|
Factory Direct
This Factory Direct was submitted by Peak Audio. Live Sound makes every effort to eliminate any use of marketing inspired hyperbole.
As convergence implies, these networks are used to carry many different types of data including audio, video, lighting, security, point-of-sale information, as well as general office computer data. The world has truly become a networked world. So what's all the hype about? Networks afford a great deal of benefit to the user, including flexibility, greater channel carrying capacity, redundancy and fault tolerance schemes, dynamic routing and integrated monitoring capability. Until recently, the audio system had not been a part of this data infrastructure. Rather, it was always a separate, typically analog, wiring system. This need for a separate audio infrastructure can result in a great increase in costs of materials, labor and time.
When using a network to route audio, flexibility is greatly increased. In a properly designed network, any audio input can be routed to any audio output without having to physically move wires or make use of patchbays. This routing capability allows for different system configurations and for future changes to be accommodated without having to make hardware or wiring changes. Redundancy and fault tolerant designs can be implemented using audio networks. A system can be designed so that if any piece of equipment or wire on the network should fail, an automatic switchover can take place. In a traditional analog system, if a wire is cut someone has to discover this through troubleshooting and then it has to be repaired, replaced or "patched around". Through network management, a centrally located computer can monitor all network products and alert the system operator if a fault should occur anywhere in the system. USING STANDARD FAST ETHERNET The CobraNet protocol uses a standard Fast Ethernet network to distribute real-time, uncompressed, digital audio. It has quickly become the standard for audio networking, with close to 40 professional audio manufacturers who license the technology. Combined, these companies have product offerings ranging from standard input/output boxes to amplifiers, powered loudspeakers and networked signal processors.
Most CobraNet devices can transmit and receive data simultaneously, which is known as full-duplex operation. With full-duplex operation, the bandwidth available is 100 Mbps in each direction (100 Mbps for transmit and an additional 100 Mbps for receive), expanding the available bandwidth to 200 Mbps. This is equivalent to 128 channels of audio over a single CAT5 cable or pair of fibers! Some CobraNet devices contain two Ethernet ports. Although these ports are never active simultaneously and therefore do not increase the available bandwidth, they do provide added fault tolerance. If the active network connection were to fail for some reason, such as a damaged Ethernet cable or failed switch port, the secondary port would be automatically activated, allowing network transmissions to continue. Analog audio I/O on a CobraNet device may be in the form of mic-level, line-level or speaker-level, as in the case of a CobraNet-enabled power amplifier. Digital audio (AES/EBU) connections are also available, and we are now beginning to see a number of bridging devices that allow CobraNet to interface to other audio networking protocols.
It is important to understand that CobraNet technology itself is not a control protocol, but a medium for the transport of these protocols. As such, multiple control schemes from different vendors may be carried on the same network infrastructure without seeing or interfering with one another. Any bandwidth not consumed by audio data is available for carrying control data. One of the greatest benefits of an audio network is the reduction in the amount of wiring, while at the same time increasing flexibility. The ability of CobraNet protocol to transport not only 100+ channels of audio over a single cable, but control data as well, further illustrates this benefit. CobraNet devices also contain a series of indicators that provide information about the status of the network, which can be very useful during network setup. These may include indicators for power, link to the Ethernet network, transmit and receive activity indicators, as well as various error/fault indicators. BUILDING A NETWORK There are many different network technologies available today, but Ethernet is by far the most commonly used, with over 100 million nodes installed worldwide. In fact, approximately $2.65 billion was spent in worldwide sales of Ethernet switch ports in the first quarter of 2002 alone! Thus, when Peak Audio began development of the CobraNet protocol, Ethernet seemed the obvious networking technology choice. New features and advances are appearing more and more often in the Ethernet world, and this is one of the luxuries of piggy-backing on the computer industry. Many large companies are out there competing for top market share. As these companies advance the state of the art in Ethernet networks and continue to drive prices down, we in the audio industry can sit back and enjoy the benefits! Because the CobraNet protocol runs on a standard Fast Ethernet network, you don't need specialized or proprietary infrastructure devices to construct your network. Off-the-shelf products from many different networking equipment manufacturers are available, and many of these may be purchased over the Internet or from your local computer retailer. You don't need to be an Ethernet expert to build a CobraNet network, but it does help to know a little bit of Ethernet background information. A CobraNet network may be constructed of Ethernet switching hubs or repeater hubs. You'll sometimes hear people refer to both of these devices types as "hubs," which can be very misleading, as the functional differences between repeater hubs and switching hubs are quite dramatic. For brevity, this article will refer to repeater hubs as repeaters and switching hubs as switches.
