Ji Kim
Visual, Interaction, Experience Design
Currently open to new projects
2024
Personal Project
Final Project at UXD for the ARVR Certificate, NYU Tandon
Role
UX/UI, Interaction Design, Visual Design, Prototyping in Unity,
Overview
Terrarium is a spatial video gallery exploration that investigates how video browsing behavior changes when freed from flat screens. Rather than treating VR as a backdrop for 2D players, this project explores video as spatial objects in which depth, proximity, and movement serve as interaction primitives.
This work began with a 2D web-based exploration and evolved into an XR prototype built and tested in Unity. The goal was not to build a feature-complete product, but to understand how short-form videos behave when browsing becomes spatial.
01. Problem: Video Browsing is Still Flat
Most VR video players replicate familiar 2D metaphors
-
• Rectangular panels floating in space
• Theater-like environments
• One video at a time, enlarged for focus
This approach works well for long-form media.
But it breaks down for short-form, memory-driven videos because users want to
- • Browse many clips
• Compare moments
• Mentally recreate the atmosphere rather than focus on a single narrative
The core question quickly became,
“What if video browsing in XR was spatial instead of sequential?”
Current UX Problems & Pain Points
Most spatial video players enlarge the video dimensions, mimicking a theater setting. Often, users are too close to the video surface. A 360 video player stretches the media to texture-map it onto a sphere. It degrades the asset’s resolution and can cause motion sickness.
Most spatial video players adhere to the traditional 2D video player design principles. Bottom line is that most existing video players are designed to accommodate long-form media like movies.
Most spatial video player apps use a realistic or skeuomorphic screening environment as their base space, such as a movie theater. This reproduction of the physical world limits the ability to watch only one video at a time.
User Behaviour Analysis on Different Video Contents
Do users behave differently depending on what they watch? What are the differences?
Browse the contents briefly. Watching individual content takes longer.
Occupy the space by sitting or lying back.
Experience is semi-public or public.
Browse content longer. Watching individual content takes less time.
The player's device is handheld.
Experience is personal.
02. Precursor Exploration: When 2D Hit Limits
Before moving into XR, I explored this concept through a web-based experiment, www.sceneries.site
The site allowed multiple videos to coexist on a single plane, emphasizing simultaneity rather than playlists.
Although visually compelling, this experiment revealed the hard limits of 2D interaction.
• Hierarchy had to be flattened into scale and overlap
• Browsing and focus faceted within the same plane
• Navigation relied entirely on scrolling and clicking
This clarified that the challenge was not visual—it was structural.
The experience required depth, proximity, and embodied movement, which directly led to the exploration of the concept in XR.
03. Core Insight: Short-form Video Behaves Differently in Space
User behavior research revealed a key distinction.
• Long-form Video: Sit, settle, focus
• Short-form Video: Browse, skim, loop, compare
Short-form video naturally supports,
• Simultaneous playback
• Non-linear attention
• Ambient presence rather than full focus
This insight justified a spatial approach in which multiple videos coexist and are available for individual attention upon request.
UI Analysis on Different 2D Video Platforms
Video player platform UI and Key interactions
Click/Touch GUI
Focus on watching and interacting with the selected content
Gesture forward, fewer GUIs
Focus on navigating through content
Competitive Analysis on Video Player Layouts
Spatial Video Player Key Interaction Sketches
04. Spatial Design Principles
Based on early exploration and testing, the project was guided by the following principles.
-
Multiplicity Over Sequence
Multiple videos are played simultaneously to simulate memory density.
-
Depth as Hierarchy
Near videos invite focus; distant videos remain ambient.
-
Surface Shape Matters
Elliptical video surfaces reduce edge distortion and feel portal-like rather than screen-like.
-
Environment as Context
Passthrough grounds digital content in the user’s physical space, improving comfort and presence.
- Physical
Movement Over the Controller
Room-scale movement replaces joystick rotation to reduce disorientation and increase embodiment.
05. Prototyping & Iteration in Unity
The project evolved through four rounds of Unity builds, each driven by qualitative user testing.
• Rectangular video surfaces
• Skybox environment
• Stationary boundary
Outcome
Felt detached and screen-like. Reinforced 2D habits.
• Elliptical video players
• Reduced environmental framing
• Simultaneous playback
Outcome
Stronger spatial pull. Videos felt like portals, but sensory overload became an issue.
• Passthrough background (screen-recording video above doesn’t capture it)
• Room-scale boundary
• Interactable video objects
• Sound muting for sensory control
Outcome
Users felt grounded, playful, and more willing to explore. Spatial interaction became intuitive rather than novel.
06. User Testing Highlights
Qualitative testing surfaced several key insights.
• Elliptical shapes invited the user's approach and exploration
• Black void environments felt detached and flattened the space
• Passthrough increased emotional connection and comfort
• Audio requires a hierarchy to avoid overwhelm
• Users wanted focus modes without breaking simultaneity
These insights directly informed the final prototype decisions.
07. Conclusion
Learnings
• Spatial UI requires restraint more than features
• Depth, distance, and scale communicate hierarchy better than an explicit 2D GUI
• Audio is as important as visual hierarchy in spatial systems
• Familiar metaphors slow down spatial understanding
• Prototyping in XR is essential—many issues only appear in motion
Next Steps
• Introduce focus modes without collapsing simultaneity
• Explore hand-gesture input for more intimate interaction
• Support annotation through voice and text
• Extend the system to Apple Vision Pro