Edgelooping is “modeling so that all surface contours naturally flow into each other, the way they do in the human body. This ensures the skin deforms in the direction the muscles naturally flow.” So, a good model that will deform well must have the edgeloops behaving along the same surface contours of the muscles. This is the only way for the geometry to move as it should. It takes reading the whole book, plus watching Tarek’s freedom-of-teach.com DVD to really get an understanding of how to model well and is beyond the scope of this book. If you are not familiar with modeling, please take a moment (or month) and take a look at those resources. For those of you that have modeled some, let’s take a closer look at one way edgeloops can make or break your rig.
The rules of edgelooping put forth by Murdock and Allen are as follows:
1.) Avoid large quads (they will have rendering issues).
2.) Edgeloops run along crease lines (wrinkles, bends).
3.) Avoid edgeloop termination (better a quad than a pole).
4.) Wisely position weak deformation areas: hide non-quad, noncomplete edgeloops and terminating edgeloops.
5.) Enclose areas of maximum deformation : shoulders, hips, elbows, joints.
6.) Define the form outline.
7.) Three edges define a curve.
8.) Avoid long thin Quads.
9.) Final rule: Ignore these rules. “Good models are made by talented artists that know the rules, and when to break them.”
The most important points for our purposes in this book are #2 and #5. Edgeloops should run along crease lines (wrinkles and bends) and should enclose areas of maximum deformation such as the shoulders, hips, elbows, and joints. My favorite example to display these concepts is the finger. In Figure 2.17 you see three fingers with differently created edgeloops. Finger A has edgeloops that are straight up and down; they do not contour along where the bend should happen. This is a very common type of edgeloop creation for the new modeler. Note what happens when that finger is bent. The wide surface area that was created by those straight edgeloops gets pulled in and collapses into a concave rubbery fold.
Finger B has edgeloops that taper towards the area of the bend and widen on the knuckle. They enclose the area of maximum deformation. You can see that the smaller surface area creates a tighter wrinkle/bend on the finger. There is nothing else special done to these fingers. They are both deformed the same way and have the same contours, only the edgeloops are different. The difference is the surface area in the bend. I like to paint this analogy. Have you ever worn kneepads? Maybe when putting in carpet or playing volleyball? Kneepads that are just plain cylinders and have big wide bands behind the knee do not bend well and ultimately cause great pain to the wearer since they bunch up and have too much surface area to deal with, whereas knee pads that have surface area behind the knee that is smaller than that on the front of the knee will not bunch up and bend nicer at the back of the knee. Don’t believe me? Go put on some plain cylindrical knee pads and crawl around on the ground for about 20 minutes. You’ll know what bad edgeloops are and possibly have bruises to prove it. (See the “Random” link on my blog: http://coffeediem.wordpress.com/ for an image of this analogy.)
Finger C goes just a little further and adds a circle pad of edgeloops at the top for the knuckles (a favorite of mine). This gives a flat area that mimics that of the fleshy areas on knuckles. The center vertice was selected and chamfered. Look up chamfer to see how to get a quad from a single vertex. I like that little quad for elbows and patellas (knee caps).
Another example to show you edgelooping would be the torso shoulder area. This is sometimes referred to as a “cape” since the edgeloop resembles a cape wrapping around the shoulders. See Figure 2.18 , which shows a low poly creature with a nice edgeloop following along the bottom of the pectoral muscle and up to the deltoid muscle. The pec actually attaches underneath the deltoid. Often beginners create their shoulders and arms by extruding them straight out of the sides of the torsos causing square deltoids that do not flow over the pec muscles. A method learned from Tarek’s modeling is to extrude the shoulders out from the torso, then extrude the arms down from that deltoid. Lastly, rotate the extruded arm vertices up as if they had been extruded out of the side of the shoulder. This results in a nice cape line that will deform well.
Another modeling and rigging note: advanced models should have the arms modeled pointing down about 45 degrees and the shoulders rounded forward slightly to equal the natural planes of the body. (Watch “The Making of Ryan” by Chris Landreth for more explanation on this.) However, in the beginning stages we will model with flat arms in the traditional T pose.
That winds up this chapter. Make sure you take your time to get these basic concepts and really understand them. We’ll build upon them and then go into more advanced concepts. Sometimes these concepts are all you need to get by though. Wonder what we’re going to do next?
Excerpt from Rig it Right! Maya Animation Rigging Concepts by Tina O’Hailey © 2013 Taylor & Francis Group. All Rights Reserved.