Customer Reviews

3D Games: Real-Time Rendering and Software Technology, Volume 1 (With CD-ROM)

5 star:		(4)
4 star:		(6)
3 star:		(0)
2 star:		(5)
1 star:		(3)

 

 

Some of the criticisms are valid, but lets get it straight right here. This is simply the best 3D book out there from all I've seen. In fact, from what I've seen, its the only 3D book worth buying right now for the intermediate<->advanced programmer. You'll have to do a lot of reading with the source code though--but all good info.

The good

1.) No API's are covered. Anyone can learn OpenGL/DirectX, and many tutorials cover those topics quite well. However, this covers the math, algorithms, and 3D side of things (for most part). Some code snippets are included, though--but its not a teach yourself OpenGL in 21 days book (thank god). It also has the best coverage of BSP tree's I've seen aside from dedicated algorithm books. And the coverage on PVS seems a tad brief, but the SDK src code makes up for it.

2.) The src code on the CD is great-- its a load of code to sift through-- not the best code, not the most bug-free--indeed, but good code to learn from. OpenGL stuff is in there, stuff that uses STL and modern C++ techniques, and BSP and PVS code. What more can you ask for in a book? This book comes with a working 3d engine and its source code. I can't think of any book that covers the topics well because the guy implemented what he was writing about. Thats why the book is above average, you get the juicy details along with theory.

3.) Has some cool pictures and visuals of varying effects. Not exactly needed for the professional 3d coder, but sometimes a little eyecandy and visuals doesnt hurt in an otherwise intense book.

The ugly

1.) The last 190 pages cover the SDK he wrote. This could have been better left offline, I think - on a CD-ROM. It gives you some insight on how a 3d API *might* look, or how file formats might look, but who cares. I want more 3d stuff in there, next time fool.

2.) The PVS section could have had a tad more detail. I want the most detail on the hardest topics, and the most briviety on the eastest topics.

3.) The collision detection could have had a more detail. I think the coverage was average, and by finding a few references in the back that should complete it. But, I'm willing to bet if you can implement BSP trees, then collision detection should be easy, I mean as far as math goes. Therefore, the higher level overview is sufficient probably for the coder to get the rest working on his own. But still Watts, its important and you should know it gets neglated way too much. Stop negletating the essentials people!

4.) What the heck is that image analysis, DCT and FFT thing about? Seems like a waste of space, doesnt cover it in enough detail dude. Come on, FFT's are discussed in a digital signal processing book. You cant do it justice there, I dont think.

5.) Networking code, umm, waste of space. People can buy a seperate book for that, or leave articles on CD-ROM.

Summary 1.) More coverage on PVS and collision detection

2.) More coverage on modern games and the techniques they use

3.) No references or crap on FlySDK, who cares? Leave it on CD-ROM

4.) More math, use extra space to discuss the various 3d topics in detail, and how about some stuff that 3d level editors need? More coverage on CSG, for example. And polygon tesselation, and maybe algorithms necessary to convert models to various formats.

I had eagerly awaited this book for months. Watts previous '3D Computer Graphics' was one of the best ever written on the subject and I had high expectations. I have several problems with the book :-

(a) The books text is made up of (to a major extent) text and diagrams from Watts other books. As I mentioned, I like one of Watts other books but I don't appreciate paying for old information. I can appreciate that there are those that do not have the old texts and it would be necessary to some extent to have it included in this book...but how about some new color plates at some point? Some of these plates and text have now been used in three books!

(b) The additional information in the book that isn't based on the older books isn't anything particularly ground breaking. There is very little treatment of new games technologies, even in the field of graphics/rendering.

(c) After reading the texts about graphics programming that I mentioned in (a) we get to the game programming side of the book which unfortunately is the worst part of it. The collision detection part of the book is appalling and no where near the size or detail it should have been given the authors backgrounds. The same could be said of the Behavior and AI chapter at around 30 pages. This sort of overview hardly makes the book a bible on software technology.

(d) How did chapters 18 and 19 get into this book? This whole section covers '2D technology' but nothing presented here is anything that anyone has ever used in a game. It's possible someone might find something useful here, but also quite a rare chance and such chapters catering for the minority or non games programming reader are taking up pages that might have been better used for something else.

(e) The final section of the book is three chapters on 'software technology'. There are an interesting few pages at the start of this on multiplayer technology but it soon changes to a reference about the fly3D sdk.

(f) The layout/typesetting of the book is awful. This all started with Watts ghastly 'The Computer Image' and unfortunately was used in the last revision of '3D Computer Graphics' as well as this book. The early 90's versions of '3D Computer Graphics' had a perfect setting - why change?

(g) The flySDK is reasonable...but its not wholly up to date in terms of games graphics technology. With Watts knowledge I would expect it to be pushing the envelope and offering a guiding light for games programmers but sadly this is not the case.

The cover mentions this as 'volume 1' so I hope the next volume is much more along the lines of what this should be. I have mentioned about my high hopes for this book and hinted on the true genuis of the authors, but if anything this book has proved to me that non games programmers shouldn't attempt to write games programming books. I am sorry to have to say that but I honestly feel that the authors might wish to leave out the unnecessary and undetailed chapters and concentrate on updating their earlier books. I can see a potentially hot title '3D Computer Graphics for Games' based on an update of the earlier book, making considerations to non opengl based rendering systems and console graphics...as well as the definitive text on collision detection covering intersections and collision resolution.

