ShaderX

wolfgang

Newcomer
Hey guys,
I am curious on your comments on the ShaderX series -espeically the last one- ShaderX5. You guys are CORE target group of the books :) What do you feel is right or wrong with the books? I am starting ShaderX6 now and I want to hear from you what we -the editor team- can make better this time.

Thanks in advance,
- Wolf

P.S: Proposals are due by April 1st, 2007. Please send them to wolf at shaderx.com. Writing guidelines and a FAQ can be requested from the same e-mail address. The schedule is available on www.shaderx6.com.
 
I found the ShaderX series to be a good reference for ideas.
I personally don't use the book as a copy/paste algorithm resource, but it's worth it's weight in gold due to the way ideas are presented and how ideas get implemented.

I found some of the algorithms in the book, especially on shadows and lighting to be a little rough but overall they gave me ideas as to what could be considered acceptable compromises in regards to quality/performance.

In any case, it's a very pragmatic book. I myself am more into big giant theory books, overall I found the ShaderX series useful.
 
The main thing that's wrong with it is that it is heavy big book. That's the only thing I never bought any of these. I can't carry on so many books with myself every time I need to move. Ebook pdf format would not help, I need that ebook reader device that looks like piece of paper finally.
 
I just got most of the ShaderX series (3, 4, 5) recently, and I have found it to be a great resource. I am currently working on the parallax occlusion mapping and the PRNG noise chapters in volume 5. The source codes are great and work exactly as expected and they actually compile without much effort! I saw a lot of other great stuff in the book but haven't gotten around to reading them in detail yet. I will be one of the first to buy ShaderX6 when it comes out :D. I am sure it'll cover some of the next-gen programming techniques on the G80 which as of now is not being pushed to its full potential on my machine. Also, I think you should rerelease ShaderX 1 and 2, I can't find it on amazon. I saw a lot of great sample chapters from the ATI developer's website.
 
Also, I think you should rerelease ShaderX 1 and 2, I can't find it on amazon. I saw a lot of great sample chapters from the ATI developer's website.

Hi!

I've posted this in your earlier thread, but it seems you missed it:

Not trying to push a site here, but have you tried http://www.abebooks.com ? It's where I usually get my books from if they're out-of-print, etc.

EDIT: Yep, there's a ShaderX pt.1 book on sale here
Shader X2: here
 
Hey Wolf, Rys actually news posted this thread on the frontpage in hope of getting some extra exposure - not sure how much of an effect that'll have, but we figured it'd be worth a shot.

Overall, I've personally only got good things to say about ShaderX5, and the ShaderX Series in general (I got 3, 4 and 5). It's especially nice that you're sticking to a regular schedule for them, so that it represents the cutting-edge more than GPU Gems or articles in other game programming books. It would be nice if the delay from idea-to-publishing was shorter, but that's obviously not really up to you! :)

One thing I'll admit not to like too much is the organization fo the book. I'm not asking for a GPU Gems-like coloured frontpage with numbers - it's nice, but arguably is overkill. The problem for me is more how the different chapters are organized in a single topic, and also their naming scheme. It's fundamentally impossible to guess if a chapter represents an idea that's usable for a game, or only cinematics, or even only large-budget films given how slow it'd be to render.

One "easy" fix for that might be to order the chapters in one category based on how "fast" the technique is likely to be and usable it'd be in real games coming in the 2008-2009 timeframe. Or potentially just rename some chapters to clearly indicate it's "For Games" or "For High-Quality Interactive Rendering". Things like that. Anyway, this entire comment might be arguably just be a small thing, but it's still the biggest thing I could think of, heh!

As for content, one very normal trend in recent ShaderX books is that chapters are more and more specialized. This is both a good thing and a bad thing; as 3D programming further extends and perhaps complicates, specialization is to be expected. However, I know I'm personally still a big fan of chapters proposing a solution to a specific problem (water; terrain; sky; particles; etc.) rather than a really new technique.

