GDnD Wiki

GDnD Wiki

This site is for Ulster University students and staff. Enter the class access password to continue.

Game Interface Types: Diegetic, Meta, Spatial and Non-Diegetic

15 min read

Summary

Game interfaces can be sorted into four types by asking two questions of any interface element. First, is it part of the game’s fiction: could the character see it or know about it, or does it exist only to help the player? Second, is it located in the game’s 3D world, or is it drawn flat on the screen? Combining the answers gives diegetic, meta, spatial and non-diegetic interfaces (course diagram on interface types, see source-four-types-of-game-interface). The two questions come from Fagerholt and Lorentzon’s 2009 study of first-person shooter interfaces (Fagerholt and Lorentzon, Beyond the HUD, see source-beyond-the-hud), and the four labels are the standard vocabulary for discussing HUDs and in-world UI. The thesis itself uses slightly different names and six categories rather than four (see In the original thesis below). It complements the broader treatment of interface design on ui-design.

Key ideas

The two-question grid

In the 3D worldNot in the 3D world (on screen)
Part of the fictionDiegetic — the character can see or use itMeta — reflects the character’s experience, drawn on the screen
Not part of the fictionSpatial — placed in the world for the player onlyNon-diegetic — a classic HUD for the player only

The four types

  • Diegetic — part of the fiction and placed in the world. The character perceives it just as the player does. Example: the health indicator on the protagonist’s suit in Dead Space, which replaces a separate health bar.
  • Meta — part of the fiction, but presented directly on the screen rather than as an object in the world. It usually expresses what the character is experiencing. Example: a crack, distortion or blood effect across the screen when the character is injured. (In the original thesis, the closest example is a phone the character owns shown as a screen overlay; full-screen damage effects belong to a separate category, meta-perception. See below.)
  • Spatial — positioned in the 3D world, but not something the character is assumed to see. It helps the player understand where things are or where to go. Example: a floating quest marker above a non-player character’s head, or an icon pointing towards an objective.
  • Non-diegetic — outside both the fiction and the world; drawn on the screen purely to inform the player. Example: a conventional health bar, ammunition counter, minimap or score in a corner of the screen.

Presentation, not content

The categories describe how information is presented, not what is presented. A single objective, “find the key”, could be delivered in all four ways (course diagram on interface types, see source-four-types-of-game-interface):

TypeHow “find the key” might be shown
DiegeticThe character checks a physical watch that displays a mission update
MetaThe screen briefly flashes a warning effect when the character is in danger near the key
SpatialA floating marker shows which building contains the key
Non-diegeticA quest list in the corner of the screen reads “Find the key”

This is the most useful habit the model teaches: when designing a piece of UI, decide what the player needs to know first, then choose the presentation type that best suits the intended experience. The thesis makes the same point: it classifies the UI element as an “informational vehicle”, so a row of ammunition digits is non-diegetic even though the bullets it counts exist in the fiction (Fagerholt and Lorentzon, Beyond the HUD, p. 47, see source-beyond-the-hud).

In the original thesis

Fagerholt and Lorentzon first tried a grid with a diegesis axis and a spatiality axis. They dropped diegesis as an axis because it produced grey areas. Is Killzone 2’s blood on the screen diegetic? Is Mirror’s Edge’s Runner Vision, which turns usable objects red, non-diegetic? In their final model the two questions are “Is the UI element existing in the fictional game world?” and “Is the UI element visualized as a part of the 3D game space?”, and “diegetic” becomes an area of the grid rather than an axis (Fagerholt and Lorentzon, Beyond the HUD, pp. 47–51, see source-beyond-the-hud).

They then sorted the interface elements from their game studies into six categories on that grid (pp. 51–52, 73–75). Later writers reduced these to the four familiar labels. Marcus Andrews’ 2010 Gamasutra article “Game UI Discoveries: What Players Want” uses diegetic, non-diegetic, spatial and meta and credits the terminology to the thesis; it has not been ingested into this wiki as a source.

