphics is to understand how to represent points
using coordinates and how to perform useful geometric transformations to these points.
You very likely have seen similar material when studying linear algebra, but in com-
puter graphics, we often simultaneously use a variety of different coordinate systems
and as such we need to pay special attention to the role that these various coordinate
systems play. As a result, our treatment of even this basic material may be a bit different
than that seen in a linear algebra class.
In this chapter, we will begin this task by looking at vectors and linear transforma-
tions. Vectors will be used to represent 3D motion while linear transformations will be
used to apply operations on vectors, such as rotation and scaling. In the following few
chapters we will then investigate affine transformations, which add the ability to trans-
late objects. We will not actually get to the coding process for computer graphics until
Chapter 6. Our approach will be to first carefully understand the appropriate theory,
which will then make it easy to code what we need.

2.1 Geometric Data Types

Imagine some specific geometric point in the real world. This point can be represented
with three real numbers,
 
x
 y 
z

which we call a coordinate vector. The numbers specify the position of the point with
respect to some agreed upon coordinate system. This agreed upon coordinate system
has some agreed upon origin point, and also has three agreed upon directions. If we
were to change the agreed upon coordinate system, then we would need a different set

9
 CHAPTER 2. LINEAR

7.3
 
p̃  −4 

v = p̃ − q̃ 12

v

q̃
f
t

Figure 2.1: Geometric data types: points are shown as dots, and vectors as arrows.
Vectors connect two dots and do not change when translated. A frame represents a co-
ordinate system, and consists of one origin point and a basis of d vectors. A coordinate
vector is a triple of real numbers.

of numbers, a different coordinate vector, to describe the same point. So in order to

specify the location of an actual point, we need both a coordinate system, as well as a
coordinate vector. This suggests that we really need to be careful in distinguishing the
following concepts: coordinate systems, coordinate vectors, and geometric points.
We start with four types of data, each with its own notation (See Figure 2.1).

• A point will be notated as p̃ with a tilde above the letter. This is a geometric
object, not a numerical one.

• A vector will be notated as ~v with an arrow above the letter. This too is a non-
numerical object. In Chapter 3 we will discuss in more detail the difference
between vectors and points. The main difference is that points represent places
while vectors represent the motion needed to move from point to point.

• A coordinate vector (which is a numeric object made up of real numbers) will

be represented as c with a bold letter.

• A coordinate system (which will be a set of abstract vectors, and so again, non-
numeric) will be notated as ~f t with a bold letter, arrow, and superscript “t”. (We
use bold to signify a vertical collection, the superscript “t” makes it a horizontal
collection, and the arrow tells us that it is a collection of vectors, not numbers.)
There are actually two kinds of coordinate systems. A basis is used to describe
vectors, while a frame is used to describe points. We use the same notation for
all coordinate systems and let the context distinguish the type.

Below, we will define all of these object types and see what operations can be done
with them. When thinking about how to manipulate geometry, we will make heavy use
of, and do symbolic computation with, both the numeric (coordinate vectors) and the
non-numeric (vectors and coordinate systems) objects. Only once we establish all of
our necessary conventions in Chapter 5 will we be able to drop the non-numeric objects
and use the the numerical parts of our calculations in our computer code.

Foundations of 3D Computer Graphics 10

S.J. Gortler

MIT Press, 2012

CHAPTER 2. LINEAR

2.2 Vectors, Coordinate Vectors, and Bases

Let us start by clearly distinguishing between vectors and coordinate vectors. In this
book, a vector will always be an abstract geometric entity that represents motion be-
tween two points in the world. An example of such a vector would be “one mile east”.
A coordinate vector is a set of real numbers used to specify a vector, once we have
agreed upon a coordinate system.
Formally speaking, a vector space V is some set of elements ~v that satisfy certain
rules. In particular one needs to define an addition operation that takes two vectors and
maps it to a third vector. One also needs to define the operation of multiplying a real
scalar times a vector to get back another vector.
To be a valid vector space, a bunch of other rules must be satisfied, which we will
not dwell on too closely here. For example, the addition operation has to be associative
and commutative. As another example, scalar multiplies must distribute across vector
adds

α(~v + w)
~ = α~v + αw
~

and so on [40].
There are many families of objects which have the structure of a vector space.
But in this book we will be interested in the vector space consisting of actual motions
between actual geometric points. In particular, we will not think of a vector as a set
of 3 numbers.
A coordinate system, or basis, is a small set of vectors that can be used to produce
the entire set of vectors using the vector operations. (More formally, we say that P a set of
vectors ~b1 ...~bn is linearly dependent if there exists scalars α1 ...αn such that i αi~bi =
~0. If a set of vectors is not linearly dependent, then we call them linearly independent.
If ~b1 ...~bn are linearly independent and can generate all of V using addition and scalar
multiplication, then the set ~bi is called a basis of V and we say that n is the dimension
of the basis/space.) For free motions in space, the dimension is 3. We also may refer
to each of the basis vectors as an axis, and in particular we may refer to the first axis as
the x axis, the second as the y axis and the third as the z axis.
We can use a basis as a way to produce any of the vectors in the space. This can be
expressed using a set of unique coordinates ci as follows.

ci~bi
X
~v =
i

We can use vector algebra notation and write this as

 
h i c1
ci~bi =
X
~v = ~b1 ~b2 ~b3  c2  (2.1)
i c3

Foundations of 3D Computer Graphics 11

S.J. Gortler

MIT Press, 2012

 CHAPTER 2. LINEAR

The interpretation of the rightmost expression uses the standard rules for matrix-matrix
multiplication from linear algebra. Here each term ci~bi is a real-scalar multiplied by an
abstract vector. We can create a shorthand notation for this and write it as

~v = ~bt c

where ~v is a vector, ~bt is a row of basis vectors and c is a (column) coordinate vector.

2.3 Linear Transformations and 3 by 3 Matrices

A linear transformation L is just a transformation from V to V which satisfies the
following two properties.

L(~v + ~u) = L(~v ) + L(~u)

L(αv)
~ = αL(~v )

We use the notation ~v ⇒ L(~v ) to mean that the vector ~v is transformed to the vector
L(~v ) through L.
The class of linear transformations is exactly the class that can be expressed using
matrices. This is because a linear transformation can be exactly specified by telling us
its effect on the basis vectors. Let us see how this works:
Linearity of the transformation implies the following relationship

ci~bi ) = ci L(~bi )
X X
~v ⇒ L(~v ) = L(
i i

which we can write in vector algebra notation using Equation (2.1) as

   
h i c1 h i c1
~b1 ~b2 ~b3  c2  ⇒ L(~b1 ) L(~b2 ) L(~b3 )  c2 
c3 c3

Each of the 3 new vectors L(~bi ) is itself an element of V , it can ultimately be written
as some linear combination of the original basis vectors. For example, we could write
 
h i M1,1
L(~b1 ) = ~b1 ~b2 ~b3  M2,1 
M3,1

for some appropriate set of Mj,1 values. Doing this for all of our basis vectors, we get
 
h i h i M1,1 M1,2 M1,3
L(~b1 ) L(~b2 ) L(~b3 ) = ~b1 ~b2 ~b3  M2,1 M2,2 M2,3  (2.2)
M3,1 M3,2 M3,3

Foundations of 3D Computer Graphics 12

S.J. Gortler

MIT Press, 2012

CHAPTER 2. LINEAR

v

bt

Figure 2.2: A vector undergoes a linear transformation ~v = ~bt c ⇒ ~bt M c. The matrix
M depends on the chosen linear transformation.

for an appropriate choice of matrix M made up of 9 real numbers.

Putting this all together, we see that the operation of the linear transformation op-
erating on a vector can be expressed as:

 
h i c1
~b1 ~b2 ~b3  c2 
c3

  
h i M1,1 M1,2 M1,3 c1
⇒ ~b1 ~b2 ~b3  M2,1 M2,2 M2,3   c2 
M3,1 M3,2 M3,3 c3

In summary, we can use a matrix to transform one vector to another

~bt c ⇒ ~bt M c

(See Figure 2.2.)

If we apply the transformation to each of the basis vectors, we get a new basis. This
can be expressed as

 
h i h i M1,1 M1,2 M1,3
~b1 ~b2 ~b3 ⇒ ~b1 ~b2 ~b3  M2,1 M2,2 M2,3 
M3,1 M3,2 M3,3

or for short
~bt ⇒ ~bt M

(See Figure 2.3.)

And of course, it is valid to multiply a matrix times a coordinate vector

c ⇒ Mc

Foundations of 3D Computer Graphics 13

S.J. Gortler

MIT Press, 2012

 CHAPTER 2. LINEAR

bt

Figure 2.3: A basis undergoes a linear transformation ~bt ⇒ ~bt M

2.3.1 Identity and Inverse

The identity map leaves all vectors unchanged. Its matrix is the identity matrix

 
1 0 0
I= 0 1 0 
0 0 1

The inverse of a matrix M is the unique matrix M −1 with the property M M −1 =

−1
M M = I. It represents the inverse transformation on vectors. If a linear trans-
form happens to map more than one input vector to the same output vector, then the
transform will not be invertible and its associated matrix will not have an inverse. In
computer graphics, when we chose 3D to 3D linear transforms to move objects around
in space (as well as scale them), it will seldom make sense to use a non-invertible trans-
form. So, unless stated, all of the matrices we will be dealing with in this book will
have an inverse.

