With it having been several years since I last worked with Direct3D (DX9), I wanted a book as a refresher in the DirectX way of doing things when I decided to return to computer graphics. What I got, was largely unspectacular.
Practical rendering is by no means a poor book. It authors are Microsoft DirectX Most Valuable Professionals. This means the material presented is accurate and well written, but it fails on too many fronts to be considered great. The first half of the book is dedicated to explaining the Direct3D 11 Pipeline or at least it tries to. What you get is ultimately a regurgitation of the freely available DX documentation. The authors do little to actually explain the behind the scenes workings and I have a feeling if it is your first foray into DX you will be quickly lost. The one bit of explanation they routinely throw at you is through the use of images to explain concepts. This sounds excellent until you realize what it really means. You get images like a cube with six exploded sides demonstrating a cube map (which is sadly one of the better images) and my personal favorite, an image of a sphere in three different positions to demonstrate translations. This examples may sound petty, but if you read this book you will constantly roll your eyes at the ridiculousness of these listings. Code listings for the book's first half are no better. They are literally ripped from Microsoft's documentation and dumped on the page in an unremarkable matter.
The book improves in it's second half with more concrete examples of the concepts. They are actually interesting reads and very well explained compared to the first half. Unfortunately, here is where the book's biggest problem comes in. The authors have elected to use Jason Zink's Hieroglyph 3 engine as the basis for all of their examples. While I'm certain Mr. Zink's engine is of a high quality, it is a huge mistake. The justification for it's use is so we as readers are not bogged down in minutia when it comes to initializing Direct3D and Win32. In practice, it fails to allow us experience in initializing Direct3D. This is a fairly important component of using the API and it's dismissal is absurd. You will be forced to return to the documentation of the DXSDK in order to find anything of use, unless you want to be locked into the Hieroglyph engine. The biggest problem with authors using their own engines is in the changes that occur over time. Including raw DX and Win32 code allows future use even through subsequent DXSDK changes with a minimal of rewriting. The Hieroglyph engine is already changing from the version when the book was published just a few short months ago. Nowhere is this more apparent than in the books appendix stating that the Boost libraries are required for building the engine. On the engines homepage, this dependency has already been removed. This isn't a big deal for now, but does speak to the rate at which libraries tend to change overtime. It is entirely possible in the future the engine will have changed so much it's usefulness will suffer. Because of the use of the Hieroglpyh engine, all of the examples focus on shader code and leave everything else up to the engine itself. This is not particularly useful when you want to learn how to code in D3D11 from the ground up.
While the authors have presented a few useful chapters, the book fails to deliver consistently. If you are looking for anything other than a few shader code examples of trendy topics, you will have to look elsewhere. I recommend picking up Frank D. Luna's Direct3D 10 book to learn the fundamentals of DX programming. Afterward the Direct3D 11 documentation will be more than sufficient at highlighting the differences in the older and newer APIs. If you want the examples this book offers, I would suggest a GPU pro or ShaderX book as they are considerably heavier on content and will provide many more examples than this book provides. Again, it is not a bad book and if I were looking for strict documentation this would be high on my list. It's weakness however is in striking a balance between documentation like theory and cohesive examples of implementation.