I don't know how other people feel about this, but I do really like well done chapters that deliver a good summary of the state of the art in a specific domain and an incremental improvement to go with it. Of course, this depends completely on people actually proposing you to write that kind of chapter - but if that was the case, I think it might be worth it if you are confident in that person's ability to write on the subject.
 
I enjoyed ShaderX1 when it came out. Some chapters are obviously more interesting to me than others, but on the whole I liked the offering. My one comment from that book is that I'd prefer if there was fewer code segments, leaving those for the CD. Diagrams, discussion and results are much more useful to me in text than code samples for the most part. When I'm going to implement something, I'll check the CD code if necessary. Ideally however, I can implement it directly from the chapter description without needing sample code :)

ShaderX5 is in the mail, so I'll post back when it arrives.
 
Indeed it is a very good book.

The only negative point I can see are the drawings/screenshots. Most of them are done as if they would be color-printed. As they are only grayscale, most of them aren't meaningful anymore and doesn't help the reader at all.

So I would suggest you to make sure the authors create them with those limitations in mind. Or print the book in color ;)
 
I will try to open source ShaderX and the ShaderX2 books (it all depends on if I can get all authors to sign an agreement that they allow me this). I knew that someone will bring up the color argument :) ... as far as I know NVIDIA is paying for this. Our publisher does not want to do this. On the other side ShaderX is cheaper than GPU Gems :).
Regarding the pictures: ShaderX -the first book is now five years old- we tried to improve since then :)

ShaderX1 was a uphill battle. With ShaderX2 I tried to make things better but ended up in getting absorbed into the book process. I spent a considerable amount of spare time on this book ... and the job that pays our bills suffered a bit at the time. ShaderX3 we used the first time the section editor idea and this worked out well. This way the time budget for each of us is low enough to do it in the spare time.

Naming conventions: yes this is difficult. We usually only try to focus on game programming related areas, but sometimes a technique looks good for future hardware, then we also include this. It is hard to predict the speed of techniques on different hardware with different geometry ... e.g. a simple demo showing a technique might not be vertex bound, but any real game implementation is easily vertex bound ...

<<<
However, I know I'm personally still a big fan of chapters proposing a solution to a specific problem (water; terrain; sky; particles; etc.) rather than a really new technique.
<<<
I am thinking about this for a while. A book that is divided like this would be cool, but then you would need to pick the core algorithm that solves the problem the best. For example there are numerous ways to do terrain and each one is only valid for a certain type of game ... an isometric game will use terrain different than a fps game etc.
The same is true for water: you need different ways to do water to mimic small rivers, large rivers, the ocean or just a sea ...
 
grayscale printing is not that much of a problem, as long as the pictures are created to be grayscale-printed... of course color is nicer, but I understand the financial reason for that.

Beyond that, the series is excellent. with volume 5 being my bedside book for about one month ;) So, keep up the good work!
 
Beyond that, the series is excellent. with volume 5 being my bedside book for about one month ;) So, keep up the good work!
Did you have any nightmare so far? :)
Kidding, I received my copy a few days ago and I'm reading articles before going to bed as well..
 
nAo >> rather magic dreams :LOL: the problem is, I get a lot of new ideas while reading, but have forgotten everything the next morning ;)
 
<<<
However, I know I'm personally still a big fan of chapters proposing a solution to a specific problem (water; terrain; sky; particles; etc.) rather than a really new technique.
<<<
I am thinking about this for a while. A book that is divided like this would be cool, but then you would need to pick the core algorithm that solves the problem the best. For example there are numerous ways to do terrain and each one is only valid for a certain type of game ... an isometric game will use terrain different than a fps game etc.
The same is true for water: you need different ways to do water to mimic small rivers, large rivers, the ocean or just a sea ...
The bolded part is mine. That's all I need to say or suggest wrt the ShaderX series.

It all depends on who you target wrt the ShaderX series.
 
Back
Top