Chapter 8: Advanced Character Modeling using Polygons and Subdivision Surfaces

Subdivision modeling has become the standard way that character modeling is being handled in most video and film environments. Maya has implemented a powerful and user-friendly way to use subdivision surfaces. The purpose of this chapter is to show the user a proven method to create a complex character using subdivision surfaces. In this case, the example will be a character’s face.

When an artist can focus on the human face, they are focusing on what possibly could be the most easily recognizable, familiar, and expressive 3D shape known to humankind. This shape is also notoriously difficult to get just right. There are subtle nuances and details that everyone sees everyday numerous times, and can be spotted very easily by anyone. The exercise of creating a human face from a conceptual sketch will give the 3D artist the ability and tools to tackle any model that uses organic forms and shapes.

This tutorial uses the concepts and artwork of Krishnamurti Costa, a prominent and talented 3D artist. His techniques and processes will be documented in detail, explaining every step he uses to create his artwork.


Defining Loops and Areas of Detail

Building a Polygon Cage

Converting Polygons to Subdivision Surfaces

Subdivision Levels Overview


The character will be stylized, but have human features to make insure that there are more different modeling situations that can be explored for this tutorial.

Instead of creating a 3D model from scratch, artwork is used to illustrate how a model would be created in a professional environment. Very seldom will an artist be given a project to work on that he has complete creative control over. When an artist is shown how to model from a piece of artwork, they can easily use that technique in a professional situation where they will be asked to create models from conceptual art.

Laying Out the Views in 2D

These drawings are excellent examples of the level of detail needed in the artwork to create a detailed subdivision model.

Figure 8.1: Front View front view

Figure 8.2: Side View side view

A minimum of 2 views are necessary to create an accurate 3 dimensional model from 2 dimensional artwork. These views, called orthographic views, need to have the overall dimensions line up between the drawings. The overall length and width and the placement of the details, like eyes, nose and mouth need to be aligned between the 2 drawings.

Figure 8.3: Two Views Aligned two views

Before modeling can begin, it is important to know something about modeling in general. There are certain practices and guidelines that should be understood before the first modeling tools are used. These concepts, once they are thought about beforehand, will guide the modeler through the modeling process easier and faster.

Modeling Concepts

These concepts are universal concepts that can be applied to any model. The way that a model will deform, animate and render will all depend on how the model is built. The flow of the geometry requires planning, and the planning of the model begins before the first curve is drawn.

Different modelers and studios use different processes to create polygonal models. Any given process will create different results. Many modeling tutorials will suggest an approach where the modeler will use primitives to start a model, and fill in the detail later. Some studios will use scanned data to digitized data to begin modeling.

The approach that will be taken in this tutorial uses carefully placed rows of polygons to create a very easy to edit polygonal cage. This technique is called “poly-by-poly modeling”. This polygonal cage is the basis for the subdivision model. If the geometry flows naturally along the surface of the model, it will make texturing, lighting, rigging, and animation much easier later in the production process.

Before modeling takes place, it is important to understand the model that you will be creating. How will it move? What kind of detail will be required on this model? Where is the detail going to be? How will it be textured? What kind of lighting will be applied to it?

The production requirements for a model can be very considerable, and may change during the time the model is being created and used in production. It may be impossible to know exactly what will be required of the model before it is created, but a good modeler will attempt to make educated assumptions about these requirements.

An experienced modeler knows that there are 2 things that will probably happen when they are given a model to build for a project. When they hear that the model will not have to be too detailed, it is almost certain that the model will appear full screen at some time. And, when the model has been detailed to support any amount of close up scrutiny, the rendering deadlines will require that the model’s level of detail be taken down to the absolute lowest levels it can go to expedite production.

Luckily, subdivision modeling supports both of these scenarios. When the model is carefully planned out, and executed using levels of detail in the subdivision, the model can hold up to careful scrutiny, and can have the detail reduced quickly and easily to render out fast and easy.

Modeling Standards

Modeling standards change from studio to studio. In some cases, it is fine to have a model have places where there are a lot of triangles in the model, and in some cases, it is strictly forbidden. Generally, if it looks good in animation and lighting, there is some freedom allowed in the process used to create the model.

It is important to note that modeling processes have changed over time, and what was considered fine practice a couple years ago is now not allowed. Two specific cases in point are the use of triangles in a subdivision model, and the occurrence of a condition called a “star”.

A “star” is a place where 5 edges come together in one 5-sided corner. This condition can cause many problems in animation and lighting. They cause irregular flashing and twisting when the object is used in production. The model shown here has been built to the standards specified here, and was produced using just four-sided polygons. “Stars”, on the other hand, are harder to avoid because are caused by the edge loops crossing. “Stars” can be controlled and should be put in places where there is little or no deformation. When it is impossible to avoid a 5-cornered condition, always avoid putting it in a place where there will be a great deal of movement.

