next »

[Carl Maxwell]

From reading an article on Gamasutra by Rami Ismail, I've become convinced that making a game a week is a good exercise for practicing game development, so, I'm taking up the challenge. I'll update this post with the games as they come.

To be clear, the games will be more like prototypes or tech demos than real full games. It's for practice, learning, testing out concepts.

Feedback welcome!

The Rules

• I'll the post games each saturday for as long as I continue doing this.
• I will not develop the same game for two weeks in a row, though I might return to a game for a later week, to add additional layers of polish, or resolve a problem that I figured out the solution to.
• I'll try to develop a game each week, some weeks I might end up not having an actual game to show at the end, in that case I'll just describe what I tried to do with the game, what I got done before the deadline, and then move on to the next game. In future weeks I may return to failed attempts, after I figure out what went wrong and how to do them.

One of the big things I'd like to experiment with is using different "tones", distinct 'game feel', different forms of 'polish' elements like camera shake, particles, &c. I'd like to focus on games that use a first person perspective, but I may deviate from time to time.

The Games

• Lane Assault - first person shooty shooty game.
• Barun Dun Dunt - first person shooty stealthy-ish-kindof-almost game.

update:

I've gone on a hiatus from this project for the time being. I'm going to work on a longer term project, though I think this game a week thing definitely has merit. Even though I didn't do it for very long I definitely learned some stuff. Hope to find time to do it some more eventually.

[melos]

Good luck! that sounds like a good idea for figuring out waht you like...it's definitely harder if you're steeped in longer projects. But if something appeals to you it's definitely worth exploring at length further than a week!

[Carl Maxwell]

Lane Assault is a first person shooty shooty game where you dodge giant slow-moving bullets.

The controls are mouse + wasd + left click to shoot.

The game doesn't do much of anything to explain the rules to you, so:
you have infinite ammo,
you can take three hits,
you heal to full health when you reach a new level.

It's got a tutorial and two levels, takes about 5 minutes to play through... If you never die.
I don't think the difficulty is balanced very well though, so, good luck with that.

Download

I developed it in Unreal 4. Feedback would be appreciated, particularly surrounding the interface and difficulty.

It takes influence from the bullet hell genre, Nuclear Throne, and Hets. Originally I wanted it to have procedurally generated levels, but I quickly decided to just make a few static levels as a sortof proof of concept. If I hypothetically developed the game more I would probably prefer having procedurally generated levels, more variation among the enemies, upgrades between levels, less linear levels, etc.

I had a lot of problems with Unreal 4's lightmass system and with UV unwrapping, I learned a bit about those things but never really got them to behave the way I wanted to. Mostly what I learned was just getting in practice with my workflow for modeling things, which is like sketchup for basic modeling -> then remaking the mesh in blender and doing the UV unwrapping on that -> importing into Unreal 4, position &c. I also got Allegorithmic Substance and learned some of the most basic things that can be done in it.

Good luck! that sounds like a good idea for figuring out what you like...

Thanks!

[Carl Maxwell]

So far this week I've spent most of my time playing with Allegorithmic Substance, I'm starting to get reasonable at it, so I'm hoping to have all bespoke textures for this game. Right now roughly half of the textures in the game are custom, the others are ones that come with Unreal 4.

I'm calling this game Precursor, and am going with a more painting-inspired visual style for it. It's proving much more taxing than the low-resolution visual aesthetic of Lane Assault. The levels are being procedurally generated, I've only got three buildings in right now though, aaand it's looking like that'll be all for the first version. All the little nitpicky details of buildings takes a lot of time, fixing little mistakes. I realized today I have to go through each building and scale the "sidewalk" around them, they're twice as tall as they should be.

[Carl Maxwell]

Barun Dun Dunt is a first person shooty stealthy-ish-almost game where you need to steal three golden statuette thingies.

The controls are spacebar to jump + mouse look + wasd + left click to shoot.

The main mechanic of the game is greenflame.
You start with four greenflame, if you get hit you expend it like hit points, you also expend it when you shoot.

You can recharge your greenflame at the greenflame lampposts spread throughout the level.

Also, enemies can only be damaged if you hit them from behind.

It's only got one level, it's really short.

Itch.io page

I developed it in Unreal 4. Feedback would be appreciated.

Feelings about the game

I don't feel like this game really came together well at all. It's missing sound effects, camera shake, all sorts of juice effects are missing. There are also performance issues, and a generally unfinished feeling to the game. It's a prime candidate for returning to later.

But I'm ok with that, it's only the second game a week I've done, the first few of these were bound to be bad.

It takes influence from stealth type-ish games like Dishonored and Assassin's Creed. It has like a half-heartedly procedurally generated level, where the buildings in the level are randomized between three prefabricated buildings, and their position & rotation are shifted about a bit.

I wanted to expand on the Greenflame mechanics, with stuff along the lines of:

• when a watchman discovers an unlit lamppost he'll go to a lit lamp post and take some of its flame to the unlit lamp to relight it.
• watchman having only one shot, recharging at lamp posts.
• etc

But I ran out of time.

