Video over IP
May 1, 2009 12:00 PM, By Brad Gilmer
New standards allow broadcasters to use video over IP for video transport.
As network speeds and reliability improve, it is becoming more common for broadcasters to install and maintain transport circuits that employ video-over-IP technology. Moving video over packetized networks is not without issues, and the industry has responded by providing several standardized approaches.
In the context of this article, video over IP refers to a professional application in which video is transmitted over IP networks. The application is most often either contribution or distribution. With contribution, video originates at a remote location and is transported back to a central broadcast network facility for manipulation, branding and packaging. In the case of distribution, video is sent from a central broadcast network facility to other broadcast facilities or cable systems for ultimate transmission to the end user.
There are several challenges in transporting professional video over IP, including:
- Accurate characterization of losses on the network;
- Dealing with error characteristics of IP networks;
- Preserving timing relationships; and
- Developing a standardized approach to the treatment of network errors.
Network loss profiles
When putting together a video-over-IP transport solution, it is critical to understand the loss profile of the network. A loss profile tells you that errors occur, as well as whether the errors are single bit losses sprinkled randomly in time, or whether they are groups of errors lasting for tens or hundreds of milliseconds. It is important to understand the loss profile of the network in order to develop a strategy to deal with these losses.
It may be difficult — if not impossible — to get solid information about losses on the network without measuring it yourself. Fortunately, manufacturers have studied the loss profiles of various networks and determined the best strategies for dealing with these losses.
Error characteristics of IP networks
Internet Protocol is a self-routing protocol containing both the source and destination address in each packet. On nonmanaged networks, each packet is transported across the network individually, without knowledge of what route previous or future packets may take. At the IP layer, there is no concept that these packets are somehow related. Because of network congestion and other factors, packets may arrive out of order, they may be lost, or they may even be duplicated in the network. If you are using a closely managed, dedicated network that has been specifically designed to carry video and audio, in all likelihood, the network was designed to eliminate many of these issues.
Preserving timing relationships
Because packets may take different paths through a network from source to destination, and network switches and routers employ buffers that can have variable transit delay, any timing relationship between data in one packet and data in another packet is lost. Of course, video is a time-sensitive medium. Destroying the timing relationship between parts of a video frame can result in artifacts or no image at all. For this reason, all video-over-IP transport solutions include the carriage of timestamp information so that receivers can recreate the video timing relationship that was originally present at the video input at the output.
Standardized treatment of errors
Given that packets will be delayed, lost, reordered or duplicated, it is important that manufacturers come up with standardized ways to deal with errors. Some manufacturers have developed standards that address the transmission of video over IP, including treatment of errors. The SMPTE and the DVB have both described error correction methods.
Of course, because this is video, it may not be necessary to correct every bit error that occurs in transmission. In fact, some manufacturers use error concealment techniques to mask errors rather than repair each error.
A closer look at standards
As mentioned earlier, a SMPTE standard has been written for video-over-IP transmission. SMPTE 2022 is a multipart standard, meaning that it has different parts that may be used alone as separate standards, or the parts may be used together. SMPTE 2022 Part 1 is titled “Forward error correction for real-time video/audio transport over IP networks.” As the title implies, the document describes a FEC method that can be used to protect video-over-IP streams. SMPTE 2022 Part 2 is titled “Unidirectional transport of constant bit rate MPEG-2 transport streams on IP networks.” This document describes how to map video presented as an MPEG-2 compressed transport stream onto IP networks. To sum up, 2022-1 deals with FEC for MPEG-2 constant bit rate (CBR) streams on IP networks. 2022-2 deals with the mapping of CBR MPEG-2 streams onto IP.
| Want to use this article? Click here for options! |  |