Figure 8.4: Star Condition in Polygon Layout star

One important point to keep in mind is that subdivision models work much better with four sided polygons than three sided polygons. In this exercise, this model will have no triangles whatsoever. With careful placement of edges, and by using the tools explained here effectively, it is possible to create high quality models that have no triangles in them at all.

Animation Flow

The first concept to understand when planning out the flow of geometry in a model is animation flow. Animation flow is an idea that basically states that the flow of the geometry must support the way the model will move. A model must have geometry that can be moved along the lines that the model will be required to move.

The areas that are required to move in this model are the eyes, mouth, and jaw. The flow of animation in these areas will surround these areas. By outlining the general vicinity surrounding the areas of movement, we can visualize how the geometry will need to flow along the model.

Figure 8.5: Areas of Animation animation flow

This geometry flow is simple to understand. Geometry must flow radially away from areas of movement. The careful arrangement of radial polygon rows placed along the paths of motion will allow the model to be easy to rig and animate. Rows of polygons that surround the animation areas will create uniform movement in the model, will define better the anatomic forms and will allow the model to render properly as well.

Figure 8.6: Animation Flow Diagram animation flow

Edge Loops

A more complex idea that relates to construction is the concept of edge loops. Edge loops are related to animation flow, because they follow the muscle construction in the actual face, and the face will naturally move along the path of these muscles. But the edge loop concept takes the idea of animation flow and expands on it to cover the construction of muscle groups, bone areas, and areas of fat along the surface of the face.

Muscle edge loops are placed around areas of the face that will move. These places are the eyelids, the lips the brow and the jaw line. Muscle areas must be carefully modeled to avoid awkward intersections and triangles. These areas will require the closest scrutiny when animating, and will have the most problems if not modeled correctly.

It is important when defining the edge loops of the muscles to try to integrate the different muscles by defining lines that not only radiate parallel to the areas of movement, but also define the lines perpendicular to the movement where different muscles will connect. In the figure shown below, notice how the lines flow around the lips, but also lines will flow through the lips as well. These lines allow the flow of the geometry of the lips to align with the flow of the geometry of the cheek and jaw muscles.

Figure 8.7: Muscle Edge Loops muscle loops

Fat edge loops define the areas of the face where movement takes place, but it is passive movement that is not driven by a direct force. This movement is caused by the muscles. Fat areas must be as carefully modeled as muscle areas because they will move almost the same amount. Fat areas do not really have to interact with other fat areas. Fat areas exist as small islands of passive flesh that allow the muscles to move freely.

Figure 8.8: Fat Edge Loops fat loops

Bone edge loops are the hard areas of the face where the bone comes closest to the surface. These areas do not have much flexibility and will not have much movement within these areas. For this reason, it is convenient to use these as places where the model can have transition areas like 5 cornered intersections. Because these areas do not move much, some geometric flaws can go unnoticed if constructed carefully.

Figure 8.9: Bone Edge Loops bone loops

The entire edge loop layout is used to plan the placement of the geometry during construction. Using careful planning and geometry placement, a model can be built to be easy to edit and easy to animate. Even if the model will not be used for animation, a good edge-loop layout can help your model to have a good definition of the anatomic details, just because bones, fat and muscle loops are part of the same main volume.

Figure 8.10: Edge Loop Layout all loops

Modeling the Face

Once there is a basic understanding of the flow of the geometry, the process of building a wire cage can begin. The modeling process is much like any other kind of artistic endeavor. The artist will work on the piece until the work is finished, and throughout the course of the work, many operations will take place. The artist will build something, tweek something, erase something else. These steps will build up over time until the artist has done quite a lot of small things that amount to a big piece of art the artist is proud of.

The problem with modeling in the Maya is the computer will remember everything you did. The number of history steps can be set to a small number to mitigate this overhead, but over time it is important to optimize the workflow by performing certain steps during the modeling process.

1. Save Often. Save your file many times over the modeling process to unique and sequential names. These names are usually something like: workFile.0001.mb, workFile.0002.mb, etc…

2. Delete History Often. History will slow the modeler down to the point where simple things will take a long time to do. As the modeler gets more experience, they will recognize quickly when the history starts to get to be too much and will delete it.

3. Subdivide The Model Judiciously. Subdivision is a great way to create fast and very high quality models. The subdivision process itself, however, is computationally expensive. In this exercise, the model will be built almost entirely using polygons only until the very end when subdivision modeling needs to take place. During the construction process, it is important to view the subdivision model periodically to make sure the model is progressing correctly. In these cases, the model is subdivided to preview the progress, and then the subdivision operation is undone so the clean and light polygon modeling process can continue.

