Interesting ways to create levels

[phobosdeimos1]

In this thread, post your ideas for methods of creating levels using an individual or unusual method,

tried or tested, failed or yet to be done, any method you've heard, thought or seen in use.

Here's some examples to start off:

- Creating a level through starting with the exit and working backwards.
Pro's: You always know where the level is leading and you don't rush the exit out of excitement of nearly finishing the level.
Con's: Playtesting properly is impossible until you have finished the level's base layout.

- Creating a level via drawing a pre-sketched map on paper.
Pro's: When you come to the editor you're never stuck for ideas.
Con's: Creating the level in the editor is boring and feels like work instead of creative output.

- Creating a SP level via making a Deathmatch level as a base.
Pro's: Not sure yet, i'm still making it
Con's: So far, for some reason it's seeming alot more time consuming.

Get the jist? Post away, my dead wives.

[Mithran Denizen]

WadC is sort of related.

I've played around with it a bit in the past, but was never patient enough to get anything playable out of the efforts. It's an interesting/unusual way to design levels though.

[hex11]

Using very old DM maps as a base for SP maps:
http://www.doomworld.com/vb/wads-mods/57184-undeath-94-a-new-vanilla-doom-episode/

[Phobus]

#Making a map in layers (as in, you start with the base layout, then you add heights, then you add map functionality such as doors and a sense of progression, then you do texturing/lighting, then gameplay and then aesthetic additions)
Pros:
+The mapper thinks about functionality and playability before looks, allowing for open layouts and interesting progression much more easily.
+It's also really fast to get the layout done and then it's just a case of filling it in.
Cons:
-Due to the way the maps start off so abstract it can be a bit overwhelming to then implement a texture theme or do much in the detailing stage.
-The maps can feel finished once gameplay is in if the textures are done alright, so you're tempted to not bother with the visuals, leading to maps that are actually quite ugly.
-The other end of the stick is that detailing ends up going in for details sake to avoid the abstract areas looking bland.
-Gameplay testing can also be pretty tough if you do it before texturing due to the missing textures and the like.

This was my approach to speedmapping for 32in24-11 - I found it was much quicker if I had a solid idea in my head rather than just a couple of vague ideas. E1M2 of ZPack was also done this way.


#Making self-contained areas that you then stitch together in a patchwork map.
Pros:
+Allows for transitions between themes in a map easily - helps make areas visually distinct.
+Sections of maps are easier to make well than one whole map at a time.
+A good way to maximise the use of clever/novel ideas as each section can focus around a gimmick.
+Easy to make each of the self-contained areas non-linear.
Cons:
-Easy to make a map that's a bit too long or feels like it could have been ended at each of the sectional transitions.
-Can make the overall map feel very linear.
-"visually distinct" with a variety of themes can actually just feel disjointed.

The majority of my maps look like they were made this way, but in reality only ph_quik2.zip, ZPacks E2M1 and E2M0 and my CC4 MAP10 contribution "Hoedown" have made use of this technique. The first three dodge the cons by explicitly having the player move between different realms and themes whilst Hoedown sticks to one theme loosely throughout and makes use of backtracking from the self-contained section to the rest of the map.

[phobosdeimos1]

Great reply Phobus, I had actually got the idea for this thread from a post you once made stating you'd tried alot of different techniques and found the results never quite matched up to doing it as you go along from start to finish.

I've personally tried the methods you outline here and hideously failed, especially at the first one, finding it an extremely daunting and time consuming task to go back and add heights and textures.

The patchwork map seemed like a good idea to me after looking at Tim Willit's maps, finding that they all seem to be large rooms with about three different 'floors' (drastic height changes in the floor), in which the rooms are joined by corridors and doors.
Like you said it also allows for gimmicks, and spawned an idea I have about writing down random descriptions for rooms e.g Mancubus cave and Moving floor computer room then patching these together with little corridors and such.
This didn't really work for me as it just didn't feel natural, and like with nearly every method I've tried, it slowed me down so much.

