Level design techniques
by Marvin Hawkins · in General Discussion · 07/26/2005 (1:31 pm) · 8 replies
I know that it's wierd that i've been a member here since 2000 but the questions i ask seems pretty elementary. Well the reason behind this is i'm working on my first two titles, while i've been playing around and expirementing with tge and 3dgs (the engines which i'll be using) but now i actually want to get results from my development. So i want to use any efficient process i can for completing work in a specific time window. One of my concerns is level design. I've tried a few different ways to design a level but they all seem to take hours or days. can someone please share thier processes that they use when creating levels/assets?
Do you start with a 2d paper design?
Do you draw 3d views of prominent areas? (or areas that are of greater importance)
what decisions go into determining how long a level is/its flow for you? etc.
Do you gather references before you build a level?
Do you start with a 2d paper design?
Do you draw 3d views of prominent areas? (or areas that are of greater importance)
what decisions go into determining how long a level is/its flow for you? etc.
Do you gather references before you build a level?
#2
If it's more than a single-room interior, I do start with a 2d paper design. I won't necessarily adhere to the paper design, but it gives me a foundation and a plan. I also gather whatever references I can. The Internet is great for grabbing pictures of real-world locations to use as reference material.
I haven't drawn 3d views of prominent areas. However, I worked with several artists on larger games, and they made use of concept art for levels extensively (whether they did it themselves, or used concept art from another artist). I kinda use the reference materials for concept art for the most part, but that's probably an iffy proposition because of potential style differences.
There's another thread on level design, but for the levels I've done so far, they have been smaller locations more for mood / setting / drama than game flow. I think a lot depends on the game. For deathmatch-style FPS games, EVERYTHING centers on having a level where players can use to tactical advantage, where there are no "perfect" camping positions, etc. So level designers will design and test a level to death without textures, making sure the gameplay is perfect before taking the time to texture the level (or add any other details that would be time-consuming to change, like sound).
07/26/2005 (1:51 pm)
I'm a beginner at the process myself - but since I'm a learner, I figure I'll share what I've learned so far. (And it still takes hours / days for me to do stuff...)If it's more than a single-room interior, I do start with a 2d paper design. I won't necessarily adhere to the paper design, but it gives me a foundation and a plan. I also gather whatever references I can. The Internet is great for grabbing pictures of real-world locations to use as reference material.
I haven't drawn 3d views of prominent areas. However, I worked with several artists on larger games, and they made use of concept art for levels extensively (whether they did it themselves, or used concept art from another artist). I kinda use the reference materials for concept art for the most part, but that's probably an iffy proposition because of potential style differences.
There's another thread on level design, but for the levels I've done so far, they have been smaller locations more for mood / setting / drama than game flow. I think a lot depends on the game. For deathmatch-style FPS games, EVERYTHING centers on having a level where players can use to tactical advantage, where there are no "perfect" camping positions, etc. So level designers will design and test a level to death without textures, making sure the gameplay is perfect before taking the time to texture the level (or add any other details that would be time-consuming to change, like sound).
#3
07/26/2005 (1:58 pm)
I rough the general layout on paper (as well as the concepts of how I would like the "final"--haha--product to look). Then I rough the necessary pieces in Quark. I don't worry about details initially. Once I get the orientation of things the way I like them, I begin the detail work. This takes a LONG time. Fleshing out the structures and texturing them carefully will take the hours that you're talking about. Tossing blocks together to get an idea of the size and spatial organization shouldn't take that long, though.
#4
Would you do this after locking down gameplay for the level. (IE get to Zone x to complete the level)?
07/26/2005 (2:24 pm)
Yeah thats the point i've gotten to so far. for example the test level will actually be a level in the game of course minus the test elements. but i've got a basic room going but at what point do you add the final art assets? Would you do this after locking down gameplay for the level. (IE get to Zone x to complete the level)?
#5
I usually start with a few concept designs. I focus on the some key areas, roughing out the general feel. In my Untitled Project, a main area is going to be a large underground complex (think Cheyene (sp?) mountain complex(I look at SG:1 tv series for the feel), since I doubt I'll ever get a look inside). The key areas that I see is the entry area, corridors (I focused on a 4-way intersection) maybe a room or two.
I quickly sketch these areas. Then I look for details that would become to fiddly to model with Cshop/Quark (in this case, pipes, light fittings. access control panels, other props, beds, desk) These things I'll model in Lightwave.
