Topology should actually be considered when building your primitive shells. It’s how the mesh flows in an optimal fashion. It’s a challenging stage and one that can eat up a lot of time. For example: when you build the bicep, how does that connect with the forearm and the shoulder, in terms of points and polygons?
A Holy Grail of mine is to try and keep the mesh in quad polygons but more often than not the necessity to insert a tri-poly connecting key areas is inevitable. I find it best to leave the joint areas of the character as quads although I have found that a well-placed tri-poly on a three-way joint can work well in certain situations. The example here shows an example of the creature’s wrist and how it was stitched together. Note how there are more points on the edge loop of the hand compared to that of the forearm. I have still managed to keep the band of the wrist as quads, merely changing the shape of the quad to suit the stitching.
Why take time doing left and right when you can concentrate on one side and let the computer aid in the other, in essence that’s mirror modelling. Although we now have the ability to mirror our modelling as we build point by point I generally work on one half and then use a mirror function at stages to check the overall form.
Once I’m happy with one half I mirror stitch to create the overall form and then continue using a mirror modelling function to work on finer details of the mesh.
Using this process we can get a perfectly symmetrical mesh whilst only doing the work to half the mesh. Asymmetry, which is one of our nuances and often (if not always) seen in characters, can be introduced as a last stage – mirror modelling allows us to get to this stage quicker.
THE 'SETUP' POSE AND CONSIDERATION FOR OTHERS
This is usually my last stage in the modelling process and quite an important one when working in production. If you are to look forward the next stage in the pipeline will be rigging so the animators can start breathing some life into your character.
Get talking to the rigger, find out his/her needs. This can be as simple as splitting your geometry into particular layers or posing the character in a certain way that suits a skeletal structure or adding another edge loop to aid in a better deformation.
Communicating with the rigger will ascertain what’s required for the next stage and more importantly for the modeller will eliminate having to do this further down the line once you’ve moved onto something else.
Truth is I find this quite a tedious stage, but one that dictates care and attention.
There can be a lot dependent on your UV map and it’s best to think ahead for its uses further down the line.
For example if we know the modelling process will be furthered by sculpting this will dictate no overlapping UVs. If I know this isn’t a requirement and know the texturing can be mirrored I may opt to overlay UVs to cut down the painting process.
My preference would always be to try and layout the UVs with no overlaps. However if I feel a UV map can over lay itself to cut down my time further in the process I will opt for this route (see the example given here). I try not to do things for the sake of it. Whatever approach you take I find the biggest consideration is picturing your three dimensional shape in a two dimensional form. I usually think about where the most inconspicuous places I would want my texture seams. A crude example of this is if I were laying out UVs for a foot I would opt for the seam to be on the bottom of the foot, the area that’s usually less in view.
Another aspect I like to check about my UV map is undesirable stretching across the topology. Generally the tools available today unwrap and lay with a minimum of stretching but I find the more you may have to manipulate the UV points by hand that stretching can occur. The best approach I find for checking whether this is happening is to actually apply a texture to the geometry in the form of a numeric grid to indicate portions of the mesh in isolation as it relates to the geometry.
TEXTURING AND SHADING
Texturing and shading are also areas that will dictate your approach dependant on the job at hand. An example is if I were looking to create a photo real model I’ll try and collect as much photographic material as possible. If the project is not photorealistic I’ll perhaps see if there’s any thing of use in the concept material even if it’s just for a template.
In the case of the creature I’ve used for the article I pretty much took the concepts and constructed the bulk of my texture. This allowed me to get the exact line art as per the concept without spending too much time.
3D mesh painting software is also very useful in certain situations. Besides many advantages I particularly like there ability to allow you to place paint exactly where you need it, eliminating the process of going between 3D and paint packages.
As well as your texturing you also need to consider the shading properties of your surface and how it should ‘feel’ to the viewer.
Consider what components make up the surface. Is it a hard surface with high gloss highlights or a soft surface with low gloss sheen? Is it transparent? Is it reflective? By analysing the surface to this degree allows you to breakdown what you actually need to do.
ADVANCEMENTS IN TOOLS
In time some of the steps above will become easier as tools available to the artist advance. Advancements create new challenges as the old ones become easier, with some enhancing your work and allowing you to do operations that were unachievable in the past.
Sculpting tools are one such advancement, giving the modeller the ability to add that extra level of detail which is sometimes a necessity rather than a luxury.
Regardless of the advancements, the best advice I can give is ‘Don’t be afraid of them, embrace them!’