SomaSpace · M.A. Thesis, HXDI
A visual environment for returning to the present
SomaSpace reads your body in real time and reflects it back as gentle, generative visuals. Move quickly and the field comes alive, slow down and it softens with you, so grounding emerges through perception instead of instruction.
Final demo, the installation running live.Watch on YouTube: youtu.be/31B_y4Q9nCs
Earlier prototype, motion driving the particle field.Watch on YouTube: youtu.be/8ATLH7qLYsY
In 30 seconds
Wellness tools add steps and instructions at the exact moment attention is already exhausted.
A solo M.A. thesis installation where body movement, read by webcam, drives generative visuals and sound. No buttons, no menus, no instructions.
Two gestures only: hand distance shapes the visual and the volume, and opening the left palm switches between three graphics.
Refined through two build-and-test iterations, then exhibited at CSULB's Duncan Anderson Gallery, where visitors slowed down without being told to.
Why I made this
Starting from my own overstimulation
Like a lot of people, I spend most of my day on a screen, and I noticed how rarely I actually feel present in my own body. Notifications, deadlines, and endless feeds pull attention outward until tension shows up as shallow breath or tight shoulders, usually unnoticed until it becomes fatigue. Most tools meant to help ask the mind to work harder: follow these steps, breathe on cue, track your progress. I wanted to try the opposite. SomaSpace began from one question: what if grounding did not need instructions at all, and could instead emerge from the body itself, through movement, perception, and a space that quietly responds? That question became my thesis.
Abstract
Grounding through the body, not the mind
SomaSpace is my M.A. thesis in Human Experience Design Interactions. It explores how embodied interaction can foster grounding in an age of constant stimulation. Real-time body tracking translates movement and stillness into generative visuals that mirror the participant's state, so technology becomes a reflective companion that listens and responds rather than directing behavior. The project was built through the Double Diamond, and this page follows the structure of the paper: Discover, Define, Develop, and Deliver.
Design process
The Double Diamond
Before the project detail, here is how it was made. SomaSpace followed the Double Diamond, two cycles of opening up and narrowing down. The first diamond explores and frames the problem, Discover then Define. The second explores and delivers the solution, Develop then Deliver. The rest of this page walks through those four phases in order.
Explore widely: research, literature, and precedents on body-based grounding.
Narrow to a clear problem and the person it serves.
Prototype and test the visuals, sound, and mechanism.
Refine, install, exhibit, and reflect.
Understanding body-based grounding
The Discover phase asked a single question: what if the path to grounding started not with instruction, but with sensory experience? Most wellness tools begin with the mind, explaining how to relax, and add another layer of demand at the exact moment attention is already tired. SomaSpace instead treats the body as the interface, where small changes in posture, pacing, and pause carry meaningful signals about internal state.
Literature review
Four studies anchored the concept, each answering one core question about grounding, awareness, and nonverbal interaction.
Grounding is an embodied process, visible through posture, weight, and rhythm, and can be read without words. (Engelhard et al., 2021)
Awareness is a lived, ongoing process shaped by environment, so systems should support noticing rather than directing. (Kjolstad et al., 2022)
Attending to internal sensation reduces cortical load and lets awareness arise without language. (Farb et al., 2023)
Gentle, non-continuous visual feedback heightens awareness of action and guides attention without instruction. (Yu et al., 2024)
Case studies
Eight precedents, from art installations to everyday objects, showed a common pattern: grounding is most effective when a system responds gently to the body and invites slow, sensory attention.
Reflection guides grounding when it responds gently, not evaluatively.
Responsive light synchronizes a group's perception and supports calm.
Grounding strengthens when a system responds to presence, not performance.
Balanced frequency and soft rhythm ground people without cognitive tasks.
Predictable, repeated motion releases restlessness and restores focus.
Texture invites slow, hands-on attention that reduces tension.
Steady tempo and tactile counting deepen presence over time.
Shifting attention from thought to sensing anchors a person in the moment.
Comparative reflection
Across all eight, one pattern held: grounding happens when systems respond gently to the body and invite slow, sensory attention. Whether through visual reflection, rhythmic light, tactile repetition, or simple noticing, each example reduces cognitive pressure and reconnects people to themselves through perception rather than instruction.
Finding of the Discover phase
Grounding begins when people sense their own posture, movement, and rhythm, without needing words.
People settle when cognitive load is low; simple sensory cues beat verbal steps.
When the environment mirrors movement softly, people feel grounded without pressure.
Framing the problem and the person
Modern life is filled with constant visual and cognitive stimulation, and most tools meant to help require guidance, steps, or progress to measure, adding effort at the exact moment users have little attention left. The problem is not a lack of tools but the high interactional load built into them. There is very little between tools that over-direct and tools that ignore the body entirely, so SomaSpace sets out to be a listening tool: the body leads and the environment follows.
