Slashdot Mirror
OpenGL Distilled
Martin Ecker writes "Until now, if you were looking for an introduction to the OpenGL graphics API, the "OpenGL Programming Guide" (also known as the Red Book) was your best bet. Now Addison Wesley Publishing provides a new alternative that is easier to digest than the all-encompassing Red Book with its more than 800 pages. Paul Martz's "OpenGL Distilled" concentrates on discussing only the important fundamentals you need to program 3D graphics using OpenGL 2.0 and provides a concise introduction to the most important cross-platform graphics API currently available." Read the rest of Martin's review.
OpenGL Distilled
author
Paul Martz
pages
266
publisher
Addison-Wesley Publishing
rating
8/10
reviewer
Martin Ecker
ISBN
0-321-33679-8
summary
A concise introduction to the OpenGL graphics API
Before going into more detail and reviewing the chapters of the book I have to disclose that I was a technical reviewer of the book before it was published.
"OpenGL Distilled" is aimed at people interested in learning the basics of OpenGL. The reader should already be familiar with programming in C++ and have a basic grasp of linear algebra, in particular vector and matrix algebra. Familiarity with other 3D graphics APIs, such as Direct3D, is an advantage, but not a necessity. The book does a good job of presenting only the fundamental aspects of OpenGL and 3D graphics programming in general and never overwhelms the reader with too much unnecessary detail. The author draws a good line between what to discuss and what is beyond the scope of the book. More advanced features of the API are only mentioned shortly with appropriate references to more in-depth literature. Some deprecated features, such as the feedback buffer, that are no longer commonly used are left out as well. In my opinion an unfortunate omission, is shader programming with the OpenGL Shading Language, which is only briefly mentioned in an appendix. A full chapter introducing the basics would be a nice addition to the book. Until then the reader is referred to the Orange Book, which discusses shader programming in OpenGL in detail.
One thing I highly appreciate about "OpenGL Distilled" is the introductory section of each chapter, which contains a "What You'll Learn" and a "What You Won't Learn" bullet list. This makes it clear what the chapter is about and - more importantly - what it is not about. Especially the latter is mostly missing in other books.
The book has a total of 8 chapters and 4 appendices. The first chapter explains what OpenGL is, talks a bit about setting up a development environment on the most common operating systems to actually develop OpenGL programs, and immediately gets your feet wet with a first simple example program. The chapter is concluded by a whirlwind tour through the almost 15-year history of OpenGL and its predecessors.
Chapter two focuses on drawing primitives, such as lines and triangles, and the various ways supported by OpenGL to specify vertex data. In particular, vertex arrays and vertex buffer objects (VBOs), a fairly recent addition to OpenGL to allow high-performance rendering, are the focus of this chapter. Additionally, a first overview of the OpenGL pipeline that a primitive passes through until it finally ends up in the framebuffer is presented. A more detailed discussion of this pipeline, in particular with regard to coordinate transformations, follows in chapter three. The various coordinate systems used in OpenGL programming, such as object, world, eye, and clip coordinates, are presented and discussed in detail in this chapter.
Now that we can render primitives we need to light them to make them look more interesting. Chapter four sheds some light on this by discussing the lighting and material model used in OpenGL's fixed-function pipeline. The best part of this chapter is the section titled "Debugging Lights", which gives some insights and helpful advice on how to debug OpenGL programs that use lighting. Many other books only describe the features of OpenGL lighting but do not explain common debugging techniques that can be applied when all you get is a black window instead of a nicely lit scene.
Chapter five is about pixel rectangles, in particular how to read from and write to the framebuffer. Some performance considerations are also discussed, which is a good thing since reading from the framebuffer is a costly operation. Again, this chapter concludes with a nice section on debugging techniques. The explanation of the raster position in this section is probably the easiest to understand that I have read to date.
Chapter six is a comprehensive chapter on 2-dimensional texture mapping that also discusses some more advanced applications of the technique, such as light maps and depth maps. Also using cube maps as environment maps is introduced. 1-dimensional and 3-dimensional texture mapping was omitted from the discussion.
Chapter seven deals with detecting the feature set of the OpenGL implementation, in particular, determining the version of the OpenGL specification the implementation adheres to and the available extensions. This chapter also discusses how to retrieve and use entry points for available extensions.
