Warning: This is going to be a long, rambling post because I want to sort through some ideas and get some feedback before writing a short, concise post later.It's time to get serious about the graphics in GH2. This is a big issue- or, rather, a tangled collection of big issues all knotted together. I want the game to look as good as it possibly can, and to do this I'm definitely going to need outside help. Before I can get that help, though, I need to know what I'm asking for.
One of the issues that needs to be dealt with is graphics content. Currently the game uses a mix of 3D backgrounds with 2D sprites. If the 2D sprites were removed and replaced with 3D meshes, things would probably look much better. The down side is that then I'd need to get a whole lot of meshes (which would probably cost money*). If the 2D sprites could be made to look decent, it'd be so much easier to build up lots of content, but as mentioned there's probably no way to make this look as good as a fully 3D world.
The current sprites are 64x64 pixels and contain eight rotations of the model at an oblique angle. They don't work particularly well- the format was chosen just because that's the format used in GH1.
Today I decided to try an experiment: paper dolls. (This took about 20 minutes to complete, just to let you know.)
![\"\"](\"http://i137.photobucket.com/albums/q211/pyrrho12/GearHead/paper_model_test.jpg\")
Instead of a 64x64 sprite with multiple directions, this is a 128x128 sprite that's always viewed from the front. Direction (and later on, possibly speed as well) is indicated by the octagon at the model's base. Seeing everything from the front was a bit odd at first but I got used to it quickly. The increased sprite resolution looks good; the jaggies of the 64x64 sprites are mostly gone. It was harder to determine an enemy's facing from a distance, but this is a problem that can probably be fixed.
Why perform this experiment? Just to see how it looks, mostly. Completely flat paper dolls like this represent one extreme in the spectrum of possibilities. They'd definitely be the easiest sort of graphics to add to the game. On the minus side, they're probably the most unnatural representational style that I could use without getting into abstract expressionism. Back on the plus side, they'd make it fairly easy to add a paper doll equipment system for the main character.
The important thing about the interface is to convey all the necessary information to the player, and to do it in a reasonably attractive manner. There are all kinds of trade-offs involved in deciding how it's going to work.
How many graphical elements does GearHead need? All the terrain, scenery, buildings and props need to be modeled. There must be at minimum 9 different mecha models (one for each form), though ideally each mecha should have its own model (though similar designs such as the BuruBuru and Daum might be able to get away with having different skins on the same mesh). There should be at least 28 character images (14 job categories, male and female of each) and I don't know how many monsters.
*I'm not against spending money on GearHead, but if I were to do this I'd want to earn back at least some of the investment, and monetizing the game sounds like a giant pain in the arse. As much as I'd love to see GearHead with higher production values my main reason for working on this project is that it's the opposite of a pain in the arse... let's just drop that metaphor here.
