Interactive Cubes Workshop

From Noisebridge
Jump to: navigation, search
Sifteo cube teardown.jpg

Software + Hardware Hacking, let's dig deep into the Sifteo cubes


An amazing explanation by Micah of sifteo's insane design hacks is on adafruit blog.

But as a quick intro, Sifteo cubes can be seen as the best hackable platform, here is a summary of what they have:

  • an accelerometer
  • a touch screen (128x128)
  • 4 proximity sensors (to recognize other cubes touching their 4 sides)

They are wirelessly connected to a base with:

  • an STM32 (cortex M3)
  • a mono speaker
  • a usb connection to log anything to a computer

The base can communicate with up to 12 cubes (maybe 24 if you ask nicely).

It can be programmed with a simple and intuitive C++ API, check out the SDK.


At least during the noisebridge 30c3: December 26th (to 30th) might lead to a hackathon with Sifteo prizes soon too.

Hack ideas[edit]

Here are a few examples of that could be done during these sessions:

  • magic light juggle: acceleration change the color/brightness and the cubes disappear in the air
  • music & light: implement a BPM counter to blink color screens on the music rythme by listening to the vibrations sensed by accelerometer (like here)
  • siftephone in: listen to a smartphone vibrations with Sifteo's accelerometer (like above)
  • siftephone out: talk to a smartphone using sound with audio system such as (a 1st proof of concept can be found here)
  • midi controller: play music using the cubes as midi instruments with Ableton or Renoise (a 1st proof of concept can be found here)
  • internet of things: control anything remotely using python to forward Sifteo logs to an email or a server such as heroku or
  • turn up the volume!: follow this awesome instructable!
  • UART communication: simple hacks involving microcontrollers can be tried too but the project must be worth it ;p (it's a bit more complicated)
  • Gesture recognition: cubes accelerometer data can be sent to a computer (w/ USB log) and gestures can be recognized with toolkits such as the MIT-GRT. A lighter implementation can even be ported to the cubes: dollar Unistroke Recognizer.


If interested, please don't just show up, your participation/preparation is important:

Interhacktively yours,


PS: For more info talk to J.C. or Drix.