The process for modeling the face consists of placing the images planes in 3D, creating a wire cage that outlines the details of the face, and creating a polygonal cage to create the subdivision surfaces from.

Laying Out the Image Planes in 3D

As a part of the design process, a wire frame needs to be created that represents the 3 dimensional model. The images need to be loaded prior to any manipulation of the view ports takes place. If there is any camera movement in the view ports before the image plane is imported, use the View Default Home button located at the top of each view port to reset the view. The two drawings are loaded into the Maya view ports using View Image Plane Import Image Plane button located at the top of each view port.

Figure 8.11: Import Image Plane Button image plane

Continue loading the images to get the front and side images located in the correct vied ports. The images should be arranged so the front image is in the front window, and the side image is in the side window. This will create a perspective view where the images can be seen positioned at 90 degrees to each other. The perspective view is where a lot of the wire frame modeling will be take place.

Figure 8.12: Images Loaded into Viewports images loaded

Now we are ready to begin laying out the model in 3D. The first thing that you can do to make the modeling process easier is to build a rough wire cage using splines that will be the basis for the polygonal modeling process. The polygonal cage will be used to create the final smooth model that will be animated and rendered. The wire cage will be the skeleton for the model.

Creating a Wire Cage

The wire cage should be built using curves that are easy to control and still give an accurate representation of the form that you are trying to create. Using the Create EP Curve Tool Options button, create degree 1 Edit Point NURBS curves with Uniform Knot spacing. These curves are the most similar to Polylines (an entity type that is not supported by the Maya modeling software). Since we are building a polygonal cage from the wire cage, it is simpler to use degree 1 curves.

Figure 8.12: Create EP Curve Tool Button ep curve button

Figure 8.13: Create EP Curve Tool Options ep curve options

Start by drawing curves in the front view. Begin tracing the details of the drawings using fairly simple curves. Remember, the more points you use now, the more points you will need to edit later. Capture the outlines around the eyes, cheeks, chin, and the rest of the face as shown. Lay out the curves roughly around the areas of animation so the model can be built along these lines later.

Figure 8.14: EP Curves Drawn on Face curves on face

Once the front view is finished, the curves in that view all rest in one place at “0” (zero) in the “Z” axis. When the front view curves have been drawn, the curves in the front view need to be edited to match the side view. Begin, for example, with the curve in the lower eyelid. Translate the curve outwards in “Z” to get the curve positioned correctly. Edit the curve point-by-point to get the curvature correct along the “Z” axis.

Figures 8.15: Lower Eyelid Curve Edited lower eyelid

Continue this process with the rest of the face. The Curves will all need to be edited before any curves are added to make additional details for the side view. There will be curves like the bottom of the chin and the brow that may need additional information to be added to the curves so the curves flow correctly along the “Z” axis in the side view. This can be done by adding curves in the side view that extend the front view curves beyond their original extents and joining the curves later during the modeling process.

Figure 8.16: Front and Side View Wire Cage front side cage

Figure 8.17: Perspective View Wire Cage persp cage

Work on the wire cage until it is a clean representation of your concept. This is a critical step in the modeling process. This wire cage will allow you to visualize your design in 3D before you begin creating polygons and subdivision surfaces using easy to edit degree 1 curves. It is also important to remember that this wire cage can be edited later as well, so do not waste unnecessary time finessing it. This is a guide, the real modeling will be done later. Another important thing to have in mind is that we are working with orthographic views, front and side, and the 3D model needs to be adjusted to fit in the human eye perspective view (50mm).

Creating a Polygonal Cage

Most artists will use their favorite tools to create their artwork. The same is true in modeling. Not all polygonal modeling tools that exist in Maya are necessary to create a great model. The trick is to get good at the tools that you really need, and know when to use them.

This section of the chapter will focus on how different tools will be used at different stages of the modeling process. The specific tools that are used relate to the specific modeling process that is taking place. As with all modeling projects, the model will start with simple forms and will move into more complex and refined shapes later.

The tools that are used at these phases of construction reflect the way that the construction will go from simple to complex. There will be many tools that are used throughout the process of constructing the polygonal head, but the main tools covered here are:

1. The Polygons Create Polygon Tool will be used with the Polygons Append to Polygon Tool to create the entire main form of the head from the first rows of polygons around areas of the head where the animation will flow until areas including the eyes, mouth and jaw areas.

2. The Edit Polygons Split Polygon tool is used to cut faces to add more details or change the edges direction of the model.

