Updated July 10, 2005
HOME PAGE > REBEL ACT DEVELOPPERS > JAVIER ROLLON MORAN
3d
Artist: Javier Rollón Morán
Before becoming a part of
RAS I was in the TV industry making some 3D advertisements with Alias power
Animator, and later, animations for the Cheshire cat from Alice in Wonderland
for DisneyQuest in a company called SiliconArtist; as some internet stuff (web pages, VRML,
etc.). Before getting into RAS I got in touch with video game developers in
This is the way we work in
3D Studio Max 2. I like this layout; 3 Windows: The upper one is the track view
and lets us easily see where the keyframes are. The
dots are the keyframe of every object of the body,
and as you can see they are only in the rotation field (The center object, also
motion field). The keyframes are the position (in
this case an angle) of the object in one moment of an animation. For most
animations we put a frame rate of 20 frames per second (From 0 to 20 represents
1 second of movement, from 0 to 40, 2 seconds....). The Black lines are only
markers to make an easy way of moving those keys.
The
bottom ones is the 3d-views, and we always put the left (or right) and front
views of the model in this 3d-view. RAS have made some plugins
for MAX, not for working but for exporting animations, models or importing maps
(to take reference where our characters are putting the feet).
The most time consuming
part, is animating a model. Making a model takes two weeks, while animating it
can take 2 moths. In this game, nowadays we have 1500 animations so, making them all by hand is a huge work to do. That is
why we chose the motion capture system, as well as the feeling it gives to an
animation, making it more realistic.
When a model is finished,
we make the skeleton of that model. This skeleton defines the form for the
pieces that cannot be deformed. You can see these pieces in the picture when
you look at the Dwarf, and you can see that there are other faces that are not
there. The engine draws these faces later. The visible pieces are linked to
each other forming a tree, where its parent is the center of the body, and the
children are the head, hands and feet. The body of the model consists of 25
parts, and two parts for the swords. We use two parts for the swords because
they guide us in fitting both the hands well.
Later, we go to a motion
capture studio where an actor put a suit with a lot of ball-markers, and 10
cameras record all his movement. With a plug-in for 3DStudio Max, made by Angel
Cuñado, we import this movement file into the main
character (in this case, the Dwarf). From this point we have two types of
movement: looping movements (walking, running, relax...) and non-looping
(deaths, strikes, jumps...).
In RAS we have to make them
loop-able and clean the imperfections in the movement. Some times, when Jose
Luis Vaello (Art Director) doesn't likes
a movement, we have to correct them, and sometimes it's better to make it by
hand completely, than try to use that motion capture movement. There are also
enemies that cannot be done by motion capture: Birds, Wyverns, spiders, demons
and other monsters that I cannot tell (surprise...heh)
Sometimes, other models can
share an animation if they have the same body proportions. If not, we have to
convert that animation to the others. For example: in one hand weapon
movements, that movement don't depend in the type of sword the character has,
its always the same, and some movements of the Barbarian (one hand movements)
are shared with the Dwarf, or Amazon. But in our game each character has a
speciality: The Knight is expert in the usage of one handed swords, so he is
the only owner of a special set of animation and combos. The Barbarian is the
only one that can make well movement with swords held by both hands, etc. What
happens if I choose the Dwarf (it's my favorite.
Looks funny), and I find a DeathSword (name of the
weapon in the picture)? I can carry it, but when I hit the fire button, it
doesn't show a well-done movement, it makes a sluggish movement. That is the
attack in the picture. That's why it is called "Dwf_g_bad_sword".
The yellow lines form the connection between parents and children in the link hierarchy. The "+" is the pivot of each object, and is where each object can be rotate around. All objects can only have rotation keyframes except the center one that can also have moment keys. The yellow lines are only visual helpers in Max.
From the beginning of my
work here in RAS, the only work I have done is animation. Well, at the end of
the game I am taking part of the model of some characters in the game, in high
polygon count for marketing. Luis Miguel Quijada and
Daniel Delgado are the people who have modeled most of the enemies and
characters. Daniel has been here from the beginning of Blade and he has done
both modeling and animating. Luis Miguel modeled the Dwarf.
The right panel is the
Inverse Kinematics (IK) menu. IK is a way of moving objects in hierarchies.
When I apply IK and I move a foot, the parents of that foot moves accordingly,
until the parent stops, by putting a Terminator modifier to it (in the right
panel). Imagine that you have a character in a relax movement, and his leg come
out bad and introduce it into ground. How do you fix this? Easy! Using that
panel I put the numbers of the start and end of the animation where you want
the leg to stay in place. Mark the center as a Terminator, and the foot as Bind
Position and Bind Orientation. Then press the Apply IK button, and it's
fixed!!! Bind means to make an object stay in place and orientation
When we finished the
movement in MAX (or think that we have end the movement), we export the
movement to the game, and test it. Now we have Direct3D and OpenGL, but before
we needed two monitors to put it into Glide.
The blue line, with the red
boxes, is showing the line of motion for an object. The boxes are measuring the
amount of time between the frames. The boxes are only shown for the motion
keys, none for the rotation keys; each box is put in every key of the animation
(20 keys, 20 boxes), and when a lot of boxes are together then the movement is
slower. Sometimes, if we have 100 frames with 2 keys, one at the 0 frame and
the other at the 100, but the position is very different, the animation can be
slowly. It is a great help for us to make the movement smoother.
Normally we don't work in
shaded mode. We use wireframe mode (it's quickly),
but we use the shade mode to see that none of the body parts move inside other
parts of the body.
As an
anecdote must say that when we were in