Modelling techniques – limitations of different surface types

Describe different 3D modelling techniques used in digital art… 800 words… reference and cite at least 4 scholarly texts to support your findings…. Elements to discuss may include common toolsets, common terminology, modelling workflow options. limitations of different surface types or typology requirements for deformation.

Modelling requires a balance of quality and speed. The more complex the model the longer it will take to build, the more polygons it has the longer it will take to render (Danaher, the Complete Guide to Digital 3D Design, 2004)  Movie models can be generated with millions of polygons using several different techniques at once. Gaming models have to be more efficient. Lower poly assets will render more smoothly and give better frame to frame renders during game play.   (Dargie)

Maya uses 3 types of modelling polygons, NURBs and subdivision surfaces. (Derakhshani, 2011)

NURBS is short for “Non Uniform Rational Bézier Spline”. (Linden Research, Inc., 24) Unlike polygons NURBS are resolution independent, they are controlled by fewer points called control points.  (Danaher, the Complete Guide to Digital 3D Design, 2004)

The splines (for drawing curves) within NURBS modelling can typically be controlled easily and manipulated, and their structures are influenced by control points, drawing the curves in particular directions. One of the primary advantages to using NURBS is that smooth curves can be achieved (Smith)

They are particularly suited for modelling in 3D, because they provide excellent continuity with a minimum number of control points (Autodesk Maya, 2012)

You can create a NURBS model only 1 way with a grid of NURBS curves, like moulding a wire mesh. You can’t extend this rigid topology surface like a polygon one, but you can add extra geometry, blend between 2 surfaces, adding a new surface between with both ends matching perfectly with the originals. You can project curves onto surfaces to deform and build off them. However you can’t cut holes in them because they are so rigid, only the tessellated version of them can appear to have a hole. (Danaher, the Complete Guide to Digital 3D Design, 2004)

Polygonal Modelling is a versatile technique. Polygons can be either simple shapes, such as polygonal primitives, or complex models built from the various polygonal tools. A polygonal object can be closed, open, or made up of shells, which are disjointed pieces of geometry. Also known as mesh. (Autodesk Maya, 2012)

The Polygon Mesh is the most common modelling method used in 3D computer graphics. The process involves the direct manipulation of polygons- triangles and quadrilaterals (the one to avoid if possible is the ngon (a polygon with more than four sides) – these can cause problems in some programs and lead to issues if they’re subdivided). (tips-and-tricks-for-organic-modelling, 2012). The process also involves the manipulation of faces, vertices and edges to produce the desired shape. (Linden Research, Inc., 24)

The modelling process involves the combining, dividing and extrusion of these polygons to create a structural model. An advantage to using polygonal techniques is that computer programs are generally able to render them effectively and fast. A disadvantage to polygonal modelling is that curves cannot be accurately represented. (Smith)

While the arbitrary nature of polygonal surfaces provides great freedom and flexibility it can lead to invalid or non-manifold geometry in creating surface topology.

A non-planar face (where at least 1 of the vertices is on a different plane to the others) may not render correctly or transport well to a game engine. (Danaher, digital 3D Design- the use of 3D applications in digital graphic design, 2001)

Manifold geometry is faces that can be cut and folded, non-manifold can’t. The 3 types of non-manifold are: T-shaped- 3 faces share a common edge, Bowtie- 2 faces share a vertex but not an edge, Reversed normals-2 faces sharing an edge have opposing face normals (Danaher, digital 3D Design- the use of 3D applications in digital graphic design, 2001)phone 067a

Lamina faces share all the same vertices and edges, they are essentially laminated together but are incorrect geometry.
Mesh – cleanup– a tool set option, is an excellent tool for dealing with non-planar, non-manifold geometry, lamina faces and other unwanted polygon issues. (Danaher, digital 3D Design- the use of 3D applications in digital graphic design, 2001)

Subdivision Modelling involves techniques that are applied to an already existing model and has some of the features of both polygonal and NURBS modelling.

The outer cage acts like a poly object so all the poly tools can be used on it but it needs only a fraction of the polygons to create a very high resolution, smooth surface with lies under the cage. SubdDs allow you to subdivide a mesh to add detail only where you need. (Keller, 2010)

If you start with a NURBSs or subdivision surface in the end it is preferable to convert everything to polygons. (Derakhshani, 2011) At render time all geometry of every type is converted to poly triangles. Tessellation refers to the number and placement of the triangles on the surface when the scene is rendered. Objects with low tessellation will appear blocky but will have a shorter render time than hi tessellation objects.  (Keller, 2010)

Generally speaking it’s best to model in quads. Quads (quadrangles – four-sided polygons) are predictable and subdivide in a clean way, leaving less chance for bad artefacts at render time. (tips-and-tricks-for-organic-modelling, 2012)


Autodesk Maya. (2012). Autodesk Maya help. Retrieved 12 3, 2012, from Autodesk Maya:

Danaher, S. (2001). digital 3D Design- the use of 3D applications in digital graphic design. London: Cassel and Co.

Danaher, S. (2004). the Complete Guide to Digital 3D Design. Cambridge: the Bex Press Limited.