2.3.2 Matrices for Basis Changes

Besides being used to describe a transformation (⇒), a matrix can also be used to
describe an equality (=) between a pair of bases or pair of vectors. In particular, above
in Equation (2.2), we saw an expression of the form
~at = ~bt M (2.3)
t
~a M −1
= ~bt (2.4)
This expresses an equality relationship between the named bases ~at and ~bt .
Suppose a vector is expressed in a specific basis using a specific coordinate vector:
~v = ~bt c. Given Equation (2.3), one can write
~v = ~bt c = ~at M −1 c
This is not a transformation (which would use the ⇒ notation), but an equality (us-
ing the = notation). We have simply written the same vector using two bases. The
coordinate vector c represents ~v with respect to ~b while the coordinate vector M −1 c
represents the same ~v with respect to ~a.

Foundations of 3D Computer Graphics 14

S.J. Gortler

MIT Press, 2012

CHAPTER 2. LINEAR

2.4 Extra Structure

Vectors in 3D space also come equipped with a dot product operation

~v · w
~

that takes in two vectors and returns a real number. This dot product allows us to define
the squared length (also called squared norm) of a vector

k ~v k2 := ~v · ~v

The dot product is related to the angle θ ∈ [0..π] between two vectors as

~v · w
~
cos(θ) =
k ~v kk w ~ k

We say that two vectors are orthogonal if ~v · w

~ = 0.
We say that a basis is orthonormal if all the basis vectors are unit length and pair-
wise orthogonal.
In an orthonormal basis ~bt , it is particularly easy to compute the dot product of two
vectors (~bt c) · (~bt d). In particular we have

~
bt c · ~bt d = ( ~bi ci ) · ( ~bj dj )
X X

i j

ci dj (~bi · ~bj )
XX
=
i j
X
= ci di
i

where in the second line we use the bi-linearity of the dot product and in the third line
we use the orthonormality of the basis.
We say that a 2D orthonormal basis is right handed if the second basis vector can
be obtained from the first by a 90 degree counter clockwise rotation (the order of the
vectors in the basis is clearly important here).
We say that a 3D orthonormal basis is right handed if the three (ordered) basis
vectors are arranged as in Figure 2.4, as opposed to the arrangement of Figure 2.5. In
particular, in a right handed basis, if you take your right hand opened flat, with your
fingers pointing in the direction of the first basis vector, in such a way that you can curl
your fingers so that they point in the direction of the second basis vector, then your
thumb will point in the direction of the third basis vector.
In 3D, we also have a cross product operation takes in two vectors and outputs one
vector defined as

~v × w
~ :=k v k k w k sin(θ) ~n

Foundations of 3D Computer Graphics 15

S.J. Gortler

MIT Press, 2012

 CHAPTER 2. LINEAR

y

x

z (coming out)

Figure 2.4: A right handed orthonormal coordinate system. The z axis is coming out
of the page. Also shown is the direction of a rotation about the x axis.

y
z (going in)

x

Figure 2.5: A left handed orthonormal coordinate system. The z axis is going into the
page.

where ~n is the unit vector that is orthogonal to the plane spanned by ~v and w
~ and such
that [~v , w,
~ ~n] form a right handed basis.
In a right handed orthonormal basis ~bt , it is particularly easy to compute the cross
product of two vectors (~bt c) × (~bt d). In particular, its coordinates with respect to ~bt
can be computed as
 
c2 d3 − c3 d2
 c3 d1 − c1 d3 
c1 d2 − c2 d1

2.5 Rotations
The most common linear transformation we will encounter is a rotation. A rotation is
a linear transformation that preserves dot products between vectors and maps a right
handed basis to a right handed basis. So in particular, applying any rotation to a right
handed orthonormal basis always results in another right handed orthonormal basis. In
3D, every rotation fixes an axis of rotation and rotates by some angle about that

Foundations of 3D Computer Graphics 16

S.J. Gortler

MIT Press, 2012

CHAPTER 2. LINEAR

axis.
We begin by describing the 2D case. We start with a vector.

i 
x
h
~v = ~b1 ~b2
y

Let us assume ~bt is a 2D right handed orthonormal basis. Suppose we wish to

rotate ~v by θ degrees counter clockwise about the origin, the coordinates [x′ , y ′ ]t of the
rotated vector. can be computed as

x′ = x cos θ − y sin θ
′
y = x sin θ + y cos θ

This linear transformation can be written as the following matrix expression

h i x 
~b1 ~b2
y
i  
cos θ − sin θ x
h
⇒ ~b1 ~b2
sin θ cos θ y

Likewise, we can rotate the entire basis as

h i
~b1 ~b2
h i  cos θ − sin θ

⇒ ~b1 b2~
sin θ cos θ

For the 3D case, let us also assume that we are using a right handed orthonormal
coordinate system. Then, a rotation of a vector by θ degrees around the z axis of the
basis is expressed as:
 
h i x
~b1 ~b2 ~b3  y 
z
  
h i c −s 0 x
⇒ ~b1 ~b2 ~b3  s c 0   y 
0 0 1 z

where, for brevity, we have used the notation c := cos θ, and s := sin θ. As expected,
this transformation leaves vectors on the third axis just where they were. On each fixed
plane where z is held constant this transformation behaves just like the 2D rotation just
described. The rotation direction can be visualized by grabbing the z axis with your
right hand, with the heel of your hand against the z = 0 plane; the forward direction is
that traced out by your fingertips as you close your hand.

Foundations of 3D Computer Graphics 17

S.J. Gortler

MIT Press, 2012

 CHAPTER 2. LINEAR

Figure 2.6: By setting the rotation amounts on each of the three axes appropriately we
can place the golden disk in any desired orientation.

A rotation about the x axis of a basis can be computed as

 
h i x
~b1 ~b2 ~b3  y 
z
  
h i 1 0 0 x
⇒ ~b1 ~b2 ~b3  0 c −s   y 
0 s c z
Again, the rotation direction can be visualized by grabbing the x axis with your right
hand, with the heel of your hand against the x = 0 plane; the forward direction is that
traced out by your fingertips as you close your hand (see Figure 2.4).
A rotation around the y axis is done using the matrix
 
c 0 s
 0 1 0 
−s 0 c

In some sense, this is all you need to get any 3D rotation. First of all, the compo-
sition of rotations is another rotation. Also, it can be shown that we can achieve any
arbitrary rotations by applying one x, one y, and one z rotation. The angular amounts
of the three rotations are called the xyz-Euler angles. Euler angles can be visualized by
thinking of a set of gimbals, with three movable axes, with three settings to determine
the achieved rotation (see Figure 2.6).
A more direct way to represent an arbitrary rotation, is to pick any unit vector ~k as
the axis of rotation, and directly apply a rotation of θ radians about that axis. Let the
coordinates of ~k be given by the unit coordinate vector [kx , ky , kz ]t . Then this rotation
can be expressed using the matrix
kx2 v + c
 
kx ky v − kz s kx kz v + ky s
 ky kx v + kz s ky2 v + c ky kz v − kx s  (2.5)
kz kx v − ky s kz ky v + kx s kz2 v + c

Foundations of 3D Computer Graphics 18

S.J. Gortler

MIT Press, 2012

CHAPTER 2. LINEAR

where, for shorthand, we have introduced the symbol v := 1 − c. Conversely, it can be

shown that any rotation matrix can be written in this form.
We note that 3D Rotations behave in a somewhat complicated manner. Two ro-
tations around different axes do not commute with each other. Moreover, when we
compose two rotations about two different axis is, what we get is a rotation about some
third axis!
Later in this book, we will introduce the quaternion representation for rotations,
which will be useful for animating smooth transitions between orientations.

2.6 Scales
In order to model geometric objects, we may find it useful to apply scaling operations
to vectors and bases. To scale any vector by factor of α, we can use


h i x
~b1 ~b2 ~b3  y 
z
  
h i α 0 0 x
⇒ ~b1 ~b2 ~b3  0 α 0  y 
0 0 α z

To scale differently along the 3 axis directions we can use the more general form
 
h i x
~b1 ~b2 ~b3  y 
z
  
h i α 0 0 x
⇒ ~b1 ~b2 ~b3  0 β 0   y 
0 0 γ z

This kind of operation is useful, for example, to model an ellipsoid given that we
already know how to model a sphere.

Exercises
Ex. 3 — Which of the following are valid expressions in our notation, and if valid,
what is the resulting type? ~bt M , cM , M −1 c, ~bt N M −1 c?

Ex. 4 — Given that ~at = ~bt M , what are the coordinates of the vector ~bt N c with
respect to the basis ~at ?

Foundations of 3D Computer Graphics 19

S.J. Gortler

MIT Press, 2012

 CHAPTER 2. LINEAR

Ex. 5 — Let ~0 be the zero vector. For any linear transformation L, what is L(~0)?

Ex. 6 — Let T (~v ) be the transformation that adds a specific non-zero constant vector
~k to ~v : T (~v ) = ~v + ~k. Is T a linear transformation?