Thesis categoryFiction?3D space?Thesis examplesPopular label
Non-diegetic (HUD) elementsNoNoConventional overlays: health, ammunition, mapsNon-diegetic
Geometric elementsNoYesLeft 4 Dead teammate outlines, Mirror’s Edge Runner Vision and checkpoint beacons, Splinter Cell: Conviction objectives projected on wallsSpatial
Meta-representationsYesNoGrand Theft Auto IV phone shown as an overlay, GPS devices, terminalsMeta
Meta-perceptionUsually no; can be yesNoRed screen tint, tunnel vision, heartbeat audio, pulsing vibration; Killzone 2 blood on the “lens”Usually also called meta
Diegetic elementsYesYesDead Space spine health tube and holograms, Far Cry 2 hand-held map, GPS and wristwatchDiegetic
Signifiers (a sub-group of diegetic)YesYesSmoke from a weapon about to break (Call of Juarez), blood trails (Dead Space), fire on a barrel about to explodeDiegetic

Two points matter for students. First, “spatial” is the thesis’s “geometric”: it lies in the 3D world but outside the fiction. Second, “meta” merges two different ideas. Meta-representations are fictional objects drawn off-world. Meta-perception is the full-screen damage or health effect, which the authors say is “conventionally” not connected to the fiction. They treat it as making up for the broken perceptual link between player and avatar (p. 52). The common teaching example of meta UI, blood or a red vignette when hurt, is therefore meta-perception in the thesis, and fictional only when the effect is grounded in the world, as with Killzone 2’s blood.

In practice

Unity

Each type maps onto a familiar Unity implementation route.

TypeTypical Unity approach
DiegeticA World Space Canvas, mesh, emissive material or light parented to an in-world object (the character’s suit, a weapon, a wristwatch, a terminal)
MetaA full-screen overlay image, or a post-processing effect such as a URP Vignette or lens-distortion override driven by game state
SpatialA World Space Canvas placed at the target, or a screen-space icon positioned with Camera.WorldToScreenPoint each frame
Non-diegeticA Screen Space – Overlay Canvas (UGUI) or a UIDocument built with unity-ui-toolkit

In the thesis’s terms, the post-processing route in the Meta row builds meta-perception. A fictional device shown as an overlay, such as a phone or a GPS screen the character owns, is a meta-representation, and you would usually build it as a screen-space Canvas panel.

A spatial objective marker that tracks a world position on a screen-space canvas:

using UnityEngine;
 
// Spatial UI: keeps a screen-space icon over a world-space objective.
// Put this on a manager object, not on the marker itself (the marker is hidden when behind the camera).
public class ObjectiveMarker : MonoBehaviour
{
    [SerializeField] Transform target;          // the objective in the world
    [SerializeField] RectTransform marker;      // icon on a Screen Space - Overlay canvas
    [SerializeField] Vector3 worldOffset = new Vector3(0f, 2f, 0f);
    Camera cam;
 
    void Awake() => cam = Camera.main;
 
    void LateUpdate()
    {
        Vector3 screen = cam.WorldToScreenPoint(target.position + worldOffset);
        bool inFront = screen.z > 0f;
        marker.gameObject.SetActive(inFront);
        if (inFront) marker.position = screen;
    }
}

A meta damage effect using a URP Volume whose profile contains a red Vignette override:

using UnityEngine;
using UnityEngine.Rendering;
using UnityEngine.Rendering.Universal;
 
// Meta UI: the vignette strengthens as the character's health falls.
public class DamageVignette : MonoBehaviour
{
    [SerializeField] Volume volume;
    Vignette vignette;
 
    void Awake() => volume.profile.TryGet(out vignette);
 
    // healthNormalised: 1 = full health, 0 = dead
    public void SetHealth(float healthNormalised) =>
        vignette.intensity.Override(Mathf.Lerp(0.5f, 0f, healthNormalised));
}

Other engines

  • Godot: CanvasLayer nodes for non-diegetic and meta overlays; a SubViewport rendered onto a mesh for diegetic screens; Camera3D.unproject_position() to place spatial markers.
  • Unreal Engine: UMG widgets added to the viewport for non-diegetic HUDs; a WidgetComponent in world or screen space for diegetic panels and spatial markers; post-process materials for meta effects.

