LightByte: Animated Shutters


Here’s another interesting project to come out of the MIT Media Lab — it’s called LightByte, and it’s all about interacting with sunlight and shadows in a new, rather unorthodox way.

We suppose its technical name could be a massive interactive sun pixel facade, but that’s a bit too much of a mouthful. What you really want to know is how it works, and the answer is, a lot of servos. We weren’t able to find an exact number but the hardware behind LightByte includes well over 100 servos, and a matrix of Arduinos to control them. While that is quite impressive by itself, it gets better — it’s actually completely interactive; recognizing gestures, responding to text messages and emails, and you can even draw pictures with the included “wand”.

We love anything mechanical like this — it’s just something about mechanical shutters that make them so awesome. Of course, reverse-engineered flip dot displays are pretty cool too! Or massive home-made flip-dot displays like this one…

[Thanks Alexander!]

28 thoughts on “LightByte: Animated Shutters

  1. yes…I would like to return the poor (performing) pixel display I spent $$ to buy 100+ servos and Arduinos…Interesting Idea, but I expect this out of USC…not MIT…

    1. $300 will buy you a nice projector that will outdo this project in virtually every way. The idea isn’t to make some new Earth-shattering technology — it’s basically art.

      1. I’ve got a chance to see this thing in front of me during my business trip. Several hundreds of moving pieces responding to my input all at once, quick, precise, very impressive! Too bad this video only demonstrated a few interactions and the secret sauce was hidden. it’s definitely worth a visit just for this! one of the best things I’ve seen lately.

  2. This is a good concept but leaves so much undone – if each “pixel” (shadel?) is run by proportional servos, why not have each pixel be (reliably) capable of rotation/density/color change rather than using the servos as simple solenoids?

  3. Once again, I’m left mystified at the reputation the famous Media Lab has gotten. Is this really the caliber of work they do now, or did somebody forget to check the date on an April Fools post?

    1. I usually don’t like to criticize the projects on HAD since they’re usually works in progress and very (very) experimental, but this is an interesting idea poorly done and badly presented. It reeks of a minimum-effort undergraduate senior project wrapped up in an ostentatious artist’s statement ( ) and hiding behind Media Labs’ broad skirts. Worse, somehow they’ve managed to snag a design award for it, which says someting unflattering about design awards. Admittedly playing with shadows is fun, and it’s an intriguing idea, but the execution here is dismal.

      1. I don’t get all the hate here. It seems like interesting investigation into what a world of internet connected things and nature may look like. No where does it claim to be a finished product for consumption or even about the object itself, rather it is an investigation into a concept of augmenting nature. I think it is about design concept rather than engineering and instead of CGI like most designs end at, this one was actually built.

  4. It looks like they are using 16 rows X 24 columns which results in 384 servos. With some adjustments, its possible to complete this exact same process with only 40 servos.

    Although their existing set-up allows them to trigger almost every servo at roughly the same time, the proposed changes to use only the 40 would require an algorithm to ensure that the proper cells in each row were being opened/closed.

    The process is similar to using a keypad format where you have power coming into the columns and you only trigger the rows to change. Cycling through the columns and rows, the servos would adjust accordingly. The algorithm would have to account for where your attempting to trigger multiple rows and columns at the same time, you cant trigger a combination where you have something that should not be adjusted.

    Taking it a step farther, applying current to a cell will toggle it open or closed, regardless of its current state. This would greatly increase the speed to open/close various cells at the same time.

    Overall, the existing concept is kind of cool.

    1. I also thought that they were using too many servos :-)

      My thought would be to use one per row OR one per column, depending on which is less. The servo would move a rod that reached all the way along the axis, with electromagnets mounted along its length at each pixel point. Each pixel would have a permanent magnet mounted on an arm, allowing the rod to “engage” or disengage the pixel as required. Then, movement of the rod would move the engaged pixels.

      Drawbacks would be that you would need a sufficiently powerful supply to drive all the electromagnets, as well as that much stronger servos to move all the pixels of the row simultaneously.

      1. Intriguing!given the density, the electromagnects may interfere with adjacent ones though. Also, if any component needs to be replaced, it can influence the whole roll and other columns.

        There is always more than one way to do it, but I think they had each flap controlled by one servo for a good reason.

        Very cool project!

    1. The interesting thing here is they created a screen-less interaction to interface with the physical world. using LCD panels would simply not be the same. very cool design with wood, which makes the patterns and sound that digital displays can’t beat.

  5. It can be very challenging to control hundreds of servos at once. Also, putting together all the pieces of tiny components to achieve the kinetic behavior is not an easy task. plus the gesture sensing and light and shadow mapping… quite an engineering challenge!

    This is Very well done!!

  6. I find this interesting because it represents something that has been implemented and is not just some crazy concept video, It will probably be refined but it’s cool to see that it works with different modes of input

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.