Finally, chapter eight deals with the platform-specific interfaces required to hook up an OpenGL program with the underlying operation system. These platform-specific interfaces are called GLX for Unix, WGL for Windows, and AGL for Mac.
The book has four appendices, which deal with a quick overview of advanced features, best practices, performance-related issues and debugging tips and tricks. Especially the latter two appendices on performance and debugging contain a lot of insights invaluable to programmers just starting out with OpenGL.
The book is printed in black and white throughout. Having some color plates in the book itself would have been a welcome addition considering that the topic is computer graphics. However, you can download some color plates from the books website at where you will also find the source code to the example programs in the book.
In conclusion, "OpenGL Distilled" is an excellent introduction to OpenGL, not only for someone new to 3D graphics programming but also for those that have worked with other 3D graphics APIs in the past that wish to get up to speed with OpenGL quickly. The book omits advanced and deprecated features that would unnecessarily overwhelm a beginner with OpenGL and is a good companion on the way to becoming an experienced OpenGL programmer.
The review author has been involved in real-time graphics programming for more than 10 years and works as a games developer for arcade games. In his rare spare time he works on a graphics-related open source project called XEngine http://xengine.sourceforge.net./
You can purchase OpenGL Distilled from bn.com. Slashdot welcomes readers' book reviews -- to see your own review here, read the book review guidelines, then visit the submission page.
Before going into more detail and reviewing the chapters of the book I have to disclose that I was a technical reviewer of the book before it was published.
"OpenGL Distilled" is aimed at people interested in learning the basics of OpenGL. The reader should already be familiar with programming in C++ and have a basic grasp of linear algebra, in particular vector and matrix algebra. Familiarity with other 3D graphics APIs, such as Direct3D, is an advantage, but not a necessity. The book does a good job of presenting only the fundamental aspects of OpenGL and 3D graphics programming in general and never overwhelms the reader with too much unnecessary detail. The author draws a good line between what to discuss and what is beyond the scope of the book. More advanced features of the API are only mentioned shortly with appropriate references to more in-depth literature. Some deprecated features, such as the feedback buffer, that are no longer commonly used are left out as well. In my opinion an unfortunate omission, is shader programming with the OpenGL Shading Language, which is only briefly mentioned in an appendix. A full chapter introducing the basics would be a nice addition to the book. Until then the reader is referred to the Orange Book, which discusses shader programming in OpenGL in detail.
One thing I highly appreciate about "OpenGL Distilled" is the introductory section of each chapter, which contains a "What You'll Learn" and a "What You Won't Learn" bullet list. This makes it clear what the chapter is about and - more importantly - what it is not about. Especially the latter is mostly missing in other books.
The book has a total of 8 chapters and 4 appendices. The first chapter explains what OpenGL is, talks a bit about setting up a development environment on the most common operating systems to actually develop OpenGL programs, and immediately gets your feet wet with a first simple example program. The chapter is concluded by a whirlwind tour through the almost 15-year history of OpenGL and its predecessors.
Chapter two focuses on drawing primitives, such as lines and triangles, and the various ways supported by OpenGL to specify vertex data. In particular, vertex arrays and vertex buffer objects (VBOs), a fairly recent addition to OpenGL to allow high-performance rendering, are the focus of this chapter. Additionally, a first overview of the OpenGL pipeline that a primitive passes through until it finally ends up in the framebuffer is presented. A more detailed discussion of this pipeline, in particular with regard to coordinate transformations, follows in chapter three. The various coordinate systems used in OpenGL programming, such as object, world, eye, and clip coordinates, are presented and discussed in detail in this chapter.
Now that we can render primitives we need to light them to make them look more interesting. Chapter four sheds some light on this by discussing the lighting and material model used in OpenGL's fixed-function pipeline. The best part of this chapter is the section titled "Debugging Lights", which gives some insights and helpful advice on how to debug OpenGL programs that use lighting. Many other books only describe the features of OpenGL lighting but do not explain common debugging techniques that can be applied when all you get is a black window instead of a nicely lit scene.
Chapter five is about pixel rectangles, in particular how to read from and write to the framebuffer. Some performance considerations are also discussed, which is a good thing since reading from the framebuffer is a costly operation. Again, this chapter concludes with a nice section on debugging techniques. The explanation of the raster position in this section is probably the easiest to understand that I have read to date.
