LED-based pixel wall for animations and tetris visualisation. Provides 16x24 RGB pixels on 80x120cm. LEDS are controlled with an Arduino, a Raspi provides input from arbitrary image or stream inputs including showing live Tetris games from NES input.

Complete RGB-Tetris-LED-Wall


  • WS2811 5050 LED strips (60leds/m, 20ics/m)
  • 12V power supply (HP 750W HSTNS-PL18), provides 62.5A/12V
  • Arduino Mega
  • Raspberry PI 3 B
  • EasyCAP USB grabber with Syntek STK1160 chip set
  • Microphone: line in, perhaps with amplifier Iduino SE019
  • Switches, Buttons, Photo-resistor,
  • Case made from HDF, with cells of size 5x5x5cm, top layer is Transparenzpapier + 3mm clear acrylic

Energy supply:

  • worst case need: 0.02x24x16x9 = 69.12A
  • mostly organises via 4×2 molex connectors
  • 12 -> 5V converter for Arduino + Raspi + Controlleds (WS2812b)



Purpose: direkt control of all LEDs via the FastLed library.

  • controlled via buttons: mode (animations, images, beat dedection, NES Tetris stream, pixelflut), submode (e.g. which animation), speed, brightness
  • adaptive brightness via a photo-resistor (with voltage divider via 5k1ohm resistor for input pulldown)
  • annimations: library of 16+ animation patterns: rainbow, fire, plasma, ...

Communication with Raspi:

  • SPI, 487500 bps
  • Arduino sets sync_pin if ready for receiving data.

Libraries (C):

Raspberry PI:

Purpose: Proprocessing streams of input images, beat detection, NES tetris streams, and pixelflut.

  • Image processing: cycles through all images that are located in /images
  • Beat detection: Microphone measures sound for simple beat detection that can get visualised
  • NES Tetris stream: reads the composite signal of an NES via the Syntek STK1160 grabber. All frames are analysed for the content of the Tetris game: game state, next block, score, level, number of lines
  • Pixelflut (tm)

NES Tetris stream:

  • NTSC or PAL possible
  • 12-15 fps feasible

Libraries (python):

  • base64
  • numpy
  • Pillow
  • pigpio
  • pyserial
  • PyV4L2Camera