Level Design Pipeline

This article sumerizes how I take on levels when working on a project, this description is linear but in reality the different stages and actions most often overlap eachother. Furthermore, it should be mentioned that this is my general way of working and every pipeline for each project can differentiate a lot from each other.

Establish the context of the level, what setting the player will experience and how far in the game the level is. Gather references over the setting and related settings, pile up the images in a folder to get the ideas going. Get the note book, look at the reference folder and start storming out ideas on paper without any reconsideration on whether they are technically possible or even convenient for the game. Do the second pass on the notes, categorize and eliminate impossibilites and bad ideas, consult programmer(s) how crazy the ideas are and no matter the answer – archive everything.

Make a compiled document of all good and reasonable ideas, present it to the team for further feedback. When ideas stop coming, move on directly.

Reference Questions
What’s most interesting thing about this setting?
What do player expect from this setting?
What doesn’t players expect from this setting?
What emotion(s) do I want the player to experience?
What’s the target playtime?
What does this setting Really look like?

Once a strong basline for the level is there, start hatching out a LLD. If there are previous level design documents made in the project they can be used as template and it is not a bad idea to even establish a template LDD for the project. This step and the next are very transparent and it is not a bad idea to switch between these steps repeatedly.

This part is pretty straight forward, start sketching out the level and in order to get some concrete visual representation. When you feel satisfied with a top-down, put it away and make another with a new angle and repeat this until you have a couple of drafts. Take the drafts and discuss with the team, compile the good things into one draft. Before passing onto the next stage, make sure to get some feedback from team members for each draft and make revisions if needed.

Reference Questions
How many levels(floors) of height will there be?
Where does the level end and start?
How many paths are there available?
What focal points exists in this level?

When the top-down sketching have taken shape, start out with blocking out the major shapes of the level to get the distances in place. In this phase I tend to build paths more narrow than I expect them to be, simply because it is easier to carve out where you want to walk than filling up void in the level. During this phase it is very important to constantly have a reference object to represent a player avatar or trying it out in-game as often as possible. It should also be mentioned that if the game includes vision-based mechanics line of sight should be overviewed constantly, just moving around the map in-game to see and feel the layout. As far as the design goes this is just the first draft of the layout and everything is subject for change.

As a level most contains a vast amount of assets and objectives a priority list should be made together with notes on what they are to be used for. When I make this lists I reconsider why they are important and how they will be used. At the very top the so called Hero Props can be found, these are props that stick out from everything else in the level and has a strong identity, ranging from a 3d model of a huge windmill to the sound set of a raging waterfall. When I sit down to make this list I try to cover all senses available to experience gameplay to enchance the level of immersion; shapes, colors, temperature, sounds, social interations et cetera. With time budget in mind it is important to spend a fair amount of time in this phase and thoroughly think through what assets you need. During this stage it also a good idea to make some very basic lighting to cover up the most important areas.

Together along the Hero Props you will find assets that tie-in with the objectives of the level, varying from levers/buttons that opens doors, vending machines that gives you soda, medical boxes where you find the vaccine et cetera. Once again this is very related to the setting of the game yet it is important to know that these props are on top of the priority. In many cases objective related assets have higher priority than Hero Props, but that is case specific.

During this phase assets needed for blocking are created as dummy-assets which uses correct scale so that they can be replaced easily when actual assets are brought to light.

This phase is something that never really starts or ends, it is done consecutively for the full length of the pipeline yet before the Level Dressing starts rolling on full speed. The game should be play-tested in order to find issues that needs fixing or improvement as it can severely impact not only the layout but also the assets you need for the level.

Placement of dummy art assets have now been going on ever since ‘Assets and Objectives’-phase and it is not until now they are getting replaced with actual art pieces.

Near the end of the pipeline there is the packaging phase where the level needs to be wrapped up, all functions have to be working properly, all important assets needs to be in there and a couple of art passes have already been done. It should be mentioned that the level still needs quite a bit of work to be called finished but if anything the level should at this point be defined as ‘stable’.

Upon reaching this phase the level will not reach another one and you never really declare it entirely ‘finished’ as there is always something that could use some more love. Play-testing to hopefully only do minor changes if possible, art assets that needs trimming,lighting that needs tweaking, post processing effects which enchances the atmosphere – all of those things and more is applied and looked over in this final phase. At any point when the level have made a change it is important to do package it into another stable version and then continue polishing until the time budget is cut or simply depleted.