Foundations of 3D Computer Graphics 20

S.J. Gortler

MIT Press, 2012

Chapter 3

Affine

3.1 Points and Frames

It is useful to think of points and vectors as two different concepts. A point is some
fixed place in a geometric world, while a vector is the motion between two points in the
world. We will use two different notations to distinguish points and vectors. A vector
~v will have an arrow on top, while a point p̃ will have a squiggle on top.
If we think of a vector as representing motion between two points, then the vector
operations (addition and scalar multiplication) have obvious meaning. If we add two
vectors, we are expressing the concatenation of two motions. If we multiply a vector
by a scalar, we are increasing or decreasing the motion by some factor. The zero vector
is a special vector that represents no motion.
These operations don’t really make much sense for points. What should it mean to
add two points together, e.g., what is Harvard Square plus Kendall Square? What does
it mean to multiply a point by a scalar? What is 7 times the North Pole? Is there a zero
point that acts differently than the others?
There is one operation on two points that does sort of make sense: subtraction.
When we subtract one point from another, we should get the motion that it takes to get
from the second point to the first one,

p̃ − q̃ = ~v

Conversely if we start with a point, and move by some vector, we should get to another
point

q̃ + ~v = p̃

It does makes sense to apply a linear transformation to a point. For example we

can think of rotating a point around some other fixed origin point. But it also makes

21
 CHAPTER 3. AFFINE

sense to translate points (this notion did not make sense for vectors). To represent
translations, we need to develop the notion of an affine transform. To accomplish
this, we use 4 by 4 matrices. These matrices are not only convenient for dealing with
affine transformations here, but will also be helpful in describing the camera projection
operation later on (See Chapter 10).

3.1.1 Frames

In an affine space, we describe any point p̃ by first starting from some origin point õ,
and then adding to it a linear combination of vectors. These vectors are expressed using
coordinates ci and a basis of vectors.

 
c1
c2 
h i
ci~bi  ~t
X
p̃ = õ + = ~b1 ~b2 ~b3
 c3  = f c
õ 
i
1

where 1õ is defined to be õ.

The row
h i
~b1 ~b2 ~b3 õ = ~f t

is called an affine frame; it is like a basis, but it is made up of three vectors and a single
point.
In order to specify a point using a frame, we use a coordinate 4-vector with four
entries, with the last entry always being a one. To express a vector using an affine
frame, we use a coordinate vector with a 0 as the fourth coordinate (i.e., it is simply a
sum of the basis vectors). The use of coordinate 4-vectors to represent our geometry
(as well as 4-by-4 matrices) will also come in handy in Chapter 10 when we model the
behavior of a pinhole camera.

3.2 Affine transformations and Four by Four Matrices

Similar to the case of linear transformations, we would like to define a notion of affine
transformations on points by placing an appropriate matrix between a coordinate 4-
vector and a frame.
Let us define an affine matrix to be a 4 by 4 matrix of the form
 
a b c d
 e f g h 
 
 i j k l 
0 0 0 1

Foundations of 3D Computer Graphics 22

S.J. Gortler

MIT Press, 2012

CHAPTER 3. AFFINE

We apply an affine transform to a point p̃ = ~f t c as follows

 
c1
c2 
h i
~b1 ~b2 ~b3 õ  
 c3 
1
  
a b c d c1
h i e f g h   c2
~b1 ~b2 ~b3

⇒ õ 
 i j
 
k l   c3 
0 0 0 1 1

or for short
~f t c ⇒ ~f t Ac

We can verify that the second line of the above describes a valid point, since the
multiplication of
 ′    
x a b c d x
 y′   e f g h   y 
 ′ =  
 z   i j k l  z 
1 0 0 0 1 1

gives us a coordinate 4-vector with a 1 as the fourth entry. Alternatively, we can see
that the multiplication of
 
a b c d
~b′ ~b′ ~b′ õ′ = ~b1 ~b2 ~b3 õ  e f g h 
h i h i 
1 2 3  i j k l 
0 0 0 1

where 0õ is defined to be ~0, gives a valid frame made up of three vectors and a point.
Also note that if the last row of the matrix were not [0, 0, 0, 1] it would generally
give us an invalid result.
Similar to the case of linear transform, we can apply an affine transformation to a
frame as
 
a b c d
h i h i e f g h 
~b1 ~b2 ~b3 õ ⇒ ~b1 ~b2 ~b3 õ  
 i j k l 
0 0 0 1

or for short
~f t ⇒ ~f t A

Foundations of 3D Computer Graphics 23

S.J. Gortler

MIT Press, 2012

 CHAPTER 3. AFFINE

p̃

v
f
t

Figure 3.1: A linear transform is applied to a point. This is accomplished by applying

the linear transform to its offset vector from the origin.

3.3 Applying Linear Transformations to Points

Suppose we have a 3 by 3 matrix representing a linear transformation. we can embed

it into the upper left hand corner of a 4 by 4 matrix, and use this larger matrix to apply
the transformation to a point (or frame).

 
c1
h i c 
~b1 ~b2 ~b3 2 
õ 
 c3 
1
  
a b c 0 c1
e f g 0   c2
h i
~b1 ~b2 ~b3

⇒ õ 
 i j
 
k 0   c3 
0 0 0 1 1

This has the same effect on the ci as it did with linear transformations. If we think
of the point p̃ as being offset from the origin õ by a vector ~v , we see that this has the
same effect as applying the linear transform to the offset vector. So, for example, if
the 3 by 3 matrix is a rotation matrix, this transformation will rotate the point about
the origin (see Figure 3.1). As we will see below in Chapter 4, when applying a linear
transformation to a point, the position of the frame’s origin plays an important role.
We use the following shorthand for describing a 4 by 4 matrix that just applies a
linear transform.
 
l 0
L=
0 1

where L is a 4 by 4 matrix, l is a 3 by 3 matrix, the upper right 0 is a 3 by 1 matrix of

zeros, the lower left 0 is a 1 by 3 matrix of zeros, and the lower right 1 is a scalar.

Foundations of 3D Computer Graphics 24

S.J. Gortler

MIT Press, 2012

CHAPTER 3. AFFINE

3.4 Translations
It is very useful to be able to apply a translation transformation to points. Such transfor-
mations are not linear (See Exercise 6). The main new power of the affine transforma-
tion over the linear transformations is its ability to express translations. In particular, if
we apply the transformation
 
c1
~b1 ~b2 ~b3 õ  c2 
h i 
 c3 
1
  
1 0 0 tx c1
~b1 ~b2 ~b3 õ  0 1 0 ty   c2 
h i  
⇒  0 0 1 tz   c 3 
0 0 0 1 1

we see that its effect on the coordinates is

c1 ⇒ c 1 + tx
c2 ⇒ c 2 + ty
c3 ⇒ c 3 + tz

For a translation we use the shorthand

 
i t
T =
0 1

where T is a 4 by 4 matrix, i is a 3 by 3 identity matrix, the upper right t is a 3 by 1

matrix representing the translation, the lower left 0 is a 1 by 3 matrix of zeros, and the
lower right 1 is a scalar.
Note that if c has a zero in its fourth coordinate, and thus represents a vector instead
of a point, then it is unaffected by translations.

3.5 Putting Them Together

Any affine matrix can be factored into a linear part and a translational part.

    
a b c d 1 0 0 d a b c 0
 e f g h   0 1 0 h   e f g 0 
 =  
 i j k l   0 0 1 l  h i j 0 
0 0 0 1 0 0 0 1 0 0 0 1

Foundations of 3D Computer Graphics 25

S.J. Gortler

MIT Press, 2012

 CHAPTER 3. AFFINE

Figure 3.2: Left: Shape in blue with normals shown in black. Middle: Shape is shrunk
in the y direction and the (un-normalized) normals are stretched in the y direction.
Right: Normals are re-normalized to give correct unit normals of squashed shape.

or in shorthand
    
l t i t l 0
= (3.1)
0 1 0 1 0 1
A = TL (3.2)

Note that since matrix multiplication is not commutative, the order of the multipli-
cation T L matters. An affine matrix can also be factored as A = LT ′ with a different
translation matrix T ′ , but we will not make use of this form.
If L, the linear part of A, is a rotation, we write this as

A = TR (3.3)

In this case we call the A matrix a rigid body matrix, and its transform, a rigid body
transform, or RBT. A rigid body transform preserves dot products between vectors,
handedness of a basis, and distance between points.

3.6 Normals
In computer graphics, we often use the normals of surfaces to determine how a surface
point should be shaded. So we need to understand how the normals of a surface trans-
form when the surface points undergo an affine transformation described by a matrix
A.
One might guess that we could just multiply the normal’s coordinates by A. For
example, if we rotated our geometry, the normals would rotate in exactly the same
way. But using A is, in fact not always correct. For example in Figure 3.2, we squash a
sphere along the y axis. In this case, the actual normals transform by stretching along
the y axis instead of squashing. Here we derive the correct transformation that applies
in all cases.
Let us define the normal to a smooth surface at a point to be a vector that is orthog-
onal to the tangent plane of the surface at that point. The tangent plane is a plane of

Foundations of 3D Computer Graphics 26

S.J. Gortler

MIT Press, 2012

CHAPTER 3. AFFINE

vectors that are defined by subtracting (infinitesimally) nearby surface points, and so
we have

