Wednesday, February 22, 2012

Making of Radar demo #2: Map Design

Ah, another making of post, and a bit delayed as usual. Anyway, where were we? Oh yes, map-design. Designing your worlds, like a game God.

1. Mapping: Common sense versus bullshit
Ideas often start with pictures, snapshots from other movies/games, or weird (day)dreaming. Or being drunk. Smart professors found out that being drunk and/or tired makes you switch over to the creative parts of your brain instead of the logic portion. That's why I keep "idea-doodle" lists. If I saw something on TV or imagined freaky stuff just before going to sleep, I write it down on a paper. These lose, random fragments are not directly suitable for a game though. Nor are rusty factory photo’s from the internet. The Canadian/American army didn't build that Radar Station as a secret party spot for kids that want to get drunk. Neither did the design count on Tower22 or any other action-shooter maps to be made in the future.

Most buildings focus purely on a certain task. Sheltered, providing all required facilities, storage space, and maybe some comfort if the wallet allows. Now concrete structures aren't exactly meant to be cozy. And unless you're talking about Hitler’s bunker, don’t expect a lot of secret narrow halls, illogical stairs, or spectacular vistas. As for Soviet apartments flats, there is not much to expect other than a central corridor and apartments... over and over again, each store. If you are lucky, there might be an office section, or shops at the ground-floor. But all in all, a skyscraper is one of the most boring places to start a game ;)

So one of the first things you got to do as a map-designer, is to let go the logics. The player isn't counting toilets or inspecting your building on structural integrity, running water, if the air-ventilation makes sense, if the apartments are up to the ISO norms, or presence of salmonella. He or she wants cool places to explore, have walls and obstacles to cover or hide, find secrets, be amazed by stuff you don't see every day.

Sounds easy, but that's dang difficult, especially for Beta persons. I'm used to fix (program) things with logical thinking, not with bullshit. It's easy to design a place that makes sense, as you have design-rules and countless of references all around you. Just look in your house, a factory, school, or whatever. But if the map relies on fantasy(nonsense), it gets difficult to make a picture. Imagining one crazy room with weird shapes, floating guts and dancing dragons isn't that difficult. But filling a whole game with it...

* If you like, use realistic structures / locations as a framework, but fill them up with your own sick ideas.
* Write down all ideas, store all crazy pics, and always keep a doodle paper next to your bed.
* There are many photo collections about old buildings and stuff on internet.
* Or just grab the camera and have a look in a factory. Do your homework.
* Dare to experiment! Add or remove rooms you would normally expect, enter a room via a secret corridor instead of a frontdoor, and don't forget to reserve space for bigger "wow!" scenery.

* Respect the theme. Whether your map needs to be realistic or not, you can't mix up spaghetti with choocolate ice-cream. Variation is needed, but within a certain boundary. Save bizarre ideas that are hard to place for later usage.

The very first room in Doom2, which takes places on Earth apparantly, doesn't make sense already.

Don't even know what I'm looking at here (Silent Hill). Someone had a very bad dream here.

2. Mapping: Stitching socks
Rooms need to be connected somehow. And in such a way that it suits your type of game. Action games require lot's of coverage, plenty stuff to blow up, and locations that lends itself for spectacular (scripted) scenes. A Metroid or Zelda game needs to be more open, contain loads of secrets, blocked routes that can be unlocked later, and fantasy architecture that remind you you are in Hyrule or on Zebes, or whatever those planets are called. Tower22 will be about puzzling and exploration, so count those elements in.

To make things a bit more difficult, it needs to be scary in our case. The Heinz blood-ketchup tool only works for a short while. Your structure needs to be fundamentally f*cking scary to be horror worthy. One dull hallway spoils the climax. Bad lay-out that gives you a sense of safety, and also don’t make things too easy or the fear-factor will be ruined.

The bottom line, you can't just spray random ideas over your mapcharts. All the places need to be linked with each other in such a way that it challenges the player. Whether that is blasting things up, running for monsters or puzzling your way through. You need to think back- and forward. Room1 may influence room45 later on in the game, or vice-versa.

