Rain road is based on the storybook, Rain Road, written by Michaella Moon and illustrated by Joanne Kim.
The story follows the character Jane. Jane worries about her crops in the drought but manages to find a way to convince their friend Cloud to help them water their crops. As they skip along the dry desert; Cloud, overflowing with rain, trails behind Jane mesmerized by their singing.
This installation invites the audience to interact with the five umbrellas hanging upside down from the ceiling. Due to orientation of the umbrellas, the whole installation is flipped upside down. The sky, being projected closer to the space’s floor and the desert scenery towards the ceiling. Each of these umbrellas has its own photocell sensors, a light-responsive sensor, and arduinos built onto the clear plastic umbrellas. When a light source is shined into the tops of the umbrellas, it triggers several elements in the installation to change. This one action sets into motion both an aural and visual change. Each umbrella has its own characteristic audio effect that is added to the song from a playlist of rain-themed songs.


The playlist currently includes 11 songs:
BJ Thomas- Raindrops Keep Falling On My Head
Blind Mellon- No Rain
Creedence Clearwater Revival- Have You Ever Seen The Rain
Gene Kelly- Singing In The Rain
Guns N’ Roses- November Rain
Lady Gaga- Rain On Me
Led Zeppelin- The Rain Song
Mariah Carey- Through the Rain
Prince- Purple Rain
The Beatles- Rain
Tim McGraw- She’s My Kind Of Rain

With the same light shining motion, there is a change in the projection mapped, AI-generated scenery. These images were created using a VQGAN+CLIP AI image models. Through this model, one is able to generate images using carefully selected texts and sentences. Different sentences and words generated different images. This installation uses four images generated by a VQGAN+CLIP model. There are two sets of these images that represents the before and after of dry desert. Even though similar keywords were used to create the images, one set of images included more greenery and a sense of abundance than the other two. The audience interaction switches between the two pairs.

This installation uses multiple programs. Arduino, to control the sensors and LED strips. Processing, to control and facilitate the data (as a hub that connects the sensors to all the other programs via OSC). MaxMsp, to control the playlist of songs and audio effects. TouchDesigner, for all the visuals and projection mapping.