~n · (p˜1 − p˜2 ) = 0

for the normal ~n and two very close points p˜1 and p˜2 on a surface. In some fixed
orthonormal coordinate system, this can be expressed as
   
x1 x0
   y1   y0 
nx ny nz ∗   z1  −  z0  = 0
    (3.4)
1 1

We use a ’∗’ in the forth slot since it is multiplied by 0 and thus does not matter.
Suppose we transform all of our points by applying an affine transformation using
an affine matrix A. What vector remains orthogonal to any tangent vector? Let us
rewrite Equation (3.4) as

  
x1 x0
  −1
 y1   y0 
nx ny nz ∗ A  z1  −  z0
(A     ) = 0

1 1

If we define [x′ , y ′ , z ′ , 1]t = A[x, y, z, 1]t to be the coordinates of a transformed point,

and let [nx′ , ny ′ , nz ′ , ∗] = [nx, ny, nz, 1]tA−1 , then we have

x1′ x0′
   
  y1′   y0′ 
nx′ ny ′ nz ′ ∗ 

 z1′  −  z0′  = 0
   

1 1

and we see that [nx′ , ny ′ , nz ′ ]t are the coordinates (up to scale) of the normal of the
transformed geometry.
Note that since we don’t care about the ’∗’ value, we don’t care about the fourth
column of A−1 . Meanwhile, because A is an affine matrix, so is A−1 and thus the
fourth row of the remaining three columns are all zeros and can safely be ignored.
Thus, using the shorthand
 
l t
A=
0 1

we see that

nx′ ny ′ nz ′ l−1
   
= nx ny nz

and transposing the whole expression, we get

Foundations of 3D Computer Graphics 27

S.J. Gortler

MIT Press, 2012

 CHAPTER 3. AFFINE

nx′
   
nx
 ny ′  = l−t  ny 
nz ′ nz

where l−t is the inverse transpose (equiv. transposed inverse) 3 by 3 matrix. Note
that if l is a rotation matrix, the matrix is orthonormal and thus its inverse transpose
is in fact the same as l. In this case a normal’s coordinates behave just like a point’s
coordinates. For other linear transforms though, the normals behave differently. (See
Figure 3.2.) Also note that the translational part of A has no effect on the normals.

Exercises
Ex. 7 — I claim that the following operationP is a well defined operation on points:
α1 p̃1 + α2 p̃2 for real values αi , as long as 1 = i αi . Show that this can be interpreted
using the operations on points and vectors described at the beginning of Section 3.1.

Foundations of 3D Computer Graphics 28

S.J. Gortler

MIT Press, 2012

Chapter 4

Respect

4.1 The Frame Is Important

In computer graphics we simultaneously keep track of a number of different frames.
For example, we may have a different frame associated with each object in the scene.
The specifics of how we use and organize such frames is described in Chapter 5. Be-
cause of these multiple frames, we need to be especially careful when using matrices
to define transformations.
Suppose we specify a point and a transformation matrix, this does not fully spec-
ified the actual mapping. We must also specify what frame we are using. Here is a
simple example showing this. Suppose we start with some point p̃ as well as the matrix
 
2 0 0 0
 0 1 0 0 
S=  0 0 1 0 


0 0 0 1

Now lets fix a frame ~f t . Using this basis, the point can be expressed using some
appropriate coordinate vector as p̃ = ~f t c. If we now use the matrix to transform
the point, as described in Chapter 3, we get ~f t c ⇒ ~f t Sc. In this case the effect of the
matrix is to transform the point by a scale factor of two from the origin of ~f t , in the
direction of the first (x) axis of ~f t .
Suppose we instead pick some other frame ~at , and suppose that this frame is related
to the original one by the matrix equation ~at = ~f t A. We can express the original point
in the new frame with a new coordinate vector p̃ = ~f t c = ~at d, where d = A−1 c.
Now if we use S to perform a transformation on the point represented with respect
to ~at , we get ~at d ⇒ ~at Sd. In this case we have scaled the same point p̃, but this time
we have scaled it from the origin of ~at in direction of the first (x) axis of ~at . This is
a different transformation (see Figure 4.1). Figure 4.2 shows the same dependence on

29
 CHAPTER 4. RESPECT

f
t ASA−1 c =
at SA−1 c

p̃ = f
t c =
at A−1 c f
t Sc =
at A−1 Sc

f
t
at = f
t A

Figure 4.1: The scaling matrix S is used to scale the point p̃ with respect to two different
frames. This results in two different answers.

frame when rotating a point using a fixed matrix R.

The important thing to notice here is that the point is transformed (non uniform
scaling in this case) with respect to the the frame that appears immediately to the left
of the transformation matrix in the expression. Thus we call this the left of rule. We
read

p̃ = ~f t c ⇒ ~f t Sc

as “p̃ is transformed by S with respect to ~f t ”. We read

p̃ = ~at A−1 c ⇒ ~at SA−1 c

as “p̃ is transformed by S with respect to ~at ”.

We can apply the same reasoning to transformations of frames themselves. We read

~f t ⇒ ~f t S

as “~f t is transformed by S with respect to ~f t ”. We read

~f t = ~at A−1 ⇒ ~at SA−1

as “~f t is transformed by S with respect to ~at ”.

4.1.1 Transforms Using an Auxiliary Frame

There are many times when we wish to transform a frame ~f t in some specific way
represented by a matrix M , with respect to some auxiliary frame ~at . For example, we

Foundations of 3D Computer Graphics 30

S.J. Gortler

MIT Press, 2012

CHAPTER 4. RESPECT

f
t Rc =
at A−1 Rc

p̃ = f
t c =
at A−1 c

f
t ARA−1 c =
at RA−1 c

f
t
at = f
t A

Figure 4.2: The rotation matrix R is used to rotate the point p̃ with respect to two
different frames. This results in two different answers.

may be using some frame to model the planet Earth, and we now wish the Earth to
rotate around the Sun’s frame.
This is easy to do as long as we know the matrix relating ~f t and ~at . For example
we may know that

~at = ~f t A

The transformed frame can then be expressed as

~f t (4.1)
= ~at A−1 (4.2)
⇒ ~at M A−1 (4.3)
= ~f t AM A−1 (4.4)

In the first line, we rewrite the frame ~f t using ~at . In the second line we transform the
frame system using the “left of” rule; we transform our frame using M with respect to
~at . In the final line, we simply rewrite the expression to remove the auxiliary frame.

4.2 Multiple Transformations

We can use this “left of” rule to interpret sequences of multiple transformations. Again,
recall that, in general, matrix multiplication is not commutative. In the following 2D
example, let R be a rotation matrix and T a translation matrix, where the translation
matrix has the effect of translating by one unit in the direction of the first axis and the

Foundations of 3D Computer Graphics 31

S.J. Gortler

MIT Press, 2012

 CHAPTER 4. RESPECT

rotation matrix has the effect of rotating by θ degrees about the frame’s origin. (See
figure 4.3).
We will now interpret the following transformation
~f t ⇒ ~f t T R

We do this by breaking up the transformation into two steps. In the first step

~f t ⇒ ~f t T = f~′ t

This is interpreted as: ~f t is transformed by T with respect to ~f t and we call the resulting
frame ~f ′t .
In the second step,
~f t T ⇒ ~f t T R

or equivalently
t t
f~′ ⇒ f~′ R
This is interpreted as: ~f ′t is transformed by R with respect to ~f ′t .
We can also interpret the composed transformations in another valid way. This is
done by applying the rotation and translation in the other order. In the first step
~f t ⇒ ~f t R = ~f ◦t
~f t is transformed by R with respect to ~f t and we call the resulting frame ~f ◦t . In the
second step,
~f t R ⇒ ~f t T R
~f ◦t is transformed by T with respect to ~f t .

These are just two different interpretations of the same final composed transforma-
tions. 1) Translate with respect to ~f t then rotate with respect to the intermediate frame.
2) Rotate respect to ~f t then translate with respect to the original frame ~f t . At times
it will be more convenient to use the first interpretation, and at other times it may be
more convenient to use the second one.
These types of interpretations are often summarized as follows: If we read transfor-
mations from left to right, then each transform is done with respect to a newly created
“local” frame. If we read the transformations from right left, then each transform is
done with respect to the original “global” frame.

Exercises
Ex. 8 — Using the definitions of Section 4.2, draw two different sketches illustrating
the transformation: ~f t ⇒ ~f t RT (Compare to Figure 4.3).

Foundations of 3D Computer Graphics 32

S.J. Gortler

MIT Press, 2012

CHAPTER 4. RESPECT

f
t f
′t f
t f
′t

(a) Local translation (b) Local rotation

f
◦t f
◦t

f
t f
t
(c) Global rotation (d) Global translation

Figure 4.3: Two ways of interpreting the expression ~f t T R

Ex. 9 — Suppose ~f t is an orthonormal frame, and we apply the transform ~f t ⇒ ~f t ST ,

where S is a matrix that applies a uniform scale by a factor of 2, and T translates by 1
along the x axis. How far does the frame’s origin move, measured in the original units
of ~f t ?

Ex. 10 — Given the two orthonormal frames ~at and ~bt shown below

bt

d4

d3
d2

at θ
d1
with distances given by the positive quantities di . What are the matrices R and T such
that ~bt = ~at T R? What are the matrices R and T such that ~bt = ~at RT ? (Note: do
this without using trigonometric terms in the matrix T .)

Ex. 11 — Given the three frames below

Foundations of 3D Computer Graphics 33

S.J. Gortler

MIT Press, 2012

 CHAPTER 4. RESPECT

bt

θ
d d

φ

at φ
ct
~
Suppose that b = ~a N and ~ct = ~at M . Express the matrix M using only the symbols
t t

N and θ.

Foundations of 3D Computer Graphics 34

S.J. Gortler

MIT Press, 2012

Chapter 5

Frames In Graphics

Now that we have covered the basics of points and matrix transformations, we can
describe how they are typically used in computer graphics. We will then discuss various
modeling manipulation and imaging operations.

