Any articles on the engineering behind the Sphere? How do they get the data to each individual pixel?

Let's assume 10 pixels per inch. One square foot would be 14,400 pixels. A PC can easily drive 1920x1080, or 2 million pixels. At our dpi, that would be 140 square feet. The Sphere's display is 160,000 square feet. So we could split that up into 1100 nodes, where each node is 140 square feet. It would be realistic to have each node be a single ARM computer with enough storage for a couple of hours of video. Push the data to the node over ethernet before the show, then have a clock signal going to each node so they stay synchronized.