3. The Edit Polygons Extrude Edge tool will be used to extend the rows of polygons to cover large areas of the model quickly. The Edit Polygons Merge Vertices tool is used to collapse vertices and close holes in the mesh.

4. The Edit Polygons Sculpt Polygon Tool and Modify Soft Modification Tool to create subtle and very controlled modifications to the geometry using a “faded deformation”.

5. The Edit Polygons Collapse tool will be shown as a way to reorganize the polygonal layout of the model surface by collapsing different edges in only one.

6. The Edit Polygons Extrude Face tool will be shown as a way to create very interesting areas of detail that maintain excellent geometric layout as well as beautiful detail.

Creating and Appending Polygons

Polygonal modeling begins with the creation of a single polygon. Using the edge loop layout, polygons can be placed strategically and then duplicated into rows using the Polygons Create Polygon Tool. Polygons are drawn using the snap to curve function, or by pressing the “C” button while drawing the polygon. The points of the polygon will snap to the wire cage that was constructed earlier.

Figure 8.18: Create Polygon Tool create polygon

The next step is to append the polygon into a string of polygons using the Polygons Append to Polygon Tool. This tools works much like the Polygons Create Polygon Tool, but the user must first pick the polygon that needs to be appended, then the edges of the polygon that are beyond the polygon are selected using the “C” key to snap the new edges to the wire cage curves that were drawn earlier. The Polygons Append to Polygon Tool makes clean polygons that are already merged to the original polygon, and do not require cleanup.

Figure 8.19: Append Polygon Tool append polygon

Proceed working around the eye, then press enter to complete the process. For the next row of polygons, the Polygons Append to Polygon Tool is used again, but the lower edge of the polygon is used to create the next row of polygons. From the sequence of images that are shown below, it is apparent how the Append to Polygon Tool can be a very useful way to create a fast network of polygons.

Figures 8.20: Append Polygon Inner Rows inner rows

Figures 8.21: Append Polygon Outer Rows outer rows

Once there are an adequate number of rows to begin visualizing the shape of the eye area, you may need to duplicate the eye area across the “X” axis. When you do this, it is important to use the Edit Duplicate Options, then use the Scale -1 and Geometry Type Instance. This will allow the model to update instantly as one side is changed.

It is important to note that the modeler will continue to use this instanced geometry to view the model throughout the entire modeling process. Only when most of the modeling is done, and the artist is about to convert the entire model to subdivision surfaces, can the instance be merged with the rest of the model to make one piece of geometry.

Figure 8.22: Instance Across “X” Axis across x

The next area that will require attention is the mouth and chin area. This is done the same way as the mouth area. The rows of polygons are started as single polygons, then strung across using the append tool. Because the character’s left side of the face is the original, the character’s right side of the face is automatically updated. The mouth is laid in first, creating a template edge for the next row of polygons. The chin is completed, and then the upper lip area is laid in later.

Figure 8.23: Mouth and Chin Area mouth chin

The forehead and the surrounding areas of the face are tied in using the same process. When joining different areas of polygons, the Polygons Append to Polygon Tool will require the snap to vertecies option, or press the “V” key. Snapping to polygons across the span will cleanly create a bridge where the model will form one smooth mesh.

Figure 8.24: The Forehead Area forehead

Splitting Polygons

The wireframe image shows the simplicity of the construction, and how the polygonal layout adheres to the original edge loop layout. At this stage, many areas of detail have been added.

Since Maya first introduced the polygon modeling tools, the Edit Polygons Split Polygon tool has been one of the most valuable tools in the modeler’s toolbox. This tool works extraordinarily well to maintain geometric integrity, while allowing the modeler to create significant changes to the topology of the model. While using the Edit Polygons Split Polygon tool, the UV’s of the model are not deleted, the polygons that are split are not detached, and there are no spurious duplicates of the polygons that sometimes occur when using other tools.

The Edit Polygons Split Polygon tool will be used through the modeling process to create areas of detail and to re-route some of the polygons in order to avoid triangles and 5-sided intersections. Used in combination with Edit Polygons Delete Edges the modeler can keep a model’s surface intact, but can completely change the polygonal layout of the model to suit the needs of animation and production.

To create the areas of detail shown below, the Edit Polygons Split Polygon tool was used to make rows of polygons tighter in areas where natural seams occur along the face. It is important to get these details as correct as possible before creating subdivision surfaces. This will reduce the number of times the model will require different levels of subdivision.

Figure 8.25: Areas of Detail areas detail

When the model gets to the point where there are just large areas to create with very little detail, like the top of the head and the neck, the Edit Polygons Extrude Edge tool can be very helpful. The next few diagrams will show the detailed step-by-step process of creating extruded rows that will be clean and line up with the rest of the model.