Dargie, J. (n.d.). Modeling Techniques: Movies vs. Games. Retrieved 12 3, 2012, from ACMSiggraph:

Dargie, J. (n.d.). Modelling Techniques Movies vs Games. Retrieved Dec 3rd, 2012, from ACMSiggraph:

Derakhshani, D. (2011). Introducing Autodesk Maya 2012. Indianapolis: Wiley Publishing.

Keller, E. (2010). Matering Autodesk Maya 2011. Indianapolis: Wiley Publishing Inc.

Linden Research, Inc. (24, Jan 2009). Sculpted Prims: 3d Modeling Glossary. Retrieved 12 3, 2012, from Second Life Wiki:

Smith, S. (n.d.). 3-D Modeling Techniques. Retrieved 12 3, 2012, from eHow tech:

tips-and-tricks-for-organic-modelling. (2012, 07 19). Retrieved 08 18, 2012, from


Critical reflection

Critical Reflection

I like modelling now. In the beginning it was very hard with many things to learn using Maya for the first time.

My experience is that the design kept changing and improving and this was much better on paper before I started modelling. ‘It is important to conceptualise your ideas on paper first’. (Eric Miller, 2006).

Modelling a maquette was the next step it gets the shape into your head. You can photograph it and use it as the reference and even draw the muscle flow on top of it. (Eric Miller, 2006)

The first step in the modelling is planning the necessary edge rings for it to deform and animate as expected. (Eric Miller, 2006) Placing edge loops along and around muscles and fat bulges is best done by following anatomical references (Eric Miller, 2006) I have some anatomical edge flow on Abby’s torso and need to do the limbs however as she is a little girl she will not have defined muscles,  but will still need the flow to be correct for when the model deforms.

The 2nd key modelling concept is uniformly spacing the geometry, this aids texturing too so it will not stretch unevenly (Eric Miller, 2006) this was evident on my models.

The third key is to maintain quads and eliminate triangles and n-sided polygons or n-gons (Eric Miller, 2006). This was hard for me when I was trying to do edge flow at the same time. But when I realised you do the flow first and fix that up later I got the hang of it.

Box modelling, starting with a poly cube and using image planes with reference images at the back and side was fun and straight forward. I made the images the same size  and lined them up by making sure the top and bottom are equal, lining up places I was modelling if the drawings weren’t perfect. Using the front view then side view with half of the model making the outside edges the same as the drawing is fun.

You can easily create complex models by adjusting the simpler poly objects with the poly editing tools. (Derakhshani, 2011) The model developed by inserting and moving edges, moving and merging vertices and extruding.

I became familiar with the menu options using the shift key and control key with edges, vertices or faces selected eg poke face, bevel, interactive split tool, insert edge loop too, edge ring and split, merge vertices etc. The poke tool is useful for creating detailed sections of the mesh (poly surface)and inserting an eye. The bevel tool rounds sharp corners (Derakhshani, 2011) and I used it on a cube for Abby’s head to make it a sphere without triangles. I used the cut faces tool after selecting1 face when I could not get an edge to go in, I liked the deform tool lattice for changing the shape of the whole limb.

Then I rigged it to test the geometry, made a skeleton and bound it to the skin, made controls and IK handles. I constrained  joints and IK handles to the controls with point, aim, orient, parent and pole vectors. I found it quite difficult to work out the best shape the joint capsules should be formed into because sometimes a vertex would jump to the floor level and pull the geometry with it making long spikes, then the geo would deform wierdly when making allowances for that. I am bad at that. I found when the models moved a limb or head a part of the torso would deform unnaturally. I need to learn this better to know whether flow has something to do with it

Overall I learned a lot by doing it a lot and trying new things.


Derakhshani, D. (2011). Introducing Autodesk Maya 2012. Indianapolis: Wiley Publishing inc.

Eric Miller, P. T. (2006). Maya Techniques – Hyper-Realistc Creature Creation. Toronto: Alias Systems Corp.

Organic Character Modelling Project stage 3: Refining model and testing geometry

I saw I needed to improve the flow and even out the size of the quads so the next step was to delete half of the body again and move vertices and edge loops so that when it bends it will deform naturally along muscle lines and places where the body creases like joints. I wasn’t sure whether to add 1 or 2 extra edges at joint places like top and bottom of the neck and the top of the arm, around the ankle and knees. It seemed to added creases so I took some out and moved the remaining ones apart and used it for adding to the smooth transition of the connection, say of the head to the neck. I extracted the head from the body because I saw that the number of vertices at the top of the neck and base of the head didn’t match.
I enjoy making a human body, knowing anatomy from my previous studies helps me know where muscles and fat  bulges etc are and the shape of them. I found it quite difficult at first moving edge loops and vertices because I was always trying to make sure there were no triangles or polygons (n-gons) at the same time. I found it easier when I realised you take care of that later. Using the polygon selection constraint, in the polygons menu set under select and select using constraints menu, shows triangles and n-gons so it is easier to see them and fix them.
Poor Little Abby’s limbs got all twisted up when I constrained the wrist or ankle to the PV controller, the mistake was that I centred the pivot at the wrist/ankle not the controller
 TnA18  TnA19
TnA12 TnA10 TnA11 teddy TnA17
On further model refining I realised Abby needed more fat at the inside of her upper thighs and under the upper arm and a smaller chin area. Teddy needed longer arms for when he rubs his nose in the animation. I have seen that when a model looks like the right dimensions in the angular view, number 1, when smoothed, number 3, is smaller and looks too skinny, so I have to widen parts that shrink too much on smoothing.
TnA13  TnA14  TnA15  TnA16
Testing the model by inserting a skeleton and binding it to the skin then moving it saw terrible squeezing and stretching in places it shouldn’t have. Experimentation fixed some of these problems but often a vertex from, say, the hair or neck would jump to the floor and stick there when I adjusted joint capsules in, say, the hip or neck, resulting in long pointy bits of the model dragged down to the floor. Better rigging techniques will help this and more refined models with edge loops and edge rings arranged correctly along and around the body’s bending and flat areas will help it deform more normally.

