Nevion AES-IP-MUX, AES-IP-MUX-SFP, ADA-IP-MUX, ADA-IP-MUX-SFP User Manual

nevion.com
AES-IP-MUX/-SFP
ADA-IP-MUX/-SFP
User Manual
Revision: A
2016-04-01
Contents
1 Revision History 5
2 Product Features 6
3 Introduction 7
3.1 Top view 7
3.2 Product Description 7
3.3 The Nevion IP audio concept 9
3.4 AES67 9
3.5 Packet times and latency 9
3.6 Unicast broadcast and multicast 10
4 Specifications 12
4.1 Copper ethernet 12
4.2 Optical ethernet 12
4.3 AES3 12
4.4 Sample buffers 12
4.5 Audio Latency 13
4.6 DARS output 13
4.7 AES67 streams 13
4.7.1 Packet times 13
4.7.2 Audio word length 13
4.7.3 RTP timestamp offset 13
4.8 PTPv2 14
4.8.1 Domain 14
4.8.2 Modes 14
4.9 Other network protocols 14
4.10 Power 14
4.11 Multicon 15
5 Configuration 16
5.1 iPath mode 17
5.2 Multicon Matrix mode 17
5.3 Fixed Matrix mode 18
5.4 DIP switch descriptions 18
5.4.1 Base Addresses SW1 1-8 18
5.4.1.1 Matrix mode 18
5.4.1.2 iPath mode 19
5.4.2 Audio channel direction SW2 1+2 20
5.4.3 Packet time SW2 3+4 20
5.4.4 Audio word length SW2 5+6 20
5.4.5 Operational mode SW2.7 21
5.4.6 DIP config mode SW2.8 21
5.4.7 Example 1:Point to point 21
5.4.8 Example 2: Multicon matrix module Outputs 25 to 32 21
5.4.9 Example 3: iPath control Unicast 22
5.4.9.1 Source 22
5.4.9.2 Destination 22
5.4.10 Example 4: iPath control Multicast within an ethernet segment 23
5.4.10.1 Source channel 23
5.4.10.2 Destination channel 23
5.4.11 Ethernet media control 23
5.5 Multicon controls 23
5.5.1 IP addresses 23
5.5.2 Target system latency 23
5.5.3 DARS output format 24
5.5.4 Automute 24
5.5.5 Audio packet parameters 24
5.6 Multicon GYDA 24
5.6.1 Hot-swap 24
5.6.2 Operational modes 25
5.6.2.1 Multicon matrix mode 25
5.6.2.2 iPath mode 27
6 Connections 29
6.1 Standard ethernet backplanes 29
6.2 SFP option backplanes 30
7 LEDs 34
7.1 Status LED 34
7.2 LAN link LED 34
7.3 PTP LED (marked EDH) 34
7.4 Optical option LED 35
7.5 On-site re-programming. 35
8 General environmental requirements 36
9 Product Warranty 37
Revision History 5
AES-IP-MUX/-SFP ADA-IP-MUX/-SFP User Manual Rev. A
1 Revision History
Revision Date Comments
A 2016-04-01 First revision
Product Features 6
AES-IP-MUX/-SFP ADA-IP-MUX/-SFP User Manual Rev. A
2 Product Features
IP audio infrastructure scalable to enormous networks
Digital audio transport and routing on gigabit LAN
Uses existing CAT5 cabling and SOHO ehthernet switches
Low latency AES67 linear audio (down to 250us packets)
Transparent AES3 mode
Network PTP clock reference
Unicast or multicast
AES-IP-MUX:16 AES3 ports per module, configurable direction
Digital audio sync output. AES11 or wordclock.
Optical network connection on SFP option
ADA-IP-MUX:8AES3 ports permodule, 8 analogue audio channels. Three direction com-
binations.
Introduction 7
AES-IP-MUX/-SFP ADA-IP-MUX/-SFP User Manual Rev. A
3 Introduction
3.1 Top view
3.2 Product Description
The AES-IP-MUX and ADA-IP-MUX are used to transport a large number of digital audio signals over a dedicated IP local area network. They may be used for:-
Large scale audio infrastructure (iPath mode)
Smaller scale matrix type infrastructure up to 128x128 stereo (matrix mode)
Point to point links over IP (fixed matrix mode)
The ADA-IP-MUX product is a combination of an AES-IP-MUX, one of three audio converter modules and a common backplane. Most of this manual refers to the AES-IP-MUX. The audio converter modules are ADC-AES8, DAC-AES8 and the ADDA-AES8. The -AES8 converter digital audio channels are connected to the IP-MUX channels 7-10. This is to allow the use of the ADDA-AES8 converter which has two digital inputs and two digital outputs.
