Cat 5 Audio Is Coming - Part 1

05-26-2014 08:28 AM

You have probably heard an increasing amount about new technologies for distributing pro audio over ethernet recently. James’s review of the Focusrite Rednet system hints at the way ahead. However there is more to audio over Cat5 cable than just the Rednet system. This approach is far from new and most people working in live sound will have come across such systems already but in studios the advantages of audio over ethernet are less compelling and consequently awareness of these technologies in studios is less advanced. So what systems are available and why do you need to know about them?
The one thing all of these systems have in common is they all use the hardware associated with ethernet. I often refer to “audio over ethernet” but I have found during my research into this article that even finding an accurate name for these technologies is a little fraught. The safest description would probably be “audio over Cat5”. Cat5 refers to the hardware (i.e. the cables and connectors associated with ethernet) but even that isn’t strictly accurate with most cabling actually being Cat5E or Cat6 these days.
This kind of detail is unnecessary for the purposes of this article. This is an overview of these technologies, not a networking tutorial and I’m an audio person not a network person. Before I started writing this piece I was aware that while I knew of lots of different Cat5 systems, I was unsure of the differences between them and I found it difficult to find a general primer offering an overview of this rapidly developing area.
Requirements Of A Network For Pro Audio

Regardless of terminology there are some basic requirements which must be met for a network to be a viable alternative to using conventional copper cables and multicores.

  • The network must have high bandwidth in order to carry multichannel, uncompressed audio.
  • It must have low latency.
  • Its performance should be both deterministic and reliable in terms of latency and channel count.

There are two alternative approaches to achieve this. The first is to use the hardware associated with ethernet but to sacrifice the flexibility of a network of many interconnected devices and run a proprietary protocol incompatible with ethernet/IP but maximising performance using discrete, point to point connections, an example of this approach would be HyperMAC. The second approach is to use both the hardware and software layers of the ethernet standard to build a high fidelity, low latency version of voip (a sort of Hi Fi Skype similar to Source Connect) which is compatible with ethernet and importantly, can use existing network infrastructure. An example of this approach would be Dante.
OSI Layers

We can’t discuss the differences between the various systems without referring to network layers. This refers to the Open Systems Interconnection Seven Layer Model. Fundamental to networking, this model defines the framework of modern computer networks. There are many people far more qualified than me to explain this and the detail is unnecessary for our purposes. More information on this is only a Google search away but the first three layers are significant to us.

  • Layer 1 - Physical Layer. This is the hardware, the cable, connectors and ports. For example Cat5 uses RJ45 connectors whereas USB uses the various USB plugs. Systems such as A-net and AES50 use Layer 1 running a proprietary format to maximise performance for point to point connections.
  • Layer 2 - Data Link Layer. This layer defines the procedure for data transfer and while it can include error correction does not include any traffic management features. Examples of systems operating on layer 2 would be AVB and Cobranet.
  • Layer 3 - Network Layer. The network layer introduces data packets and IP addresses. At this layer the data can be distributed across a network alongside conventional network traffic. Examples of systems which use layer 3 would be Dante and Ravenna.

In part 2 I’ll look at the various systems currently available. Some are well known, others less so. I’ll consider what the differences mean in terms of performance and flexibility and what is driving these developments.