While folks over at Slashdot discuss the topic of “VGA and DVI Ports To Be Phased Out Over Next 5 Years” there is active research at the Intel Visual Computing Institute* (IVCI) and DFKI that might present a few other interesting video connectors alternatives. Instead of having various different video input and output methods the idea is to use something that already exists and could potentially do the same tasks: the good, old Internet Protocol.
Today’s TVs have numerous connectors to get video and audio into the device. Many TVs (however not many monitors yet) also have an Ethernet port to enable browsing or stream videos from online media streaming providers. From a purely technical point of view without the copyright issues, using IP connectivity to transfer data could just be the only input/output method we need. Considering the bit rates that current HD videos uses this could also be done over WiFi easily. For instance, Intel’s WiDi already demonstrates in the 1 device : 1 display setup to stream HD content wirelessly. As for handling the received, compressed video stream it is important to note that many SmartTVs already have the ability to efficiently decompress such data.
Variety of connecters on the back of a TV
Going back to the numerous connectors scenario there is another limitation we have lived with and got used to. Today we are mostly limited to 1:1 connections. One laptop connects with one cable to one TV. A second output source can be added by one more cable. Some graphics cards support multiple monitors at the same time but each monitor must be connected by a separate cable to the video card in the computing device. As the number of displays increases beyond what’s supported today by popular video cards, connecting these display devices becomes tricky. There is only complicated special solutions to deal with that the problem.
As the displays and the way consumers interact with their computing devices evolve, new scenarios arise. Think about using TVs, monitors and tablets together as output device?
The easy solution to all the above mentioned scenarios is the “Display as a Service” project (under the lead of Alexander Löffler and Jochen Miroll at the IVCI). There is the concept of a networked, virtual frame buffer (NetVFB). The NetVFB is used to coordinate between the image generators and all the displays. An application draws into that virtual frame buffer and the user specifies on his machine where and how (e.g. what size, rotated etc) it should be shown on the available displays.
1:n setup: App renders into virtual frame buffer. Using the network to transfer the image data to output on n displays.
“Display as a Service” has also been extended from the 1:n setup (1 driving machine, n displays) to an m:n setting – having more machines calculating an image and getting it merged together at the output displays.
m:n setup: Multiple apps or machines render into the virtual frame buffer. Output on n displays.
Of course synchronization becomes a very important topic when dealing with numerous displays. Researchers at the IVCI solved this by having a master screen that automatically synchronizes itself with the other displays over network. It works so perfectly that it can also be used with active stereo screens. And exactly such a stereoscopic display wall is what we would be happy to present you at CeBIT 2012, hall 26, booth F34.
* “Intel Visual Computing Institute” is a collaborative research institute at Saarland University, co-funded by Intel