By selecting contiguous edges manually, or by selecting one edge, and using the Edit Polygons Selection Select Contiguous Edges tool, select a row of edges where the model needs to have a new row of polygons added.

Figure 8.26: Select Contiguous Edges contiguous edges

The default extrusion behavior in Maya is to separate every polygon. In order to make sure that the polygons stay together it is necessary to select the Polygons Tool Options Keep Faces Together option. This will force the extruded polygons to keep the edges together during the extrusion process.

Figure 8.27: Keep Faces Together Option keep faces together

Select the Edit Polygons Extrude Edge tool to create the row of polygons. This selection will display the extrude manipulator that can be used to translate, rotate, or scale the leading edge of the extruded polygons. When extruding the row of polygons, make sure that the row is built as close to the desired final location as possible. The manipulator makes this process as simple as it can be. Any editing that cannot be achieved without using the manipulator will need to be done by hand after the extrude operation is complete.

Figure 8.28: Extrude Edge Tool extrude edge

Figure 8.29: Extrude Edge Manipulator edge manipulator

Figure 8.30: Extrude Edge Complete edge complete

When the extrude operation is complete, it will be necessary to weld the vertecies at the edges of the polygon row that was just created. The last vertex of the row will be disconnected from the remainder of the model, and will need to be integrated.

The two vertices that need to be welded should be selected, then Edit Polygons Merge Vertecies Options selection should be executed.

Figure 8.31: Selected Vertices selected verticies

Figure 8.32: Merge Vertices Tool merge verticies

Merge vertecies will weld the selected vertecies within a given tolerance. In order to insure the selected vertecies are welded, the option for the weld distance is set fairly high. In this case, the options were set to a level of 10.0 to insure the vertecies are welded.

Figure 8.33: Merge Vertices Options merge verticies options

After the options are set, the MergeVertex button is used to execute the merging process. The finished extruded row and subsequent merged vertex can be seen below.

Figure 8.34: Merge Vertex Complete merge complete

Modifying the Mesh

There are several tools in Maya that allow the user to move geometry in a subtle way that moves the geometry like it was clay. The main ones that were utilized on this model are the Edit Polygons Sculpt Polygon Tool and the Modify Soft Modification Tool.

The Edit Polygons Sculpt Polygon Tool is a great tool for gently pushing and pulling geometry in and out, but the one of the most useful options in this toolbox is the smooth tool. The smoothing tool works by averaging the values of the effected vertecies and blends irregular areas out into one smooth area.

Figure 8.35: Sculpt Polygon Tool sculpt polygon tool

Figure 8.36: Sculpt Polygon Smooth Option sculpt polygon tool options

Another tool that was introduced in Maya version 6.0 is the Modify Soft Modification Tool. This tool uses proximity around the selected vertex to control the falloff of the transformation. The curve of influence from the selected vertex to the last effected vertex is clearly drawn out so the user can adjust the falloff and the amount of intensity the falloff has from the center to the outer edge.

Figure 8.37: Soft Modification Tool soft modification

Figure 8.38: Soft Modification Tool Options soft mod options

These tools work exceedingly well when used in combination. The smoothing option in the Edit Polygons Sculpt Polygon Tool prepares the surface by distributing the geometry more evenly, and the Modify Soft Modification Tool moves the geometry ina soft controlled way that other tools cannot do as well.

Collapsing Polygons

This technique will allow the modeler to modify the number of polygons without having to cleanup edges and vertecies by welding them. The process is simple, but should be thought out carefully before using Edit Polygons Collapse.

Begin by identifying a problem area in the model. These are places where the model has polygons that are not four sided or where the flow of the geometry does not coincide with the flow of animation or edge loops. The reorganization of a models polygon layout should be done to avoid problems later on in production.

Select the edges of the problem areas in the model, then execute the Edit Polygons Collapse tool.

Figure 8.39: Select Problem Edges select problem edges

Figure 8.40: Collapse Polygon Tool collapse polygon tool

The resulting geometry shows how the polygons in that area have been collapsed. The polygons surrounding the area do not have problems with duplicate vertecies or duplicate edges. The collapse tool takes care of the cleanup that is usually associated with this kind of cleanup.

Figure 8.41: Resulting Collapsed Area collapsed polygon

Extruding Polygons

For fast and easy construction of detailed areas, one of the most surprisingly useful tools is Edit Polygons Extrude Face. In most organic models there are areas that undulate in and out, making conventional modeling very difficult. In this section there are several examples of how this tool can benefit a modeler when creating detailed areas around the head. \

In the example shown below, the muscles of the neck are modeled in simple polygons using the Polygons Create Polygon Tool and the Polygons Append to Polygon Tool.

