Having created character turnarounds in Adobe Illustrator I
decided to begin modelling my character to rig and animate. Unfortunately, due
to my characters head being an umbrella I had not considered a top view and was
unable to use this as reference.
To use the turnarounds
as useful guides, I selected the right viewport and selection configure in the
option menu. Within the configure menu I selected the Back menu and loaded in
the side view image I had created. Following this I selected the front viewport
and repeated the same process with the front view image.
To ensure the images were lined up correctly, I then created
a null object which I used a base to add reference points. However, due to the
accuracy caused by drawing in Illustrator the images were positioned correctly.
If by some chance my reference imagery was not aligned, I
would change the numerical data in the Offset X and Offset Y boxes so that the
front image was aligned with the side image.
Now that the guides for my model were
successfully in place I added a cube to the project which I shrunk down to the
size of the toes of the characters foot using the scale tool. Now with the cube
in place correctly I made the cube an editable object, allowing me much more freedom
to manipulate its shape and attributes.
From this stage I entered point mode, a setting which
changed the visable object so that all the corners have points which can be
used for manipulation. Then using the direct selection tool and the move tool
(ensuring that only the select visible points box is unchecked) I selected the
points and adjusted them so that the edges of the cube followed the guidelines
of the image to some extent.
It found that it was a good idea at this point to switch to
the front viewpoint and ensure the points are aligned with the image in that
view as well before the model became more complicated.
Once the first object had been moulded, rather than create
multiple objects creating a slower, less efficient model, I chose to extrude
the object.
To do this I changed
the object mode from point mode into polygon mode and selected the face to the
left of the object. After switching the viewpoint to one where the perspective
view is visible (for easier polygon viewing) I selected the Extrude Tool and
used this to extend the shape by dragging on the screen nearby the object.
By doing this, the object is extending allowing more
editable points and helping to maintain the structural integrity of the mesh.
Following this process, I carried on extruding and aligning
until the foot was complete stopping when the object reached where the heel
joint should be.
Having rigged the
foot into the vague shape I wanted it to be, I decided to use Hypernurbs or
subdivision surfaces. Hypernurbs are basically a generator object which
subdivides an object and smooths out angular edges when the object it placed
inside it, for instance a cube inside a hypernurb would appear to be a sphere.
Above: Without hypernurbs. Right: With hypernurbs.
To achieve a more accurate curvature in lines of the foot,
once the foot was placed in the hypernurb I reselected the foot object and used
the point tool to adjust the curvature of the lines.
Above: Example of using points to manipulate line curvature
Whilst doing this however, often I selected the hypernurb
itself and began to edit. Doing this caused a permanent subdivision which
causes significant problems in modelling the rest of the character as picture
below.
It is important to try and keep the objects mesh as square
as possible as if a triangular mesh is placed in a hypernurb, the subdivision
will not work effectively and various problems will occur.
Having repeated this modelling process for one half of the
body I began to work out how the joints would work for my character. As they
were based vaguely of a mannequin I decided that where a joint would appear I
would create a sphere and that hopefully when it came to rigging the sphere
would create the illusion of a ball and socket joint for the character.
After modelling the legs, I decided that since this
technique had appeared aesthetically pleasing, I would use a more enclosed
variation of it for the elbow joint.
Having attempted a basic rig I realised that the enclosed
nature of these elbow joints was unrealistic and would physically not be able
to rotate enough for a conventional walk cycle. To counter this, I decided to
flatten the part of the leg or arm that would be interfering with the natural
movement.
To further replicate natural movement, a hierarchy is
needed. A hierarchy is the placement of objects inside other objects for
instance a hand inside a forearm. Doing this means that every time the forearm
is moved, the hand moves with it stopping and pieces of model being lost and
helping with the cleanliness of the project.
Having successfully modelled one side of my characters body,
I then began creating the other half. Rather than doing this manually and
wasting a lot of time, I decided to learn to use the symmetry object. Having
learnt about the volatile nature of hypernurbs, I chose to remove the object
which was being cloned from the hypernurb before attempting any symmetry.
A
symmetry object works in a similar way to a hypernurb in that it affects the
object placed inside it in this example a leg.
Despite
it appearing on screen, the second leg is not its own object yet. To separate
the two objects I selected the symmetry object in the objects menu, right click
and select current state to object.
Doing
this will create a duplicate of both of the legs where both the legs are
edittable as one. Once this is done, I removed the original leg from the
symmetry object and simply delete the copy of the original leg using the point
mode and the selection tool.
However
as the torso/body will need to be one solid object, the approach of using the
symmetry tool was different.
Before
placing the torso into the symmetry object, I used polygon mode to select and
delete the side which would be joining to the other side, making the torso
hollow. After this was done I placed the torso inside the symmetry object.
Using the move tool along the X axis, I made sure that there was a small gap
between the two sides of the torso before selecting current state of object as
before with the legs.
With
both parts of the body now edittable, I entered point mode and decided to weld
the parts together. To begin welding, I selected two opposite points using the
direct selection tool (this time with the only select visable elements box
checked).
I
then chose the weld tool which is located in under the mesh menu, create tools
and then select weld. As this is a repetitive process I decided to create an on
screen window allowing me to access recently used tools. I did this by clicking and holding the mouse
button over the weld tool logo, dragging the mouse over the dotted line at the
top of the menu and releasing.
With
the weld tool selected, I hovered the mouse in between the two selected points
until a white point appeared in between the two selected points. Clicking this
white point causes the two previously selected points to join allowing both
sides of the torso to become connected.
After
repeating this process all around the torso I had successfully created the
whole 3D model I intended to use except for the head. The method I eventually
found to create the head consisted mainly of using MoSpline rigs and therefore
will be covered in the rigging development guide.
