Customer Reviews

Texturing and Modeling, Third Edition: A Procedural Approach (The Morgan Kaufmann Series in Computer Graphics)

5 star:		(5)
4 star:		(1)
3 star:		(4)
2 star:		(2)
1 star:		(0)

 

 

This is an update of a classic book on procedural texturing and modeling by the main founders of the field. The book presents lengthy discussions of classical procedural texturing using various noise functions of the sort originated by Ken Perlin (one of the authors). It discusses newer texturing techniques such as cellular texturing, which can be used, for example, to create convincing stone patterns. Other chapters focus on animating solid textures (e.g. marble forming, volumetric gasses, etc.), fractal terrain generation, and tips for utilizing existing graphics APIs and hardware for realtime procedural texturing. This is only a sampling of the topics covered.

Code samples in C and RenderMan are given throughout, although most algorithms are given in only one of those languages. This can be a bit of a problem, as many readers will probably not have access to a RenderMan implementation. Nevertheless, it is not too difficult to translate the RenderMan code into C code in many instances.

The biggest drawback to this book is its lack of rigorous technical coverage. The decision to omit many mathematical details was a conscious choice on the part of the authors. Instead the book is mostly prose discussion of the techniques and the coarse descriptions of the underlying concepts. Although the prose is mostly clear, many times I felt myself in need of more specific, technical details. Fortunately, the book's authors are the primary researchers in this field and most of the ideas in the book have been published in academic journals. It was very easy to supplement the book with these primary sources.

Overall I found this to be a very interesting and useful book, with many algorithms essentially ready-to-run right out of the book. It would get five stars, except for the lack of technical and mathematical details mentioned above. Every serious worker in graphics needs to have this book on their shelf. I use mine often.

This book has several serious problems.

The most glaring is that a significant number of the examples are coded in the "Renderman shading language". This language serves, in this book, to hide detail, detail specifically related to producing textures. Of course, if you know the language, you're fine - but most won't know the language and so this is a grievous error.

By way of welcome contrast, other examples in this same book are instead presented as C code fragments or functions. That's just the ticket - using a broadly known, freely available, relatively low-level language with no recourse to unknown hidden graphics functionality is precisely the way to go when explaining ideas in the domain of those this book is intended to convey.

The second problem is one of content. While being concise to the level of a math text is not desirable, this book contains a very sparse field of useful information considering the number of pages. The margins are too wide, the text too large, the form factor of the book too small, and the authors too wordy to possibly convey a good basis for texturing in general - it is a broad and fascinating field, touched only in the briefest and most unsatisfying manner by this book.

I do take issue with the reviewer who complained about the exposition on how to make a brick texture; that area of the text, while it may be already quite familiar to many who are interested in texturing, contains precisely the level of detail that needs to pervade a book of this type, and detail about steps that underly critical basic texturing ideas. Without understanding those basic texturing tools, a novice misses the first step on the stairs and fall on their face. The problem is, this approach is not consistent for more complex ideas in this volume, few as those actually are.

The book is entitled "Texturing and Modeling". While there is a moderate amount of texturing information in it, whatever you do, don't get it if modeling is your goal. It is very nearly devoid of modeling information, and what there is (smoke, a planet simulation, a few other items) is very basic indeed.

Finally, as a general critique, the authors (all of them) need to learn the basic idea that when presenting a function in any language to a new audience, one should precisely define the domain (and rationale) of the inputs and outputs of the function. As an example, one might encode the function for Perlin noise, and have no idea whatsoever as to what values to feed it to get particular types of results. For those of us who can read and understand what the function is actually doing (which is esoteric, make no mistake about it) the answers will eventually be illuminated by careful study of the function. However, this is very advanced material, and I am absolutely certain that many readers will be unable to figure out how to effectively use this function without a great deal of trial and error. You can also read that as "wasting a great deal of their time." That is because they won't be learning anything that could not have been conveyed by the author(s) in a single short paragraph of domain information.

In summary:

The 2 stars is because I didn't think this was a very good book. On the other hand, it is one of the very few books that deals with the subject at all, and for that reason, you should definitely own it if textures are an interest of yours.

If you're newly interested in textures, this will give you a basis for further exploration. It won't give you a cookbook by any means.