Next I get my graph paper out and start to draw the layout(top view). If I come accross a feature that needs more detail than can be shown in the topview I'll draw or look at my reference sketches.
From the sketches/topview layout and my own skills I can evaluate my time frame to complete a section/feature. One thing I have noticed if you're trying to learn Torque scripting, and the code base. Plus implement new features into your project on your own.... well it does screw up things...for me anyway..
One thing I forgot to mention is texture mapping. I already have quite a large collection of my own stock images. If I get stuck with how/what a section could look like. I just browse through my images. I find a nice image and go "gee this will make a great wall" or "this grungy wood texture, will make a good pipe texture with a little work". Having a collection of images can save you alot of hassle. Just grab your digital camera, and take a photo of the pavement, a wall, a puddle of water, whatever you see while going to work, shopping, etc. It saves the problem of trying to find royalty free art.
There is my 5cents.
07/26/2005 (5:05 pm)
@Marvin : I'm a still a noob when we comes to the engine/scripts but I have a reasonable level design workflow. IMHO.I usually start with a few concept designs. I focus on the some key areas, roughing out the general feel. In my Untitled Project, a main area is going to be a large underground complex (think Cheyene (sp?) mountain complex(I look at SG:1 tv series for the feel), since I doubt I'll ever get a look inside). The key areas that I see is the entry area, corridors (I focused on a 4-way intersection) maybe a room or two.
I quickly sketch these areas. Then I look for details that would become to fiddly to model with Cshop/Quark (in this case, pipes, light fittings. access control panels, other props, beds, desk) These things I'll model in Lightwave.
Next I get my graph paper out and start to draw the layout(top view). If I come accross a feature that needs more detail than can be shown in the topview I'll draw or look at my reference sketches.
From the sketches/topview layout and my own skills I can evaluate my time frame to complete a section/feature. One thing I have noticed if you're trying to learn Torque scripting, and the code base. Plus implement new features into your project on your own.... well it does screw up things...for me anyway..
One thing I forgot to mention is texture mapping. I already have quite a large collection of my own stock images. If I get stuck with how/what a section could look like. I just browse through my images. I find a nice image and go "gee this will make a great wall" or "this grungy wood texture, will make a good pipe texture with a little work". Having a collection of images can save you alot of hassle. Just grab your digital camera, and take a photo of the pavement, a wall, a puddle of water, whatever you see while going to work, shopping, etc. It saves the problem of trying to find royalty free art.
There is my 5cents.
#6
I usually start with a few concept designs. I focus on the some key areas, roughing out the general feel. In my Untitled Project, a main area is going to be a large underground complex (think Cheyene (sp?) mountain complex(I look at SG:1 tv series for the feel), since I doubt I'll ever get a look inside). The key areas that I see is the entry area, corridors (I focused on a 4-way intersection) maybe a room or two.
I quickly sketch these areas. Then I look for details that would become to fiddly to model with Cshop/Quark (in this case, pipes, light fittings. access control panels, other props, beds, desk) These things I'll model in Lightwave.
Next I get my graph paper out and start to draw the layout(top view). If I come accross a feature that needs more detail than can be shown in the topview I'll draw or look at my reference sketches.
From the sketches/topview layout and my own skills I can evaluate my time frame to complete a section/feature. One thing I have noticed if you're trying to learn Torque scripting, and the code base. Plus implement new features into your project on your own.... well it does screw up things...for me anyway..
One thing I forgot to mention is texture mapping. I already have quite a large collection of my own stock images. If I get stuck with how/what a section could look like. I just browse through my images. I find a nice image and go "gee this will make a great wall" or "this grungy wood texture, will make a good pipe texture with a little work". Having a collection of images can save you alot of hassle. Just grab your digital camera, and take a photo of the pavement, a wall, a puddle of water, whatever you see while going to work, shopping, etc. It saves the problem of trying to find royalty free art.
There is my 5cents.
07/26/2005 (5:37 pm)
@Marvin : I'm a still a noob when we comes to the engine/scripts but I have a reasonable level design workflow. IMHO.I usually start with a few concept designs. I focus on the some key areas, roughing out the general feel. In my Untitled Project, a main area is going to be a large underground complex (think Cheyene (sp?) mountain complex(I look at SG:1 tv series for the feel), since I doubt I'll ever get a look inside). The key areas that I see is the entry area, corridors (I focused on a 4-way intersection) maybe a room or two.