* Don't start building blindly. Write down the global game-flow first and use that as a red line.
* Make lists of the items, monsters and puzzles you like to add in the game. So you can pick and mark them on the list.
* Split your world up in sections, chapters, levels, area's. or whatever you like to call them. Making puzzles and routes per sub-section is easier than for the whole game, plus it's also easier to scrap or adjust sub-sections.
* Later on, you can link the sections with bigger-scaled puzzles (find itemX in section1 to proceed in section3).
* Tease. Show interesting stuff that can’t be accessed yet. Reason to keep playing!

Found this cool pic here. It's the floorplan from the Resident Evil mansion. (Virtual) Architect George Trevor sure had weird ideas about making a comfortable home.

3. Mapping: Don't Enlarge your penis
Another important note about map-design, but certainly horror in particular, is the size of your maps. Bigger != Better (that's Nerdish for bigger isn't better), with the exception of Grand Theft Auto or WoW maybe. Development teams often agitate the public by telling "Our world is 8, EIGHT! times bigger this time, covering 100 billion acres in total!!!". And? So? Bigger is only better if the map is filled with useful things. Zelda Twilight Princess had a big world all right, a big empty world. Secrets were too spread, and because you traveled so much I never felt a cozy bond with the location. Which is usually the big charm in Zelda games for me.

Several war/action-games have big environments too... filled with the same houses, trees and rocks over and over again. At first you're curious what's inside such a house so you explore the area. Later on you just keep running forwards. Screw those details, nothing to do there anyway. I just shoot a few terrorists if I need ammo.

What I liked so much about Duke Nukem 3D (1995), was its realism. Realism?! Hey don't forget that all other games of that time took place in dark dungeons or claustrophobic space-stations. Duke Nukem actually had streets, trashcans, cars, shops, and bars. Recognizable elements from real life. But don't mistake, the DN3D maps didn't make that much sense either if you think about it. Streets with 2 dead ends, spaceships, secret tunnels behind posters, green alien goop area’s, flooded cities. They dosed these fun-realistic elements very carefully.

When I did my first DN3D maps with the brilliant “Build” editor (my start as a “game maker”!), I always filled the maps with as much as possible. 4 houses, each with toilets, kitchens, bedrooms and everything. 3 bars, 2 parking lots, 10 skyscrapers, and so on. Exploring the first house and finding the toilet & mirror is fun. Seeing it for the second time is less of a surprise. Seeing it three or more times just isn't fun anymore. That's (probably) why the realistic "wow!" sights in DN3D were carefully spread over all the ~30 maps. One level had a cinema. Another level had a bar. Yet another level had a livingroom + toilet. Level X had a sewer-system. Et cetera.

The moral: don't repeat your ideas too often. If that means you have to reduce the overall map-size, then so be it. Especially horror-scenery can quickly become dull and predictive if you repeat it. If you look at a game like Resident Evil 1, you won't see 5 gardens, 40 bedrooms, and 6 laboratories either. In fact, the game is pretty small. If Jill had a (working) flamethrower and all keys, she could invest the whole building within 30 minutes, like a real Japanese can finish Super Metroid within 30 minutes.

I hate short games though, so instead of adding more floors with the same stuff to Tower22, the puzzles, monsters and interesting sights should stretch the length.

* List all your main ideas (example, I want a lava, forest and ice world)
* Note the sub-ideas, make an "idea tree"
* Try to equally spread the ideas. If one section has too much while the other is empty, shuffle. Or just remove the boring parts if there is no other option.
* Don't start too big. Adding area's can be easier than removing parts.
* Keep closed sections in your maps. A cheap way to grow in size, but you can also fill them later on if you need to stuff more ideas in the same space.

One of my all time favourite game vacation spots is Clocktown (Zelda Majora's Mask). A small town, but you can spend many hours here. Finding out what all characters do on a particular time of the day, finding hidden passages, or just have a stroll on a rainy day.