Figure 8.42: Polygons of the Neck Muscle neck muscle

These polygons are selected, and the Edit Polygons Extrude Face tool is executed. The default behavior for Edit Polygons Extrude Face is to divide the faces into separate entities. The separating polygons are not usable, and require a lot of cleanup of used by mistake. It is important to make sure the default behavior of the polygon modeling toolset is the Polygons Tool Options Keep Faces Together option described earlier in the preceding “Extruding Rows of Polygons” section.

Figure 8.43: Extrude Face Tool extrude face tool

Once the extrude tool has been executed, the faces will appear highlighted with a manipulator present so they can be translated and scaled to fit into the appropriate detail. When the Edit Polygons Extrude Face tool is used, the polygons remain cleanly connected to each other. If the Polygons Tool Options Keep Faces Together options are not correctly set, and the faces separate, undo the operation and start over with the correct options set.

Figure 8.44: Extruded Faces of Neck Muscle extruded neck polys

There are many instances where this modeling technique would be beneficial. Below are two more examples where the Edit Polygons Extrude Face tool can be utilized in the neck area.

Figure 8.45: Selected Faces in Neck Area 1 faces neck area

Figure 8.46: Extruded Faces in Neck Area 1 extruded faces neck

Figure 8.47: Selected Faces in Neck Area 2 selected neck faces 2

Figure 8.48: Extruded Faces in Neck Area 2 extruded neck polys 2

Now that the basic tools have been described, the specific techniques for building individual details of the model can be explained. The parts of the model that make all the difference when viewing the final work are the trickiest to do. Much of the process that is involved in the finishing touches of any model are more technique than technology.

Details of the Head

This section will use images that have been collected showing a step by step approach to building out each part of the model. The basis for all of the rest of the construction is in place using the processes that have already been covered, but the tools are used in different combinations.

When these areas are created, it is important to know how these elements are going to look when they are subdivided. To preview the subdivision surface of the model, the Modify Convert Polygons to Subdiv tool can be used. This tool is used at this point as a way to previsualize the subdivision surface, and will not be used to commit to a subdivision model until the polygonal cage is completed. At the finishing stages of the model, it is very important that the subdivision is previewed often. Only by testing the subdivision can the modeler see what the finished version of the model will look like.

Back Area

The back area is constructed by building a single polygon first. Use the Polygons Append to Polygon Tool to create a polygon that goes from the base of the head to the edge of the lower back. This polygon is split down the center using the Edit Polygons Split Polygon tool.

Figure 8.49: Single Polygon Split in Half back poly split

The edge that was created by the Edit Polygons Split Polygon tool is selected and translated the edge to the center of the back. This will serve as the base for the rest of the construction.

Figure 8.50: Edge Translated to Center of Back vert centered

The Edit Polygons Split Polygon tool is used again make the center polygon have the same number of divisions as the edge row of polygons. This sets up the mesh construction for the rest of the back.

Figure 8.51: Center of Back Split into Sections centered poly sectioned

The Edit Polygons Merge Vertices tool is used to join the vertecies along the vertical polygon rows. This creates a simple mesh along the back that will become a more complex sculpted mesh.

Figure 8.52: Joining the Rows of Back Polygons sectioned polys joined

When all of the remaining edges of the rows are joined using the Edit Polygons Merge Vertices tool, the bottom polygon is split using the Edit Polygons Split Polygon tool. Now the surfaces is ready to become a more complex shape.

Figure 8.53: Back Polygons Joined in Center back joined

Using the Edit Polygons Split Polygon tool, the polygon mesh in the back is modified and made into a surface that can be sculpted easily by selecting points and pushing and pulling them into an organic shape. The sculpting process can also be done using the Edit Polygons Sculpt Polygon Tool or Modify Soft Modification Tool.

Figure 8.54: Back Surface Completed back done

Ear Area

The ear is one of the trickiest parts of the human body to visualize in any medium. Perhaps the most difficult medium to capture the complex form of the ear is 3D. Using digital medium or traditional sculpting methods, this area of the face has given artists problems. The only way to really get it right is to turn work from reference. For the initial construction phase of the ear, the polygonal model’s visibility is turned off, and the image planes are turned on. In the side view, the image planes are used to trace the outer edge of the ear using the Polygons Create Polygon Tool and the Polygons Append to Polygon Tool.

Figure 8.55: Outside of Ear is Traced ear traced

The rest of the details of the ear are blocked in by creating a single polygon inside the ear and using the Edit Polygons Split Polygon tool to edit the shape. This creates a 2D template for the inside of the ear.