The AES-IP-MUX has 16 AES3 audio ports which may be used as inputs or outputs, sources or destinations in the network. The AES-IP-MUX uses a central timing reference (PTPv2) and all digital audio signals are locked to this reference. The AES-IP-MUX encodes AES3 linear PCM audio into AES67 packets. The packet times sup­ported are 0.25ms, 0.5ms, 1ms and 4ms. The sample frequency supported is 48 kHz. The audio signals are transported either in a standard 16 or 24 bit packet format or completely bit transpar­ently AES3. The audio transport has a minimum latency of just over the selected packet time plus the network delay. The network delay is less than 0.1 ms in a normal gigabit LAN. The encoding parameters are set on the AES3 input (source) ports and are automatically detected on the output (destination) ports.
Introduction 8
AES-IP-MUX/-SFP ADA-IP-MUX/-SFP User Manual Rev. A
The module is intended to be used in a centrally managed system with ’out of band’ management. i.e. The Multicon frame controller is used for routing and configuration which may be connected to a separate LAN. The AES-IP-MUX does not use ’client to client’ managed routing or session management. Larger systems, spanning multiple audio LAN segments may be realised with Video iPath.
There are two operational modes for the modules.
1. iPath mode.
2. Multicon Matrix mode.
Large systems may be realised using Video iPath and modules in iPath mode. This mode uses Multicon GYDA but the settings of the modules are controlled directly from iPath.
The Multicon Matrix mode can provide up to 128 2-channel audio connections. Routing control may be dynamically controlled with hardware router control panels or web panels, or may be static for point to point links.
The modules may also be equipped with an SFP providing optical link capability.
Three backplane type are presently available with a standard network ethernet connector.
AES-IP-MUX-C1 has two DB-25 connectors (16 AES3). Sync output on DIN 1.2/3.
ADA-IP-MUX-C1 has two DB-25 connectors (8 AES3, 8 analogue audio)
ADA-IP-MUX-C2 has 16 Molex KK connectors (Flashcase, 8 AES3, 8 analogue audio)
The SFP option adds an SFP cage and short fibres for connection to the backplane SC connectors. Standard MSA-SFPs with DOM (digital optical monitoring) may be used. The Nevion SFPs range may be found at:-
https://nevion.com/products/gbe-sfps-range
Four single width backplanes in the VMUX range may be used when the SFP option is present and no electrical ethernet connection is required.
AES-VMUX-C1 has a single DB-25 connector (8 AES3)
AES-VMUX-C2 has 16 Molex KK connectors (Flashcase, 16 AES3)
ADA-VMUX-C1 has a single DB-37 connector (4 AES3, 8 analogue audio)
ADA-VMUX-C2 has 16 Molex KK connectors (Flashcase, 4 AES3, 8 analogue audio)
The optical input provides an alternative network interface but the actual network connection is chosen automatically. The input switches if a suitable signal is not present, or if the signal disappears. The optical interface has priority if both interfaces are present on power-on.
The following dual width backplane is available with both optical and electrical ethernet connec­tors.
AES-MUX-C1 has two DB-25 connectors (16 AES3) and dual optical connectors. Sync output on BNC.
Introduction 9
AES-IP-MUX/-SFP ADA-IP-MUX/-SFP User Manual Rev. A
3.3 The Nevion IP audio concept
The Nevion AES-IP-MUX produces a standard AES67 audio stream. The packets are always 2 channel with a sample rate of 48 kHz. The packet times are from 1/4ms to 4 ms and the audio word lengths may be 16 or 24 bit or transparent AES3 as used in Ravenna networks. The Sample rates and packet times are limited to reflect the typical usage in broadcast infrastructures requiring low latency, high quality and few ’flavours’. An infrastructure format should be one supported by ALL connected devices.
The audio data is packed into the AES67 packets in either a standard RTP L16 or L24 bit packing or an AES3 transparent packing mode using 32 bits per channel. The standard RTP audio packet does not support the V, U and C bit of AES3 audio, so Nevion has used the same AES3 packing format as the Ravenna IP audio network, which is based on a format described in firewire audio.
The AES3 sample clock is derived from the PTP time reference. The AES-IP-MUX is to be used in a synchronous broadcast infrastructure.
The AES-IP-MUX is intended for use in a network where gigabit ethernet links are available with­out unpredictable traffic causing packet loss or congestion. The high network bandwidth means that it does not need any of the bit-reduction codecs or forward error correction schemes, reducing the cost of the modules and keeping the latency of the transport low.
3.4 AES67
The AES-IP-MUX uses AES67 compliant streams. The audio is packed into RTP (Real time proto­col) packets. The PTP (precision Time Protocol) is used as a centralized ’house’ synchronisation clock. AES67 covers more than the stream format.
