The AES67 interoperability standards allow IP audio streaming to multiple devices simultaneously and efficiently. The standard specifies essential configuration parameters and each manufacturer has it's own mechanism to manage and configure AES67 settings. For example, Tieline uses the Toolbox Web-GUI and Wheatstone uses Navigator software to configure WheatNet-IP AES67 settings.
To connect an AES67 audio stream a node requires session information for stream connection management. The AES67 and ST 2110-30 standards supports Session Description Protocol (SDP) for configuring the number of audio channels per stream, encoding format, bits per sample, sampling frequency and number of samples in a packet. Session discovery is not included in the standard.
The AES67 standard provides a way for devices transmitting streams to encapsulate all the multicast address, IP address, port, packet timing and format information about a particular stream using an SDP file. This file provides information in a standardized way and is supported by Tieline codecs. SDP information can be generated by WheatNet-IP Blades and Telos xNodes and simplifies the process of configuring AoIP streams by copying data from one device into another.
A unique multicast destination address is used to identify a stream over an AoIP network. Multicast routers replicate packets and send them to network devices that want to receive them. There is a large number of multicast IP addresses available and Tieline 'Source' multicast addresses use the range 239.3.3.0 to 239.3.3.7 to send 8 streams by default.
Normally the devices streaming audio to a Tieline codec will provide a destination multicast address to enter into the codec for each stream. By default, Tieline codec destination streams are populated with multicast addresses in the range 239.33.1.0 to 239.33.1.15. Note: It is not currently possible to receive a stream on multiple Destinations using the same multicast IP address.
It is important to create a multicast address plan to allow simple identification and management of multiple streams. In a studio there will likely be many AES67 devices and it is very important to document which devices are transmitting and receiving streams from which addresses. Different manufacturers and devices use different auto-assigned multicast addresses. For example, Wheatnet uses a second octet of .192 in multicast addresses, so they look appear as 239.192.xxx.xxx.
Important Note: The default destination multicast address will normally be configured by the device streaming audio to the Tieline codec. Verify this address with the network IT administrator and then place the correct address into the Address text box in the AES67 Destinations panel. |
In the AES67 standard unicasting is supported using SIP for connection management as defined in RFC3261. SDP is used to configure information about the stream, including the encoding format, bits per sample, sampling frequency and number of samples in a packet. Note: Not supported in current firmware release.
IGMP is a communications protocol used by codecs and other hosts and adjacent routers on IPv4 networks to establish multicast group memberships to IPv4 routers. IGMP support allows network filtering of undesired multicasts. The codec supports both IGMPv2 and v3. The default is IGMPv3. Select IGMPv2 if an AES67 network requires this configuration. IGMPv3 should be selected for ST 2110-30 network compatibility. IGMP snooping is the process of listening to IGMP network traffic for delivery of IP multicasts. Network switches supporting IGMP snooping maintain a map of which nodes require IP multicast transmissions. In this way multicasts can be filtered across networks to conserve bandwidth on links between nodes.
The AES67 standard recommends using port 5004 as the default port for RTP data streams. Port 5005 is recommended as the default RTCP port. Devices sending and receiving data may use other ports from 1024 to 65535 as well.
Managed switches are normally used and recommended for AES67 network routing of audio streams. These switches allow QoS DSCP/DiffServ packet prioritization and can usually be configured using a web interface. DiffServ is a framework for classification and differentiated treatment of network traffic. DiffServ operates on classes of traffic rather than on individual traffic flows. Per-hop behavior (PHB) is fundamental to and at each router or switch, traffic is classified and retransmitted according to that classification. When network congestion occurs, higher priority traffic is retransmitted promptly while lower priority traffic waits in buffers inside network equipment and may be discarded altogether.