Arcane University: Level Design
This is the main page of the Beyond Skyrim Arcane University Level Design tutorials.
Contents
Expectation
A Level Designer is a person who can create the physical shape of the lands and places where the stories imagined by Writing and Concept Art and team discussion take place. These creations must also be compatible with the Music Composition, they deserve their own Sound Effects and they must be compatible to the planned methods of Implementation for the quests and activities later on.
General Workflow
Hello BS members and Skyrim modders,
there are a few basic questions when working on level design. By answering these appropriately, and having some practice, it is possible to create a beautiful dungeon or world space. This tutorial is written for people who have built their own first layouts and know the basics but want to achieve more.
1) General Considerations Level design is the stage between designing the world on paper and implementing the mechanics of it; it is the intermediate step of realizing your vision.
As with all sections of development, the criteria of good and bad will be determined by how things as a whole come together and work out. And “works” means from testers/players perception, so it is always helpful to get quickly into the game tests and get third person opinions.
The ties between especially the writing – level design – implementation chain are very strong. Bad writing annihilates every chance on good level design (if the level design follows the writing) and obviously the implementation afterward. Also, “bad” level design that does not fit the written story or the implementation makes a splendid player experience impossible. For that reason, good level design needs a basic understanding of the writing and implementation, and early testing by others.
In general, the level designer should be in close communication with the responsible writer and implementer, or himself being the same person; and at least another equally (or more) experienced person should check the work and add feedback, suggestions, constructive criticism. This requires a respectful but honest and rational communication. On the process level, it also requires running through stages as a team, like brainstorming, filtering and taking decisions. On sub level, the level design itself also follows a chain of stages. This can be done by either a responsible group or a professional director or department lead who is aware of all details. The optimum would be, if all the department members checked it, everyone knows about the background, and all responsible persons; and after a detailed discussion, a decision can be achieved that convinces everyone and takes inspiration from all suggestions.
2) Steps of Level Design The level design workflow follows stages. These could vary from each person, but to give you a general idea, I took a preference:
a) interior
Concepting Phase (story, the concept of playstyle, enemy stories, type of dungeon) Layouting Phase (having a closed layout without holes and getting stuck) Basic Cluttering Phase & basic light sources/lighting template raw (adding first objects to get a better impression of the design) Test and basic implementation (if the functionality is not yet proven) detailed clutter (going into a certain direction of design after having enough input, repairs and reworks at this point might improve things) encounters, traps, activators (making the interior functional for the test, adding the interactive elements) Navmesh, Lighting Template fine-tuning, Sound; load optimization, north-marker, World-Hook Up, Name of the cell, coc marker, Markers etc (adding the full game integration data) extended tests and polishes (expect this to take a lot of time until it feels right) b) world space
height map district layout regional layout basic clutter regional weathers, music interactive elements (as above) detailed clutter 3) Planning the Design Most work happens on a paper with a pencil, or a google doc, or your head (if you know what you are doing, but being able to discuss this with your team always helps). Standing on the brink of creating a new interior for example (but this also applies to exteriors) the first questions you should ask yourself are the following:
What do I want to portray? -> Story, Peoples, Quests, Atmospheres. you need some game environment (we have Skyrim’s) and a world/lore to realize at this point. the main challenge is to find a way to appropriately introduce this topic later to the player, as you need to consider the player’s conception; how he enters and leaves a place, how it feels from the first-person view, what relations he could build (or cant) with NPCs in your location; and how immersive that is and so on, you see this is already a heavily writing-connected topic… Most importantly, note that you can only transmit a limited amount of information; so relevance is key. Who would be inside when the player enters? -> “residence topic” consider exactly who or what needs to be placed inside and what these encounters require you to do from a level design point of view; if you have a huge monster, for example, having narrow sewers is no option, as the beast would get stuck and this loophole would quickly get exploited by players; which would ultimately break the balance and immersion of your aspired game experience.
Who created that Interior, what was the applied process, which tools have been used? -> in your game and your world history your dungeon is somehow connected to its environment. It was created by certain natural forces, peoples or beasts, and this should be reflected by its composites and its structure. This can greatly aid you in getting forward with a meaningful, realistic and believable design.
What happened to the Interior until then? -> the next point is logical: when the player arrives, things happened. Make sure you portray these things as part of the environmental storytelling by structures, objects, and damage or intertwining materials (Aging, weather, wars, Dragon attacks…).
