TITLE: Rendering 45
NAME: Werner Versluys
COUNTRY: Belgium
EMAIL: werner.versluys@skynet.be
WEBPAGE: None
TOPIC: Unbelievable
COPYRIGHT: I SUBMIT TO THE STANDARD RAYTRACING COMPETITION COPYRIGHT.
JPGFILE: 45wv.jpg
RENDERER USED: 
    Autocad R14's Photo Raytrace

TOOLS USED: 
    Autocad R14 for modeling, texturing, lighting and rendering
            Microsoft Internet Explorer 4.0: for finding the background image
            on the World Wide Web
            iPhoto Express 1.1 (Ulead Systems Inc.): for conversion
            from 1.407 kb BMP to 233 kb JPEG

RENDER TIME: 
    about 1 hour 30 minutes

HARDWARE USED: 
    Pentium 200 MMX, 64 mb RAM


IMAGE DESCRIPTION: 


Imagine a conflict, somewhere in cyberspace. Opposing fractions are pictured
RED and BLUE and their battlefield, obviously, is a chess-board. This is not
by chance, for chess, to many people a game of strategy and superior tactical
insight, is essentially about winning or losing. And how are they doing?
Well, many scenarios may have lead to the current situation, where the white
queen threatens the black king. Apparently, the BLUE figures are pushing the
white pieces around, since one of them is leaning casual against the attacking
queen. This, while a black tower, only recently slain, is being removed by a
RED player. It's RED's turn now, and they are running out of options. They
cannot move their king to safety, nor threaten the attacker; they can only
block it's path by means of pawn E6, F7 or horse G8. To move one of these will
close the trap and lead inevitable towards a victory for BLUE. And BLUE knows
it. They appear confident and relaxed, pointing out the situation to RED and
studying the possibility of a new game.
As for RED... well, they simply can't believe it !

Note: 
The tactical situation on this board has been altered many times for reasons
of composition. If, therefore, the chess players amongst U see another possible
outcome for the situation I described above, please do not hold it against me,
for I am only a modest player myself.
Let's just assume that RED isn't very good at chess either !


DESCRIPTION OF HOW THIS IMAGE WAS CREATED: 


Concept:

The most basic idea I had was to participate in this competition, which was
quite a big step since I never actually put a scene together before. I did find
it stimulating to work towards a goal and a deadline, meaning that you have to
let your creation go at some point. And for the best, I guess, especially since
my family is lining up behind me for use of the computer. So I better send it
in before they shut down my rendering operations altogether.

Modelling and composition:

I made the chess-board and pieces some time ago, but without having an actual
use for them. Now I did, so they had to be in the scene. The are all made as
solids and exact copies of the real thing, meticulously measured. I'm quite
pleased with that.
Next, there was this K'NEX-figure, a toy I borrowed from my children, that had
been standing alongside my computer for some time, waiting to be modelled. Now
was a good time and again it turned out to be a fairly good copy.
So, I had a chess-board, K'NEX-figures and a topic which I find to be open to
interpretation if I have ever seen one. I came up with this: "Two groups having
a game of chess with life-size pieces and experiencing a checkmate situation
which one group just can't believe". I did lose the checkmate situation along
the way, but I find the overall composition to be more important to me.
(That is: when I noticed, there was no way back.)

To start with, I needed a point of view towards my chess-board, and I wanted to
get away from the isometric pespective, affraid it might seem to easy. But
every other angle I tried looked weird. Then I realised that the board would
seem randomly filled with pieces, chaotic even, so I went back to the isometric
perspective: tight geometry as a contrast to (seemingly) random positioning.
The actual viewpoint is (-1,-1,0.7) I had to change something :)

Next, I needed the tactical layout, something I tried out on the real board
first, modifying a situation found in a book to create gaps for the figures to
be placed in. Basically, there are two vertical open lanes, one on the left
half
of the image, one on the right. The center is an open square, marking the
vertical
line between the white queen and the black king and therefore drawing attention
to the main event.

