Environmental HDR Demo App

Demo and Session App for Google IO '19

Environmental HDR is a light estimation API that released in ARCore version 1.10. It uses a deep learning-based method to infer 360 HDR lighting from a limited field-of-view LDR image captured from a smartphone’s camera.

I had the opportunity to create a demo app to show off the new feature which ended up being presented on stage at Google IO ‘19. I was in charge of the project and ultimately responsible for delivering the app. On the front-end side my work consisted of designing the UI and implementing the behavior of the buttons.

  • Company :

    Google

  • My Role :

    Technical Artist

  • Year :

    2019

App in Action on-Stage

The app opens with a windowed AR view and three main buttons. The presenter, Konstantine, taps on the screen to spawn the rocket. The three buttons on the bottom are for: opening the light drawer, toggling ground-plane grid, and clearing spawned objects, respectively from left to right.

the Lighting Drawer

The light drawer (row of icons on the left) allows the user to toggle the three components of Environmental HDR: directional light, ambient spherical harmonics, and HDR cube map, respectively from top to bottom. The "light" button toggles between the new Environmental HDR lighting method and the standard lighting method in previous ARCore versions.

in the Wild

This is a screenshot of the app in action. The lighting of the rocket convincingly blends with the soft lighting of the interior.

in the Wild

This is another screenshot of the app in action. The sharp lighting of the sun is imitated and the direction of the shadow is matched.

the Keynote