concise but clear and contains all the useful info, developing the photon mapping technique with a practical attitude through the course of the book, viewing related ideas along the way. its probably best if you already have some understanding of how BRDFs, monte-carlo, and raytracing work. one of the appeals of this branch of graphics is that the core concepts are simple and general - particles of light moving in straight lines and bouncing off things - something that is easy to grasp. but where this is so often clouded with equations when reading research papers, this book doesnt forget that computer graphics is about algorithms, and it explores those basic concepts in that way. mathematics just provides some checkpoints. at the end its become quite a complex system, but still not perfect. by then though, youll be wondering about your own possible extensions or variations... hmmm... if youve had an inclination to investigate the ray/path/particle-tracing side of graphics, or if you want to develop a high-quality renderer then this is a must have. (i think the marble bust is artemis. hail artemis!)
The book is quite well laid out, and guides you through an introduction to photometry, the data structures and algorithms involved, and so forth. There is quite an extensive bibliography, which I found quite helpful. All in all, we found this guide was invaluable while implementing a renderer using photon mapping. On the other hand, his sample code is atrocious, with well-thought-out variable names like "foo" and "foo_photon" and nary a comment to be seen. Although you can pick through the worst of it, you're almost better off not bothering - just pick up your algorithms text from university and just start reading. Quite a disappointment after reading an otherwise excellent book.
Photon mapping is an extremely powerful global illumination algorithm. It is probably the most important thing that happened to rendering in the last ten years, in my humble oppinion.
Its inventer, Jensen did an exceptional job at explaining it with this book. The way the book is put togather is sublime. It goes into detailed explanation of the foundations and previous works in computer graphics to be pertinent rather than tedious. If you understand the math that he presented to you, great! But what's really cool is that the math will tend to make more and more sense intuitively as you read and implement the algorithm. You will probably need some form of computer science background (analysis of algorithms and at least one computer graphics course) to fully appreciate it. But anyone should be able to implement a photon mapper in C++ after reading the book. Jensen even gives you the source code for a photon map, all you need is to write a ray tracer.
I would definitely recommend this book for anyone who is interested in computer graphics, especially if you are interested in rendering. Every CS student in rendering should write a hemicube radiosity renderer and a photon mapper... or at least own this book.
This book provides a excellent introduction to the subject of ray-tracing. It is well organised, and covers a lot of ground. To fully understand the technicalities you should be comfortable with intergrals, goniometry and linear algebra. However, this mostly applies to chapter two, and most material is quite easy to understand intuitively, I believe. After reading the book I felt that ray-tracing really isn't such a complex matter as I though it would be (although it is far from trivial!), and felt like it would be fun to try and write a ray-tracer as a hobby project. Regarding such a project, the book is all-inclusive, even providing some source code and data for some glass objects. Some more example images of ray-traced scene, even if black&white only, would make the book even more fun to read.
The content of the book is not a problem and is the definitive place to look if you want to implement photon mapping. But the physical quality of the book is so bad that it's literally down-right pirated book quality. You can get the same printing quality by just sending the original hard cover version of the book (which is unfortunately out of print and is sold at $270 by some sellers on Amazon) through a copy machine, and even then you still get better paper. I expected the paperback version of the book to have lower quality than the hardcover one, but didn't expected to be this low.