Evidence

  • The course diagram defines the two questions and the four categories, gives one example for each, and works through the “find the key” scenario. It stresses that the categories are “design tendencies, not fixed rules” and that the central lesson is to choose the presentation that best serves the intended player experience (course diagram on interface types, see source-four-types-of-game-interface).
  • The diagram credits the classification to Fagerholt and Lorentzon (Beyond the HUD: User Interfaces for Increased Player Immersion in FPS Games, MSc thesis, Department of Computer Science and Engineering, Chalmers University of Technology, 2009, written with EA DICE). The thesis confirms the two questions, phrased as whether the element exists “in the fictional game world” and whether it is “visualized as a part of the 3D game space”. It names six categories rather than four: non-diegetic (HUD) elements, geometric elements, meta-representations, meta-perception, diegetic elements and signifiers (Fagerholt and Lorentzon, Beyond the HUD, pp. 48–52, see source-beyond-the-hud).
  • The thesis gives the Dead Space spine tube as a “fully diegetic” health bar and the Far Cry 2 map, held in the character’s hands in real time, as a diegetic alternative to a pause-menu map (pp. 36–38).
  • On immersion, the authors write that “a fully diegetic user interface is not automatically more immersive” than one that uses elements outside the diegetic area (p. 75). In their small user tests (five participants per test set), no participant said non-diegetic or meta-perceptual elements such as Runner Vision, the BioShock compass arrow or full-screen damage effects broke the experience, and genre conventions were given as the reason (pp. 63–64). Participants did object when information did not fit the fiction: the BioShock protagonist somehow knowing where to go, and Dead Space’s in-world screens referring to controller buttons (p. 57).
  • The authors tie the choice of category to the game. Narrative-led games should lean on diegetic elements and signifiers, while competitive, skill-focused games need the unambiguous information that non-diegetic and geometric elements give (p. 77). They also describe their guidelines as unverified and in need of testing with purpose-built prototypes (p. 97).
  • The course lecture series uses the same four categories, with examples such as the Dead Space spine health bar, the paper map in Far Cry 2 and screen-edge blood in Call of Duty (course lecture series, see source-player-psychology-lectures). Those examples are collected on ui-design.

Implications

  • Immersion versus clarity is a trade-off, not a ranking. Diegetic interfaces can strengthen immersion by making information feel part of the world (see presence-and-immersion and immersion-models), but they can be harder to read than a conventional HUD. A fast-paced game may need non-diegetic information precisely because it can be read at a glance. Fagerholt and Lorentzon go further: they argue that diegetic UI is not automatically more immersive, and that the level of competition, the fiction’s plausibility and the setting decide how much diegesis is worth (Beyond the HUD, pp. 75–77, see source-beyond-the-hud).
  • Guidance can be too explicit. In the thesis’s navigation test, every participant preferred finding the way through Mirror’s Edge with no navigational aid, even though aids made navigation easier. The authors suggest diegetic signifiers and guidance that becomes more detailed only when the player asks for it (Beyond the HUD, pp. 62, 82–84, see source-beyond-the-hud). This is one small test, so treat it as a prompt for playtesting rather than a rule.
  • Meta interfaces are a cheap route to the character’s state. They communicate injury, panic or intoxication without needing a physical object, though heavy screen effects can obscure play and raise accessibility issues (motion, flashing, colour-only cues).
  • Spatial interfaces are direct player-guidance. Waypoints and floating markers solve navigation problems quickly, but they can also train players to follow markers rather than read the level. Many designers therefore pair them with environmental cues or make them optional.
  • Non-diegetic HUDs depend on layout discipline. Because they sit outside the world, their readability rests on hierarchy and grouping (see gestalt-principles-ui and ui-design).
  • Mixing types is normal. Most games combine several types, for example a non-diegetic ammunition counter, spatial objective markers and a meta damage vignette. The model is a design vocabulary for deciding which element goes where, not a style that a whole game must adopt.
  • Virtual reality changes the defaults. Arguably, a flat screen-locked HUD is awkward in a head-mounted display, which pushes VR designers towards diegetic and spatial presentation (see virtual-reality-fundamentals).