Project statement
To create a peaceful digital space that helps people feel more grounded and present, users interact with a screen that changes in real time based on how much they move. As they slow down and become still, the visuals gently respond, helping them feel calmer, more relaxed, and more connected to the moment.
Target audience and expected effects
The audience emerged from the research: people carrying constant stimulation who need simple, sensory relief. The expected effect is a sense of relief that comes from quiet, perceptual interaction rather than direction, so they can reset attention and return to their body without feeling judged.
People who move fast through the day and want short sensory pauses that need little instruction.
Students, designers, and knowledge workers seeking brief resets between tasks.
People in galleries or on campus who welcome a quiet moment they can enter and leave easily.
Persona
- Age
- 27
- Work
- Visual content designer
- Location
- Los Angeles, CA
- Traits
- Reflective, empathetic, occasionally anxious
- Tech
- High comfort
"I don't want the app telling me how to calm down. I just want to feel connected again."
Needs
- A brief way to reset attention between tasks
- A sensory moment with no steps or mental effort
- Space to simply notice, without being evaluated
Behaviors
- Juggles apps, tabs, and client messages
- Takes short breaks but usually scrolls or checks email
- Loves ambient light, color, and texture, and visits gallery installations
Frustrations
- Overloaded by guided wellness apps
- Finds most digital tools too task-oriented
- Little time or attention for anything demanding
User journey map
Mapping Jennifer's journey traced her arc across six moments, from curiosity and hesitation, into playful exploration, and on to slowing down, a grounding moment, and reflection, with her emotions rising as the visuals began to mirror her.
Data collection
A preference study asked how people actually want to settle, across visuals, sound, and their own reflections. The results were directional, and they set the course for every design decision that followed.
Visuals. Participants preferred graphics that expand and contract at a calm, steady tempo, with long fades, smooth transitions, and low-contrast shifts. Cool sea green and muted blue supported stillness, while soft golds and purples offered warmth and renewal.
Sounds. Many preferred natural, environmental sounds, described as steady and easier to relax into than spoken guidance. That pointed to a low-frequency layer between 80 and 100 Hz that anchors attention without demanding it.
Thematic analysis. Participant reflections clustered into five themes: immersion (the space felt absorbing and quiet), connection (movement felt mirrored, as if the space responded with empathy), awareness (users noticed their movement more clearly), trust (the system felt patient and nonjudgmental), and release (people felt clarity and steadiness).
Storyboard
The target experience, from a rigid app to a field that answers her movement and settles as she settles.
Finding of the Define phase
People are overstimulated, so keep the interaction simple and prioritize perception over steps.
Brief sensory moments that let people reconnect without being told what to do.
People settle when sensory cues mirror their presence softly and naturally.
Building the system
The Develop phase turned the research into a working system. Technology in SomaSpace is an extension of perception: it translates the presence of the body into visual behavior without friction, so users feel grounding through the visuals, not by learning an interface.
Inspiration
Inspiration connected the psychology to a visual language of texture, motion, light, and feeling: opacity gradients for depth, low spatial frequency so large soft forms fill the field, and temporal smoothness so long fades replace abrupt cuts. References like diffused glass, cloud chambers, underwater light, and long-exposure smoke, where motion reads as continuity rather than displacement, guided the look. Every iteration was checked against a small emotional vocabulary.
Long fades and slight motion blur replace abrupt transitions, so nothing competes for attention.
Motion reads as a continuous stream, drawn from diffused glass, cloud chambers, and smoke.
Low spatial frequency lets large, soft forms fill the field while small details stay quiet.
Technology: the pipeline
A lightweight, frictionless chain reads the body and renders a response in real time, tuned so users feel presence rather than technology. It exists to translate the body into visual behavior with no learning curve.
Captures live frames of the upper body and hands at chest height. A sensor for shape and motion only, with no identity recognition.
Detects pose and hand landmarks, outputting 3D keypoints for shoulders, elbows, wrists, and finger joints.
Smooths jitter, measures velocity, hand distance, and stillness, then maps those to visual parameters.
Renders the scene at room scale on a matte wall, with luminance tuned for low light so visuals feel soft.
Graphic visuals: three modes
Rather than commit to one aesthetic, I built three responsive systems, each mapping movement to motion differently. Users switch between them live by opening and closing the left palm.
A continuous stream that grows expressive with fast movement and softens as you slow. The baseline perceptual anchor.
Renders the body as a field of particles and leaves a stain-like trace, so continuity of movement becomes visible.
A loose, cloud-like field that disperses and reconvenes around subtle motion. Preferred by 80% of participants.
Fluid: the one mode with two color families
The fluid graphic is the only visual built in two tone families. Research split almost evenly between warm and cool for stress reduction, so instead of choosing one, I made both and let each person settle on whichever soothes them. Cool turquoise, sea green, and muted blue support stillness and inward attention. Warm amethyst, soft gold, and amber offer gentle activation and re-emergence.