If you're looking for cookbook and "how to" approaches, get on the web and the newsgroups.

Finally, if you're considering writing a good book about creating textures, by all means, please do. The world needs a good one: this most definitely isn't it. I'd be delighted to be one of the first owners of your new book.

Yippeee!!!
Finally, a book that explains Fractal noise, Bump mapping, textures and Terrain modelling in a simple way that doesn't boggle your mind

I've always wondered how to make my own Smoke effects, how 3D Fractal Terrain is created, how to make Clouds and a whole lot more...
This Book is really cool! It does feel a little *academic*, but its a fascinating read and even a dummy like myself can follow the text! (Hey, I hate Maths!)

Excellent!

It doesn't get better than this. If you're doing more with graphics than drawing circles and boxes, you NEED this book. The writing style is academic yet conversational. A solid math background is recommended.

This book has five authors, and that is exactly the way it is written. It changes character practically at the beginning of every chapter. It is obvious that the authors have a great deal of knowledge of procedural textures and modeling, however, it is all just lumped together into one big hard-to-sift-through mountain. You will have to do a great deal of work to extract algorithms from this book. Only because there is good information there do I give this book three stars instead of two. Truthfully, it is probably a 2.5 star book.

An example of the authors' inconsistent narrative style is this: Chapter two goes into great detail on the obvious - clamping, antialiasing, and the brick wall texture. In chapter 15 on "Fractal Solid Textures", the authors brush over the complex issue of how to produce fire, water, wind, and rocky terrains. Also, the vast majority of the time, rather than show the procedural modeling with pseudocode or with a high level language such as C, the authors choose "Renderman", which is unfamiliar to many people and makes the included code useless to those uninitiated in that language. Plus, in many cases Renderman has functions that hide the details of particular algorithms. This is counterproductive, since the algorithms are supposed to be the point of this book in the first place, or at least I thought that they were.

My advice to people interested in this subject is to skip this book unless you can find it at a greatly reduced price and look online at Elias Hugo's webpages on procedural modeling. Mr. Hugo explains the authors' techniques much better than the authors themselves do.

The first edition was outstanding for the sheer breadth of its coverage of ways of producing natural-appearing textures via computation. Coupled with an easy-to-read prose style and good examples and color plates, it'll make you think of ways of doing apparently unrelated things you'd never have thought of otherwise. I only got a glimpse of the second edition, but this has been a busy field the last five years, and I fully expect the second edition to be as useful and interesting as the first.

This book is for experienced programmers who want to understand the fundamentals of procedural graphics generation. This book is not for artists looking for practical applications of procedural texturing and modeling concepts.

The authors cover a very large array of topics in the field, including many pertinent code examples, mostly in the RenderMan shading language. It focuses on the groundwork of the field from the first texture maps in the 70's onward, with a cursory mention of the state of the art at the time of publication.

Representative text:

"The particular kind of fractal we're building is called fractional Brownian motion, or fBm for short. fBm is characterized by its power spectrum, which charts exactly how amplitude relates to frequency. Oops! Pardon me - I'll knock off the math."

"Long ago I gave this idea the wonderfully unpretentious - not! - moniker "generalized Impressionistic texture," or GIT for short. (We need more TLAs - threeletter acronyms.) The GIT matrix generator system takes the form of a time-varying swarm of color samples in a color space, usually the RGB color cube."

I found this book in a rack in a local book shop where they had kept books about Photoshop and Illustrator. I asked the book seller, he said it's about making nice images so it is there. Any way absolutely must book for serious graphics researcher. I got introduced to genetic textures, I was unaware of them.

The book does not cover reaction diffusion textures.

Unless you are a complete novice to the subject, I would avoid this book. These so called 'gurus' merely give some of the ideas they have had for making procedural textures, and suggest what you could potentially do. Then they go into great detail on subjects like beveling a brick texture! Not useful. You WONT find explict texture recipes. You WONT learn how to code a texture engine (like Bryce uses for example). There is little you cant find on the web. If this book were [less expensive] I might be singing a different tune...

This book covers all of the CG aspects needed to code material and texture applications. Many commercial CG softwares are using the essential theory and practice given in this book. This book seems as if the compilation of the best Siggraph papers of the pionering researchers of the subject.

If you need to learn the procedural textures and materials: this is THE book.