What would the basic structure be like? -> some civil engineering and material science and maybe socio-historical economy on small scale here 😉 (does it carry its own weight, what kind of material is it, how durable is it, what stretching still makes sense, would the tunnels collide, how much/deep “digging under the earth with elder scrolls lore resources and magic is realistic”?) Make sure that your small village of farmers does not possess a heavily fortified castle exterior unless you really have good reasons for that! As a reminder, you need either decades of rich lords and/or big communities or strong magic to build such structures.
what would I want to show with it? -> (feeling, mood, quest event) Your interior should clearly follow a function and a functionality. Make sure that this purpose is followed and the player intuitively can understand that.
how I connect it towards world space / other interiors? -> (load doors, connection network or “linear run through” consider what you need: a circular path, an expanding labyrinth of instances, an open world or a linear street of dungeons? Think about this carefully because it influences the density of “distribution of content” in your world. Also, obviously ice caves should not be placed in deserts; so check if things fit together, and also consider how much “player convenience” (hidden exit after the boss was killed, shortcuts, etc) you want.
Can I make a sketch and draw it? -> the first best thing to do after you have a frame for your story, is to actually sketch the layout with pencil and paper. “think in multiple ways” you can use your haptic impression with the pencil to realise “empty areas” and “choking points” or maybe you can achieve something more organic if you have a blueprint before touching very linear wall segments. Only playing with linear wall segments without a previous thought on an organic layout often leads to a design catastrophe. You want it to be relevant, unique and definitely not too big!
how does it connect to the player’s quests, how could it be interesting to him/her? -> you don’t only want your interior to be fitting to the worlds environment and story, but also to have some intrinsic motivation for players to go there. Or something unexpected, or better both. Make sure you have room for that!
Do I use the same materials as in the surrounding exterior or connected interiors? Should I use 2 or more sets? -> (cave+mine+imperial fort? Or rather stick to the main idea of the dungeon?) be careful not to visually “overload” your dungeon with various assets that don’t look very much like they belong together. The visual impact should be scored by an organic unique shape and a (fake) variety of assets along with unexpected and before unseen progression, not a big pit full of various rubble. In fact, monotony if done right can amplify certain aspects of the player’s awareness of the level design.
Where can I place events, ambushes, loot, “safe zones”, hints for the story, encounters, traps, lights, give the player nice detailed Points of Interests or points of sight? -> along with the implementation, this is maybe the most important. If your interior or world space cannot offer spots for what you need to place your gameplay, no matter how beautiful it is, it is useless. Also, remember that it is not very pleasant to enter a dungeon and be attacked immediately. This might cause scenarios where people are kept in the loadscreen and already harassed by enemies and dead by arrival. (depending on game engine and implementation)
4) Basic Rules of Execution There are some basic “unwritten rules” that you can follow if you want to avoid a bad impression: mention the scaling size! 0.5 < “your object” < 2.0 (because of texture resolution) Scaling under 0.5 will drop fps; or over 2.0 will look miserable in game. A 1024×1024 texture would practically be reduced by half. Especially objects using vanilla resolution textures will appear super low resolution and break the visual quality level, which should be held consistent during the game to address both optimisation and appeal.
Workarounds: Sometimes you have to use especially big meshes with low texture resolution to achieve a certain composition of space. There are a few ways how to work with them without making the low resolution too obvious:
If you can cover the low-resolution texture with other objects, so the low-resolution surface is noticeable.
If the surface is not seen close up, and you have some distance between the players closest possible position and the object, it is also less noticeable.
If you have a higher resolution texture than vanilla, the resolution will not be that low.
On the other hand, regions with a low object count might allow small meshes with higher resolution. Still, it is suggested to test the impacts on the engine and game stability, which you can easily do by checking fps when loading the area and running around there. This is just a general guideline like everything here and depends on your case of application. (thanks anathem7x for the hint of exceptions)
Hate Duplicates! Better do not get used to duplicates. Try to be original and unique whenever possible. The factor of composition, the combination of things allows for new designs more than you would have thought when doing the first concept sketches or thoughts. This might take time but it is worth it.
Balance the density of Objects! Don’t use too few or too many objects; you need to find a balance of amount and density of object placement. The two extreme examples are “an empty floor” and “a full room”. Both would look awkward. At least, the amount of clutter should be balanced and distributed evenly across the area. Try to stay below the suggested object count (see that percentage value at top of render window in CK?) More objects also mean more havoc chaos (unless you add the unhavok script, the checkbox does not always work) and more does not always mean better. On the other side, more objects allow more expression of layout and design and unique composition. It is a balance between detail and structure.
