In traditional facilities, SDI, MADI and AES are the common forms of signal distribution. We use a combination of X-Y routing matrices and patch-cords to distribute signals around a facility. One to one connectivity is required to maximize transfer of power between source and destination through impedance matching.

If we need to send a signal to more than one destination, such as a sound console output to multiple monitoring loudspeakers, some form of distribution amplifier is used. And using one-to-one mapping provides us with a simple over-view of which microphones, console channels and monitoring is routed together.

SDI, AES and MADI Integration is Complex

Transferring signals between different formats starts to get challenging as we interface SDI, MADI and AES signals together. If we need to extract channel two out of a MADI stream and insert it into channel one of an embedded SDI stream, then an array of embedders, multiplexers and de-embedders is required, often making a system extremely complex and difficult to visualize or remedy if an error occurs.

IP routing solves many of these problems for us as audio and video signals are treated as pure data, and one of the greatest strengths of IP is that it neither knows nor cares what type of data it is transporting in its payload.

IP Allows for Mixed Formats

Diverse types of audio and video formats can be inserted into the IP payload, for example, audio sampled at 48KHz with a bit depth of 24bits can be distributed on the same network as audio sampled at 44.8KHz with a bit depth of 16bits. The same is true of video, signals of 720p29.97 can be distributed on the same IP network as 1080i59.94, or 4K-60P-444.

There needs to be a point where the signals are converted to a mezzanine format for processing, but this only happens at a production interface leaving the audio or video signal to be stored and distributed in its native format to maintain the highest quality possible.

In IP networks, terminal devices such as servers, microphones, production switchers and mixing consoles are referred to as “host” devices. A host device must have a unique IP address so a router, or switch knows where to send the signal to.

Another often unappreciated aspect of an IP system is that the network is bi-directional, that is, signals can be sent to and from a host. It’s entirely possible to have a microphone with an IP interface that will output audio across its IP connection, but also receive configuration information from a sound console or control application. Microphones could have configurable pre-amps or equalization built directly into them to further improve quality and provide greater flexibility.

Routers are referred to as layer-3 devices and switches as layer-2. The difference is due to the level at which they transfer data packets, and the numbering system is derived from the Open Systems Interconnection model, a reference model describing how data can be distributed throughout a network. Layer-2 refers to the data-link layer, and layer-3 refers to the network layer for routing.

MAC Addresses are Universally Unique

Ethernet is usually the dominant data-link layer used in broadcast facilities, but other formats are used such as Synchronous Optical Networking (SONET) and WiFi, especially as broadcasters connect to the outside world through telecommunication providers (Telco’s).

Each host device must have its own unique Media Access Control (MAC) address so that the device can be identified. Although IP also has an addressing scheme, it differs from the MAC as its address is not absolutely attached to that device, this is to allow faulty units to be replaced without major configuration. For example, if a microphone has an IP address of 10.10.56.19, and it goes faulty, it can be replaced with another microphone with the same IP address. However, the MAC addresses will be different but the Address Resolution Protocol (ARP) is designed to deal with this.

The Institute of Electrical and Electronic Engineers (IEEE) administers the issuing of MAC address numbers, but it is the responsibility of the manufacturer to make sure that the MAC address is programmed correctly during manufacture. Once installed, the MAC address should not be changed.

Layer-2 Keeps Switching Fast

IP datagrams are encapsulated within the payload of an ethernet datagram and switching of data streams takes place independently of the IP address to keep switching speeds high. Routing at the IP address level requires the ethernet header to be decoded and then the IP header, thus increasing switcher processing times and reducing the time taken to transfer the datagram.

An ethernet switch has many ports to connect to other ethernet switches or host devices such as microphones and sound consoles. If, for example, ten microphones are connected to the first ten ports of an ethernet switch and the eleventh port is connected to a sound console, the switcher will need to know it must send the first ten microphones to the sound console, in effect the switcher is acting as a multiplexer, as ten microphones are routed to one sound console ethernet connection.

The quality and speed of the switch impacts heavily on how effective this multiplexing process is. In broadcasting, we take for granted that the throughput of an X-Y matrix guarantees one hundred percent data delivery, but in IT, this is not always the case.

Non-blocking switches guarantee that the data received on all its ports can be sent on the designated port without packet loss. For example, if a switch consists of sixteen ports, each with an input and output speed of 10Gbps, then the total input speed is 160Gbps, so the switch manufacturers guarantee that all the packets can be sent to the outputs of the ports totalling 160Gbps.

Use Non-Blocking Switches

Non-blocking switches are usually only available on high-end switches. The alternative to “non-blocking” is the “blocking” switch. As the design cannot guarantee all packet delivery, some packets will either be lost or unacceptably delayed. Blocking switches will work in some broadcast applications, but a great deal of attention must be paid when designing for data throughput, otherwise packet loss, and excessive delay and jitter will occur.

When the audio or video stream is sent to another building or district, IP routing will be used as the host will need to know where the transmission is going. For example, the sound console might have an IP address of 10.10.76.01 and its destination address might be a transmitter site with address 10.1.100.89, the IP router will have been configured to send all datagrams destined for 10.1.100.89 to the port connected to its telco.

Ethernet switching is generally faster than IP routing and if non-blocking devices are used then packet jitter and delay are greatly reduced. But to distribute signals outside of the broadcast facility, and to maintain maximum flexibility for addressing, IP routing is used, although it’s kept to a minimum to help keep transfer speeds high, and jitter and delays low.