Descent Into Darkness

About the project:

Descent Into Darkness is a first-person linear horror game created in Unreal Engine 5. Players take on the role of Azrael, a woman accused of attempted assassination, navigating a blood-soaked underground facility in a desperate bid for survival. The game focuses on creating a narrative-driven horror experience, combining immersive atmosphere, survival challenges, and puzzles as players unravel the mysterious events within the facility.

Design Process

Level Design:
I prototyped a large, complex level, refining the design through multiple iterations to perfect the game’s flow. Each iteration improved the pacing and player experience, ensuring that the environment balanced exploration with narrative progression. To guide players subtly, I used signposting, such as illuminated signs and on-screen prompts, helping players navigate the space while maintaining immersion.

Systems Development:
I blueprinted and implemented several interactive features, including doors, sound logs, and readable notes, to deepen player immersion. Additionally, I developed a simple save system, allowing players to save their progress, including variables like player location. These systems contributed to a seamless gameplay experience and allowed players to track their progress effectively.

AI Development:
I designed and implemented AI using behaviour trees and Blueprints. The AI was programmed to pursue and engage with the player dynamically, adapting to player actions such as hiding. Over time, I made the AI more complex by enabling it to patrol different areas and search for the player in hiding spots, creating a more tense and unpredictable horror experience.

Polished and Expanded:
Originally created as a master’s project for my degree, I revisited and upgraded the game by adding additional mechanics and improving the overall polish. I expanded the AI systems, added more interactive elements, and refined the level design to enhance player immersion and gameplay depth.

Additional Mechanics and Enhancements:
New mechanics were introduced to enrich the game’s experience, such as more intricate puzzles, expanded AI behaviors, and refined environmental interactions. These additions helped to create a more engaging and dynamic player experience, making the game more immersive compared to its initial version.

Features

Audio Logs

I created audio logs that the player can listen to in the game. These audio logs provide clues and background to the events happening.

Inventory System

I created an inventory system and interactable items that the player can collect and then use, or even examine.

Notes

I created notes that the player can collect to complete objectives, which also provide clues and background to the current story.

Objectives

I implemented objectives that point the player in the correct direction to progress through the game.

Blaze

Blaze is an intuitive AI created by using a complex behaviour tree. He patrols areas, pursues the player and when he catches you it's over.

Behaviour Tree

As described previously, here is an example of the code I used in my behaviour tree for Blaze.

Lockpicking

I purchased this asset from the UE marketplace to use in my game. I use it in a couple of the areas of my game. The player must find lockpicks to be able to progress and find hidden items or additional lore.

Demon child (Boy)

This little boy appears at random throughout the level, frequently causing the player damage to their health. The source of his power is not clear.

Locker

These lockers are scattered through the underground facility, mainly working as a way of being able to hide from Blaze. He can see you getting into them and if he does he will catch you and drag you out, so be careful.

Load and Save game

I implemented a load and save game system for the players location due to the size of the level.

Note examination

Players can get a closer and easier way to read notes by clicking the examine prompt displayed on screen when picking up a note.

Item Examination

Players can rotate items for possible clues, or to examine meshes. They can also find out what the purpose of the item is, an example being a bandage will health 25 health.

Post-Mortem

What Went Well:

Atmosphere and Immersion:
The game’s atmosphere, created through a combination of sound design, lighting, and level design, was highly effective in evoking tension and fear. The dark, claustrophobic environments and eerie sound cues contributed to the sense of isolation, making players feel constantly on edge.
Iterative Level Design:
Through multiple iterations, the level design struck the right balance between exploration and narrative progression. Each iteration improved the game’s flow, ensuring players felt guided without being over-directed. The use of signposting helped players intuitively navigate the environment.
AI Complexity:
The AI system, using behaviour trees and Blueprints, worked well in creating dynamic encounters with enemies. Players had to stay alert and adapt to the AI’s unpredictable behavior, which added depth and tension to the gameplay. The ability for the AI to patrol, detect hidden players, and change strategies based on player actions added another layer of challenge.

Challenges Faced:

Balancing AI Difficulty:
Finding the right balance between AI difficulty and player engagement was a significant challenge. Early iterations either made the AI too easy to outsmart or overly difficult, leading to frustration. This required numerous tweaks to ensure that encounters remained challenging but fair.
Scope of the Project:
Initially, the AI systems and environmental interactions were ambitious for the scope of the project. Some features had to be scaled back or simplified to meet time constraints while still delivering a polished final product.
Time Constraints for New Mechanics:
Although new mechanics were added during the upgrade from the original master's project, some features, such as more advanced environmental interactions, had to be postponed due to time limitations.

What Could Be Improved:

Deeper Narrative Integration:
While the game successfully conveyed a sense of mystery and tension, there’s room to expand the narrative further. More character development and lore could be added to deepen the story and player connection to the world.
Variety in Enemy Encounters:
Additional enemy types and more varied AI behaviors could further enhance the challenge and unpredictability of the game. Expanding the enemy roster would keep players on their toes and increase replayability.
Enhanced Visual Feedback:
While the visual design achieved the desired eerie atmosphere, more visual feedback during combat and key moments, such as UI indicators or special effects, could improve the overall player experience.

Key Takeaways:

Polishing Mechanics:
Working on an existing project taught me the importance of refining mechanics that are already in place before expanding on them. The upgrades to the AI and the addition of new mechanics were successful because they built on a strong foundation.
Balancing Ambition and Feasibility:
One of the biggest lessons from this project was learning how to balance ambitious features with the available time and resources. This helped me prioritise core mechanics and ensure they were polished, rather than overextending the project’s scope.
Iterative Development is Key:
The iterative process, particularly with level design and AI, was crucial in achieving a polished final product. Each round of testing and feedback allowed me to refine the game, ensuring that the player experience was consistently engaging and challenging.