4. Mapping: Review yourself
Getting started on an idea is always difficult. One day I'm creative like melting butter, the other day it's not coming any further than blocky Wolfenstein 3D corridors. But once you are rolling, it's easy to link a whole chain of ideas and plot them as a superfast matrix-printer on paper. Or Paint in my case. MS Paint? Yeah, you heard me, MS-Paint dammit. Because it boots up within a second and doesn't bother you with anti-aliasing, hundreds of tools, artistic poopbrushes or what else. I just want to draw lines and squares, that's it. Detail is for later.

Anyway, as said a few times before, you'll have to be a bit careful with your enthusiasm though. Ideas that sound better than sliced bread right now, can still suck if you review them later on critically. But if you spend many hours on drawing the whole map, it's hard to take distance of it. Some ideas just rust into your brains, and you're not open anymore for adjustments or complete different ideas. The same issue also counts for other elements in your game btw. Such as the looks of a monster, the story, et cetera.

* Start your maps simple. Don't draw all details yet, just blocky schemes so its easy to adjust shapes or scrap them if needed.
* Make your maps, then let them rest for a few weeks.
* Then look again... Still got that warm feeling about them? Ifso, good for you. If not, just redo them. Probably your second or third attempt combines the best of two worlds.
* Usually I make two or three variants of an idea. Sometimes with a few months or even years between them. Each “iteration”, I pick the best ideas of the previous attempts and mix it with the new fresh ones.

5. Mapping: Radar Station
Luckily, the Radar Station wasn't part of the game. Meaning I didn't have to follow those guidelines that strictly. So why am I telling that whole story. Dunno :p The Radar didn't have to be playable in puzzling and action terms. The monster down below wasn't even planned originally. Hence, it was meant for testing new objects & physics remember? The idea was to throw a few barrels in the water to show some physics, and have a quick fly-through the building.

Yet, we got a bit more ambitious... if we want to record a movie here, it should at least breath the T22 feeling a bit. At the same time, this map is meant as a playground to test objects and physics. So I wrote down the requirements. Uh...

- Needs to be pretty big so you can render larger objects as well
- Stairs for (ragdoll) physics
- Water & ice for physics, and because it looks cool
- The camera has to be able to take 1 smooth path through the whole building. Not entering a room and turning 180 degrees each time.
- Not only boring concrete and metal s'il vous plait

But on top, don't forget we're dealing with a radar. As said, the building didn't have to make perfect sense. Hell, I don't even know how a real Radar structure looks from the inside, other than from a few photos and N64 Goldeneye. Yet I could do some suggestions of course:

- Control room: Charts and computers and stuff
- Storage space: Small warehouse
- High building, so... stairs, elevator
- Personal: restrooms, beds, dressrooms
- We wanted ice and water, so... open roof maybe?
- How about... a radar dish?

With that in mind, it was only a matter of making rooms for each of those ideas and linking them in a somewhat logical, yet playful way. Got to mention that I didn't have a real good idea of how each room would look yet. Just draw a square in Paint and write "Barracks" or "Control Room". The texture setup and micro-details such as junk or rusty decals were for later concern.

All in all, the Radar isn’t exactly the scariest building ever made. You could find such a thing in reality maybe. Yet did we follow the 4 rules mentioned above? Logic mixed with nonsense, check. The building has no front door, there is a hole in the roof, some places aren’t accessible, and no idea why you would need fuel drums in a radar. Connecting the rooms properly? Played with stairs, ladders and dark tunnels… Size of the map? Not too big indeed. All ideas we’re only used once, so that makes each room interesting to explore. Did we review ourselves? Erh… I did, while writing this ;)

After the floorplans were made a bit, I made a simple mesh so the others could have a better look as well. 2D floorplans are fun, but it's hard to tell what someone has in mind without seeing any 3D shapes and colors.

No comments:

Post a Comment