Take Me to the Dancing — Gargi Guchhait

The problem

Friends who live far apart can dance together on a video call or in VR. But the friend stays inside a screen. You never feel them in your own room.

My solution

I put the friend on the floor. A sensor reads their feet and a projector draws their footsteps on your floor as they move. We call it sharing the same ground.

A living room at home with a friend’s footsteps projected on the floor. AI-generated concept image.

About the project

Many people live far from their friends and partners for study or work. Calls keep them talking, but doing something together is hard through an upright screen.

I chose dancing because it is social, it is good for body and mood, and it needs no words. This is my research at Keio University Graduate School of Media Design.

I left out VR and AR on purpose. Headsets can make people feel sick and screens make them tired. The room you already feel safe in becomes the dance floor.

How the concept would look at home: a friend’s footsteps projected on your own floor. AI-generated image, not a photo of the prototype.

How the work progressed

Five steps, from a survey at a university exhibition to a tested prototype.

1Step 1 of 5

I asked how people would want to dance with a distant friend

In May 2022 I surveyed 25 visitors at a university exhibition. I asked which sense they would rely on, sight, touch, sound or smell, and where they would want to dance.

Two needs came out of the survey and the reading. People need to see a mark in the room to believe someone is there. And they need privacy: presence without sending video of themselves.

18 of 25chose sight as the sense they would rely on

21 of 25wanted to dance indoors

2Step 2 of 5

The idea came from footprints at home in West Bengal

Where I am from, we draw footprints on the ground and treat them as the presence of a god. A bride steps in red ink and leaves red footprints as she enters her new home.

A footprint says that someone is here. So I asked what would happen if a friend’s footprints appeared on my floor.

I also saw that the related work all used an upright surface, a screen or a wall. I chose the floor.

Footprints, the starting point for the idea

An alpona, a ground drawing from West Bengal

3Step 3 of 5

A first prototype, moved by my finger

Before building any sensing, I tested the idea by hand. Two white dots were projected on the floor. I moved them with my finger on a phone while I watched the dancer and answered their steps.

35 people tried it. All of them accepted that someone remote was there, because the dots moved in an irregular, imperfect way, like a real person.

They also wanted proof that the dots were their friend, so voice had to be part of the final version. 19 people preferred a footstep shape to an abstract one.

The first prototype: two white dots projected on a floor

Pre-test at the university forum, November 2022. The two dots on the floor follow my finger on a phone.

4Step 4 of 5

The final prototype reads real feet

A Microsoft Azure Kinect captures each dancer’s ankle and toe positions. TouchDesigner turns that data into two footprints on a rectangular stage and projects them on the partner’s floor.

For the test I built two private rooms from two tents in one hall. Each tent had a projector, a speaker, a tablet and a Kinect.

The tents were linked by cable, not the internet, so the footprints had no visible delay.

Setup: two tents, with the Kinect sensor and projector in each, linked through the laptops in the middle.

Kinect sensor Projector Laptop

Two tents, 1.8 m apart. The sensor in one tent drives the projector in the other.

Inside one tent, with the Kinect and projector on a pole and a tablet on a chair

The final user test. Left: the two tents. Middle: inside one tent, with the Kinect and projector on the pole and a tablet on the chair. Right: the two laptops that linked the tents.

5Step 5 of 5

28 people danced in three ways

14 pairs took part. They were non-dancers aged 20 to 55, plus one four-year-old. Some pairs were friends and some were strangers.

Each pair danced three times. Afterwards I interviewed every participant and they filled in a questionnaire.

Video call only: They saw and heard each other on Google Meet.
Footprints and video call: Footprints on the floor, with the call still running.
Footprints only: No camera and no voice.

Two participants dancing during the user test

A participant dancing on the projected footprints

What happened

People felt guilty when they stepped on their partner’s footprints.

In the interviews, participants said they felt bad about treading on the projected feet, and asked for a bigger stage so they could avoid it. I read this as co-presence: they were treating the footprints as a person.

22 of 28

A video call alone lost people in 20 seconds

22 participants were unhappy with the video call within the first 20 seconds.

13 of 18

Voice mattered. Video did not.

18 people felt closer once call and footprints ran together. 13 of them said hearing their partner helped. 8 stopped looking at the video and danced with the footprints only.

Strangers

Pairs who did not know each other chose footprints only

With no camera and no voice they felt more comfortable. Six people preferred dancing with no sound at all.

How participants rated each way of dancing

Each participant scored the three scenarios from 1 to 10. The video call alone averaged 2. The two scenarios with footprints averaged 4 and 8.

Video call only

2 / 10

With footprints, lower of the two scenarios

4 / 10

With footprints, higher of the two scenarios

8 / 10

“When I am involved in dance, I prefer to move in various directions. I find it limiting to view my partner fixed in one spot.”A participant, about the video call

“I favor the footprint feature without the video since I find dancing in front of a camera somewhat uncomfortable.”A participant in his 20s, paired with a stranger

What the test taught me, and what is still missing

What I learned

Sound is essential for feeling someone is there. Video is not.
Real, imperfect foot movement feels like a person. That is what separates it from dancing with a digital character.
Being anonymous helps people who met online or do not know each other well.
It only works when at least one of the two is willing to start dancing.

Limits of this prototype

The two tents were joined by cable, 1.8 m apart. It has not run between two real locations.
It ignores the height of the foot, so complex steps are hard to read.
Each pair danced for 2 to 3 minutes. I do not know how it feels over a longer time.
People could not change the room light or the music volume, and the picture sometimes glitched.

What I would do next

Give each foot its own colour, as participants asked.
Make the stage bigger, so nobody has to step on their partner.
Build sound into the system.
Connect two locations over the internet.

Non-dancers told me it would be a private way to learn to dance, for people too shy for a class. Couples told me it suits close, wordless communication.

Take me to
the dancing

My role

Scope

Team

Timeline