Each K'NEX-figure was then saved in a separate file (about 5 mb each) to create
the right pose. The file-size made me worry about the final drawing-size (I
wanted
six figures from the start) so I made a copy every time a figure was added to
the
scene. If my computer would choke on it, I could always go back one step, but
fortunately, that didn't happen. I also had to scale the figures down. They
were
1.5 times bigger originally. Their positionning in the scene was made
deliberately
symetrical because they would draw attention before anything else by means of
their color. Therefore they had to be evenly spaced-out around the center.
There
are two in the upper left, two in the upper right, one down left and one in the
down right quarter of the image.

Finally, I slightly changed the positions and orientations of the pieces
compared
to the squares they were standing on. Until then, they were all nicely centered
and therefore looking completely unnatural.

Textures and mapping:

I went somewhat easy on the textures. First, because I imagined that most
testing
would go to composition and lighting. Secondly, because I had no need for
stunning
textures in this image (all objects are at some distance from the viewer).
Thirdly, since I'm only just learning the basics, I have one rule: keep it
simple
(I had a chance to try out 3Dstudiomax once, and the material editor still
makes
me wake up sweating at night). And finally, I am quite happy with the textures
that came with Autocad R14 (for now, that is).
So, the wood-textures (both white and brown) are standard textures, without
changing anything. The K'NEX-figures looked already plastic with their drawing
colors, but I assigned a plastic material (grey, blue and red) so I could
change
the reflection and roughness a bit.

All the pieces are mapped in the same way, but the king's cross looked funny. So
I
made them into separate objects and mapped them differently. They look better
now,
though it's barely noticeable in the final image.

One thing that still bothers me is that the K'NEX-figures don't really seem to
be 
"on" the board. They could easily be a few inches above it (they are not, trust
me,
but it appears that way). I noticed this after the first test-renderings and I
solved it by making the board reflective (mirror, actually). At first, it didn't

work because the blue and red from the bottom figures was reflected right on
the
edge, throwing the whole composition off balance. I then made only the squares
mirror, and just a little bit so the reflected legs were barely noticable. It
was
beautiful and exactly wat I wanted. Unfortunately, all these tests were done
with
a black background. Once I used the red star-picture (which I had tested with
the
empty, non-reflective board before) the red background was reflected as well,
throwing the whole color-scheme off balance. A moment of frustration, indeed.
Especially, since I couldn't solve the problem. I tried some other backgrounds
(the typical black-space/white-star images) but they just didn't have the
impact
this one had. So I abandoned the reflective board, knowing it was the finishing
touch but believing that the first impression of an image is probably more
important (isn't it?)
So another week had gone by, without any significant progress. I did learn
something though: making computer generated images is about suffering.

And then there is the image within the image. The idea was to repeat the final
picture on the map this K'NEX-figure is holding (not an original idea but I
like
these kind of twists). During all test-renderings I used this previously made
background test. Once I replaced it with the final image, it was a
disappointment.
You could barely recognize the K'NEX-figures, so I changed it back to the
test-map
and it is much better this way.

Lighting:

This was my major worry because I had no idea how to tackle this. So I explored
some books about photography and found this very basic setup. There is a
keylight
(pointlight) about 45_ to the left, from the viewer's point of view, and 45_
up.
To soften the shadows, I put a second pointlight about 15_ to the left, a bit
lower
and only half the intensity. I then added a backlight (spotlight) about 105_ to
the
left and 45_ up again to produce highlights (same intensity as the keylight).
And
even a fourth light (pointlight) was added, way above but with low intensity,
to
make the scene seem evenly illuminated. This should also explain the absence of
shadows on the board. I tested with shadows but, believe me, this scene does
not
need shadows. Once again, it completely disturbed the composition and since I
worked hard on that, I wasn't going to ruin my composition with a simple
effect.
Making computer generated images is about courage, too.
Now, since the lights are way up and all the K'NEX-figures are slightly bending

forward, there were dark shadows on the inside. I therefore put an extra
spotlight
on each one of them, shining up to soften these shadows. You don't really
notice
the presence of these spotlights, but I they weren't there, you would find
something to be missing.
It then took me several test-tenderings to find the right intensity settings,
and
even then, I wasn't satisfied. Until I discovered you could turn the ambient
light
off (intensity=0), and suddenly it looked so much better: the pieces seemed to
have a lot more volume.

And finally:

So that's about it. The drawing file turned out to be 34 mb, which is why I
didn't
include it. The final rendering was number 45, hence the file name. Most of
them
took more than an hour. Making computer generated images is mostly about
waiting...