Chapter six is a comprehensive chapter on 2-dimensional texture mapping that also discusses some more advanced applications of the technique, such as light maps and depth maps. Also using cube maps as environment maps is introduced. 1-dimensional and 3-dimensional texture mapping was omitted from the discussion.
Chapter seven deals with detecting the feature set of the OpenGL implementation, in particular, determining the version of the OpenGL specification the implementation adheres to and the available extensions. This chapter also discusses how to retrieve and use entry points for available extensions.
Finally, chapter eight deals with the platform-specific interfaces required to hook up an OpenGL program with the underlying operation system. These platform-specific interfaces are called GLX for Unix, WGL for Windows, and AGL for Mac.
The book has four appendices, which deal with a quick overview of advanced features, best practices, performance-related issues and debugging tips and tricks. Especially the latter two appendices on performance and debugging contain a lot of insights invaluable to programmers just starting out with OpenGL.
The book is printed in black and white throughout. Having some color plates in the book itself would have been a welcome addition considering that the topic is computer graphics. However, you can download some color plates from the books website at where you will also find the source code to the example programs in the book.
In conclusion, "OpenGL Distilled" is an excellent introduction to OpenGL, not only for someone new to 3D graphics programming but also for those that have worked with other 3D graphics APIs in the past that wish to get up to speed with OpenGL quickly. The book omits advanced and deprecated features that would unnecessarily overwhelm a beginner with OpenGL and is a good companion on the way to becoming an experienced OpenGL programmer.
The review author has been involved in real-time graphics programming for more than 10 years and works as a games developer for arcade games. In his rare spare time he works on a graphics-related open source project called XEngine http://xengine.sourceforge.net./
You can purchase OpenGL Distilled from bn.com. Slashdot welcomes readers' book reviews -- to see your own review here, read the book review guidelines, then visit the submission page.
My only complaint is that the binding broke in a couple of weeks, and I had lost the receipt. So hold onto it.
Some of us hate reading stuff on the screen. It's starting to piss me off that some hardware manufacturers are just supplying pdf versions of their catalogs or manuals and they expect me to print them out.
The owls are not what they seem
This is the sort of book that may become quite useful in the future. Over at OSNews there have been some rumors that the AMD/ATI combo will be releasing a high-quality OpenGL 2.0 implementation to take advantage of the many OpenGL-related advancements there have been for X recently. Supposedly they want to be at the cutting edge of technologies like Xgl, Compiz, glitz and so on.
I really don't know if this is idle speculation or there is some basis to it, but it is something interesting to consider. Perhaps nVidia should step up and release an OpenGL 2.0 implementation before AMD/ATI manages to. Last I checked, the implementation provided with the nVidia drivers for Linux was for OpenGL 1.4.
I'm reading this right now, and it is substantive and helpful. The downside is that it can get rather terse at times, without as much explanation as I felt was deserved. I suppose that's why they called it, "distilled". :)
Anyhow, the bulk of the book is not for the faint of heart. Be dialled when you read this, you're not getting any hand-holding.
It's interesting the review above was written by one of the technical reviewers of the book, the result being more of an overview than a pure objective review (with positives, negatives). Or maybe I'm asking too much.
Did you print out his post before you replied to it?
--
WHO ATE MY BREAKFAST PANTS?
Nope, but then again it didn't span several pages. I do print out man pages, e-mails and manuals because a hardcopy is easier to store and read.
The owls are not what they seem
Yea, I understand that. I don't mind them printing sections of the code in order to explain them, but when they print whole source files especially when they include the long EULA statements at the tops of the headers, it's very annoying. I've never run into any hardware that didn't provide me with a printed manual, but most of my equipment has PDFs available for download in addition to the included printed manual, which I find convenient. I would be pretty annoyed if they didn't provide me with a printed version as well though.
This sounds like an interesting book to read over coffee or something. But does it cover that weird thing where the RGB values are 16-bit in some places and 24-bit in others? That really confused me in the past.
A hardcopy is easier to store? How many hardcopies do you have? How do you dispose of your hardcopies? Do you have hardcopies lying around the office? And is the case of your computer about the size of a file cabinet? Or four file cabinets? Maybe I should just stop asking questions?