AES67 was created primarily as a standard for equipment interconnect. These connections require ’end client’ management to manage audio content and connections. The Nevion IP audio concept is simpler than AES67 in that a central ’out of band’ management (control network is not the same network used to convey the audio) is used to control the network. This is closer to a traditional broadcast facility audio infrastructure.
The AES-IP-MUX is simpler to use and more secure than most AES67 interfaces as it does not require the session control protocols or the automatic discovery mechanisms mandated by AES67.
3.5 Packet times and latency
A default packet time of 1 ms has been chosen as it is the mandated packet time for AES67 and results in a packet which is far below the normal Maximum-Transport-Unit (MTU) of 1500 octets. This is the normal maximum packet size on ethernet. 4 ms of 48 kHz AES3-packed bit stereo audio gives a packet payload of 1544 octets. Modern switches and the AES-IP-MUX have no problems handling the larger packets but some PC interfaces have problems with packets larger than 1500 octets. All other combinations of sample rates and packet sizes will give smaller packets and lower network efficiency due to the ethernet and IP wrappers.
The AES-IP-MUX has only one sample rate and four packet times as it is intended for broadcast use. This also reduces the complexity of the module. Low packet times result in low latency but they reduce the efficiency of the network. 1 ms AES3-packed packets only contain 384 octets of payload. All RTP-based audio packets have a packet overhead as shown in the following table.
Introduction 10
AES-IP-MUX/-SFP ADA-IP-MUX/-SFP User Manual Rev. A
Overhead type octets
Interframe gap 12
Preamble 8
Ethernet header 14
IP header 20
UDP header 8
RTP 12
Frame checksum 4
Total = 78
The network overhead ratio is 78 / (384 + 78), about 17% of the traffic with 1ms AES3 packets at 48kHz. The total latency of audio over AES-IP-MUX modules is mostly due to the AES67 source buffer. This buffering collects the samples to fill each packet. The transmission of the packet to the remote module takes a very short time as long as the network is not busy. The module at the destination may begin to output the audio samples immediately after reception.
The speed of sound in air is about 340m/s so a 1ms delay is equivalent to the sound travelling
0.34m. The total delay of the audio traversing a gigabit IP network with 1 ms packets is only just over 1ms which has an equivalent audio distance of around 0.34m. The distance from audio sources to microphones is often larger than 1m. Foldback of audio over the IP network to commentators for example will give comb-filter effects but will not be detectable as echoes. The limit for echoes is about 20ms. 4ms packets will still give acceptable latency for most usage while 1ms packets may be preferable for signals that will be used for foldback.
3.6 Unicast broadcast and multicast
Unicast is the standard method of addressing packets with both ’to’ and ’from’ ip addresses. The routing information is sent to both modules involved. i.e. Both source and destination modules must be told the other endpoint address of the connection.
Multicast is used to send the audio stream to all destinations on the current network segment. It is set when the audio source is to be sent to more than one destination. The modules have a packet filter so that any unneccessary multicast packets may be discarded without blocking the interface. Multicast produces more traffic than unicast but does not require the source to send more than one packet per packet time period which greatly simplifies the module. A gigabit network can support more than 256 source channels all set to multicast.
Multicast is also used when the audio is to be sent to several destinations on other network segments. The multicast stream will be broadcast on the local network segment but edge routers will be aware of the stream and announce this to the network beyond the source segment. The audio will not
Introduction 11
AES-IP-MUX/-SFP ADA-IP-MUX/-SFP User Manual Rev. A
be sent until another router communicates that it wishes to receive the stream. This negotiation is handled by the IGMP protocol (This will be available in a later firmware release).
Routing may also be acheived with the OpenFlow protocol which results in a centrally managed switched network. This network has a predictable latency, switching and traffic behaviour. This is used in the Nevion Video iPath control system.
Specifications 12
AES-IP-MUX/-SFP ADA-IP-MUX/-SFP User Manual Rev. A
4 Specifications
4.1 Copper ethernet
Standard 1000-Base-T
Auto-negotiation On
Duplex Full
MDI Auto MDI-X
4.2 Optical ethernet
Standard 1000-Base-X
Optical range depends on the output power of the transmitter, the sensitivity of the receiver of the receiver module and the attenuation of the eventual multiplexing filters. See the Nevion SFP datasheets available from support or the Nevion web site.
4.3 AES3
Inputs and outputs according to AES3-2003
Physical interface 110 ohm transformer balanced
Sampling frequency 48 kHz
Double sample rates using AES3 Single channel double sampling frequency mode is supported. i.e. 96kHz
Mono signal using both sub-frames of the AES3 signal.
AES3 Outputs may be configured to switch off completely if there is no stream received (automute active) or generate a silent AES3 stream.
4.4 Sample buffers
The audio sample buffer is 512 stereo audio samples per channel irrespective of packet word length.
Loading...
+ 25 hidden pages