During my first year as an undergraduate at Cornell, I worked as a cameraman for the athletics department on the weekends, filming sports games so they could be streamed online. Basketball was one of the sports we covered, and we didn’t have a reliable way of overlaying live statistics from the game (score, clock, fouls, etc.) onto our video stream in real time. I figured we could try to decode the signals coming out of the scoreboard controller driving the main arena scoreboard and convert them into a data format that we could then use to develop a graphic overlay on our live video stream.

The basketball stadium uses a Fairplay MP-69D scoreboard controller, which I reverse engineered by analyzing its output signal on a digital oscilloscope.

Once I worked out the nature of the protocol I was able to retrieve all types of scoreboard data, using an Arduino Uno to process the raw signal. In conjunction I also originally developed a simple Python GUI application to run on a Raspberry Pi, that takes in serial data from the Arduino over USB and displays and creates a scoreboard graphic with live data overlaid on top of it.

The Raspberry Pi has an HDMI and component video output, which could theoretically be fed into any video streaming application. I have been working on writing an additional Python script to integrate my hardware with NewTek’s LiveText graphics system. LiveText is an application that is used with NewTek’s high-end TriCaster video system, and can receive data dynamically from text files and/or database files to display in graphic templates. My goal is to use the Arduino to decode the signal coming from the MP-69D controller, send the serial data to a laptop, and have a Python script running that reads the incoming serial data, updates a static text file, and, in turn, updates LiveText (it can pull data from a text file).

I eventually had a printed circuit board made up that plugs onto an Arduino and has 1/4” jack inputs right on the board for the Scoreboard and Timer outputs of the MP-69D. Here is the code used to process the input and here is a live demo of the decoder in action.