Multiple Description Coding: Compression Meets the Network

Your paper on the influence of medieval ornaments on contemporary art is due tomorrow. Luckily you have the latest wireless modem for your laptop, and hundreds of pieces from the Metropolitan Museum of Art collection are displayed on its web site. But as you examine the pictures, your web browser repeatedly gets stuck with partially loaded web pages. You see a reliquary, but an empty box sits where a scepter should appear; you find a Modigliani, but compression artifacts cause a Magritte to appear wrongly cubist. A lot of things could be going wrong; thus, many technological improvements could save you. Having more antennas on your laptop would make multipath a virtue instead of an impediment and could result in higher throughput and better reliability. More cellular base stations imply smaller cells and could lead to fewer conflicts with other users in your cell. Your wireless service provider could have higher capacities in the wired connections to its base stations. The whole wired infrastructure could be better, with fewer packets lost due to buffer overflows. The museum web site could handle more simultaneous connections or could be cached closer to you. Each of these changes could improve your browsing experience. This article focuses on the compressed representations of the pictures. The representation does not affect how many bits get from the web server to your laptop, but it determines the usefulness of the bits that arrive. Many different representations are possible, and there is more involved in the choice than merely selecting a compression ratio. The techniques presented here represent a single information source with several chunks of data (“descriptions”) so that the source can be approximated from any subset of the chunks. By allowing image reconstruction to continue even after a packet is lost, this type of representation can prevent a web browser from becoming dormant.