I agree it's easier to read things on paper than on screen, but I think that storage of hardcopies is not easier than on your computer. Even Windows Search works better than me trying to comb through a file cabinet. You must be the most organized person in the world.
Quake/Doom , Unreal Tournament, etc... use OpenGL. Many workstation apps use it too.
Because $34.99 isn't cheaper than $34.99?
You don't have to be the most organized person in the world when you file printouts in an inverse chronological order and mark the most important printouts. Folders are cheap, my office is roomy and we've got plenty of room in the archive room anyway.
The owls are not what they seem
Not to mention SketchUp, recently re-released by Google, AutoCAD & every other major package for architects (one called ArchiCAD uses openGL to accelerate the display of 2d linework), and any 3d modelling program (MAX, MAYA, Blender). Hell, there is even an openGL renderer for Flash somewhere.
Maybe the parent was thinking about StarFox for SNES?
Dammit Otto, you have lupus.
Stop spreading these myths about OpenGL not being supported on Vista. OpenGL is will remain alive and well for Vista.
It's true, most users of OpenGL today aren't game developers. We're doing far more serious and practical work. What we're talking about here are medical imaging, military simulations, petroleum exploration, and advanced CAD design.
OpenGL likely powers the design tools that were used to develop many of the components in your computer, your vehicle, your cell phone, and your television. OpenGL is used by the simulation software your father's doctor uses to track the spread of your father's penile cancer. OpenGL is used in the process of locating the oil which becomes the gas you pour into your car.
OpenGL is used all around you. You're mostly just oblivious to it, since your world doesn't extend outside of Halo 2.
You're in a losing battle, sir. Those pesky books are going to be become error-prone as more information makes its way to real-time. Do you print the addendums and errata and file those too? Do you scan for software updates and print the release notes? Howabout the common bugs and patches? Post-release common forum questions and answers? How do you index and search such material?
You'd be better to keep a room full of DVD's and print on demand, given the sheer metrics of security, space, fire hazards, power usage, paper and time consumption that bulk printing entails. Perhaps you only buy product from companies about to go belly-up and thus no information is available anywhere else on earth?
Eh. Then again, in a few years, you could open a museum. I'll visit - with a printout of our little chat here.
I'm in the market for a new computer, and I'm going to be writing games for it.
What's the deal with developing with OpenGL on linux at the moment? Last time I tried (2002 maybe?), there was something called Mesa, another thing called DRI or some such, another thing called GLX. It all got a bit confusing. I ended up using cygwin for a bit, then finding something else to do.
Is it the case now that I can just 'apt-get install opengl-devel' and have stuff just work?
Why, God, why did I burn my mod points yesterday modding up "...welcome our new overlords" and "3. ??? 4. Profit" jokes?
I badly want them back so I can mod this whole thread OT.
To be fair, you could store a cdr/w in the center Yucca Mountain and I still wouldn't consider it a safe place. (hint: by place I'm not referring to Yucca Mountain...)
--
WHO ATE MY BREAKFAST PANTS?
OpenGL is a open source 3D graphics rendering standard, sort of like direct x without sound and I/O support.
Mesa is a open source linux OpenGL compatible library, you will find that other venders make Linux openGL compatible librarys like ATI. Example if you install the Official ATI linux video drivers you also get their version of openGL library along with it.
DRI (Direct Rendering Infrastructure) this open source video driver system allows direct access to the video card which in turn improves OpenGL preformance.
GLX an improved x-windows openGL interface for applications
cygwin an linux style interface for windows
Wise men speak because they have something to say, Fools because they have to say something!!!!
If movies taught me anything its that The Red Book is NSA Trusted Networks and doesn't fit on shelves.
Obviously Unreal Tournament, the Quake/Doom series, Warcraft III, and that obscure game known as 'World of Warcraft' don't use OpenGL. Unpopular games like Half-Life/Counterstrike don't use OpenGL. There are no games for the Gamecube, and there will be no games for the Wii or the PS3.
And obviously all serious visualization work isn't done in OpenGL.
Go back to your cave, troll.
