Independent · Technical Audio
Crowd Audio Manager
Unity + Wwise
A runtime crowd-audio system that changes how sound is represented as crowd size increases.
Watch on YouTube (opens in a new tab)At a glance
- Type
Independent technical audio.
- Engine
Unity.
- Middleware
Wwise.
- Core system
A crowd audio system that changes representation as crowd size grows.
- Control
A single Wwise
Crowd_SizeRTPC (0–100) scales behaviour; a MidChatter layer joins beyond 25.
The problem
Spatial detail at a crowd scale.
Individual spatial playback provides localised activity and definition while a crowd is small. As the crowd grows, the system changes its representation instead of simply adding more individual voices.
What I built
One system, changing representation as the crowd grows.
A runtime crowd-audio system that plays individual spatial emitters for a small crowd, then switches to grouped Left / Centre / Right crowd audio as the crowd grows — controlled at runtime through a single Wwise Crowd_Size parameter.
My contribution
Design, implementation and iteration.
- Audio-system design and sound design
- Unity implementation and Wwise setup
- Runtime parameter behaviour and spatial-emitter logic
- Testing, debugging and iteration
How it works
Representation by crowd size.
As the crowd grows, the system stops treating every character as a separate audio source and switches to grouped crowd audio.
≤10
Individual spatial emitters
Individual looping voices retain localised activity and granular positioning for a small crowd.
>10
Left / Centre / Right
Three grouped spatial emitters replace the individual representation and follow the crowd’s distribution.
>25
MidChatter
An additional chatter layer increases perceived density at larger crowd states.
Small crowds
Each character keeps its own looping voice while the crowd is small enough for that to stay manageable.
Transition
Past the individual-emitter range, the system switches to grouped crowd audio instead of continuing to add more individual voices.
Grouped crowd
Left, Centre and Right are spatial emitters, not static stereo channels — their positions are informed by the runtime crowd distribution and geometry, so the grouped representation still moves with the crowd. Beyond 25 members, an additional MidChatter layer increases perceived density on top of the three grouped emitters.
Wwise control
Unity crowd state and count are normalised into a single Crowd_Size parameter (0–100) and sent to Wwise as an RTPC, which drives crowd audio behaviour and intensity at runtime.
Result
Movement remains spatial as the representation changes.
The system preserves spatial crowd movement while changing representation as the number of active crowd members grows. This keeps the system manageable as the crowd grows instead of playing a separate looping voice for every character.