Ryan Stec

Infinite Sunshine is a single-channel generative audio-visual work about sunlight and its interplay with water, wind, and sky. Ultra slow-motion video, music, and field recordings are combined using generative principles to produce an endless, non-repeating experience, one that unfolds differently with each viewing.

The generative logic driving the work draws from the compositional strategies of Steve Reich, Brian Eno, and Terry Riley, particularly the concept of phasing, in which simple loops of different lengths produce near-infinite combinations. A large bank of long video clips cycles and slowly cross-fades, two or three at a time, creating a shifting image that moves between the qualities of a melting abstract painting and a fragment of a lake at the end of a perfect summer day. The audio layer combines water field recordings with sparse classical string accompaniment, structured through the same generative logic as the video components.

The footage itself is produced using a combination of high frame-rate cameras and machine learning interpolation systems. Cameras capable of capturing at 240fps allow for HD slow-motion with extraordinary temporal detail. Machine learning systems then extend this further, increasing both frame rate and resolution up to 8K. The results sit at an unusual perceptual threshold, the hypnotic complexity of light on water captured in the real world, rendered with the hyper-resolution of synthetic image production.

The work is built and distributed using TouchDesigner, a real-time visual programming environment that handles both the generative engine and presentation layer. This allows the piece to adapt across different contexts, from a single HD monitor with headphones to larger projection configurations.

Infinite Sunshine is designed as an enveloping and restorative experience; simple in its elements, complex in its unfolding.

---

Below is a tour via screen capture of the audio video system. The video navigates through some elements of the TouchDesigner file. It shows the maintenance screen version which shows a representation of the EQ system on the left of the screen that is controlled by the luminance value of different quadrants of video. It also show the multiple layers of video that make up the final composition.

The audio is composed of the distinct groups, two musical pieces focus on very basic harmonics and ambient feel and one field recording. Each bank of the three elements contains between 5-10 different recordings.

At different moments, I manually refresh different elements of the system to show some variations. The system however is automated and continues to evolve in perpetuity if left playing.

Below are several low res video only test outputs from the system.