• entries
    422
  • comments
    1540
  • views
    488809

Yay to getting up at 7am on a Sunday

Sign in to follow this  
jollyjeffers

139 views

Just to demonstrate how much work I'm having to put into this book that hopefully you'll all go and buy (at least three copies each please [wink]) I set my alarm for 6am on a Sunday and was sitting at my desk typing away by 7.15am.

Three hours later and I've finally finished the third chapter from my section of the book. [grin][grin][grin]

That means I've now finished the 1st, 3rd and 4th chapters - ten days to finish the 2nd, 5th, 6th, 7th, 8th, 9th and 10th. Not sure whether'[lol]' or '[bawling]' is more appropriate...

The breakdown of "Techniques for Per-Pixel Lighting":


  • 11,361 words
  • 38 pages of A4 (~55-60 pages in the final book, I think)
  • 3 techniques for fetching normals
  • Code for generating tangent-space coordinate systems. Including runtime in a GS
  • Complete discussion of formats for storing normal maps and their trade-offs
  • Simple normal mapping
  • Parallax Mapping with Offset Limiting
  • Ray traced (aka Relief Mapping, aka Parallax Occlusion Mapping)
  • Ray traced with self-shadowing


Pictures:



Simple Normal Mapping


Parallax with Offset Limiting


Ray Traced


Ray Traced with Self Shadowing
Sign in to follow this  


10 Comments


Recommended Comments

Quote:
Original post by jamesw
Holy crap dude, you bet I'll buy the book.


Hell yes!

Since I'm a beginner in so many things, I need to suck up as much knowledge as possible. From the looks of your book, it should provide me with enough knowledge for 6-8 months.

Share this comment


Link to comment
Heya!

First off, thanks for moderating the DirectX forum, always appreciate the responses. I'll certainly be buying your book when it's published!

A few questions about the book:

What PS spec are you writing it to, 3.0? 4.0?
Are you going to cover Geometry Shaders at all?
Are you going to cover any of the relief mapping optimizations, like Cone-Step mapping?

And one last comment:

Your last self-shadowing relief mapping looks a little strange. While I believe the ray-tracing looks like it is calculating the shadows correctly, it seems like it's replacing the shadowed pixels with a flat gray color. Of course, I'm on a laptop in a bright room, so I may just not have the contrast to see the shadowed/shaded pixels correctly.

Either way, keep up the good work!

Share this comment


Link to comment
Quote:
Can I buy it with MVP bucks?
heh, would be nice if we could get other books with those wouldn't it! [smile]

Quote:
What PS spec are you writing it to, 3.0? 4.0?
Purely SM4 as that is the only model exposed for Direct3D 10.

Quote:
Are you going to cover Geometry Shaders at all?
The book does cover the GS, but my particular section doesn't use it much - there aren't that many opportunities for it with conventional lighting models.

Quote:
Are you going to cover any of the relief mapping optimizations, like Cone-Step mapping?
No, not really. I simply didn't have the time or page-count to fit everything in. I've provided a basic POM/RM implementation with a few notes on how to extend it (e.g. Tartarchuk's soft shadowing technique in the ATI sample)

Quote:
it seems like it's replacing the shadowed pixels with a flat gray color.
You are correct in your observation! The main evaluation returns a value between 0.0 and 1.0 as you'd expect, but the final part of the PS adds an ambient term of 0.25, which is most obvious in the shadowed areas. No particular reason for this - just thought it looked better [smile]

Cheers,
Jack

Share this comment


Link to comment
You know, this is just me thinking out loud here, but if an artist went up to me and said that he'd like to basically make the 3rd image there using pixel shader raytracing as opposed to just doing some extra geometry+parallax mapping, I think I'd PROBABLY kick him in the nuts.


THE ALIASING! MY EYES; THE GOGGLES DO NOTHING!

Share this comment


Link to comment
Glad I'm not an artist working with you then [lol]

It would be interesting to run a full analysis of brute force geometry, proper displacement mapping and POM/RM - see if there really is that much advantage to using tricks over genuinely creating the geometry...

Jack

Share this comment


Link to comment
I seem to recall we discussed this before, but I'd guess the main advantage of using POM (or the like) over regular geometry is scalability as it helps reduce the art requirements (both in dev time and memory-wise). You can use the same geometry for all target platforms, ranging from shoddy per-vertex lighting for low-end platforms, to normal mapping, to POM by throwing some relatively simple textures into the mix.

Does that somewhat intelligent reply in any way warrant an evaluation copy of your book? ;)

Share this comment


Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now