loader image

UMass Virtual Tour

Campus tours meet the Web3: What if exploring a university could be gamified to the point that it feels like you’re actually walking around in-person?

Scroll to
View More

mouse

Traditional virtual college tours are done using technology similar to Google Street View: 360° images coupled with point-and-click arrows. We wanted to bring this concept to Web3 with real-time computer-generated imagery, photorealistic graphics, branching dialogue trees, and digital twins. However, this presented a challenge: how can such a large and open area be digitized and showcased realistically while being performant on multiple devices, and have legroom to run additional gamification mechanics.

Furthermore, I wrote a branching dialogue script, cast several voice actors, recruited models for the visual imagery of the tour guides, sculpted the terrain topology, designed the tour path, and managed a team of five people that developed some of the more intricate mechanics. We also implemented the Interactiva Engine in the second iteration of the project that helped us adjust the quality level to the user’s hardware, add tutorials, responsive UI, and other interactable elements.

Real Actors
We decided against using 3D characters due to the inherent intricacies of driving detailed, talking, and realistic-looking avatars that showed emotion. Instead, we opted for static snapshots of actors in diverse positions and emotive gestures. The result was a dynamic, RPG-inspired, guided tour with character that let their personality go beyond the tech.
1:1 Layout, props, and landmarks canción
The entire central part of campus had to be modeled, including the WEB Du Bois Library and the Campus Pond. Terrain details were sculpted by hand using visual references of the topology, while vegetation was replicated in SpeedTree and placed using terrain tools.
AI Ducks
The ducklings waddling and swimming around the campus pond are one of most important highlights while visitng campus. We replicated this in the digital twin by programming an AI duck that could swim, walk, avoid the player, and even play recorded sound effects of their real-life counterparts.