In early 2012, my friend Merche Blasco an I worked on a project called Radio. In the most basic sense, it is a large wooden map of America that you can use to listen to the different radio stations broadcasting across the country by navigating a little knob on the surface. As another friend put it, it’s like “eavesdropping on America.” I never got a chance to post about it, so I will now.
Our interest started with the idea that much of what a local culture is can be revealed to us by what is being broadcast on its radio local frequencies. America is a large country with a diverse people: the song playing in Massachusetts will be different from what the residents of New Orleans listen to. The differences between advertisements, talk shows, game shows, etc. will be interesting as well. With Radio, you can tune in by isolating an area or by mixing different radio stations together and hearing the bigger picture.
Because we wanted to keep the geographical component of the project was so important, the interface obviously had to be a map, so that part was easy.
We also wanted to evoke the feeling of an old-fashioned radio, so a nice hardwood surface came to mind. Naturally, time and cost restrictions came into play, so we settled some with the leftover material from the woodshop, and used a CNC to cut out a shape of America and carve out small holes to mark the cities.
You can barely make it out in the picture above, but there’s a crack in the middle where we used to connect the two cut wooden parts together like puzzle pieces — it runs along where the Mississippi is. A nice little nod to Radio history, where stations that are west of the Mississippi have a call sign starting with W, and East have call signs starting with K.
For the prototype, we relied on camera vision to detect where the knob was on the map at any given time. We placed infrared LED in the knob and a security camera mounted above to track the invisible light.
An XML file was created with some map points that corresponded to each state’s capital cities and the 5 most popular radio stations in the area that were streaming on the internet.
We used openFrameworks to read the map points from the XML file and plotted them on a grid that was calibrated to the actual physical map’s size and position. When the knob’s IR LED was within the vicinity of a map point, a random radio station from that area would begin to play. For this, we used openFrameworks to communicate with Terminal, which then triggered VLC to access the streaming radio station.
Here’s what I think: I love this project. I think it’s simple, elegant, and offers a rich experience. It’s a new way to listen to the radio, and a new way to think about the country.
When you love a project, it’s hard to see where it can be improved, but thankfully that’s what exhibits and shows are for. We discovered, while watching users interact with our piece, that most of what needed to be tweaked involved the user experience: when the user first approaches the installation — made up of two things, a wooden map laid out like a table and a metal knob on top of it — does she know how to use it? Does she know what she is listening to? And within those questions, things like: how to handle the knob; displaying city names; and of course the responsiveness of the software.
The Knob: It was important that the user DRAG the knob across the map, instead of lifting it up and moving it somewhere. There’s a technical reason for this (our software was looking for the LED around a last-known-location) but there is an romantic aspect to “traveling” from one location to another, instead of just hopping around (in my mind, the equivalent of mashing keys together to see what this or that button does). But inevitably, users were impatient and chose to pick up the knob and place it in the city of their choice.
So how then do we prevent this behavior? Some thoughts:
- Weight. If we made the knob heavy enough, with a nice padded bottom to make dragging across the map it a smooth, sliding action, people would instinctively know to slide, not lift.
- Pull. If there were tiny magnets under the cities, then the user would “feel” a gentle weight when it approaches a city, and release as the knob moves away — a fun detail that can’t be experienced with just dropping the knob down in any location.
- Audio. To reward sliding it across further, the audio would get louder as you approach a city, and decrease as you move away, or doing a simple fade in/ fade out based on distance. The fun bit is that you effectively “mix” two stations together when you are in between cities, controlling their levels as you move around the board.
Name The City. In the beginning, we were quite adamant about not having any text around the map — no city names, no station names, nothing. The conceit was that the focus needed to be on the discovery of radio stations as you “travelled” across America, and text would just be clutter.
However, as more people used it, we saw that a big part of their enjoyment came from being able to identify the cities. People enjoyed tuning in to their hometown’s favorite station, or tried to guess what kind of music would be playing from a given town (and vice versa — “Let’s go find some country music.”)
So in the next version, there will be a tiny screen somewhere on the map — maybe near Florida — that will display the current city as well as the currently playing radio station.
Hardware. Using computer vision worked great as a prototyping tool, as there were pretty much no wires to run or solder, but it was also fairly unstable. Calibration took a while, and IR light can come from anywhere and disrupt your data (for example: an open window). For the next version, we’re going to give some hardwired sensors with an Arduino or Raspberry Pi a go.
Hawaii, Alaska, and Manhattan Finally, there are those areas that simply couldn’t be included in the installation — in this regard, we are still flabbergasted. If anyone has any ideas, we’d love to hear them.