Yeah I’m just going to go ahead and probably strip the “weekly” part from the titles of these posts.  But hey, I’ll make it up to you with videos!  Because everyone prefers looking at video than reading text, right?

Right.

What got done

I figured out why the file size of the exported release builds were so damn high.  Apparently the “complexity” of the image matters when saving JPGs.  I was always under the delusion that a pixel is a pixel – it didn’t matter what color it was or how it was different from the pixels around it, but apparently it does matter.  After simplifying many aspects of the level design (graphically-speaking), the file size plummeted.  I can now probably fit three or four levels into two megabytes very easily.

I also implemented a system for foreground images (the trees and the leaves on the ground from the level in the video, for example).  They’re not part of the background (since they appear on top of your car) and are really just treated as tiles.

That’s what all of my times was practically spent on these past two or three weeks.  As minor as those two additions are, they were very complex to get implemented and working (efficiently, that is).  Because of what I’ve done, however, the process is now streamlined for the rest of the levels in the game.

Oh, and I added some simple enemy AI, just today.

What needs to get done

Next up on my plate is to get the AI to be have a little more I.  As it stands, the way the AI turned out after just an afternoon of programming is really good.  However, it’s still a little rough around the edges turns.  The cars still act a little funky when approaching certain turns and sometimes just get outright stuck behind a wall while flooring it to the next way-point.

After that, I want to get the HUD functional.  This implies that other things will need to be functional, too.  For example, I’ll need to register laps (every time the player car drives over the starting line) as well as have all of the cars seemingly loop around the track (if you didn’t notice, the track isn’t actually circular, it just loops on the y-axis).

I also want to finally implement the final weapon into the game, all of the art assets have been drawn-up and animated for it since August, I just got lazy and didn’t implement it (plus I kind of needed to do a few things before I could implement it anyway).

So that’s that.  Oh, and happy almost-Thanksgiving to all of you fellow Americans out there.  Who’s ready for 24/7 Christmas music?

No video this week.  I was planning on it, but I have all of the required software all configured on my desktop, and due to the power flickering on and off, I’m not going to risk turning it on.  (Hurricane Sandy and all, pretty cool stuff.)

Nevertheless, here is the weekly update.  This week consists of good news and bad news.

First, the good news.

This week I was able to get a lot of minor things accomplished.  Weapons (both projectiles and “drops”) accurately collide with the cars in the level.  I also programmed in functionality to have the affected car undergo some torque when, for example, colliding with a mine or being hit with a grenade.  It efficiently spins the car and sends it out of control for a moment, really giving some fun realism to the gameplay.  Machine guns, grenade launchers, and mines also all produce explosions, shake the camera, etc.

I also finally fixed the issue with the camera leading the car (so you, as the player, can see ahead where you’re driving).  Before it was very shaky for some reason, but that’s been nicely fixed.

The final big change this week was the design and implementation of the HUD (heads-up display).  It was a lot easier than I thought it would be thanks to all of the great Starling capabilities (bitmap fonts, for example).  The game sure has come a long way from where it was just two months ago.

Okay, bad news…

I’ve finally come to the conclusion that the way I’m currently building the levels is not going to work.  A release build of the game is 2.5MB.  This is just one level with no sounds or any other big graphical assets.  To put this in perspective, the absolute maximum size I would allow for this game would be 10MB, any bigger and I’d be limiting potential sponsors.

The problem lies in how I visually build the levels, numerous 1024 x 1024 textures are just too massive.

The only solution is to use tiles. (Or significantly reduce the size of the levels, which I really don’t want to do.  If the race tracks are too small they’re just not as fun.)

How is this going to work out?  I don’t know.  I have to admit, I really liked my current system of designing levels, I felt like it was going to make things very easy while still being very dynamic.  Changing the code in the game to allow for a background built from tiles isn’t going to be a big deal.  The thing I’m worried about is how to design big levels that look good.

Either tomorrow or the next day I suspect I’ll be playing in Photoshop seeing what I can come up with, so I’ll have a better idea by then.

ALRIGHT BYE.

I don’t intend on these being monthly updates, really – I don’t.  I’m just constantly busy with other things and can’t get a sufficient amount of game work completed in the week.  Once the weekend rolls around I feel like I don’t have anything worth showing.  Then, in the middle of the week, I’m usually able to get some things worked out in the game, but I also begin work on new features.  Come the weekend, the game seems un-presentable again.  Rinse and repeat.

