Jianghao Li - Sisyphus's Worst Day

Watch the Trailer

Overview

Sisyphus's Worst Day is a USC Advanced Game Project developed over 10 months. It is a third-person puzzle game that reimagines the myth of Sisyphus in a humorous, modern setting. Players take on the role of an office worker guided by the gods, who must trap Sisyphus in an endless loop of mundane tasks.

As the Lead Technical Sound Designer, I oversaw the audio production process and ensured that every sound element was effectively integrated into the game through FMOD and Unity. I collaborated closely with four other sound designers to create an immersive audio experience that complemented the game's distinctive narrative and gameplay mechanics.

Sisyphus's Worst Day gameplay screenshot

Spreadsheet showing organized audio asset list with sound descriptions and implementation notes — FMOD Assets

FMOD & Unity

The audio implementation process involved using FMOD to create interactive audio events and Unity to integrate these events into the game. I developed a structured workflow that allowed for efficient collaboration between the sound design team and the development team, ensuring that all audio assets were properly organized and implemented.

The FMOD project was managed and maintained in Perforce, allowing for version control and collaboration among fellow sound designers. We organized and categorized all audio assets in FMOD, ensuring a consistent and efficient workflow.

Interactive Music & 3D Audio

The game featureed an interactive music system that dynamically changes music based on the state of the game.

There were 2 stages in the game, both placing stage (where the player sets up the puzzle) and playing stage (where Sisyphus attempts to push the boulder). The music would do a horizontal change, meaning the two fully mixed tracks would crossfade when the transition happens. This was achieved by using FMOD's parameter system.

All diegetic sound effects were implemented in 3D space using FMOD's 3D audio capabilities. This allowed for a more immersive experience, as sounds would change based on the player's position and orientation within the game world.

Sisyphus's Worst Day 3D audio settings — 3D Audio Settings

Custom audio manager in C# for integrating FMOD events into Unity — Audio Manager Snippet

C# Development

We developed a custom master audio manager in C# to handle the integration of FMOD events into Unity.

This script managed audio playback, transitions, and parameter control, ensuring that the audio responded dynamically to player actions and game events. It provides a centralized and encapsulated system for managing FMOD related functionalities. APIs were designed so that the engineers were able to easily call the audio manager to trigger sound effects and music events without needing to understand the underlying FMOD implementation.

Voiceovers

We implemented voiceovers for multiple scenes in the game. To ensure a clear technical architecture and practice a data driven approach, we designed ScriptableObjects in Unity to manage the voiceover lines and their corresponding audio files. This allowed for easy updates and modifications to the voiceover content without requiring changes to the codebase.

The ScriptableObjects automtically linked the voiceover lines to their respective audio files, ensuring that the game system would get the correct audio file for each line of dialogue.

To manage each voice data, we also created a VoiceOver ScriptableObject that contained a list of all the voice lines for a specific character or scene. This allowed us to easily organize and manage the voiceover content, making it easier to implement and maintain throughout the development process.

We developed a custom master audio manager in C# to handle the integration of FMOD events into Unity.

This script managed audio playback, transitions, and parameter control, ensuring that the audio responded dynamically to player actions and game events. It provides a centralized and encapsulated system for managing FMOD related functionalities. APIs were designed so that the engineers were able to easily call the audio manager to trigger sound effects and music events without needing to understand the underlying FMOD implementation.

One of the key challenges in implementing voiceovers was the bark lines.

The design team wanted to have a bark line that represented different moods, which were then paired with different emojis to create a more expressive and engaging experience for the player.

To help the engineer team achieve this, we used FMOD's built-in parameter and audio callback system to dynamically look for the correct mood label and then pass it to a thread-safe queue. The script that handled such playing behavior would then dequeue the label in the Update thread and execute the corresponding emoji behavior.