News      |      Events

IP Workflow in Broadcast Environment | Webinar, 24 July 2020

In traditional workflow, broadcasters are familiar with SDI signals and SDI Routers. In SDI infrastructure, every connection is typically made using individual, separate coaxial cables. Also, a lot of Video/Audio Distribution Amplifiers have to be used in this environment. In IP workflow, although backbone equipment, such as, cameras playout units, recorders, switchers etc. still stay in place, the signal structure and the signal distributing scheme changes dramatically. From continuous unidirectional signals within individual physical cables between equipment, to packetized nonlinear data traffic within a network. In the IP world, we build up an IP network with fast, capable switches, and then we connect every item of equipment to this network with one single cable, that’s all. We don’t need distribution amplifiers, to multiply our signals and we don’t need many cables. Also, distance between equipment is no longer an issue in the IP world; equipment can be located anywhere, as long as it is connected to the network.

The motivation behind the idea of transferring workflow from SDI to IP, can be summarised as; IP is cheap, universal, flexible, fashionable and so on. However, there are some general worries about using IP infrastructure in a broadcast environment; such as the IP world being non-deterministic, can be lossy and may have more latency, Also, a timing reference is not a simple signal but a complex protocol that older broadcast technical staff are not familiar to work with.

To send data over an IP network, we have to prepare it, chose a protocol, address it, and send it step by step. This process starts with dividing the data into many parts. Then identifier headers need to be added to every part of the data. These include information about how many parts there are, and the identity of each part. Then we choose the protocol, as TCP (Transfer Control Protocol) or UDP (User Datagram Protocol). The TCP header is longer than the UDP header and it includes handshaking and confirmation procedures between sender and receiver. The UDP header is smaller than the TCP header and there is no handshaking or confirmation, it is more like a ‘fire and forget’ process.

In the studio environment we mostly prefer UDP, because handshaking and confirmations cost us bandwidth and time, both resources that are very important to us. Therefore, we add a UDP header to every data piece. Then we add the IP header and continue to the process of encapsulating our data pieces. Finally, we add a frame header at the beginning, and a frame footer at the end of each encapsulated data piece, to complete the encapsulation process. We then we send these encapsulated structures to the physical network as a stream of ones and zeros. These packets are like many cars that drive out from same garage and want to reach some other garage all together, possibly using different dynamically changing routes in crowded city traffic.

Compatibility with the IGMP (Internet Group Management Protocol) is one of the important parameters for IP switches for use in our broadcasting environment. Because, the multicast scenario we need to operate in our network is only possible with IGMP compatible switches.

Reference timing in IP workflow is done by PTP (Precision Time Protocol). It is a complex protocol when compared with the traditional genlock signals of the SDI world. PTP achieves clock accuracy in the sub-microsecond range. However, there are no cable length issues, or necessary phase adjustments for PTP.

One of the major stepping stones in transferring from SDI to IP is the SMPTE ST 2022-6 standard. In this standard, we have one big RTP flow with all the information in it, video, audio and data. In this standard, every piece of equipment needs to capture and decapsulate all the information to reach and isolate the specific data part that it needs to process. Then it also needs to encapsulate the stream to return the processed data to the network. In SMPTE ST 2110, which is currently the most popular standard, separate RTP flows are used for video, audio (for each channel), and data. For example, if you are using 1 video, 16 audio and 3 data streams, then, in SMPTE ST 2022-6, you will have 1 single RTP flow with everything in it; but in the SMPTE ST 2110 standard, you will have 20 separate RTP flows. In the SMPTE ST 2110 case, any equipment can reach and process the RTP flow that it needs and doesn’t have to deal with all the other flows. This is the main advantage of SMPTE ST 2110 over SMPTE ST 2022-6.