But not this time!

After a bunch of fumbling around and trying various different methods (even venturing into Box2D territory, but that was way too much overkill), I have finally implemented collision detection and reaction.  Check out the video below!

Of course the AI there is the absolute most basic AI you can imagine.  Their acceleration and steering depend on random values.  The walls are made up of a bunch of convex polygons, mainly rectangles.  Big shout-out to Tiled for being pretty awesome and making this level-building process pretty easy.

I also just started implementing how to detect the off-road parts of the track.  Depending on how many tires are off the road (aka on the grass, for this level), the car’s friction will increase and the rear tires will no longer produce tire skids (if they are the tires that are on the grass, that is).  It seems to function pretty efficiently, but I would like to optimize it somehow, just to prevent potential future hiccups.  (I only “drew-in” the polygons that represent the off-road near the beginning of the track, I didn’t bother with the rest of the track yet.  That’s why you can see my car producing skid marks all over the grass later in the video.  This track won’t even be in the final game and is only for testing purposes.)

In this screenshot, you can see how I’m organizing the different layers of the level data in Tiled.  The background image is just there to assist me in placing things in the correct spot.  One of the unseen features I worked on for the game was a small sub-program to read in a .TMX (Tiled) file, parse it, produce a ByteArray of all of the parsed data (polygon points, etc), and then save the file to the game library – ready to be embedded and used in the actual game.

I also like how the level doesn’t actually loop around in a circle, if you hadn’t noticed.  This way I can always have the player driving straight upward (or at least mostly in that direction).  That always seemed more comfortable.  (Speaking of this, I do actually have the code in-place to have the camera lead the car as it goes faster, so you can see ahead where you’re driving.  It’s just a bit shaky and annoying at the moment for whatever reason so it’s not active in the video.)

Alright, here’s to hoping the next update post will be here before Halloween!

Don’t fret, I didn’t lapse into another period of inactivity.  Last weekend was just a bit different so I didn’t get a chance to write up a new post.  Plus I wanted to figure out how to get some solid screen capturing done so I could upload gameplay footage, and that’s just what I did for this week’s post!

Check it out – in the video, I show off some of the weapons in the game.  As of right now (and this will probably be final), there are six weapons, each with four upgrades.  They include machine guns, grenade launchers, boosters, mine deployers, and oil slickers.  The video only shows off those five since the sixth isn’t ready yet, and will thus remain secret.

Tomorrow (or today, or yesterday, I don’t know) I intend on finally creating a “dummy” level to begin implementing collision detection.  I already have all of the “hard work” done, the separating axis theorem has been programmed into the game since near the start, I’m just not actually actively checking for collisions yet.  This should make the process go smoothly and I hope to have the player car smashing into walls by tomorrow night (or last night, or whatever).

Sometime this week I’d like to make the AI cars “solid” and have them react to weapons (as in, have projectiles blow up on contact with them).  After that, I’d like to get the sixth weapon functional and then finally begin focusing on actual AI, though now I’m getting ahead of myself.

But alright, that’s it for this week’s post.

Now get lost.

So yeah, a little over a month later and the second installment of the weekly updates for Weapons on Wheels 2 is finally here!  Seems about right.

Alright, sorry for the delays post, what with the week-long family vacation in the middle of August, college starting up at the end of August, and a Labor Day weekend full of fun-filled family get-together activities, I really haven’t had time to make a new post let alone get much work done.

The funny thing is, I seem to have been working on the same exact things over and over for the past month.  First it was designing, animating, and exporting all of the sprite sheets for the weapons – I did that twice.  Then it was getting the level background class running how I need it to run (and efficiently, at that) – I did that three times, re-writing the whole class each time (so much tedious math).  Then there was the class that needs to handle the tire skid marks, that kept needing to be revised.

But now I’m done, finally everything is running how I need it, plus I finally was able to start working on weapons (only the boosters at this point) as well as some particle effects, as you can see from the above screenshot.

I originally wanted to post a video of some of the current build and show off the four boosters and their particle effects, but after tinkering with some screen-capturing software this morning I finally gave up.

Next up I’ll start working on the other weapons, and eventually I want to create a dummy level, just something that will allow me to start adding walls (collision detection) and get some structure to get the AI up and running.  I may start on that latter part this weekend if I make enough progress.

Alright, from here on out I don’t see why I won’t be able to update every weekend, it seems all of the nonsense is finally behind me.

Back to work.