Handbook for Sound Engineers

1510 Chapter 39

but instead uses its own protocol. Aviom says on its
Web site, “A-Net manages data differently than
Ethernet does,” which has advantages as well as disad-
vantages. Aviom at this time offers two different
versions of its technology, Pro16 and Pro64. Pro16 is
limited to point to point connections while Pro64 is
more flexible. Pro64 allows up to sixty four audio chan-
nels and lets all devices see all sixty four channels. The
Aviom protocols are low latency and simple allowing
inexpensive but effective products such as its personal
monitor mixers, which have revolutionized personal
in-ear monitoring onstage and in studios. A single Cat5e
cable to a small box by each musician carries sixteen
audio channels and power, and allows the musician to
make his own monitor mix exactly as he wishes.

39.14 EtherSound

EtherSound, as the name implies, does comply with the
802.3 Ethernet standard, but EtherSound networks are
usually not built the same way Ethernet networks are.
Ethernet networks are built using a star or star of stars
topology, where each edge device connects to a switch
port. Other than the simple case of a two device
network, Ethernet edge devices do not directly connect

to each other. Other than rare Ethernet edge devices

with redundant ports, most Ethernet edge devices have

only a single Ethernet port. Ethernet devices are never

wired in a daisy-chain or cascade. EtherSound, on the

other hand, provides in and out Ethernet ports on its

edge devices, and in many cases builds networks by

daisy-chaining its edge devices. This can result in

simpler network designs, but also means that if a device

fails in the middle of a daisy-chain it splits the rest of

the devices into two isolated chains. EtherSound can

also use switches in a more conventional star topology,

but then devices downstream of the switch can’t send

audio back to the devices before the switch. The devices

can be wired in a ring for fault-tolerance, and

daisy-chain, star, and ring topologies can be mixed in

the same network.

EtherSound has low latency and can support

multiple sampling rates mixed in the same network. In

order to get this low latency, EtherSound traffic must

not be mixed with ordinary Ethernet traffic on the same

network or VLAN. In EtherSound, 96 kHz streams

occupy two EtherSound channels, while 192 kHz

streams take four.

Figure 39-57. CobraNet“ interface hardware. From the right there is the Ethernet network connection, and the isolation
transformers. The PHY is a chip that acts as the Ethernet physical interface. The MAC is the Ethernet media access
controller. The PHY and MAC chips together with the transformers constitute a standard fast Ethernet interface. The Flash
memory serves as nonvolatile storage for the DSP firmware and the management variable settings. SRAM provides memory
for the DSP processor. All the audio and Ethernet buffers are located here. The DSP (digital signal processor) is the heart of
the CobraNet“ interface and provides all control and processing functions. The sample clock is controlled by the DSP and
either serves as the master clock or locks to the Conductor’s master clock over the network. The OP timer controls trans-
mission of packets onto the network.

DSP

Flash

SRAM

Sample

Clock

Ethernet

MAC

OP

Timer

Ethernet

PHY

Ethernet

Clock

Audio

Serial

Host