Although CobraNet protocol can operate on both repeater and switched
Ethernet networks, Ethernet switches offer many more features than their
repeater counterparts. One of the most common network topologies is a star. In a star network, the nodes all connect to a central switch. In the graphic below, the boxes labeled "CN" are the CobraNet audio devices. Connections between a CobraNet device and the switch use CAT5 cable, which has a distance limitation of 100 meters. (See Figure 1) You can build on this basic topology to create a larger star of stars topology by connecting multiple switches together. If the connection between the switches shown below is a CAT5 link, then the distance between the switches can be no greater than 100 meters. If due to physical location, the distance between the two switches is greater than 100 meters, fiber optic cable can be used. Multimode fiber has a Fast Ethernet distance limitation of 2 kilometers. Single mode fiber run lengths can be even greater, but these distance limitations vary from manufacturer to manufacturer and depend on the specific fiber transceiver used. (See Figure 2) And if designed properly, there really is no limit to the size of your switched CobraNet network. (See Figure 3) BUNDLING FOR TRANSMISSION The CobraNet protocol packages all of the audio channels into groups called Bundles for transmission over the network. One Bundle can contain up to eight audio channels (at 20-bits), and for the most efficient utilization of network bandwidth, using this maximum Bundle size is suggested wherever possible. Bundles may also contain audio channels at 16- or 24-bits, although these are much less common. Using 16-bit audio, a Bundle may contain up to eight audio channels. Using 24-bit audio, a Bundle may only contain up to seven audio channels.
In Figure 4, CobraNet device "A" is configured to transmit (Tx) audio from the microphone on Bundle number 1, and only device "G" is configured to receive (Rx) audio on Bundle 1. Because this is a multicast Bundle number, the audio data is transmitted to all devices on the network, even though only device "G" is playing it back over the loudspeaker - the other devices simply ignore the data. If additional CobraNet devices were configured to receive audio on Bundle 1, those devices could also play back the audio material. Multicast Bundles are always transmitted whether or not there is a CobraNet device configured to receive that data. If device "G" was changed to receive a Bundle number other than "1", or if it is shut off (Bundle number 0), the traffic pattern illustrated by the red arrows above would still be present, and precious bandwidth would be unnecessarily utilized. A unicast Bundle supports a one-to-one routing of audio on the network. Unicast addressing takes advantage of a basic feature of Ethernet switches - the ability to read the destination address of a data packet and route it to only the device to which the packet is addressed. This prevents unnecessary utilization of network bandwidth, and allows for many more channels of audio to be present on your network. In fact, when using a switched network there is no fixed maximum number of Bundles possible - the maximum is determined by the network topology itself. A unicast Bundle number is in the range of 256 to 65,279. In Figure 5, CobraNet device "A" is configured to transmit (Tx) audio from the microphone on Bundle number 300, and CobraNet device "G" is configured to receive (Rx) audio on Bundle 300. Because this is a unicast Bundle number, the audio data is only transmitted to device "G" for playback over the loudspeaker. Bandwidth is only consumed over the links directly connecting the transmit/receive pair, making very efficient use of the total available network bandwidth. Unlike Multicast Bundles, Unicast Bundles will not be transmitted unless a receiver is requesting that channel. This allows destination controlled routing, where the receiver selects one of several possible transmitters to receive, and only the selected transmitter is activated.
January 2003 Live Sound International |
|||||||||||||||||||||