Open questions

  • Some elements resist neat classification. An “augmented vision” mode that the character supposedly possesses but that highlights objects for the player could be argued to be diegetic, spatial or meta, depending on how the fiction frames it. Fagerholt and Lorentzon raise exactly this case with Mirror’s Edge’s Runner Vision. They file it as a geometric (spatial) element, but they accept that because players interpret fiction subjectively, an element’s place in the grid “can actually differ depending on how the representation is interpreted” (Beyond the HUD, pp. 48, 96, see source-beyond-the-hud).
  • The word “meta” is also used informally for interfaces that address the player directly or break the fourth wall. That is a different idea from Fagerholt and Lorentzon’s categories; see the note on ui-design.
  • How well do the immersion findings, drawn from first-person shooters, transfer to third-person, top-down or 2D games, where the camera is not the character’s eyes? The authors expect the design space to generalise, and one of their tests used two third-person games (Dead Space and Resident Evil 5), but they did not test top-down or 2D games (Beyond the HUD, pp. 95, 97).
  • Is the four-label simplification good enough for teaching? Merging meta-perception (usually non-fictional) into “meta” (fictional) makes the grid tidier but blurs the fiction question for the most common example, damage effects.

Practice

Classify every interface element in one screenshot. Choose a gameplay screenshot with a busy interface from a game you know well, such as an open-world game, a shooter or a racing game. Number every element on it, including screen-wide effects and anything in the world that exists to inform the player. For each element, record:

  • the answers to the two questions: is it part of the fiction, and is it located in the 3D world?
  • the resulting popular label: diegetic, meta, spatial or non-diegetic
  • the thesis category: non-diegetic, geometric, meta-representation, meta-perception, diegetic or signifier
  • what information it carries

Success test:

  • every element has answers to both questions, not just a label
  • every label follows from the grid (for example, nothing that is in the world but outside the fiction is called diegetic)
  • any screen-wide damage or health effect is marked as meta-perception in the thesis column
  • at least one element is flagged as ambiguous, with a sentence on how the fiction could be read either way
  • for one element, you say which other type could carry the same information and whether it would suit this game better

Extension:

  • pick one piece of information from your screenshot, such as the current objective or health, and write a “find the key”-style table showing it in all four types; or build the spatial ObjectiveMarker from this page in Unity and point it at a target

Self-test

  1. What two questions classify any interface element?
  2. Which type is part of the fiction but drawn on the screen rather than in the world?
  3. A floating quest marker hangs above a non-player character’s head. Which type is it, and why?
  4. Why is a row of ammunition digits non-diegetic, even though the bullets it counts exist in the fiction?
  5. What did Fagerholt and Lorentzon call the type now labelled “spatial”, and which two thesis categories does the label “meta” merge?
  6. Do Fagerholt and Lorentzon claim that a fully diegetic interface is always the most immersive choice?

Answers

  1. Is it part of the game’s fiction? Is it located in the game’s 3D world?
  2. Meta.
  3. Spatial: it is positioned in the 3D world, but the character is not assumed to see it. It exists to help the player.
  4. The categories describe how information is presented, not what is presented. The thesis classifies the UI element itself as the “informational vehicle”, and the digits are drawn on the screen outside the fiction.
  5. Geometric elements. “Meta” merges meta-representations (fictional objects shown off-world, such as the GTA IV phone) and meta-perception (full-screen effects such as a red tint, which are conventionally not connected to the fiction).
  6. No. They write that “a fully diegetic user interface is not automatically more immersive”, and that the level of competition, the plausibility of the fiction and the setting decide how much diegesis is worth.

Graph View

Backlinks