5.1 World, Object and Eye Frames

When describing the geometry of a scene, we start with a basic right handed orthonor-
mal frame w~ t called the world frame. We never alter the world frame. Other frames
can then be described with respect to this world frame. If we express the location of
some point using coordinates with respect to the world frame, then these will be called
world coordinates.
Suppose we wish to model, say, a moving car in our scene. We would like to model
the geometry of the object using vertex coordinates that do not need to be aware of
the global placement of the object in the scene. Likewise, we would like to move the
car around the scene without changing any of these coordinates. This is accomplished
using object frames.
To every object in the scene we associate an right handed orthonormal object frame
~ot . We can now express the location of parts of the objects using coordinates with
respect to the object’s coordinate system. These are called object coordinates and will
be stored in our computer program. To move the entire object, we simply update ~ot
and do not need to change any of the object coordinates of our points.
The relationship between the object’s coordinate system and the world frame is
represented by an affine 4 by 4 (rigid body) matrix which we call O. That is,

~ot = w
~ tO

In our computer program, we will store the matrix O, with the understanding that it

35
 CHAPTER 5. FRAMES IN GRAPHICS

z x

(a) The frames

(b) The eye’s view

Figure 5.1: The world frame is in red, the objects frame is in green and the eye frame
is in blue. The eye is looking down its negative z-axis towards the object.

relates the world frame to object’s coordinate system using the equation above. To
move the frame ~ot , we change the matrix O.
In the real world, when we want to create a 2D picture of a 3D environment, we
place a camera somewhere in the scene. The position of each object in the picture
is based on its 3D relationship to the camera, i.e., its coordinates with respect to an
appropriate basis. In computer graphics we achieve this by having a right handed
orthonormal frame ~et , called the eye frame. We interpret the eye as looking down its
negative z axis and making a picture (see Figure 5.1). The eye frame is related by some
(rigid body) 4-by-4 matrix E

~et = w
~ tE

In a computer program, we will explicitly store the matrix E.

Given a point

~ t Oc = ~et E −1 Oc
p̃ = ~ot c = w

we call c its object coordinates, Oc its world coordinates and E −1 Oc its eye coor-
dinates. We use an “o” subscript for object coordinates, a “w” subscript for world

Foundations of 3D Computer Graphics 36

S.J. Gortler

MIT Press, 2012

 Another Random Document on
Scribd Without Any Related Topics
 Oh dreadful thought, if all our sires and we
Are but foundations of a race to be,—
Stones which one thrusts in earth, and builds thereon
A write delight, a Parian Parthenon,
And thither, long thereafter, youth and maid
Seek with glad brows the alabaster shade.

And in processions’ pomp together bent

Still interchange their sweet words innocent,—
Not caring that those mighty columns rest
Each on the ruin of a human breast,—
That to the shrine the victor’s chariot rolls
Across the anguish of ten thousand souls!

“Well was it that our fathers suffered thus,”

I hear them say, “that all might end in us;
Well was it here and there a bard should feel
Pains premature and hurt that none could heal;
These were their preludes, thus the race began;
So hard a matter was the birth of Man.”

And yet these too shall pass and fade and flee,
And in their death shall be as vile as we,
Nor much shall profit with their perfect powers
To have lived a so much sweeter life than ours,
When at the last, with all their bliss gone by,
Like us those glorious creatures come to die,
With far worse woe, far more rebellious strife
Those mighty spirits drink the dregs of life.

F. W. H. Myers (The Implicit Promise of Immortality).

It will be observed that Myers, like Swinburne, handled the old heroic couplet in
a masterly manner, undreamt of by Pope, Dryden, and their generation.
God’s works—paint any one, and count it crime
To let a truth slip. Don’t object, “His works
Are here already; nature is complete:
Suppose you reproduce her (which you can’t)
There’s no advantage! You must beat her then.”
For, don’t you mark? we’re made so that we love
First when we see them painted, things we have passed
Perhaps a hundred times nor cared to see;
And so they are better, painted—better to us
Which is the same thing. Art was given for that;
God uses us to help each other so,
Lending our minds out.

R. Browning (Fra Lippo Lippi).

For the folk through the fretful hours are hurled

On the ruthless rush of the wondrous world,
And none has leisure to lie and cull
The blossoms, that made life beautiful
In that old season when men could sing
For dear delight in the risen Spring
And Summer ripening fruit and flower.
Now carefulness cankers every hour;
We are too weary and sad to sing;
Our pastime’s poisoned with thought-taking.

John Payne (Tournesol).

 I am much engaged, an old man and out of health, and I cannot
spare time to answer your questions fully,—nor indeed can they be
answered. Science has nothing to do with Christ, except in so far as
the habit of scientific research makes a man cautious in admitting
evidence. For myself, I do not believe that there ever has been any
Revelation. As for a future life every man must judge for himself
between conflicting vague probabilities. Wishing you happiness, I
remain, &c.
Charles Darwin (Letter to von Müller, June 5, 1879).

This letter is reproduced in the Life and Letters, but evidently Francis Darwin did
not know that the “German youth” to whom he says it was written was Baron
Ferdinand von Müller, K.C.M.G. (1825-1896), then fifty-three years of age! Von
Müller was director of the Melbourne Botanical Gardens from 1857 to 1873, and
died in Melbourne in 1896. He did important work in Australian botany.
As regards Darwin’s letter, it seems to me that a sufficient reason why a great
and lovable man, who was at first a convinced believer in the immortality of the
soul, became an agnostic is given in the next quotation. The higher aesthetic part
of his brain had become atrophied.
Darwin himself thought that he had not given sufficient consideration to
religious questions and was exceedingly anxious that his own agnostic views
should not influence others,

I have said that in one respect my mind has changed during the
last twenty or thirty years. Up to the age of thirty, or beyond it,
poetry of many kinds, such as the works of Milton, Gray, Byron,
Wordsworth, Coleridge, and Shelley, gave me great pleasure, and
even as a school-boy I took intense delight in Shakespeare,
especially in the historical plays. I have also said that formerly
pictures gave me considerable, and music very great delight. But
now for many years I cannot endure to read a line of poetry: I have
tried lately to read Shakespeare and found it so intolerably dull that
it nauseates me. I have also almost lost my taste for pictures or
music.... My mind seems to have become a kind of machine for
grinding general laws out of large collections of facts, but why this
should have caused the atrophy of that (aesthetic) part of the brain
alone, on which the higher tastes depend, I cannot conceive.... The
loss of these tastes is a loss of happiness, and may possibly be
injurious to the intellect, and more probably to the moral character,
by enfeebling the emotional part of our nature.
Charles Darwin.

This is from autobiographical notes made by Darwin for his children, and not
intended for publication.

God be thanked, the meanest of his creatures

Boasts two soul-sides, one to face the world with,
One to show a woman when he loves her!

R. Browning (One Word More).

CHILDREN’S HYMN ON THE COAST OF BRITTANY.

 At length has come the twilight dim.
The sun has set, the day has died;
And now we sing Thy holy hymn,
O Mary maid, at eventide.

To Jewry, to that far-off land,

Erstwhile there came a little Child;
You led Him softly by the hand,
He was so very small and mild.

Like us, He could not find his way,

Although He was Our Lord, the King;
And so we beg we may not stray,
Nor do a sad or foolish thing.

Teach us the prayer that Jesus said,

The words you sang and murmured low,
When He was in His tiny bed,
And all the earth was dark and slow.

Hushed are the trees, and the small wise bees,

Our fathers are on the deep,—
Little Mother, be good to us, please!
It is time to go asleep.

Vincent O’Sullivan.

WESLEY’S MEDICAL PRESCRIPTIONS

For an Ague:—Make six middling pills of cobwebs. Take one a little
before the cold fit; two a little before the next fit (suppose the next
day); the other three, if need be, a little before the third fit. This
seldom fails.
 A Cut:—Bind on toasted cheese. This will cure a deep cut.
A Fistula:—Grind an ounce of sublimate mercury as fine as
possible.... (Two quarts of water to be added, then half a spoonful
with two spoonfuls of water to be taken fasting every other day), ...
In forty days this will also cure any cancer, any old sore or King’s
evil.
The Iliac Passion:—Hold a live puppy constantly on the belly.
John Wesley (Primitive Physic. Ed. 1780).

The iliac passion, now known as ileus, is a severe colic due to intestinal
obstruction.
It seems strange that so eminent a man should have believed in these absurd
prescriptions, but as a matter of fact the book generally is much more sane and
sound than one would expect from the habits and state of knowledge of the time.
For example, in his rules of health Wesley strongly advises the practice of cold
bathing, cleanliness, open-air exercise, moderation of food, etc. Also these
prescriptions are chosen for their absurdity—in each case other more sensible
remedies are offered. But Wesley in his preface says that he has omitted
altogether from his book Cinchona bark, because it is “extremely dangerous.” This
means that in regard to ague he omitted the only efficient remedy—which was
much more unfortunate than his prescribing cobweb pills.
This book went to thirty-six editions between 1747 and 1840.
 “When shall our prayers end?”
I tell thee, priest, when shoemakers make shoes,
That are well sewed, with never a stitch amiss,
And use no craft in uttering of the same;
When tinkers make no more holes than they found,
When thatchers think their wages worth their work,
When Davie Diker digs and dallies not,
When horsecorsers beguile no friends with jades,
When printers pass no errors in their books,
When pewterers infect no tin with lead,
When silver sticks not on the Teller’s fingers,
When sycophants can find no place in Court, ...
When Laïs lives not like a lady’s peer
Nor useth art in dyeing of her hair....

George Gascoigne (1525?-1577) (The Steele Glas).

