SOMA

Inspiration

In meetings and social gatherings, there are moments when the atmosphere shifts even though no one says anything. The conversation slows down, only a few people keep speaking, or an uncomfortable silence appears in the room.

Soma began with a simple question. How can we recognize these invisible emotional currents moving through a group?

The project explores whether a space could sense and respond to the balance of interaction within a group without calling attention to any individual. We looked at the idea of homeostasis, where the body maintains balance quietly and continuously.

From this perspective, Soma proposes a new kind of awareness called Social Homeostasis.

What it does

Soma is a system that senses patterns of interaction between people and interprets the overall state of a group’s dynamics. By analyzing vocal patterns such as speech overlap, silence duration, pitch variability, and speaking balance, the system identifies shifts in participation and tension within a conversation.

Instead of pointing to individuals, Soma communicates insights about the group as a whole. When needed, it gently guides interaction through behavioral prompts and spatial feedback.

Environmental elements such as lighting, airflow, scent, and music can adjust to support the flow of conversation and help restore balance in participation.

Through this process, Soma introduces the idea of Social Homeostasis, a new form of awareness that supports balance in collective interaction.

How we built it

The project began with research on social emotional signaling and environmental psychology. From there we explored how interaction patterns could be defined and sensed at the group level rather than the individual level.

These patterns were then connected to two types of response. The first involved behavioral suggestions such as asking a question, changing the topic, or breaking into smaller pairs. The second involved spatial responses including lighting, airflow, scent, and music within a smart environment.

To explore how the system might work in practice, we developed three prototype scenarios: a team meeting, a conference presentation, and a social gathering. Each scenario helped test how Soma might operate in different social contexts.

Challenges we ran into

One of the main challenges came from a basic contradiction. Emotional experience is subjective, but the signals used by a system must remain objective.

Since a group's emotional state cannot be measured directly, we needed to interpret changes in group dynamics through observable interaction patterns.

Another challenge was designing interventions that would feel natural. The system should not make people feel observed or corrected. Instead, the goal was for the space itself to shift subtly, guiding interaction without drawing attention to the intervention.

Balancing sensing, response, and ethical responsibility became the central design challenge of the project.

Accomplishments that we're proud of

One of the most meaningful outcomes of this project is proposing a new design approach for sensing group interaction.

While many emotion recognition systems focus on analyzing individual facial expressions or biometric data, Soma instead focuses on patterns of interaction across the entire group.

The system was also designed to avoid direct intervention in conversation. It never identifies individuals and it does not produce sound feedback, allowing interaction to remain natural and unobtrusive.

By combining device insights with spatial environmental responses, Soma demonstrates an interaction model where digital sensing and physical space work together to support healthier group dynamics.

This project explores how design can introduce a new sensory interface that contributes to social well being.

What we learned

During our research we found that vocal interaction patterns reveal more about group dynamics than we initially expected. Speech overlap, silence duration, pitch variability, and speaking balance can all reflect shifts in group tension or engagement.

However, patterns alone are never enough. Their meaning depends strongly on the social context in which they occur.

Another important realization was that direct intervention can easily disrupt interaction. If a system visibly reacts to what someone says or does, it can change the dynamics it is meant to support.

Because of this, Soma follows a clear principle. The system never identifies individuals and it never produces sound. Instead, it communicates only group level insights.

What's next for SOMA

In the future, Soma could be applied to a wider range of social environments.

Collaborative workplaces, classrooms, conferences, and community gatherings are all contexts where balanced interaction is important. Soma could help support healthier group dynamics in these shared spaces.

The system could also evolve by integrating additional signals beyond voice patterns, such as movement or environmental data, allowing a more nuanced understanding of collective interaction.

Ultimately, Soma points toward the possibility of smart social environments where spaces themselves help guide and support the quality of human interaction.

Built With

• aftereffect
• figma
• javascript
• midjourney