System mechanism
The system runs on a simple loop. The webcam receives the body, MediaPipe generates the motion information, TouchDesigner reads it, and the scene renders back on the wall in real time. From there, everything the user does comes down to just two gestures.
The two interactions
There are no buttons and no menus. The whole experience is driven by two body gestures: the distance between the hands, and opening or closing the left palm. The first shapes the visual and the sound at the same time. The second switches between the three graphics.
Interaction 1 · Hand distance
Hand distance is the main dial, and it does two things at once, across all three graphics. It shapes the visual: bringing the hands closer gathers the field into a smoother, more cohesive form with slower, calmer motion that invites the body to settle, while spreading them wider opens it into a looser, more energetic form with dynamic, flowing movement. It also sets the sound: the closer the hands, the smaller and quieter the tone, and the wider they spread, the bigger and louder it grows. Both the visual and the volume are smoothed between a minimum and maximum, so there are no jumps, only a gentle expansion and contraction that mirrors the breath.
Fluid, closer and wider.
Particle, closer and wider.
Interaction 2 · Open and close the palm
The second gesture changes which graphic is playing. MediaPipe detects whether the left hand is open or closed and converts it into a clean 0 or 1, so the system can treat it like a button. Each open advances to the next graphic: one open moves forward, and repeating cycles through the three visual modes, fluid, depth map, and particle. It is a deliberate, discrete switch, kept separate from the continuous hand-distance dial so the two never interfere.
Sound, in more detail
The tone itself is a low-frequency ground sound between 80 and 100 Hz, a range that can be felt physically and supports relaxed, sustained attention. It is not music or instruction but a quiet layer that mirrors breath and reinforces the link between body and visual, present enough to notice and easy to ignore.
Movement traces (depth map)
On top of hand distance, the depth map adds one more behavior of its own. When the body is still it renders a clean form, and when it moves it keeps a fading trace of the gesture, so continuity of movement becomes visible over time.
Usability testing
Two build-and-test iterations shaped the final form.
Goal: test approachability. Easy to access, but outside movement pulled focus. Insight: grounding needs a contained space.
Goal: test immersion. Participants felt safe and stayed engaged roughly twice as long, but the booth felt uncozy and pressured people to be calm. Insight: let movement feel natural while still grounding.
Across sessions, people described the experience as intuitive, immersive, and emotionally grounding. Movement-based interaction helped them release tension and hold focus through gentle motion, and real-time visual feedback supported stress reduction and strengthened body-mind awareness. Participants reported feeling calmer and more emotionally balanced afterward.
Finding of the Develop phase
Soft, fluid, patient visuals with gradients, low detail, and slow rhythm best support grounding.
Two families, cool and warm, offered as options that keep harmony across the modes.
Simple body cues, distance, pacing, and stillness, create meaningful change without instruction.
Webcam, motion tracking, and real-time rendering give enough responsiveness for immersion.
From still meditation to free movement
At the thesis technical review, SomaSpace was presented as a functioning grounding environment, but the visuals left people unsure whether to move or stay still. The original concept was built around meditation and stillness. Based on committee feedback, I added the low-frequency background sound and reframed the whole experience: users now move freely and find their own grounding moments through interaction, rather than holding still to achieve calm. Movement became the core input.
Final prototype and installation
A beam projector displayed the visuals on the wall while two webcams captured the interaction: one tracked overall body movement and the other read depth for smoother tracing and a clearer angle of detection. A chair marked the ideal detection distance, and a simple guide on the desk explained the two key gestures, so the experience stayed intuitive and self-explanatory without any verbal instruction.
Exhibition
SomaSpace was exhibited at the Duncan Anderson Gallery (DESN 102) at CSULB during the HXDI Project Showcase. Visitors stepped in and moved freely while the field responded in real time. Watching people slow down on their own, with no prompting, was the clearest signal that the interaction worked.
Future development
Distinguish several bodies in one frame, give each a visual signature, and build shared visuals that respond to group synchrony.
From a portable kit for classrooms and therapy spaces up to full multi-wall or LED installations.
Limitations
Tracking and projection depend on a dimly lit, controlled space; bright or uneven rooms reduce clarity.
Detection relies on the webcam reading contrast, so it can drop or misread when lighting is inconsistent.
Summary of the Deliver phase
The review reframed the project from enforced stillness to movement-led grounding, the refinement made movement the core input, and the exhibition showed that people explore intuitively when the installation feels open and safe. The two-camera setup improved clarity, and the design encouraged self-guided interaction without verbal explanation.
Reflection
Designing for the body reframed "no UI" as a decision
SomaSpace treated the absence of buttons, steps, and goals not as missing design but as a deliberate, testable stance. The strongest moments came when the system simply listened and the body led. The honest next steps are a longitudinal study to see whether the calming effect lasts beyond a single session, and accessibility testing for users with limited mobility, since the system currently assumes free arm movement.
Next project
Hot Chocolate: an instrument on any surface