unhavok movable objects: -> to the movable object: double-click on a reference object in render window, go in the little boxes below and at the bottom right check “do not Havok settle”; sometimes this does not work, so you can additionally turn the tabs to “scripts” and click “add script” type in the search bar of the box with all scripts “defaultdisablehavokonload” (search for *dis*havo is quicker) and double-click this script. then its added in the “scripts tab“. then you would double-click and open this script to set properties. You would click “edit properties” and leave the box unchecked, you will see that the boolean value is now set to “false”. Now you can do so for Activation, Hit, ZKey and Partnerobjets. These are 4 ways to activate the havok of the object. If everything is set to false, it will never havok (gravity ignored). then ok and ok. this item will definitely not fall around if the player wants to suddenly Fuz-Roh-Dah in the wine cellar. you may want to do this with your first movable object, and then duplicate it, so the changes are copied with it and you only have to do this manually once. Avoiding Dead Space! Get a dynamic balance in the space; use all spaces, especially heights and ceilings. People who are new to level design often are impressed by the spacial freedom it offers. They create interiors with a big empty hall above the players head. This gives a quit nice impression of spacial dimension, but for a finished game/ mod it has a few downsides that speak against a generous ceiling structure and vast halls: a) often, it is unrealistic that people would build it that way (economically, statically, storywise) b) it is unrealistic this would stay unharmed over the ages from decay (the aging process would destroy big structures quicker) c) it does reduce the tension of a narrow tomb etc (the first person impression will be very limited, certain feelings cannot be achieved anymore) d) it might force you to clutter/redo areas that nobody will use in the game (you will have to cover all the area somehow to make it appear unique) e) it might kill the visual aesthetics in general. (it might feel completely out of place compared to the environment) There are still good reasons for meaningful structures to be huge, but they need good reasons to keep a certain balance of sizes in the design consistent.
Avoiding boring Space! very close to the above, for example, be careful not to create long hallways with flat, unspecified and generic ceilings. (thanks uberman for the hint)
Fake Spaces, Choices, and Decay! Spark the player’s interest and imagination by a non-linear approach that implies that “this dungeon was actually bigger but this way is blocked” saving you object count and development time and also adding realism and non-linearity. Point out-blocked/hidden areas.
stay non-symmetric and organic! (except for places that require it) avoid any linear and symmetrical level design wherever you can (unless you work on cyrodiils palace maybe 😉 )
Style Kits don’t like kitbashing between each other Use different kits only to a believable extent in the same region/interior; (consistency, not a surrealistic world)
Optimisation load room borders, occlusion planes, check unrequired objects, remove stuff from the conception phase and clean up carefully; if you have too much to load in an interior, making a second interior is always an option, as there are limits to our beloved CK engine. Try to remove or replace small scaled objects (high texture density), FX effects where they aren’t needed and objects that are only 30% or less visible (because the rest is outside the surface).
Eternal Gravity and Out of Reach -> checking that nothing is “falling under the ground”; (resetting in the center of the cell later on load!) and that also nothing is outside of + / – 30.000 units in all x, y and z dimensions, as this can result in issues.
NAVMESH TEST with Overlay keeping at least always a walking corridor (test with navmesh). Find a good balance between the count of vertices and resolution of your navmesh map.
Additional Info acoustic spaces for different acoustic impressions; North marker for the map; coc marker, x-marker heading for teleports; consistent cell name and hooked up map markers; music playlist, trigger boxes and implementational preparations…
Light Sources make sure you use beams, spotlights, omni-emission light sources and shadow creating light sources as suggested in tutorials; as shown in the tutorials. (thanks to Dimonoider for the hint) they are specified here:
Click and drag to move
Light might also guide the player through your cave, as sounds, and it greatly influences the feeling as music does. Also, notice that there is an image space dropdown window in the cell view window. The right setting can go a long way in helping to define the proper mood for your interior. (thank you to TheBlackpixel for the hint!)
Lighting Template (not the light sources and FX, but the cells “background lighting”) be very careful when using the lighting templates. rather choose all black, if not sure, and gradually improve brightness. Because the background lighting can come off unimmersively pretty quickly. rather go with low values; especially if you plan to have realistic light sources etc.
Loot lists and items consider if the standard container with standard loot lists do fit in;
Activators & Traps consider player-interactions like lighting a fireplace or climbing a rope/ladder, mushrooms etc. That needs the appropriate space.
Build a Stage for your Encounters / Actors consider your dungeon as a stage for actors who you place later, that means, plan the encounters and patrols and ambushes (like traps and puzzles and collectibles) already at conception stage or layouting stage to distribute them consistently and make the area/dungeon feel more “smooth and natural”.
(most importantly) Stay open for inspiration, feedback, criticism even after the “duty job is finished” check all available sources of content -> historical/archeological sources and all available sources of technique like the Creation Kit Wiki. Link: https://www.creationkit.com/index.php?title=Main_Page
Quality is made by finishing a dungeon, polishing it and going deeply into the thoughts behind it. you will find that, whilst the first stages you have a lot to lose, the last stages are those where you really gain something. So you should estimate to spent more time on the later stages, even though your “duty job” is already done; but at that level it becomes fun and you can actually become creative with tweaking the overall composition as a whole.
Thank you very much for your attention! I hope this helps! Come to the Arcane University, the teachers are great. Oh and every feedback is very welcome, so we can continuously improve this little guideline! Have fun!
Here is a random selection of various artist’s work that has been shown in our previous newsletters. Check the newsletters to find out more. It might give you a first impression and some inspiration, and we would be happy to receive your application to join us.
