Herororobot Metal Justice
This is my attempt to make an old-school Golden Axe-style brawler. There is only one level lasting around 5 minutes. Had this posted up on the NAG forums, but thought it might be useful to someone to have a slightly more technically-oriented discussion here.
I'm not a great programmer, but do enjoy messing around with code and finding solutions to problems- that's where this project comes from. I went into this blind and clueless, with no idea of AI or animation, and learnt & puzzled things out as it went.
Trust me when I say that if I can do this, anyone bloody could.
It started with a pencil sketch of the main hero robot. That was scanned and redrawn in Paint.NET in a huuuge size, then scaled down to a tiny fraction and finally displayed at 2xscale for the standard pixel crunchy effect. Everything you see onscreen was done like that, using tracing paper for the crude animations.
Once a few basic art assets were in place, I set about making the framework to control animations and states. Starting by making the main guy able to throw a jab, cross, and an uppercut as a combo, a clumsy system started falling into place and ended up holding together quite robustly - a crude finite state machine!
Simply put, every actor can be in one of several states; IDLE, JUMPING, JUMPINGKICK, SHOOTING, PUNCHED01, and so on. Not sure if this was the best way for it, but it worked a charm since each state had a get-method and a set-method; this allowed me to specify rules when to allow certain states, like being PUNCHED01 only while you're not DEAD, for example.
The state machine also made animation easy, I'd just see which state a thing was in, and draw the associated image/frame.
Most importantly, the state machine made simple AI possible. NPCs try and flank you based on certain rules, or just go at you if they're alone. Bottweilers would randomly stop and bark and then maneuver into a favourable leaping attack position.
Really, once the state machine was set up, the game seemed to almost build itself.
As a full-time office worker with a life during the odd 5 hours after hours, time to work on this sort of thing is rare. Ultimately this project ran over the course of 2 years, sometimes with months of inactivity between occasional periods of crunch.
I set the goal to be a playable 1st level of a thing that feels complete in spite of its short length, added some spit & polish and set her free in the wild.
This probably won't get worked on again, but it was a massive learning experience.
Main lessons:
For solo hobbyists like me, steer clear of fighting/brawling games. They're incredibly art-intensive. While it was possible for me to draw all my own resources, animating any humanoid figure takes a heck of a lot of time and effort - and that's why my next pet project features tanks, planes, and spaceships instead.
I'm not a great programmer, but do enjoy messing around with code and finding solutions to problems- that's where this project comes from. I went into this blind and clueless, with no idea of AI or animation, and learnt & puzzled things out as it went.
Trust me when I say that if I can do this, anyone bloody could.
It started with a pencil sketch of the main hero robot. That was scanned and redrawn in Paint.NET in a huuuge size, then scaled down to a tiny fraction and finally displayed at 2xscale for the standard pixel crunchy effect. Everything you see onscreen was done like that, using tracing paper for the crude animations.
Once a few basic art assets were in place, I set about making the framework to control animations and states. Starting by making the main guy able to throw a jab, cross, and an uppercut as a combo, a clumsy system started falling into place and ended up holding together quite robustly - a crude finite state machine!
Simply put, every actor can be in one of several states; IDLE, JUMPING, JUMPINGKICK, SHOOTING, PUNCHED01, and so on. Not sure if this was the best way for it, but it worked a charm since each state had a get-method and a set-method; this allowed me to specify rules when to allow certain states, like being PUNCHED01 only while you're not DEAD, for example.
The state machine also made animation easy, I'd just see which state a thing was in, and draw the associated image/frame.
Most importantly, the state machine made simple AI possible. NPCs try and flank you based on certain rules, or just go at you if they're alone. Bottweilers would randomly stop and bark and then maneuver into a favourable leaping attack position.
Really, once the state machine was set up, the game seemed to almost build itself.
As a full-time office worker with a life during the odd 5 hours after hours, time to work on this sort of thing is rare. Ultimately this project ran over the course of 2 years, sometimes with months of inactivity between occasional periods of crunch.
I set the goal to be a playable 1st level of a thing that feels complete in spite of its short length, added some spit & polish and set her free in the wild.
This probably won't get worked on again, but it was a massive learning experience.
Main lessons:
For solo hobbyists like me, steer clear of fighting/brawling games. They're incredibly art-intensive. While it was possible for me to draw all my own resources, animating any humanoid figure takes a heck of a lot of time and effort - and that's why my next pet project features tanks, planes, and spaceships instead.
Comments
I reached a point where I had an optimal strategy of run punching the sides, returning to hit for some hadouken spam then when enemies got near a run punch to get rid of them.
The blocking felt a bit off to me, but then it normally does to me, blocking in DOA still makes me wonder if it is working half the time.
But I think I spent a good 20 minutes trying to get through on one robot, which was great fun :)
Congrats on pushing through, working on this stuff after work hours is taxing, but It's totally worth it :) Keep on keepin' on.