All our life is a meeting of cross-roads, where the choice of

directions is perilous.
Victor Hugo.
 Rose-cheeked Laura, come;
Sing thou smoothly with thy beauty’s
Silent music, either other
Sweetly gracing.
Lovely forms do flow
From concent divinely framed;
Heaven is music, and thy beauty’s
Birth is heavenly.
These dull notes we sing
Discords need for helps to grace them,
Only beauty purely loving
Knows no discord,
But still moves delight,
Like clear springs renewed by flowing,
Ever perfect, ever in them-
Selves eternal.

Thomas Campion (died 1619).

Richard Lovelace (1618-1655) subsequently wrote (Orpheus to Beasts):

O, could you view the melodie

Of ev’ry grace,
And musick of her face,
You’d drop a teare,
Seeing more harmonie
In her bright eye,
Then now you heare.

Then = than. See next quotation.

 I think the deep love he had for that sweet, rounded blossom-like
dark-eyed Hetty, of whose inward self he was really very ignorant,
came out of the very strength of his nature, and not out of any
inconsistent weakness. Is it any weakness, pray, to be wrought on
by exquisite music?—to feel its wondrous harmonies searching the
subtlest windings of your soul, the delicate fibres of life where no
memory can penetrate, and binding together your whole being past
and present in one unspeakable vibration; melting you in one
moment with all the tenderness, all the love that has been scattered
through the toilsome years; concentrating in one emotion of heroic
courage or resignation all the hard-learnt lessons of self-renouncing
sympathy; blending your present joy with past sorrow, and your
present sorrow with all your past joy? If not, then neither is it a
weakness to be so wrought upon by the exquisite curves of a
woman’s cheek and neck and arms, by the liquid depths of her
beseeching eyes, or the sweet childish pout of her lips. For the
beauty of a lovely woman is like music: what can one say more?
Beauty has an expression beyond and far above the one woman’s
soul that it clothes, as the words of genius have a wider meaning
than the thought that prompted them; it is more than a woman’s
love that moves us in a woman’s eyes—it seems to be a far-off
mighty love that has come near to us, and made speech for itself
there; the rounded neck, the dimpled arm, move us by something
more than their prettiness—by their close kinship with all we have
known of tenderness and peace. The noblest nature sees the most
of this impersonal expression in beauty, and for this reason, the
noblest nature is often the most blinded to the character of the
woman’s soul that the beauty clothes. Whence, I fear, the tragedy of
human life is likely to continue for a long time to come in spite of
mental philosophers who are ready with the best receipts for
avoiding all mistakes of the kind.
George Eliot (Adam Bede).

George Eliot would not know the preceding poem by Campion, whose lyrics had
been forgotten until A. H. Bullen revived them in 1889; and most probably also
she did not know Lovelace’s poem, as it is not one of the two or three lyrics by
which alone he is remembered.

Alas, how soon the hours are over

Counted us out to play the lover!
And how much narrower is the stage
Allotted us to play the sage!
But when we play the fool, how wide
The theatre expands! beside,
How long the audience sits before us!
How many prompters! What a chorus!

W. S. Landor.

The degree of vision that dwells in a man is a correct measure of

the man. If called to define Shakespeare’s faculty, I should say
superiority of Intellect, and think I had included all under that. What
indeed are faculties? We talk of faculties as if they were distinct,
things separable; as if a man had intellect, imagination, fancy, &c.,
as he has hands, feet, and arms. That is a capital error. Then again,
we hear of a man’s “intellectual nature,” and of his “moral nature,”
as if these again were divisible and existed apart.... We ought to
know, and to keep forever in mind, that these divisions are at
bottom but names; that man’s spiritual nature, the vital Force which
dwells in him, is essentially one and indivisible; that what we call
imagination, fancy, understanding, and so forth, are but different
figures of the same Power of Insight, all indissolubly connected with
each other, physiognomically related; that if we knew one of them,
we might know all of them. Morality itself, what we call the moral
quality of a man, what is this but another side of the one vital Force
whereby he is and works?.... Without hands a man might have feet,
and could still walk: but, consider it,—without morality, intellect were
impossible for him; a thoroughly immoral man could not know
anything at all! To know a thing, what we can call knowing, a man
must first love the thing, sympathize with it; that is, be virtuously
related to it.... Nature, with her truth, remains to the bad, to the
selfish and the pusillanimous forever a sealed book: what such can
know of Nature is mean, superficial, small.
Carlyle (Heroes and Hero Worship, III).

A little I will speak. I love thee then

Not only for thy body packed with sweet
Of all this world....
Not for this only do I love thee, but
Because Infinity upon thee broods;
And thou art full of whispers and of shadows.
Thou meanest what the sea has striven to say
So long, and yearnèd up the cliffs to tell;
Thou art what all the winds have uttered not,
What the still night suggesteth to the heart.
Thy voice is like to music heard ere birth,
Some spirit lute touched on a spirit sea;
Thy face remembered is from other worlds,
It has been died for, though I know not when,
It has been sung of, though I know not where.

Stephen Phillips (Marpessa).

Sometimes thou seem’st not as thyself alone,
But as the meaning of all things that are.

D. G. Rossetti (Heart’s Compass)

“IMBUTA”
The new wine, the new wine,
It tasteth like the old,
The heart is all athirst again,
The drops are all of gold;
We thought the cup was broken,
And we thought the tale was told,
But the new wine, the new wine,
It tasteth like the old!

The flower of life had faded,

The leaf was in its fall,
The winter seemed so early
To have reached us, once for all;
But now the buds are breaking,
There is grass above the mould,
And the new wine, the new wine.
It tasteth like the old!

The earth had grown so dreary,

The sky so dull and grey;
One was weeping in the darkness,
One was sorrowing through the day:
But a light from heaven gleams again,
On water, wood, and wold,
And the new wine, the new wine,
It tasteth like the old!

For the loving lips are laughing,

And the loving face is fair,
Though a phantom hand is on the board,
And phantom eyes are there;
The phantom eyes are soft and sad,
The phantom hand is cold,
But the new wine, the new wine,
It tasteth like the old!
 We dare not look, we turn away,
The precious draught to drain,
’Twere worse than madness, surely now,
To lose it all again;
To quivering lip, with clinging grasp,
The fatal cup we hold,
For the new wine, the new wine,
It tasteth like the old!
And life is short, and love is life,
And so the tale is told,
Though the new wine, the new wine,
It tasteth like the old.

G. J. Whyte-Melville.

The title evidently refers to Horace Ep. 1, 2, 69, 70, Quo semel est Imbuta
recens servabit odorem testa diu. “The scent which once has flavoured the fresh
jar will be preserved in it for many a day.” Moore no doubt had the same passage
in his mind when, speaking of the memories of past joys, he wrote:

You may break, you may ruin the vase if you will,
But the scent of the roses will hang round it still.

So Whyte-Melville says that when love is poured again into the heart of a man
who has lost his first love, “The new wine, the new wine, It tasteth like the old.”

I strove with none, for none was worth my strife,

Nature I loved and, next to Nature, Art:
I warm’d both hands before the fire of Life;
It sinks, and I am ready to depart.

W. S. Landor.
 The Toucan has an enormous bill, makes a noise like a puppy dog,
and lays his eggs in hollow trees. How astonishing are the freaks
and fancies of nature! To what purpose, we say, is a bird placed in
the woods of Cayenne with a bill a yard long, making a noise like a
puppy dog, and laying eggs in hollow trees? The Toucans, to be
sure, might retort, to what purpose were gentlemen in Bond Street
created? To what purpose were certain foolish prating Members of
Parliament created?—pestering the House of Commons with their
ignorance and folly, and impeding the business of the country?
There is no end of such questions. So we will not enter into the
metaphysics of the Toucan.
Sydney Smith (Review of “Waterton’s Travels in South America”).

Above green-flashing plunges of a weir, and shaken by the

thunder below, lilies, golden and white, were swaying at anchor
among the reeds. Meadow-sweet hung from the banks thick with
weed and trailing bramble, and there also hung a daughter of earth.
Her face was shaded by a broad straw hat with a flexible brim that
left her lips and chin in the sun, and, sometimes nodding, sent forth
a light of promising eyes. Across her shoulders, and behind, flowed
large loose curls, brown in shadow, almost golden where the ray
touched them. She was simply dressed, befitting decency and the
season. On a closer inspection you might see that her lips were
stained. This blooming young person was regaling on dewberries.
They grew between the bank and the water. Apparently she found
the fruit abundant, for her hand was making pretty progress to her
mouth. Fastidious youth, which revolts at woman plumping her
exquisite proportions on bread-and-butter, and would (we must
suppose) joyfully have her scraggy to have her poetical, can hardly
object to dewberries. Indeed the act of eating them is dainty and
induces musing. The dewberry is a sister to the lotus, and an
innocent sister. You eat: mouth, eye, and hand are occupied, and the
undrugged mind free to roam. And so it was with the damsel who
knelt there. The little skylark went up above her, all song, to the
smooth southern cloud lying along the blue: from a dewy copse dark
over her nodding hat the blackbird fluted, calling to her with thrice
mellow note: the kingfisher flashed emerald out of green osiers: a
bow-winged heron travelled aloft, seeking solitude: a boat slipped
toward her, containing a dreamy youth; and still she plucked the
fruit, and ate, and mused, as if no fairy prince were invading her
territories, and as if she wished not for one, or knew not her wishes.
Surrounded by the green shaven meadows, the pastoral summer
buzz, the weirfall’s thundering white, amid the breath and beauty of
wild flowers, she was a bit of lovely human life in a fair setting; a
terrible attraction. The Magnetic Youth leaned round to note his
proximity to the weir-piles, and beheld the sweet vision. Stiller and
stiller grew nature, as at the meeting of two electric clouds. Her
posture was so graceful, that though he was making straight for the
weir, he dared not dip a scull. Just then one enticing dewberry
caught her eyes. He was floating by unheeded, and saw that her
hand stretched low, and could not gather what it sought. A stroke
from his right brought him beside her. The damsel glanced up
dismayed, and her whole shape trembled over the brink. Richard
sprang from his boat into the water. Pressing a hand beneath her
foot, which she had thrust against the crumbling wet sides of the
bank to save herself, he enabled her to recover her balance, and
gain safe earth, whither he followed her....
To-morrow this place will have a memory—the river and the
meadow, and the white falling weir: his heart will build a temple
here; and the skylark will be its high-priest, and the old blackbird its
glossy-gowned chorister, and there will be a sacred repast of
dewberries.
George Meredith (The Ordeal of Richard Feverel).
 LETTY’S GLOBE

