The Walls Have Ears

A room reflecting back spoken words over time

IxDA submission
Ritika Mathur, Priyanka Kodikal
Concept development, coding, video (camera & editing)
Finalist of Interaction Awards 2014
Functional prototype in 1 week, running networked software controlling 4 tablets, 5 laptops and a projector

Benevolent eavesdropping

The ubiquity of security cameras in our world suggests we have become compliant being constantly recorded in public spaces. Yet privacy and data retention are hot topics and have made designers very much aware that their design choices affects what kind of data society we will become.

This project explores ways in which recording data in public spaces can be used in a positive, constructive and privacy-respecting way, to the benefit of the users of that space.


Whenever people enter a new space, there is curiosity to know what is happening, at that point of time as well as what has happened before.

Our interactive prototype gave visitors a sense of what it is like to have the room eavesdrop on your stories, and feeding back little snippets of it. By displaying words from the past hour on a clock face, the audience could gauge the vibe of a place: what had been talked about, were the words more positive or negative? Since the single words that are overheard are not directly linked to the other words or to the people in the room, visitors do not have to worry about their privacy.

How it works

With one week to get from idea to working prototype, we used any resources available. The audio is translated by MacOSX's inbuilt speech recognition functionality, automatically saved to file with Applescript. A Processing sketch reads the file and does some basic filtering through an online dictionary API (prepositions are less likely to be interesting than nouns or adjectives).

Each of the four recording laptops is connected through SpaceBrew to one tablet, which displays the most recently received word through an Android Processing app. The user can play with the individual letters, turning words into shapes. All computers also share the words with a central computer that picks a random word every minute, and adds it to a projected clock. The clock gives the user an impression of the kind of things spoken about in the last hour.