[Phobus]

I reckon the first method probably only works for smaller arena-like maps (preferably with one consistent theme throughout) - anything that is more of a straight line or a big, branching layout would very quickly turn into looking like far too much work. As you say though, mapping along the line of progression feels more natural (it also works better for playtesting too, IMO).

Your idea about writing down random descriptions for rooms sounds an awful lot like my usual thought process when making a map from start to finish, TBH! The most extreme examples of this method in my work are probably ph_quik3 (Justice: Infernal Mechanics) and ph_quik5 (Small Dark Twisted Computer Lab), the former because each room/area is clearly based on a gimmick and the latter because it was all based on a randomly generated map idea.

For your ideas in the OP, I have to say the only one I've tried is the map sketching one - the way round the con there for me was to only make the sketch very vague, so you end up filling in most of the map properly in the editor still and just use the sketch to collect your initial thoughts and ideas - my example here would be CC4 MAP03 "Three is a Charm". I sort of did the DM level one, but that for me was using somebody elses map as a base and then making it a ZDoom SP map (Tiny Pack 1: TINY0: Ledge-end - based on LEDGESI.WAD). Mapping from end to start sounds pretty interesting. I think I'll give that a go at some point :)


I might come back and make another reply later on some of the other approaches I've taken to mapping, as I've never really looked into explicitly working out the pros and cons before and it's quite an interesting thought exercise.

[phobosdeimos1]

I can see we're on the same page here, I'm gonna try out all these examples you're giving me, am I right in assuming I can't play Three's A Charm until the CC4 project is finished?

Is there any way it would be possible to either post here or send me personally an image of one of your sketched out maps? As it's never really worked for me before but I'm probably doing it in the worst way possible

here is a recent one (ignore the writing at the top, it's from before I drew the picture as I was going to try and write it as if it was a Walk-through from the Doom Wiki and then build it from there

http://i1190.photobucket.com/albums/z441/AmanCalledcallum/DSC00226.jpg

[Phobus]

I've not kept any of my sketches as they really have been incredibly basic, but I think I can remember the sketch for Three's a Charm, so I'll do a quick redraw of that for you, take a picture of it and upload here. You'll also receive an e-mail that is definitely not leaking my two finished CC4 maps to you ahead of the official release of the project for your perusal ;)

Your sketch actually is a lot more in-depth than mine have ever been. Closest I did to that kind of level of detail was when I was making Scourge MAP15 I think. Basically, where the beginning of the map was meant to be a copy of the end of MAP14 and I was using DEU2 (which I don't think had copy/paste), I made a scale drawing of the section of the map I intended to copy and remade it from that. Obviously these days with DB2 I'd have just copy/pasted and been done with it in about 10 seconds. The good old days, eh? :P

[Phobus]

DBLPOST:

http://i.imgur.com/j92WP.jpg

Here's my sketch-from-memory of the sketch I used to make "Three is a Charm". I did realise I probably should have done it in pen or something as I was taking the picture, but I think it's easy enough to get the gist of it from that shot. You'll get that e-mail that you shouldn't have in a minute.

[40oz]

Listen to music and just let your hands make sectors until the map is done

[EarthQuake]

Make a map without using 3D mode. You spoiled brats.

Some ideas that have been done to death:

- Create a map where all vertices are restricted to a specific grid size.
- Create a map using only use 2, 10, or some arbitrarily-low number of sectors.
- Create a map using only use 1, 2, or some arbitrarily-low number of textures.
- Create a map that fits within 512, 1024, or some arbitrarily-low size.
- Create a map using a bare bones skeleton created by someone else.
- Create a dark flooded base with a lost UAC machine.

Here's an interesting one: create a map with only three one-sided linedefs, and where every sector's ceiling texture is F_SKY. The map can have any number of sectors or two-sided linedefs.

[MajorRawne]

Design a map07 that is not a Dead Simple clone and does not feature any variation of the word Simple in its name.

[The_MártonJános]

MajorRawne said:

Design a map07 that is not a Dead Simple clone and does not feature any variation of the word Simple in its name.

Or Dead, Death, Deadly, etc.