When Letty had scarce passed her third glad year,

And her young artless words began to flow,
One day we gave the child a coloured sphere
Of the wide earth, that she might mark and know,
By tint and outline, all its sea and land.
She patted all the world; old empires peeped
Between her baby fingers; her soft hand
Was welcome at all frontiers. How she leaped
And laughed and prattled in her world-wide bliss;
But when we turned her sweet unlearnèd eye
On our own isle, she raised a joyous cry—
“Oh! yes, I see it, Letty’s home is there!”
And, while she hid all England with a kiss,
Bright over Europe fell her golden hair.

Charles Tennyson Turner.

Charles Tennyson, a brother of Lord Tennyson and author with him of Poems by
Two Brothers, took the name of Turner.
 O may I join the choir invisible
Of those immortal dead who live again
In minds made better by their presence: live
In pulses stirred to generosity,
In deeds of daring rectitude, in scorn
For miserable aims that end with self,
In thoughts sublime that pierce the night like stars,
And with their mild persistence urge man’s search
To vaster issues.
So to live is heaven:
To make undying music in the world....
This is life to come,
Which martyr’d men have made more glorious
For us who strive to follow. May I reach
That purest heaven, be to other souls
The cup of strength in some great agony,
Enkindle generous ardour, feed pure love,
Beget the smiles that have no cruelty—
Be the sweet presence of a good diffused,
And in diffusion ever more intense,
So shall I join the choir invisible
Whose music is the gladness of the world.

George Eliot.

There is an infinite pathos in these lines. Having lost her faith in a future life,
George Eliot tries to find consolation in the thought that, when she has passed
into nothingness—when she “joins the choir invisible”—she will have done
something to ennoble the minds of those who come after her. But why should
generation after generation of insect-lives waste themselves in raising and
purifying the minds of the generations that follow, if all in turn pass into
nothingness? The higher and purer men became, the more they would love their
fellow-beings and the more they would shudder at the insensate pain and cruelty
in the world—the physical torture they themselves endure, and the mental torture
both of losing for ever those they love and of seeing the sufferings of others. One
should act in conformity with one’s belief. Instead of thus adding greater pain and
sorrow to each succeeding generation, the effort should be to coarsen and
brutalize our natures, so that love, duty, and moral aspiration shall disappear, and
we shall cease to be saddened by the appalling cruelty of our existence. Our lives
should, in fact, correspond with the brutal, ugly and stupid scheme of the
universe.
This is the direct answer to George Eliot, allowing her very important
assumption that we have a duty towards others, including those who come after
us. But this assumption is logically unwarranted, if at the end of our brief years we
pass into nothingness and have no further concern with any living being. This
brings us to a familiar train of argument. Why should we be irresistibly impelled to
sacrifice ourselves for the good of others? And, apart from altruism, why should
we develop our own higher attributes—why seek to ennoble our own selves, since
those selves disappear? Why fill with jewels the hollow log that is to be thrown on
the fire? Why are we swayed by a sense of honour, a desire for justice, a love of
purity and truth and beauty, a craving for affection, a thirst for knowledge, which
persist up to the very gates of death? To take an illustration of Edward Caird’s, is
not the path of life which is so traversed like the path of a star to the astronomer,
which enables him to prophesy its future course—beyond the end which hides it
from our eyes? Otherwise, to use another simile, it is as though Pheidias spent his
life sculpturing in snow.
(This does not mean, as the sceptic usually sneers, that the virtuous man
merely desires a reward for his virtuous conduct. It is an inquiry why he is virtuous
—what is a sane view of the scheme of the universe.)
In forming the conclusion that there was no possible future for man, George
Eliot and an immense number of other thinkers of her time made also the vast
assumption that there was nothing left to discover. Blanco White’s sonnet alone
might have taught them the folly of such premature judgments. Or we may take
an illustration, used by F. W. H. Myers, namely, the discovery that, far beyond the
red and the violet of the spectrum or the rainbow, extend rays that have been
(and will for ever be) invisible to our eyes. Since George Eliot’s time the Society for
Psychical Research has during the last thirty-five years accumulated unanswerable
evidence of survival after death.
 Why are we weigh’d upon with heaviness,
And utterly consumed with sharp distress,
While all things else have rest from weariness?
All things have rest: why should we toil alone,
We only toil, who are the first of things,
And make perpetual moan,
Still from one sorrow to another thrown:
Nor ever fold our wings,
And cease from wanderings,
Nor steep our brows in slumber’s holy balm;
Nor harken what the inner spirit sings,
“There is no joy but calm!”
Why should we only toil, the roof and crown of things?...

Hateful is the dark-blue sky,

Vaulted o’er the dark-blue sea.
Death is the end of life; ah, why
Should life all labour be?
Let us alone. Time driveth onward fast,
And in a little while our lips are dumb.
Let us alone. What is it that will last?
All things are taken from us, and become
Portions and parcels of the dreadful Past.
Let us alone. What pleasure can we have
To war with evil? Is there any peace
In ever climbing up the climbing wave?
All things have rest, and ripen toward the grave
In silence; ripen, fall and cease:
Give us long rest or death, dark death, or dreamful ease.

Tennyson (The Lotos-Eaters).

See preceding quotation.

 We may well begin to doubt whether the known and the natural
can suffice for human life. No sooner do we try to think so than
pessimism raises its head. The more our thoughts widen and
deepen, as the universe grows upon us and we become accustomed
to boundless space and time, the more petrifying is the contrast of
our own insignificance, the more contemptible become the pettiness,
shortness, fragility of the individual life. A moral paralysis creeps
upon us. For awhile we comfort ourselves with the notion of self-
sacrifice; we say, what matter if I pass, let me think of others! But
the other has become contemptible no less than the self; all human
griefs alike seem little worth assuaging, human happiness too paltry
at the best to be worth increasing. The whole moral world is reduced
to a point; good and evil, right and wrong become infinitesimal
ephemeral matters, while eternity and infinity remain attributes of
that only which is outside the sphere of morality. Life becomes more
intolerable the more we know and discover, so long as everything
widens and deepens except our own duration, and that remains as
pitiful as ever. The affections die away in a world where everything
great and enduring is cold; they die of their own conscious
feebleness and bootlessness.
Sir J. R. Seeley (Natural Religion).

See the two preceding quotations.

 Death stands above me, whispering low
I know not what into my ear;
Of his strange language all I know
Is, there is not a word of fear.

W. S. Landor

LOVE-SWEETNESS

Sweet dimness of her loosened hair’s downfall

About thy face; her sweet hands round thy head
In gracious fostering union garlanded;
Her tremulous smiles; her glances’ sweet recall
Of love; her murmuring sighs memorial;
Her mouth’s culled sweetness by thy kisses shed
On cheeks and neck and eyelids, and so led
Back to her mouth which answers there for all:—
What sweeter than these things, except the thing
In lacking which all these would lose their sweet:—
The confident heart’s still fervour: the swift beat
And soft subsidence of the spirit’s wing,
Then when it feels, in cloud-girt wayfaring,
The breath of kindred plumes against its feet?

D. G. Rossetti.

Jesus saith, Wherever there are two, they are not without God;
and wherever there is one alone, I say, I am with him. Raise the
stone and there thou shalt find me; cleave the wood and there am I.
(Logia of Jesus).

This is one of the Logia or Sayings of Jesus written on papyrus in the third
century and discovered in Egypt by Grenfell and Hunt in 1897. The italics, of
course, are mine.

The first of all Gospels is this, that a Lie cannot endure for ever.

Meanwhile it is singular how long the rotten will hold together,

provided you do not handle it roughly.

There are quarrels in which even Satan, bringing help, were not
unwelcome; even Satan, fighting stiffly, might cover himself with
glory—of a temporary nature.

... Nothing but two clattering jaw-bones, and a head vacant,

sonorous, of the drum species.

Thou art bound hastily for the City of Nowhere; and wilt arrive!
Carlyle (French Revolution).
 It is interesting to learn from a correspondent of The Spectator (Feb. 17, 1917)
that Carlyle wrote two verses which he combined with Shakespeare’s “Fear no
more the heat o’ the sun” (Cymbeline iv, 2) to make a requiem, of which he was
very fond:

Fear no more the heat o’ the sun,

Nor the furious winter’s rages;
Thou thy worldly task hast done,
Home art gone, and ta’en thy wages.

