In part 1 I introduced the essential features of a network protocol suitable for the distribution of Pro Audio. In part 2 I will offer a very brief overview of the competing systems currently available offering a little detail, relating them to the seven layer network model, highlighting whether they are open or proprietary systems and offer some context on how the differences between these systems make them more or less appropriate for use in different applications.
What Systems Are Available?
Dante - Layer 3, Proprietary. Rednet is an example of a Dante system. Widely supported by manufacturers such as Yamaha, Focusrite and Lab Gruppen. Dante has established itself as a highly visible audio solution and with Rednet, Focusrite have made early inroads into the studio market.
AVB - Layer 2, an open format. AVB is a set of standards and while companies such as Avid, Sennheiser and TC Electronic have implemented AVB in their products, other Cat5 systems can use the same AVB standards in their particular implementations, for example Dante and UMAN both provide support for AVB.
HyperMac/SuperMac/AES50 - Layer 1, SuperMac is the 100 Meg AES50 specification. HyperMac is a newer Gigabit implementation. It is proprietary to Klark-Tecknik and is used by Midas, AES50 is used by Behringer
Cobranet - Layer 2, Proprietary, Cobranet has been around since 1996 and although it has relatively long latency, is popular for distribution around large installed sites. BSS soundweb is a popular system using Cobranet.
Ethersound - Layer 2, Proprietary, has been around since 2001. Low latency and widely supported in PA and studio systems by manufacturers such as Digico and Yamaha.
A-Net - Layer 1, Proprietary, familiar to many studio users as the format used by Aviom headphone systems
REAC - Layer 1, Proprietary, developed by Roland for use in their V mixer system and with Cakewalk.
Livewire - Layer 3, Proprietary, used extensively in radio, particularly in the US. Licenced by Axia Audio.
UMAN - Layer 3, Proprietary, this system is a professional media network designed for a wide range of uses, one of which is audio.
Ravenna - Layer 3, Open, primarily targeted at the broadcast market Ravenna is being seen as an open source rival to Dante. Lawo use this system extensively. Genelec and Pyramix also support this technology. AES67 is the new AES standard for audio over IP and is based on Ravenna. This ratification by the AES is a significant advantage for this format and makes it one to watch.
Waves SoundGrid - Layer 2, Proprietary. This system is well known to the Pro Tools community. it is an example of Cat5 audio being put to a slightly different use. This gigabit only system enables extremely low latencies between host and the sound grid server.
There are other formats available, examples would be Rocknet, Optocore and Omnoe. A related technology which takes an alternative approach is SonoNet which is a wifi audio distribution system used by Sonos for their wireless domestic audio products. Interestingly SonoNet is an example of a L2 wireless network which is not WiFi (802.x). It does with wifi what layer 2 formats do with ethernet, using the hardware and procedures for data transfer but running its own proprietary format. Another interesting development is MADI TP. MADI TP runs MADI over twisted pair Cat5 cable. This hybrid approach takes a proven technology (MADI) and creates a hybrid approach of MADI send and receive with IP data packets running on the same cabling and infrastructure. This allows reliable audio transmission via MADI with control data over IP down the same cable.
Compromises - Speed Vs Flexibility
In spite of the available systems all using identical hardware, most of them are incompatible with each other and depending on the particular implementation they all trade absolute performance against flexibility. Very generally speaking the systems which work on layer 1 provide the highest performance in terms of channel count and latency but provide point to point connections in a similar way to the copper multicores they might be replacing. Significant benefits of this type of system include the low cost and convenience of Cat5 cables over conventional multicores and the ease and fidelity of splitting and duplicating signals which can be problematic and expensive with analogue signals. Systems which work at layer 2 and particularly layer 3 offer the benefits of true networking with some systems being able to use existing network infrastructure already installed in buildings. This is done at the expense of absolute performance but professional performance can still be achieved with guaranteed channel counts and low latency.
The different approaches suit different needs and the performance focussed layer 1 HyperMac used by Midas on their live consoles suits their needs whereas a system like Ravenna is gaining popularity in broadcast where its openness is more important than the absolute lowest latency possible. Many of the live sound console manufacturers are offering different output cards to make their consoles “format agnostic”. This approach allows companies such as Yamaha, who offer a wide range of output cards, maximum flexibility with the minimum of commercial risk.
When I first began researching this piece I wasn’t sure what I thought about these formats. I had some knowledge of the various formats but the more I researched the more complicated it appeared to get. This is a rapidly developing area and it is difficult to really understand the differences between formats without an understanding of the fundamentals of networking. I have to thank Andrew Hingley at HHB who was an invaluable source of information. As well having an overview of the market he offered the kind of insight that only comes with experience. For example the flexibility and plug and play nature of Dante looks like an advantage but at present there is no way of restricting access to devices on the network. (N.B - this post is from 2014, this situation has since changed with the introduction of Dante Domain Manager) The very attractive idea of building large, multi-room facilities running Dante networks would have the disadvantage of unrestricted access to all parts of the network, a bit like having all the connections in a facility going through a central patch bay in the lobby. There is still a place for dedicated point to point connections as what they lack in flexibility they make up for in security.
Why Should You Care About Cat5 Audio?
The benefits of audio over Cat5 are self-evident in live sound but why should you care about this new and confusing technology if you are happily using analogue over copper much like you would have been thirty years ago, though admittedly with very different equipment connected to it.Well firstly because its definitely coming, it will allow a new generation of streaming audio and video embedded devices with price and performance previously unachievable. For example AVB is of much wider interest than pro audio. The automotive industry is a supporter of the development of this industry. An example I’ve heard more than once is that the second heaviest component in a car is the wiring loom. If it can be replaced by some Cat5 cable connecting AVB compliant devices then it will be. Avid are supporting AVB and the S3L console is an AVB system. However probably the biggest influence on the propagation of Cat5 systems and particularly AVB at the moment is video. The bandwidth demands of new broadcast formats like 4K video are forcing investment in new infrastructure and to some extent audio will be encouraged into Cat5 solutions by the video market.