AV over IP in the Intercom World
Everyone’s talking about using AoIP (Audio over IP) but what does it mean to the intercom users?
Until recently there were several standards for transporting digital audio between hardware; some of them common to the manufacturers, some standard amongst a number of different manufacturers and some just totally bespoke. Today’s standardization on Dante and AES67 helps simplify matters, but how?
Interoperability has always been part of the intercom world; analogue 2-wire beltpacks allowed different areas to work together over microphone cables or tie-lines, as an early example. Then the intercom world explored the benefits of telephony to interconnect operators across sites and even countries using POTS (analogue telephony). This was basically extending 4-wires across telephone copper cables—it was expensive and had balance issues resulting in feedback of your own voice, but it worked sufficiently well. Add simple GPIOs (general purpose interfaces, contact closures etc.) and you had some basic signaling between intercom systems. Digital Intercom had to use multiple lines to carry both voice routing and signaling data—anyone remember voice over data modems?
When ISDN came along the digital interconnection became easier to manage. It was still costly, and the ISDN lines often had to be booked in advance, but it worked in a similar way to POTS, just with embedded data and caller ID.
Within the studio, MADI grew to be the digital audio interface of choice, allowing multiple streams of AES3 audio over a single cable, rather than 64 pairs of analog 4-wires. MADI (AES10) worked well between devices and was effectively the forerunner of today’s AoIP communications.
Early adopters in IP connectivity include Trilogy Communications, which launched the world’s first IP connected intercom in 2002. The system used standards-based IP networks to allow intercommunication between intercom matrices, which permitted Trilogy to install several hundred systems world-wide that were all interconnected. At this early stage, intercom bandwidth was controlled by the use of compression codecs; in this case G.722 was the best one that could be used, with an audio bandwidth of around 7.5kHz, which has since been proven to be adequate for talkback needs.
Advances in technology saw the introduction of IP connectivity of the intercom panel, again using standards-based IP networks. This meant that panels could be sited anywhere within the Wide Area Network. Most IP panel connections shunned compression to achieve lower latency, since it’s no good telling a cameraman to switch shots and having the intercom arrive some seconds too late! IP panels were also fighting for bandwidth with other uses of the IP network, be it an office network or something also serving video, so we were introduced to the world of VPNs.
Audio over IP ‘standards’ started to appear, and the joys of a ‘plugfest’ where different manufacturers joined their hardware together to test interoperability came into style. The EBU worked first on video interoperability then on audio interoperability between intercom manufacturers. Differing standards included Ravenna, AVB, and Dante. It was obvious that a single standard was needed and various groups were working together to achieve ‘plug and play’.
The common technology below all these standards is AES67 with more sophistication built above it, like discovery and routing within Audinate’s Dante. This base level makes AES67 the AoIP choice, certainly between devices within a contained system such as between end-points within a single manufacturer’s system, where discovery and clocking is provided by the system. Dante’s rapid acceptance within the live event and performance communities meant that this was the choice for Clear-Com when providing simple connectivity to music consoles, public address, and other intercom systems.
The advantages of using IP to interconnect different audio devices include being able to leverage existing IT equipment with their diagnostics capabilities, and also use the legions of trained IT engineers—although this has brought its own problems in terms of training them to respect audio and video media. (It is perhaps more important that the audio—or video—contained within an IP packet arrives without delay than the message sent to a printer in an office!)
So, things like QoS (quality of service) and the network resilience with multiple alternate routes have become important, and the industry has created such things as ‘spine and leaf’ network layouts to provide for redundant connections with a layer of control software to ensure it all works.
AES67, with further refinements on technical options, clocking and discovery, is the common standard for audio and intercom over IP. For our customers, the ability to link multiple manufacturers’ boxes together and share multi-channel program audio, and to monitor routing and diagnostics using easily available IT tools, is releasing them from the complexities of standards conversion and hardware patching and monitoring. For these reasons, the march toward the adoption of AES67 is well under way.