Image Image Image Image Image Image Image Image Image Image

Bartlett School of Architecture, UCL

Scroll to top


No Comments

Recreating a system of audience control using coloured paddles: Part 3

Recreating a system of audience control using coloured paddles: Part 3

I decided to recreate Loren Carpenter’s experiment of audience control using coloured cards, enabling an audience to collaboratively play games. In this experiment, Carpenter observed his audience acting as an “amoeba”.

I’m interested in the effects of this kind of collaborative thinking and decision making, and the potential to create a feeling of connection in participants, as well as the potential to produce interesting results.

In recreating Carpenter’s experiment, I hoped to find some interesting observations which would inspire future work.

Set up

The auditorium was set up with plywood cards left on seats, and a camera below the screen. Light from a lamp next to the camera is bounced off reflective tape on the cards, into the camera. In Processing, the camera feed is analysed to see how many yellow and red cards are being held up. This information is then used to control various games, also coded in Processing.

The audience were not told much information about what would happen. It is worth noting that several of the audience members had taken part in a smaller test version, but they were asked not to explain anything to their neighbours.

I had some issues with the framerate of the games dropping to be very slow and this affected gameplay in some cases. Additionally, the audience had already watched an unrelated 90 minute lecture, and it was clear that their attention span would be short.



The audience is split into two teams (delineated by tape on the seating). Each team controls one of the paddles as they try to hit a ball back and forth. The movement of the paddle is controlled by the percentage of red or yellow cards. If there are 60% of red cards, the paddle will move down slowly. 90% of red cards, it will move down faster.

Unfortunately the speed of the paddle was mapped too slow initially, and I had to quickly increase it in the code and reload the game. This kind of fine control is hard to test for without an actual crowd.

Elephant Jump

Modelled on Chrome’s offline dinosaur jumping game, this game has an elephant attempting to jump over mice. If more than 60% of the audience’s cards are yellow, then the elephant will jump.

Rocket Flying

Modelled on the famous helicopter game and similar to flappy bird, this game has a rocket flying through a tunnel, trying to avoid planets. Here the Y position of the rocket is directly controlled by the percentage of red cards. If 10% of cards are red, the rocket will be 10% from the top, if 80% of cards are red, it will move to 80% from the top.



Some people try to control the group, shouting out instructions for everyone. Its interesting to see that some people lead and some follow.


Rocket Control

There was much less shouting in the Rocket Flying game, I think because it required more concentration and also because it was difficult for people to know what to instruct others to do. There is no point shouting out “red” to move down a small bit, because if everyone goes to red, then the rocket will crash into the wall. This wasn’t explained to the audience but the lack of shouting implies to me that they understood that it wasn’t controlled in the same way as the Pong paddle.

The Rocket Flying game was the most interesting to observe because the control was truly collaborative in that way. It is not possible to control the rocket without having a decent sized group of people, as if there are only a few options of percentages then the rocket can only be in a few specific positions.

I think it would be interesting to observe a crowd concentrate on just learning to control the Rocket Flying game, and see how well they develop their behaviour and game play over time.

Audience Suggestions

I told the audience I’d be interested to hear about their experience and, after the experiment was over, some of them approached me. Many people had ideas of how to make the control easier. Several of them suggested having a second screen which would show them what the rest of the group were doing so they could adjust their own behaviour accordingly.

It’s interesting that they felt that knowing what the group was doing would help the group perform better. People pointed out that those at the front could turn around and check the rest of the group’s cards, but in fact I didn’t observe people doing this very much.

Jump Game

The Jumping Elephant game seemed to be the favourite. I think partly this was due to the cute graphics.

It also was the first game they played as one team instead of two, and people seemed to prefer all working together.

The control of the game involves building up to a succeed/fail moment (successfully jumping over the mouse or not) and this seemed to be fun. It’s possible to hear the crowd building up to the jump, with different people shouting “now” at various points, and then laughing whether they succeed or not. This contrasts with the rocket game where people could fail at any point and had to keep their concentration strong throughout.

Feeling of Control

In smaller tests, people reported that they could definitely feel their influence over the Pong paddle, whereas in this large group I don’t think people felt such an agency. However, this didn’t seem to bother people, they weren’t frustrated if they felt like their input wasn’t having great influence, they were happy to play the games as a group.


The technical issues and the fact that the audience had just had a long lecture definitely affected thier attention span, which is a shame as I think it would be interesting to see their behaviour develop over time. It would be great to do this experiment with a crowd who were much more excited to take part, and without the framerate problems.

Additionally I think it would be interesting to try it in a more curated space, where it is possible to control the lighting. This would improve the performance of the card recognition as well as giving the audience a sense of a performance or occasion.

Despite these issues, it was still super interesting to watch people interact and see how audience members started chatting to their neighbours, even those who didn’t know each other.

People seemed to enjoy playing the games, and it was clear to see gameplay improve over time. In the Rocket Flying game, there were around 8 games with zero points and I was about to stop the game, accepting that it was too difficult, when suddenly the audience seemed to gain control and managed to get past 3 planets. Observing that control be learned was the most interesting part for me.

Submit a Comment