Stuff learned

I learned a lot about using Allegorithmic Substance Designer, and about UV unwrapping in Blender. I now consider myself reasonably proficient at using Substance, and not really proficient at UV Unwrapping. I should probably watch some tutorials on the latter subject.

The reason why this game came out so badly was that I was having a lot of fun working on the buildings in sketchup and the textures in substance, so I didn't end up working on the gameplay much until the last couple days, I implemented most of the mechanics on friday, and some of them on thursday. Most of the rest of the time I spent tweaking textures, adding noise to textures, fixing fiddly UV unwrapping issues, beveling edges on various meshes, trying to figure out how to properly UV unwrap beveled edges, playing around with the lighting (which is really bad right now, it was actually better a few days ago I swear, shoulda stopped messing with it, couldn't get it the way I wanted.)

It's funny, cause really the overall art of the game actually suffered from my focusing in on particular pieces of art, like the watchman have a cube for a shield and a cube for a gun because I didn't have any time left to model proper shapes for those items. They look really silly.

It was good though, I learned a lot, and I like this 'painted' style I was going for. Really hope I can get better gameplay in the next game though. I'm thinking I might put a tight limit on how much art I can do by using an artstyle that has only four colors or a limited polygon budget or something, haven't decided.

I changed the game's name from Precursor cause there's a company called Precursor Games, and I wasn't really liking the direction the lore surrounding the "precursor" storystuff was going.

[Carl Maxwell]

I've pretty much dropped the whole 'gameplay focused' thing I was thinking of, I'm thinking that for this week and possibly next week I'm gonna accept weaker gameplay, so that I can focus on getting a good workflow/pipeline setup for creating the necessary art assets.

Master Material

So, something I just started doing with this project, is having a 'master material' which I then instantiate for all the stuff in the game. Originally I was thinking that I would only have some parameters on it, color, graininess, color variation, etc, and then just have some simple 'paint strokes' textures for it. The master material just lerps between these textures according to the parameters.

'Paint Strokes' texture

That worked well enough, but then I figured out how to do texture painting in Blender and import it into unreal. That was a significant step forward for the 'handpainted' style I've been trying for... Especially since now it actually is handpainted.

<- before | after ->

There's a lot of value to this, in one way it's convenient because UV mapping isn't as important now, since the painted texture will fit to the UV coordinates, not the other way around. It's helpful with corners, letting color bleed around the edges of objects, onto other objects:

That's actually two separate objects, without texture painting I don't know how of a good way to get that wear pattern to show up there. One problem that remains is that I don't actually know how to paint, so, that may reduce the utility of texture painting.

Meanwhile, I'm still working on coming up with a good process for making this stuff. Right now it seems like the best thing is to make either the whole level in blender and paint it there altogether, or possibly just make "scenes" in different .blend files, and try to connect them together someway.

I've done the whole 'make a whole scene in blender then import it into unreal' process once, and, it was not fun.

That scene is composed of nine objects. Importing the mesh into Unreal is fine, I just export the .fbx file and then import that into unreal, but then it lists the objects as "Element 0", "Element 1", etc. So then I have to go through and pick each object and put its material object. I also have to go through and create 9 instances of the master material, export each of the painted textures from blender (which has to be done one texture at a time), import those, and attach them to the material instances. This is a lot of wasted effort, work that should be automated. There is an option for importing materials and textures with the Unreal importer thingy, I tried it but it didn work the way I expected so I didn't use it this time. I need to test that out further, I need some better way of doing this export/import process. Wish Unreal kept the names from Blender.

Also still need to learn how to properly UV Unwrap things still.

Something else I'd like to do is figure out a workflow for characters/animations. I'm currently thinking the next game a week (after this one) will be a 'walking simulator' style of game, to let me focus more on the artworky stuff.

[Carl Maxwell]

This week I did not get the game together enough to post it. First failed week.

First Sword

I really didn't like the name of this game. Should've gone with something like Elder Blade or Progenitor Blade or Kingblade or something. Anyway.

The purpose of this game was to give me a chance to try out melee combat, which is something that is quite difficult to do well in first person games. Unfortunately I did not come up with any great ideas for the combat in this game and generally had a hard time getting game mechanics together.

My way of handling melee combat was to create a "slashy slashy volume" in front of the player with a sprite of a slash, which damaged any enemies that it overlapped, moved forward at a rate of 10 cm/sec, and lasted for 0.2 seconds.

A funny thing that happened was that I haply created the player's shield as a child actor of the player, instead of just a static mesh component, so I could use the shield to push and pull the enemies around.

The art assets for this game were mostly done, although I would've prefered to have used MakeHuman or something to pull in some sort of customized mesh for the bad guys, and they needed weapons or some sort of way of indicating what attack they're using.

Next week I plan on making an art heavy Walking Simulator - style type of game thing, so that I can get my art workflow refined, maybe having some sort of dialogue based gameplay, or some puzzles, dunno.