Through the Depths of Time is a film and a virtual reality experience that moves through moments of human history as if they were memories suspended in space. Rather than presenting history as a fixed sequence of events, the work reimagines it as something fluid, unstable, and spatial, something that can be entered and experienced from within.

The work begins with AI-generated imagery, which I then transform into volumetric environments using Gaussian splatting through a custom reconstruction process. These environments are later animated in real time in Unreal Engine, where each scene shifts, dissolves, and reforms through particles, light, and motion.

I approach AI as a tool that opens up new possibilities for developing my own visual language. I am not a traditional 3D modeler, but I have always wanted to create custom environments, so I became interested in whether this technology could help me do that. Through this process, I started to treat images not as something fixed, but as something that can be rebuilt, transformed, and experienced as space.

I use Gaussian splatting and reconstruction not to scan reality, but to construct scenes that never existed, using synthetic multi-view imagery generated from AI images. This shift from capturing reality to constructing it became central to the project. Because the scenes are built as spatial environments rather than flat images, I was able to extend the work into VR, allowing the viewer not only to watch the film, but to enter the scenes and experience them from within.