Christopher Balogh - Blog Entry 2
So for my project I want to create a character. But how do I go about creating characters?
Well, first off I get an idea and then I start up GIMP. I create a rough sketch of my characters outline in a Dark Grey, Mid-Tone Grey and Light Grey. I usually use a rectangular brush with slightly rough and textured edges that turns and follows the direction of the stroke. Once that is done I get a small, hard-edged brush and paint all the details.
For my first drawing I mostly use a three-point perspective quarter view. Once I have a clean line art I erase the quick sketch and put down some base color washes. I then use addition and multiply brushes to add shading. Once that is done my concept piece is complete.
I then go on to create a model sheet. For model sheets I always use a fine brush and start with a front orthographic view. I prefer to model characters in the Y and not the T pose as I find it is easier to model the arms and shoulder blades. I then create guides at all the important places so that the side view can line up exactly with the front view.
After painting the side view, I usually duplicate the front view and erase some details and redraw some other things as it can not only speed up the process but make it match better.
Once the modelling or character sheet is completed I set up the image as a background for my front, back and side views. I use box modelling for the body and then model the head seperately face by face for extra good topology. I prefer to use Blenders hair physics system instead of texturing hair as it can be interactive and I can easily comb a different hairstyle, I do use textures for the eyelashes and eyebrows though.
Once the character is complete I start to model the clothes, I either box model the clothing, duplicate the body geomotry and edit or create a rough shape and use a shrinkwrap modifier with interactive cage edit and deform. I then use the cloth simulator and either have the cloth simulator functioning as the character moves or apply it and have it later affected by the armature. If I need to model any type of props I use primitives and then box model.
I normally focus on getting the shape and look down, even if it means getting a non-manifold mesh, and then tidying. I find it helps to just get my idea across quickly before my mind wanders aw..... Huh? Oh, sorry.
Once I have my final topology I texture the mesh. If I need to create a texture just for this object I will normally un-wrap it and bake off an AO map and paint over that. If it is a generic material, a plain stone wall, a tiled floor, dirt on the ground etc I usually use generated texture positioning.
After the texturing I do a quick lighting test to see how the textures respond to different HDRI environments. If one of the textures doesn't work I go back and do it again. After that comes the rigging.
To be blunt. I hate rigging. Which is why I decided to do a character design.
To force myself to learn it enough to properly rig a character. So skipping rigging. After that comes real lighting.
For this project I have decided to use the Cycles Render Engine, and I haven't really used it for lighting much before. But the basic fundamentals still apply. Then, as a stretch goal, comes animation.
Blunt - I hate animating. I actually know HOW to animate, use the dopesheet and f-curves and all. I just don't like it. Compositing is one of the final stages in any of my images.
I generally grab all the passes I can, you never know when you might need that Material Index or other obscure ones. One of thew things I often do in compositing is create a long string of nodesl, get a good result. Render at a full scale resolution, and only then realize I forgot to connect the Composite Node....
See Christopher's Artist Bio here.