• Welcome to Final Fantasy Hacktics. Please login or sign up.
June 01, 2020, 09:11:55 am


Don't be hasty to start your own mod; all our FFT modding projects are greatly understaffed. Find out how you can help in the Recruitment section.

Character Generator Proposal

Started by lirmont, October 11, 2013, 12:37:25 am


So, one of the things that came up in the survey (and I don't know who to credit with the idea because it's anonymous) is the lack of a character generator (re: RPGMaker has a character generator). Now, I'm not partial to the name "Character Generator", and the strictness of the RPGMaker one that comes with the free version I played with is super-restrictive. I'm also not one for making solutions to one-off problems (i.e. making one type of character over and over).

Anyway, my thought was this: what if I write something that describes the sprite formats I already have? In other words, what if I write a program that can tag certain frames (the pieces of the sprite) as arms, legs, heads, etc? And what if it could load a lot of different sprites in that same format and tag those sprites (as male, female, etc)? Then, it could use the sprites as an input, let you pick the parts you want, and output what's known in this community as a "frakensprite". Furthermore, I already have enough code related to colors from the sprite animation program that I could allow those things to be recolored pretty reliably (with very little knowledge of color theory).

This program might become part of the second award tier, followed by the sprite animation tool, followed by the control panel.


I can only vaguely begin to follow this flow chart, but my question is actually a slight aside. So lets say you wanted to add some unit variability to "generic" enemy units. You have the standard sprite, and then a couple alternate heads, a short sleeve version, and a long sleeve version. Now you could have the sprites separated into pieces in the first place and then just put the pieces together in various arrangements. This saves a lot of space and is efficient in many ways, but somewhat limits the degree of difference you can easily introduce and introduces constraints in graphical development that could be a bit of a headache. The alternative would be to just have a whole bunch of character sprites with the variations added to the full sheets. This avoids some graphical restrains, but adds a lot of manual reworking and takes up a lot of space. My main question is, which would be faster in terms of computing?
Current Projects:


I'd say it's negligible, unless you start having lots of parts on-screen at once. There are things realizing images won't do for you. For instance, you don't have the ability to rotate/resize an image in-between pixels since you're in a 2d, integer-based space. A graphics card operates in a 3d, decimal-based space, and it can do that stuff. That's probably not a concern you'd have, though.

Let's say you want to provide head-swapping via this method. What I'd suggest based on what I've seen of Mobius' goals is to follow a multi-step process for this. First, he'd just do a connected sprite for a class (re: each frame is realized right there in the image); this is the standard class sprite. Second, someone else would have to manually separate the head and the body into two different sheets. Third, you'd need a pool of other mostly empty sprites that just include heads. Now, you can just put the two sprites together.

Let's say you want to build that feature right into the sprite format (re: with the Sprite Animator). Add a named attachment where the face should be on all the composite frames. Attach whatever face you want (from an image that just contains 1 face). That simple. It even has support in the engine so you could equip expressions (like an item) that map out to face image files, and it'll change out. It's a very interesting feature.


Thank you for the thorough answer. You basically confirmed what I was thinking, even in terms of division of work load. The face expression idea is slick too. We were also thinking of having some equipment show up on characters, like a sword slung to their belt, and that would have to be a similar system anyways. Could you take this kind of thing too far though? Just thinking through it here are some potential divisions:

Simple: Head, Torso, Right Arm, Left Arm, Legs

Complex: Head, Face, Torso Front, Torso Back, Right Arm, Left Arm, Right Hand, Left hand, Legs

Then there is all the potential areas where you could be adding equipment, layering it all could be a bit complex. Though I suppose if you just start with basic sprites and then start building onto them it would become systematic after not too long.
Current Projects:


In my opinion, there's really only two scenarios you want to avoid. First, you want to avoid attaching things that are long enough to cross from the foreground into the background. Second, you want to avoid drawing the item separate from the actual pose. I'm sure some people are good enough to get the angle right, but it's easy to mess up, especially if you're using the illusion FFT's sprites use.