Hello and welcome to this making-of feature about my latest character model. This character was part of my graduation project and hopefully I can share some of the things I learned while working on/with him. This project was a huge learning curve for me (mainly from my mistakes), and I'll try and share these mistakes so you can avoid them. But mistakes and problems are one of the best ways to push yourself, so never be afraid of tackling on something that you are not sure if you can finish. That is the best way to grow.
Firstly I would like to give credit where credit is due. Kingston Chan and Tze Wei (my course facilitators), for being my mentors and for sharing their knowledge with me. Also, Kevin Lanning and his article on "Character Modeling 2", where a great deal of my workflow was adopted from. Then not to forget great forums and sites such as 3Dtotal, CGtalk and other resources on the web. This brings us to the first point.
Never work alone. Other artists and viewers are such immense help and driving force that help you to push your work to get that extra result. If you are working at home, then try to post on art forums to get feedback or just ask your friends and family how they feel about the work.
Now let's start!
Concept is something I would highly recommend you to work on before even touching your 3D software. The more you know about your character, the easier and faster it will be when you start working on the 3D model. Whether you are good at drawing or not, always try and work on the concept first.
1. When working on the concept, I like to start very loosely. Think about what you want to create before you pick up the pen. You can write down notes or just keep them in your head, but it will help you if you have a general idea of what type of character you are going for.
As of whether you should work on paper or through digital media, it is up to you. Paper tends to give you better feel, while going digital gives you speed and more room to experiment. For this character I made a quick and loose sketch on the paper. At this point, I try to capture the feel and weight of the character.
Do not worry about details at this point. One trick is to "blur" your eyes and look at the picture, this way you'll see if the general shapes are working.
2. At this point, I scanned the image and went to Photoshop to clean it up. Now, you can re-evaluate your design and make changes. For me, it did not really feel that the idea of the sword was working. Never be afraid to change and experiment. During the clean up, always think how is the character going to work, how are the joints going to move, how are the accessories going to behave while moving and so forth. I added more detail but nothing too definite.
3. After I had the line art cleaned, I started painting. At this stage I tried to capture even more of the feel and mood of the character. As this is a personal project, I'm not worried about additional angles and/or details. The more you can capture your characters feel and personality, the better. Details can always change but the soul of the character is the real driving force in my opinion.
When doing the concept, constantly keep thinking about the character - its background, motives, emotions and so on.
After my concept was done, I moved on to research. Others might work on the research first and then concept, but I want my concept process to be as free as possible so I can let the character grow without restricting the direction by reference or research. (This obviously would be different if you are trying to capture something like WW2 soldier or something that already has a historical foundation.)
One simple way is to just use Google and find reference that you feel you might need. For this character, I picked up references from military gear and other similar equipment, as well as boots, gloves, facial structures, hair, jackets, etc. Remember that the reference is not something to copy, but something to make you understand how real life applications would work. Also at this stage, it might be a good idea to check out your "competition". For example, if you are working on medieval character, you might want to check out medieval games like the LOTR or Morrowind series. This will help you to understand how others approached the same problem and it also gives you a comparison point. Always try and make your work better than what is already out there. Whether you succeed or not is not the point, but it will help you push your design even more.
There are multiple ways to get your high-detailed mesh into normal maps and onto your low polygon object.
One way is to create a low polygon object first and export it to a software like ZBrush or Mudbox, where you can work on the details, and then bake it into your normal map.
This is a fast and straight forward way. The second option is to make a high polygon object in your 3D package and then build a low polygon one on top of it to be projected. Most of my model was constructed by first working on the highpoly with Maya, then exporting it into ZBrush for fine details, folds, cracks, etc. After that, I exported the highpoly mesh to MAX and built a lowpoly on top of the character and then projected. This is a slower option compared to the first one, and you might end up with very high density meshes in the viewport, but it also gives you more control. You can use a software like Polygon Cruncher to help you chew through and reduce the useless polygons. Another advantage with making the initial highpoly in 3D package is that it might be a lot easier if you have hard surface (non man-made) models etc.
If you have ZBrush 3, try and split up your model to parts to use as subtools. This way, you can subdivide your small detail on the mesh 1 or 2 times and other larger areas more times. If you import the whole mesh and just do an equal subdivide, you will end up having tons of polygons on areas that are not needed. For example, when I started working and detailing, the first few pouches I did were just 1 mesh with approximately 2,5million polygons.
Later on I started dividing my meshes and with the use of subtools, the final polycount for each piece of object was around 100k to 400k. This is already something you can work on the viewport more comfortably. Due to my system limitations, I had to work on piece by piece, ie. one pouch at the time, vest, gloves, pants and so on.
1. I started out with a lowpoly block out in Maya. This is just to give you the direction where you are going. It helps you to see what areas might be problematic and whether the character weight is still working in 3D. This phase is quite optional, but it is something I found useful. Try not to spend too much time at this stage.
After you are pleased with the lowpoly, start replacing the parts with highpoly ones. Do not go into too high polycounts at this time, as we'll be using ZBrush to add the rest of the details. At this point, the final details and direction of the model are setting in. I treat 3D as just another stage and change the details on how I feel works the best. You can always stop at this stage and go back and sketch something on paper if you feel like you are not sure on how the details should go.