True, but by the time the cdr/w fades (provided you don't leave in in direct sunlight or a hot, humid place), those manuals will be outdated anyway. Data on a cdr/w should last a few years. And paper is also volatile if you light a match to it :P
Blizzard games use OpenGL too. Virtually all of their games (at least back to the point when they changed their name to Blizzard) have been developed with cross-platform in mind, and they usually release games on PC/Mac hybrid CD's. All of the 3D games Blizzard has put out (Diablo 2, Warcraft 3, WOW) fully support OpenGL. The Windows version will use Direct3D by default, but you can also make it use OpenGL if you want to. Doing this isn't particularly obvious (especially to someone who's never used a command line, as there's no option in the game to do this)...you just have to start the executable with the -opengl switch. This obviously makes running the Windows version under Wine/Cedega much easier (since Direct3D emulation isn't nearly as good as native OpenGL support). You don't lose ANY features by switching to OpenGL rendering (in fact for the Mac versions this is the only option)...the only way it will cause any problems for the Windows version is if your video driver has poor OpenGL support. Poor OpenGL support will become more common with Vista, as Microsoft is crippling the OS's OpenGL support, and they will probably encourage video card manufacturers to do the same.
I like OpenGL, but my one problems is the lack of ease it gives for Win32 Programmers. I know it is cross-platform, but to program a opengl app in linux and win32 are totally different, so what's the point in developing a OpenGL app if the audience you intent on targeting is small(I.E. Linux)? Why not just setup for MS's DirectX which will obviousally win your apps much more support.
N. A. Stuart
There is a devel package for mesa that provides basic OpenGL bindings (since Mesa is a software OpenGL-compatible library) and headers. It provides a (roughly) complete OpenGL 1.3 implementation (I'm pretty sure it's 1.3 with extensions), and while not explicitly licensed, is widely considered a standard implementation. It's everything you need to develop OpenGL apps.
Of course, nVidia and ATi drivers provide their OWN OpenGL libs and headers, nVidia's implementation being a full-blown version 2.0, and it includes headers and libs for development.
DRI, as I understand it, is part of the driver interface for X that supports hardware acceleration of OpenGL.
GLX is the X11-specific bindings that handle creating windows and contexts for GL output. GLX is to X11 what WGL is to Windows.
If you install mesa-devel (or whatever your distro calls it - on Gentoo, it's part of the full deal), you'll have everything you need to start hacking yesterday.
grey wolf
LET FORTRAN DIE!
Direct X is not cross platform, so is not an option for us, but Direct X is very heavy in areas that openGL is very light (in openGL, the limitation on drawing geometries is much less than the Draw index Primitives (DIP's) in directX, although state changes are heavy.
Overall, the openGL interface is much simpler to optimize than directX. That is from my experience of both (althoguh I admit my experience of optimizing direct X is much less than that of openGL), in both cases batching is incredibly important, in openGl what is going on is very transparent.
But coming from the simplicity of openGL, it has to be said the over-complicated COM interface of Direct X is hideous...
How well does the book cover the subject of stereoscopic 3D?
BTW, for anyone reading this -- does anyone know how to get Google Earth to do stereoscopic 3D?
It would be a bad idea to store anything in Yucca mountain.
I mean, they want to fill it with nuke waste; I'm sure that's bad for the life of a disc. If the radiation doesn't effect the dye layer, I'm pretty sure the heat would melt the plastic.
By the way, when did "Yucca Mountain" replace "Fort Knox" in the common vernacular?
110100 1101000 1101000 1100110 0 1101111 1101000 1100011 1
The Really Slick Screensaver collection. Blender. Jahshaka. Quake (all versions). Doom 3. Warcraft III/World of. Halflife / Counterstrike. The GameCube. The Wii. (supposedly,) the PS3. Bryce 3d. Anything by Silicon Graphics. Bindings for every language on every platform under and out of reach of the sun.
Yeah. Completely unpopular. Please do me a favor and remove your lips from Mr. Ballmer's rear end long enough to post coherently when on Slashdot.
110100 1101000 1101000 1100110 0 1101111 1101000 1100011 1
*sniffs the air*
Smell that? That's the sweet, pungent smell of GP's sarcasm.
Take it in, boy, take it in.
110100 1101000 1101000 1100110 0 1101111 1101000 1100011 1
DirectX does kick the shit out of OpenGL, but OpenGL is free.