Figure 8.56: Inside of Ear is Traced ear traced 2

The mesh that was created using the image planes in the side view now have to be translated into the correct place in 3D. Using the wire cage that was previously constructed, the ear can be placed in the approximate location where it will eventually go in 3D.

Figure 8.57: Ear is Translated to the Correct Place ear translated

The ear model then is sculpted into the correct shape using by pulling points. The details of the ear are still somewhat rough. Additional details are created using the Edit Polygons Extrude Face tool. The faces that comprise a detail of the ear are selected.

Figure 8.58: Detail Being Added to Ear detail ear 1

The extrude process is completed by scaling down the inner polygons, then translating them in towards the inside of the head. This process creates a clean detail that requires no additional editing.

Figure 8.59: Extrude Process to Make Detail detail ear 2

The ear is ready to be joined to the rest of the head. The polygonal model is made visible, and the Polygons Combine tool is used to make the two meshes into one mesh.

Figure 8.60: Combining Polysets combine polysets

Eye Area

The basic shape of the eye has already been laid out in previous steps. The rest of the eyelid now needs to be integrated into the face and adapted to an eyeball. The edges that surround the exterior of the eye are selected so they can be extruded.

Figure 8.61: Selected Edges Around the Eye edges eye

The Edit Polygons Extrude Edge tool is used to extend the edge of the eye towards the inside of the head. Normally, in an animated model, there would be an entire eye socket associated with the eye area, but this eyelid detail will suffice for now.

Figure 8.62: Extruded Edge Around the Eye extrude edges eye

Additional detail is added to the eye area using the Edit Polygons Split Polygon tool. This will create a smoother transition between the areas of relatively light detail around the eye and the tighter rows of polygons in the eyelid area.

Figure 8.63: Additional Detail Around the Eye additional detail eye

In order to finish the eyelid area, an eyeball must be put in place. This will not be the final eyeball, but the placement and diameter of this placement object will be used to create the final eyeball. Creating a simple sphere and placing it correctly will act as a guide to see if the final subdivision is working.

Figure 8.64: Creating a Sphere sphere

Figure 8.65: Placing and Scaling the Sphere position sphere

Once the eyeball is in place, the subdivision model must be previewed in order to see the progress in this area. Using Modify Convert Polygons to Subdiv, the subdivision version of this model can be visualized.

Figure 8.66: Subdividing the Model for Testing subdiv test

After previewing the model, undo the subdivision operation. Continue cleaning up the polygonal model and testing the subdivision by turning it on and undoing it repeatedly before committing to the final version of the subdivision model.

Creating Subdivision Surfaces

Before the polygonal cage is converted to subdivision surfaces, it is important to delete all of the construction history. Subdivision modeling can sometimes get slow. Additional history that is unnecessary will make it much worse.

Once the polygon cage is completed to the point where the modeler decides to apply the subdivision surfaces, the two sides of the face should be finally merged together. With both sides selected, the Polygons Combine tool is used to make the two halves into one continuous mesh.

Figure 8.67: Combining the Two Halves of the Head 2 halves

The adjacent rows of vertecies along the center should be picked, making absolutely sure that no other vertecies are picked except the ones that run up the exact center of the model. If there were some stray vertecies that crept into the selection set, this step would have to be redone until the selection set was correct.

Figure 8.68: Selecting the Center Row of Vertecies center selected

Using the Edit Polygons Merge Vertices tool, the center line between the two halves of the model is sealed. The settings are set fairly high, so if there were some errant vertecies merged together during this operation, the operation would accidentally close one or more polygons in the model. It is important that this does not happen during this step, and the results of this operation should be carefully checked before proceeding to the next step. Any accidental problem at this stage will cause many more later on.

Figure 8.69: Using the Merge Vertecies Tool merge verts

Once the two halves are combined, it is easy to see how the initial planning stages of the modeling procedure have paid off. The discussion about edge loops and areas of animation can be visualized by looking at the wireframe. The original edge loops that were laid out before modeling started can be seen clearly.

Figure 8.70: The Merged Polygonal Wireframe merged

Figure 8.71: The Merged Model with Edge Loops merged w edge loops

The subdivision step can now take place. One reason that this step is not taken until the very last possible moment is because the subdivided model is slow to work on and time consuming to update. Using the Modify Convert Polygons to Subdiv tool, the model can now be converted to subdivision surfaces.

Figure 8.72: Subdivision Surface Menu subdiv surf menu

Figure 8.73: The Converted Model subdivided

Another, more important reason that the subdivision step is not taken until the last minute is because it is at this stage of the modeling process that asymmetric detail are added to the model. Wrinkles and character lines are usually not exactly symmetrical. When these details are added, the model needs to be in a state where it is no longer creating perfectly symmetrical details across the two halves of the model. These details are added using the subdivision modeling layers and modeling options.

