Traning and Boardcast Application for NASCAR
Made with Unity
We built an application that recreation of NASCAR game with real-world data for staff training and also live broadcast.
Role on the team:
Built UI with UI Element at Runtime (Preview)Package. Using the mesh draw API in the package to create data visualization tools. Reading through the code for the package on Unity’s GitHub and access the private filed in the mesh generation class with Reflection and achieved 900% performance increase.
Filtered 5hz car position data with custom Kalman filter. Fix/impute the incorrect/missing data point in order to produce a presentable data visualization for live TV broadcasts with 0.8s buffer.
Implemented an in-game recording component that enables users to record video clips with Avpro plugin and FFmpeg. Utilized nvenc or intel quick sync based on users’ hardware to maximize the benefits from FFmpeg.
Implemented a shadow and fog controller in HDRP and dynamically adjust fog and shadow when the camera was on an extremely low FOV and looking at objects from far away.
Communicate with the team artist and guided artists through the render pipeline upgrade workflow.
Miscellaneous:
Implemented a camera system that auto-adjust its position when tracking multiple targets.
Upgraded the existing Input system to the new Unity Input system.
Implemented object spawning pool for a feature.
Live car data comparing with historical car data for FOX live stream with 0.8s buffer.