How ignorant, and blind, can a person be to make this statement? DirectX, is probably... why even try to explain how much crap DirectX is. You're definately a consumer. In the most negative sense of the word.
DirectX, will, and you mark my words, NEVER be of enough quality to be used in CGI REALISM. OpenGL can, generate very realistic renders... so realistic, it's used in all major film FX.
The reason why OpenGL games aren't as good looking, is becuase it targets the consumer; you know, morons like yourself. Who, do not have the hardware to render heavy duty OpenGL sequences. Before you jump up and down, your lame GeForce 9999999 is crap and only geared towards the most common 3D calls for GAMERS and CONSUMER LEVEL 3d applications. There is NOT ONE consumer level card, even for a scant 400-500 dollar hot release, that has hardware acceleration for 100% of ANY 3D API... even DirectX.
Now, what Microsoft did, was compose a bunch of wiz-bang API calls into DirectX, while OpenGL programmers are left to write their own. OK, does this mean DirectX is better? The only thing "better", is time for development. As far as seeing an EXACT same effect used in ALL DirectX games???? Anyways... with all the exposure of the white collar sweat shops of EA Games and other such game development houses... no wonder a lot of games on the shelf were developed under pure management and they chose DirectX and Windows as their primary market. However, a few of those programming houses actually know a thing or two about 3D programming, actually have the resources and clout to take the time to do something RIGHT. idSoftware for one, and what API do they chose? Not DirectX! And, there's a reason why. Becuase it f*cking sucks, I can't put it anymore blatant than that.
Well said, brother!
Registered Linux user #421033
Full disclosure: Paul is a friend of mine, and I helped proof this book too.
In all seriousness, this is an excellent book. Paul wrote this book to fill a serious need -- an updated, quality OpenGL book for this age. So much of what is in the canonical texts is no longer important (geometry by Begin/End), and they won't cover the new recommended practices (VBOs, Vertex Arrays, etc).
On a personal level, Paul is one of the most generous and helpful programmers I know. I owe him lots of beer for all the advice he has provided. He also participates in the open source OpenScenGraph project:
http://openscenegraph.org/
a high-performance 3D toolkit for Windows, Mac and Unix/Linux, used in hundreds of open source and commercial simulator, game and 3D visualization projects (including my company's NatureView Express tool http://3dnature.com/nv.html -- plug plug!)
-- There is no truth. There is only Perception. To Percieve is to Exist.
Mod parent troll, doesn't even consider debugging or that DirectX is a more full package and that he should be talking about Direct3D. Also mentions iDsoftware where he probably means John Carmack who said that he would probably use Direct3D nowadays if he didn't have to port.
-]Phreak Out[-
This is somewhat correct, but wrong in a big way. Before the takeover, the OpenGL ARB had to vote on features to make it into the spec, disclosing any patents they might have relating to a feature. This made things slow. You can see how long it took for OpenGL 2.0 (which started as 1.3!) to come out.
I hear the process for DirectX was more along the lines of MS asking what features card manufacturers wanted and then they were quickly put into practise. Often there was a reference rasterizer before hardware was in the people's hands.
-]Phreak Out[-
I'm currently reading this book also. I'm not so crazy about it. I have a good understanding of computer graphics, but no knowledge of OpenGL. It seems like its trying to be all things to all people, and delivers to noone. Here's an example from page 6, talking about fragments: "OpenGL implementations that support multisampling, however, store fragments in subpixel locations." And then we move on to Rasterization. What is multisampling and subpixel location storage? How do I know if my implementation supports it? Why was it important for the author to state this? If I'm already supposed to know these details, then why should I get this book? Why not just use a reference?
I'm the author of this book. It's hard to please everyone... An early technical review comment stated that "you should include complete source, mere snippets aren't very helpful". I disagreed with this, especially since conciseness was critical. If you read the full book, you'll see that snippets are the primary mode for examples. Did I use full source in a few places? Yes, guilty as charged, and required given that not everyone will have access to the example code. However, I'm constantly looking for ways to reduce the page count in future revisions, so I'll consider removing some of the source. This might make room for things like FBOs or an expanded GLSL section. No apologies for the chapter 1 material -- this is a beginner's book.