Editing the Subdivision Model

Subdivision modeling is just like polygonal modeling, except that there are levels that can be traversed throughout the model, and the menus are slightly different. Also, just like polygonal modeling there are only a limited set of tools that are necessary to create a great subdivision model. The rest of the modeling tools are great to use, and should be learned. But, only a handful of the subdivision modeling tools in Maya were used to create this model.

Most of the subdivision modeling that was done on this model was surface editing using the options in the pop up menus, and the Subdiv Surfaces Partial Crease Edge/Vertex tool.

Surface Editing

When modeling in Maya subdivisions, the right mouse pop up menu that appears when clicking on the model changes. The new options that appear allow the modeler to edit the model at different levels using different entity types. With a knowledge of polygonal modeling, and an understanding of how subdivision levels work, a model can begin to have a great deal of subtle to detailed features added very quickly.

Figure 8.74: Subdivision Pop Up Menu subdiv popup

If the modeler wants to add detail at the base level, or the subdivision zero level, the Coarser option is selected on the pop up menu. Then the modeler can select vertecies, faces or edges to edit. If the modeler wants to work at a finer level, or at subdivision level 1 or above, the Finer option is selected.

The distinction between levels going from coarsest, or level zero, to finest, which would be the highest level of subdivision that the modeler would care to set, is how subdivision modeling differs a great deal from polygon modeling.

Notice in the figure below how the vertecies do not display as simple points as they would in polygon modeling mode, they display as the numeral zero. Using a coarser setting allows the modeler to effect large areas quickly. Editing in this mode will allow the modeler to make progress quickly on large areas.

Figure 8.75: Vertecies at Level Zero subdiv zero

Similarly, the vertecies at level one display as the numeral one. Using a setting like this accommodates the creation of small details. The finer the selection goes, the tighter the details will be. Editing the surface using a level one of finer setting means that the modeler is not making big changes to the model at this point, they are creating small details.

Figure 8.76: Vertecies at Level One subdiv one

Editing the edges along a surface allows can be done the same way they can be edited in polygonal modeling mode. The display between coarser and finer levels is an obvious increase in detail that is not designated by visible numerals.

Figure 8.77: Edges at Level Zero edges zero

Figure 8.78: Edges at Level One edges level one

The results of subdivision surface editing can be seen in the figures shown below. The edited figure has many asymmetrical details and many small wrinkles that the unedited version does not have. All of these small details increase the realism and make lighting and texturing easier, without dramatically increasing the computational requirements of the model. Each subdivision detail is added to a localized region. The rest of the model that is not detailed is not effected by the additional small elements created using this technique.

Figure 8.79: Unedited Subdivision Model unedited level zero

Figure 8.80: Edited Subdivision Model edited subdiv model

Using the Partial Crease Tool

This tool creates sharp, angular areas along selected edges and vertecies within the smooth subdivision model. In cases where the model needs to get details that would normally be created by adding geometry, the Subdiv Surfaces Partial Crease Edge/Vertex tool allows the modeler to add the detail without adding more geometry.

Usually it will not be necessary to go any finer than level zero to use this tool to create creases along the skin of the model. The edges along the lines that require detail are selected first. The figure below shows the edge of the eyelid highlighted, ready to use the Subdiv Surfaces Partial Crease Edge/Vertex tool.

Figure 8.81: Eyelid Line Selected for Crease Tool crease tool

The line that was smoothed out now shows a crease where the selected edges were. This line will catch a dramatic highlight when rendered, and will create a very realistic appearance.

Figure 8.82: Eyelid Highlight Line eyelid highlight

Another area that requires a crease is the line that extends from the side of the nose to the corner of the mouth. This character line usually animates quite a bit, and is not an area where excessive detail should be added. The more complex the geometry is in areas like this, the more likely it is that the blend shapes and facial animation rigs will have problems.

For these reasons, the Subdiv Surfaces Partial Crease Edge/Vertex tool is a perfect choice for creating this detail. Like the example shown previously, the edges along the area are selected before the tool is executed.

Figure 8.84: Finished Crease Along Face crease face

When the tool is executed, the line along the face is perfectly tightened. Using conventional polygon modeling methods would make this process difficult and complicated. The Subdiv Surfaces Partial Crease Edge/Vertex tool simplifies the modeling in these areas great deal.

Subdivision modeling is a powerful tool that can create stunningly lifelike models. The ability to get the model to the professional level lies in creating convincing proportion and realistic detail. Careful attention must be paid at the very beginning to insure that the model will have the correct shape, but the ability to finish the model with realistic shapes and details will make the model come to life.

Figure 8.85: Final Model final