I quickly sketch these areas. Then I look for details that would become to fiddly to model with Cshop/Quark (in this case, pipes, light fittings. access control panels, other props, beds, desk) These things I'll model in Lightwave.
Next I get my graph paper out and start to draw the layout(top view). If I come accross a feature that needs more detail than can be shown in the topview I'll draw or look at my reference sketches.
From the sketches/topview layout and my own skills I can evaluate my time frame to complete a section/feature. One thing I have noticed if you're trying to learn Torque scripting, and the code base. Plus implement new features into your project on your own.... well it does screw up things...for me anyway..
One thing I forgot to mention is texture mapping. I already have quite a large collection of my own stock images. If I get stuck with how/what a section could look like. I just browse through my images. I find a nice image and go "gee this will make a great wall" or "this grungy wood texture, will make a good pipe texture with a little work". Having a collection of images can save you alot of hassle. Just grab your digital camera, and take a photo of the pavement, a wall, a puddle of water, whatever you see while going to work, shopping, etc. It saves the problem of trying to find royalty free art.
There is my 5cents.
#7
07/27/2005 (7:55 am)
Wow thank you everyone this is all very insightful info. feel free to keep em comin.
#8
I usually started out with thinking up or drawing a broad layout and with the area's around the flags (CTF mode in Q2WF, capture & hold in Anargy). Those areas are crucial because much of the crucial gameplay happens there. I always made sure it was interesting and that the balance between offense and defense was well done.
Sometimes I sketched the areas on paper but often I already had inspiration and idea and I would draw them roughly in the leveleditor, in 2D (floors only). So both would be topview.
After that I usually added features that would make a map scale. In practical terms this would mean adding extra paths and features that take a little more time or effort to do but worth it when the main areas would get clogged. In that stage I usually would already have a rough layout with the walls / worldlimits and some areas might already be in full detail.
I usually mixed trying out new styles and working on the layout. Some architecture features might invite adding gameplay features and visa versa.. so that usually worked well. Also if some textures would be crucial to a large area of the map, if would try to them out to see if they would work. After this all it would be finishing the maps and testplay too. I usually would make sure to get early feedback too.
Another totally different approach is starting to map from the players' start position, build a complete 'done' room and then think 'what would be fun, and where?' This is a very evolutionairy way of working but can lead to truly great maps that don't feel designed since they didn't have that overall plan hanging over everything. Of course during the process you need to keep in mind that an overall plan would work. Usually when I took this approach, I needed to balance play later by adding or changing sections but... the maps would get a unique feel with that. In real life, building also get extended and changed.. so if you want an authentic look, this usually works really well.
- Edo.
07/27/2005 (8:19 am)
I only did leveldesign for two multiplayer mods, the first was Quake2 WeaponsFactory and the second, Quake3 Anargy. (nickname: B=Ware)I usually started out with thinking up or drawing a broad layout and with the area's around the flags (CTF mode in Q2WF, capture & hold in Anargy). Those areas are crucial because much of the crucial gameplay happens there. I always made sure it was interesting and that the balance between offense and defense was well done.
Sometimes I sketched the areas on paper but often I already had inspiration and idea and I would draw them roughly in the leveleditor, in 2D (floors only). So both would be topview.
After that I usually added features that would make a map scale. In practical terms this would mean adding extra paths and features that take a little more time or effort to do but worth it when the main areas would get clogged. In that stage I usually would already have a rough layout with the walls / worldlimits and some areas might already be in full detail.
I usually mixed trying out new styles and working on the layout. Some architecture features might invite adding gameplay features and visa versa.. so that usually worked well. Also if some textures would be crucial to a large area of the map, if would try to them out to see if they would work. After this all it would be finishing the maps and testplay too. I usually would make sure to get early feedback too.
Another totally different approach is starting to map from the players' start position, build a complete 'done' room and then think 'what would be fun, and where?' This is a very evolutionairy way of working but can lead to truly great maps that don't feel designed since they didn't have that overall plan hanging over everything. Of course during the process you need to keep in mind that an overall plan would work. Usually when I took this approach, I needed to balance play later by adding or changing sections but... the maps would get a unique feel with that. In real life, building also get extended and changed.. so if you want an authentic look, this usually works really well.
- Edo.
Torque Owner David Tiernan