I have 2 models a human and a teddy. The human, a little girl called Abby, now just lies still in bed while Teddy does all the moving.

I need to work out texturing requirements on flow and to test these on the model, evenly spaced edges makes  texturing work better.

I may need her to move in the future I would need to improve surface flow so that when she moves it will deform well. It is always good to use a reference to see where they need to go.

The teddy is not human but a stitched fabric toy. In real life these can’t move because they have no muscles or skeleton etc. So I need to see how these 2 details can work together and weigh up what is right and what looks right. I will have to look at references like Ted in Toy story 3 and Woody to see how they move. Ted is a teddy and Woody has flat stitched joints like my teddy. I will then probably line my edges along the stitched joints and around bulging areas. I will have to see how stuffed fabric deforms and align edges accordingly.

I will see which poses and motions need to be tested.  I can then film someone doing these movements and test these on the models, especially extreme movements, to see how it deforms. At the moment my rigging ability is limited and may have something to do with deforming problems.

Organic Character Modelling Project STAGE 2: Preliminary modelling

I started with my character drawings. Using side and front views I modelled 1 half of the body by adding edges and dragging vertices, and extruding faces to match the drawings. This part is easy matching the box model to the schematics and reference sources. I later made changes it as I saw the need.
My first attempt was terrible and looked like a burns victim because I had tried making the facial features before we were taught how to do it and I had used a sphere for the head which is bad because it has lots of triangles at the top and bottom.
So I started again with a  cube and bevelled the edges for the head and didn’t try to do any features yet, just the general shape of the head.
I started with a cube for the body too. I didn’t have too much trouble doing this first step of matching the model to the drawing..
Extruding the leg down and the arm out the side and then shaping them was no problem either.
When I was happy with it I duplicated it and scaled it on the x axis to -1, combined both halves and merged vertices, same with the head onto the body, no trouble.

Box modelling Teddy…  using the image planes as references.. the hard bit is arranging flow  to be more anatomically correct so when the model deforms it will bend in natural places… just trying again and again to see if I can make it work so that there are minimum poles, where more than 4 edges meet at a vertex, and that they are where I might need some pinching, and no faces with 3 or more  than 5 edges. I have done 2 Teddies so far amd may do another 1 because I am not too happy with the shapes I have.


I started Teddy 3 times. I changed the joints of the limbs to be flat like it is stitched.

I had to straighten arm and legs but they can be rebent to a normal pose after the skeleton is in.

I liked it when I used a lattice to change the shape of the arm and leg, it looked better than by moving vertices 1 by 1.

The original face was terrible and the eyeballs were awful. I found it looked best when I extruded the snout out from the face


Organic Character Modelling Project STAGE 1: Character design

I decided to base my character’s face on this model from America’s top model all stars, her huge pale blue eyes, small nose and small red mouth, blonde hair and pale skin. I then decided to make her a little girl so I am basing her head shape and maybe hair on my little niece called Abby. I drew a Teddy from my imagination but got photos of a real 1 to see the way the limbs move and bend because it is different to a live being

 tumblr_lwed1s0vRP1qdzenwiphone 803teddy4


iphone 799iphone 896iphone 796mmmm paddle pop!!!


My first rough sketches were very different to where I ended up

teddy sketchescharacter 002ccharacter 002dcharacter 003bcharacter 012character 013character 014character 015character 001a

You can see the changes in my character design

Here are my final drawings for my characters. Abby will have pyjamas on rather than a dress for easier animating.



Modelling a hand

This slideshow requires JavaScript.

modelling a hand in Maya

tutorial on Youtube

by Ryan Kittleson

Ryan Kittleson

Ryan Kittleson

‘Ryan Kittleson has loved combining art with technology since he was a young child with a VIC-20. He grew up with a passion for the visually rich images by artists like Dali and Escher. He taught himself Photoshop as a teen and then 3ds Max in his early 20s. He worked as a game artist in Raleigh, North Carolina for a while before returning to school and completing a Bachelor’s of Fine Arts degree in Illustration from East Carolina University; all the while practicing the art of digital graphics with personal and freelance projects. He currently lives and works in Orlando, Florida where he teaches digital character modeling at Full Sail University.’ quoted from