Hurts thee now no harsh behest,

Toil, or shame, or sin, or danger;
Trouble’s storm has got to rest,
To his place the wayworn stranger.

Want is done, and grief and pain,

Done is all thy bitter weeping:
Thou art safe from wind and rain
In the Mother’s bosom sleeping.

Fear no more the heat o’ the sun,

Nor the furious winter’s rages:
Thou thy worldly task hast done,
Home art gone and ta’en thy wages.

It takes two for a kiss,

Only one for a sigh;
Twain by twain we marry,
One by one we die.
Joy has its partnerships,
Grief weeps alone;
Cana had many guests,
Gethsemane had none.

Frederic Lawrence Knowles.

 Byron in “Don Juan” says:

All who joy would win must share it,

Happiness was born a twin.

(Speaking of the rare and exalted nature of Dorothea, who has

adopted the normal, domestic married life) Her finely-touched spirit
had still its fine issues, though they were not widely visible. Her full
nature, like that river of which Cyrus broke the strength, spent itself
in channels which had no great name on the earth. But the effect of
her being on those around her was incalculably diffusive; for the
growing good of the world is partly dependent on unhistoric acts;
and that things are not so ill with you and me, as they might have
been, is half owing to the number who lived faithfully a hidden life,
and rest in unvisited tombs.
George Eliot (Middlemarch).

This passage, which finely expresses an important truth, is at the end of

Middlemarch. The reference is to a story of Herodotus. He says that Cyrus, the
Persian, was angry with the river Gyndes (Diyalah), because it had drowned one of
the white horses, which, as being sacred to the sun, accompanied the expedition.
He, therefore, employed his army to divert the river into 360 channels
(representing the number of days in the year). The story was probably told to
Herodotus as explaining the great irrigation system that existed in Mesopotamia.
The Diyalah flows into the Tigris not far from Baghdad.
Any sort of meaning looks intense
When all beside itself means and looks nought.

R. Browning (Fra Lippo Lippi).

Hold, Time, a little while thy glass,

And, Youth, fold up those peacock wings!
More rapture fills the years that pass
Than any hope the future brings;
Some for to-morrow rashly pray,
And some desire to hold to-day.
But I am sick for yesterday....

Ah! who will give us back the past?

Ah! woe, that youth should love to be
Like this swift Thames that speeds so fast,
And is so fain to find the sea,—
That leaves this maze of shadow and sleep,
These creeks down which blown blossoms creep,
For breakers of the homeless deep.

Edmund Gosse (Desiderium).

 The night has a thousand eyes,
And the day but one;
Yet the light of the bright world dies
With the dying sun.

The mind has a thousand eyes,

And the heart but one;
Yet the light of a whole life dies,
When love is done.

F. W. Bourdillon.

See reference to this poem in Preface.

But to come again unto Apelles, this was his manner and custom
besides, which he perpetually observed, that no day went over his
head, but what businesse soever he had otherwise to call him away,
he would make one draught or other (and never misse) for to
exercise his hand and keepe it in use, inasmuch as from him grew
the proverbe, Nulla dies sine linea, i.e. Be alwaies doing somewhat,
though you doe but draw a line. His order was when he had finished
a piece of work or painted table, and layd it out of his hand, to set it
forth in some open gallerie or thorowfare, to be seen of folke that
passed by, and himselfe would lie close behind it to hearken what
faults were found therewith; preferring the judgment of the common
people before his owne, and imagining they would spy more
narrowly, and censure his doings sooner than himselfe: and as the
tale is told, it fell out upon a time, that a shoomaker as he went by
seemed to controlle his workmanship about the shoo or pantofle
that he had made to a picture, and namely, that there was one
latchet fewer than there should be: Apelles acknowledging that the
man said true indeed, mended that fault by the next morning, and
set forth his table as his manner was. The same shoomaker
comming again the morrow after, and finding the want supplied
which he noted the day before, took some pride unto himselfe, that
his former admonition had sped so well, and was so bold as to cavil
at somewhat about the leg. Apelles could not endure that, but
putting forth his head from behind the painted table, and scorning
thus to be checked and reproved, Sirrha (quoth hee) remember you
are but a shoomaker, and therefore meddle no higher I advise you,
than with shoos. Which words also of his came afterwards to be a
common proverbe, Ne sutor ultra crepidam.
Pliny (Natural History).

Apelles, the greatest painter of antiquity. The two proverbs mean: “No day
without a line,” “A cobbler should stick to his last.” Pantofle, sandal; latchet, the
thong fastening the sandal; painted table, panel picture; controlle, find fault with.

Have you seen but a bright lily grow,

Before rude hands have touched it?
Have you marked but the fall of the snow,
Before the soil hath smutched it?
Have you felt the wool of the beaver?
Or swan’s down ever?
Or have smelt o’ the bud of the briar,
Or the nard in the fire?
Or have tasted the bag of the bee?
O, so white! O, so soft! O, so sweet is she!

Ben Jonson (A Celebration of Charis).

 Imperfection is in some sort essential to all that we know of life. It
is the sign of life in a mortal body, that is to say, of a state of
progress and change. Nothing that lives is, or can be, rigidly perfect;
part of it is decaying, part nascent. The foxglove blossom—a third
part bud; a third part past, a third part in full bloom—is a type of the
life of this world. And in all things that live there are certain
irregularities and deficiencies which are not only signs of life, but
sources of beauty. No human face is exactly the same in its lines on
each side, no leaf perfect in its lobes, no branch in its symmetry. All
admit irregularity as they imply change; and to banish imperfection
is to destroy expression, to check exertion, to paralyze vitality. All
things are literally better, lovelier, and more beloved for the
imperfections which have been divinely appointed, that the law of
human life may be Effort, and the law of human judgment, Mercy.
John Ruskin (Stones of Venice II, vi, 25).

The best of us are but poor wretches just saved from shipwreck:
can we feel anything but awe and pity when we see a fellow-
passenger swallowed by the waves?
George Eliot (Janet’s Repentance).
 The barge she sat in, like a burnished throne,
Burn’d on the water: the poop was beaten gold;
Purple the sails, and so perfumed that
The winds were love-sick with them; the oars were silver,
Which to the tune of flutes kept stroke. She did lie
In her pavilion: on each side her
Stood pretty dimpled boys, like smiling Cupids,
With divers-coloured fans....
Her gentlewomen, like the Nereïdes,
So many mermaids tended her. At the helm
A seeming mermaid steers: the silken tackle
Swell with the touches of those flower-soft hands.

Shakespeare (Antony and Cleopatra).

This and the next three quotations are word-pictures (see p. 85).

Little round Pepíta, blondest maid

In all Bedmar—Pepíta, fair yet flecked,
Saucy of lip and nose, of hair as red
As breasts of robins stepping on the snow—
Who stands in front with little tapping feet,
And baby-dimpled hands that hide enclosed
Those sleeping crickets, the dark castanets.

George Eliot (The Spanish Gypsy).

 And how then was the Devil drest?
Oh! he was in his Sunday’s best:
His jacket was red and his breeches were blue,
And there was a hole where the tail came through.

Over the hill and over the dale,

And he went over the plain,
And backward and forward he swished his long tail,
As a gentleman swishes his cane.

S. T. Coleridge (The Devil’s Thoughts).

The stanzas are reversed in order.

We walked abreast all up the street,

Into the market up the street;
Our hair with marigolds was wound,
Our bodices with love-knots laced,
Our merchandise with tansy[48] bound....

And when our chaffering all was done,

All was paid for, sold and done,
We drew a glove on ilka hand,
We sweetly curtsied, each to each,
And deftly danced a saraband.

William Bell Scott (The Witch’s Ballad).

 The above are from a series of word-pictures (see p. 85).

ON THE NONPAREIL
Naught but himself can be his parallel.

With marble-coloured shoulders—and keen eyes

Protected by a forehead broad and white—
And hair cut close, lest it impede the sight,
And clenched hands, firm and of a punishing size,
Steadily held, or motioned wary-wise
To hit or stop—and kerchief, too, drawn tight
O’er the unyielding loins, to keep from flight
The inconstant wind, that all too often flies—
The Nonpareil stands! Fame, whose bright eyes run o’er
With joy to see a Chicken of her own,
Dips her rich pen in claret, and writes down
Under the letter R, first on the score,
“Randall—John—Irish Parents—age not known—
Good with both hands, and only ten stone four!”

Peter Corcoran (The Fancy, 1820).

Randall was a pugilist of the time.

“None but himself can be his parallel” is a line from The Double Falsehood of
Louis Theobald (1691-1744), but it comes originally from Seneca (Hercules
Furens, Act I, Sc. I):
Welcome to Our Bookstore - The Ultimate Destination for Book Lovers
Are you passionate about books and eager to explore new worlds of
knowledge? At our website, we offer a vast collection of books that
cater to every interest and age group. From classic literature to
specialized publications, self-help books, and children’s stories, we
have it all! Each book is a gateway to new adventures, helping you
expand your knowledge and nourish your soul
Experience Convenient and Enjoyable Book Shopping Our website is more
than just an online bookstore—it’s a bridge connecting readers to the
timeless values of culture and wisdom. With a sleek and user-friendly
interface and a smart search system, you can find your favorite books
quickly and easily. Enjoy special promotions, fast home delivery, and
a seamless shopping experience that saves you time and enhances your
love for reading.
Let us accompany you on the journey of exploring knowledge and
personal growth!

ebookgate.com