This book is , in my opinion, the best choice for a serious amateur or a shareware games developer. Most of the games programming books now available on the market, explain, step by step, how to design a 2D or a 3D graphic engine. As a matter of fact a graphic engine is just a tool for a game developer. It is not needed to go through in detail a so complicated and boring code, to develop a game. Even some professional software houses purchase the graphic engine. My ideal game programming book should focus on graphic techniques and game logic, instead. This is what Mr Watt and Mr Policarpo do.

(1) They provide a detailed explanation of computer graphic, even exceeding the needs of a game developer

(2) a powerful graphic engine. Look at the demos,they are impressive

(3) 12 tutorials to get familiar with it.

(4) All the needed facilities. Some other books do not even supply a utility to load a .3ds file format

(5) The foundations of A.I. ,collision detection etc

(6) They use OpenGL instead of Direct X The MS library seems to become a standard for professional games developers but,in my opinion ,they are a nightmare for an amateur

(7) Last , Vol 2 has been announced by July 2001. If Mr Watt and Mr Policarpo take note of the critics of the readers who rated this book , 1 or 2 stars, a reference book for our wonderful hobby will be, at last, available

The technical information on 3D provided in the book is good, it's as decent a text as many others out there (not in the top 3 but in the top 20). HOWEVER... the SDK documentation is lacking and some basic information on using the 3DS MAX plugins, etc. is completely missing. With a lot of experimentation I managed to export levels, but such basics as how to create doors, platforms, etc. are not in either the book, the tutorial, or the Web site. Bottom line: if you want tech info on game programming with OpenGL, this is an OK book. If you want to play around with the SDK, you're better off downloading CrystalSpace from SourceForge - it's better documented!

I almost did not buy this book because of some of the earlier reviews, but since it was the only book I could find that provided a comprehensive theoretical, as well as practical, introdution to 3D games I went ahead and bought it anyway. I found that the book is, indeed, an excellent introduction to computer games. I would like to clear up issues addressed by earlier reviews.

First, this book does contain a lot of theory as well as practical information. It does assume a basic knowledge of calculus, linear algebra, and discrete mathematics--math that any graphics programmer (or any programmer for that matter) should have at least a basic understanding of. It also assumes a basic knowledge of computer graphics. Thus, if you are looking for a cut and paste guide to creating computer games and are not looking to actually understand how interactive 3D applications work, you should not buy this book. If you intend to learn the fundamentals necessary to become a professional game programmer, or serious hobbyist, you definitely should.

Second, although there are a few typos, I am sure that the "spelling errors" that have been cited are actually due to the fact that the book is written in European, not American, english. This means that, while probably 99% of the words are the same, subtle differences in spellings exist (such as colour instead of color and visualisation instead of visualization). This does not, however, detract from the readability of the book.

This an excellent book and, despite some of the negative reviews, it lived up to all of my expectations as a sound theoretical, as well as practical, introduction to 3D game programming.

The book is ok if all you want to learn is theory. If you really want something that shows you the ins-and-outs of realtime 3d processing, this book isn't for you. I felt like there were too many topics covered and the important ones were skimmed over. There is a total lack of practical example code as well - don't bother looking at the FlySDK either - it's a waste of time.

The final straw was when I had to learn BSP techniques. This book had maybe 10 pages on the topic, no real code, only a sample formula that made no sense. I did some research on the Internet and found supporting information and better formulas (for free, mind you) that helped me to understand and formulate my own BSP algorithm. I went back and looked at the section in the text again and found how much was really left out of the BSP section and it was shameful.

This book sits nicely between the uber-advanced likes of Eberly's "3D Game Engine Design" and the basic (and largely worthless) 3D programming books that seem to get churned out every year. It covers some of the same topics as Eberly, but explains things much more, and includes many diagrams to help make things clearer. It does, however, stop short of giving you good example code in some of the more advanced techniques, mainly because it comes with it's own engine API. At these points you should turn to Eberly for the nitty-gritty. Covers a good range of topics and explains them well. It duplicates a lot of stuff from other books and www resources, but it is a very nice collection of techniques in my opinion and worth the investment if you don't own others of it's type, because it is very well presented.

I've been reading game development books for more than 10 years and this is certain one of the best books i've ever read. Policarpo and Watt have created a very good 3D engine that comes in a CD with the book and they explain everything about programming 3D games with this engine and general game programming. It's really Worth Buying!

This is 800 pages of concentrated goodness, bringing together a whole package of technology needed for computer games and putting it in one place. The authors have the courtesy to give you the full story, instead of expecting you to buy their other books (which contain lots of material you probably won't need for games). There is also a peek at the emerging image-based techniques which are likely to play a part in some new games. It is a great source of teaching material, for VR as well as games material.

Criticisms? I didn't try the Fly SDK disk, which some have reported negatively on. There were some places where I had to work to get from the description to the practicalities of an implementation. But given the scope and length of this book this is only to be expected, here and there. Overall, this is an excellent addition to a sparse literature.

Although the book itself is very promising and states the basic concepts of a 3D grphics library, I still don't recommend this book at all. Why? Because if you buy a book that is built onto a SDK, you also want to USE the SDK (otherwise you buy another book that just tells you about the concepts). If now the authors provide you with tutorials and utilities to convert - let's say - 3D Studio MAX - files into their own level-filetypes things seem to be ok. But if you finally realize that the provided utilities are not working, the MAX-export-plugins have to be (a) installed in a different way they tell you and (b) still cause MAX to crash when you do not work with a simple box (how interesting!), you get really mad. After trying around for several hours I gave up and had a nice christmas-campfire.

In the past months a lot of VERY good books were released on this topic, you are better off with one of these.