Media Transport

Info
  • Data is encapsulated in Streams to use available bandwidth more efficiently.
  • Audio transport in Milan: 32bit PCM data.
  • Media Time controls playback/recording in relation to the Network Time.
  • A Stream has a configurable Presentation Time Offset in the range of 0.25ms to 2ms.
  • Bandwidth usage of Streams is explained in the Traffic Shaping section.

Audio Data

In Milan AVB, audio data is transmitted via Streams using the Audio Video Transmission Protocol (AVTP), which defines the structure of the data frames sent across the network.

The specification adopts the AVTP Audio Format (AAF) for audio transport. This is not to be confused with the Advanced Authoring Format. It mandates that each PCM sample be transmitted as a 32-bit value, with shorter samples zero-padded as needed.

A Talker in a Milan network defines the outgoing Stream format, specifying the number of audio channels in the Stream . The Listener is required to adapt to the format provided by the Talker .

The Milan Base Format specifies support for channel counts of either 1, 2, 4, 6, 8 audio channels per Stream . The support for the Base Format is mandatory for Listeners and part of the Milan certification. This ensures interoperability between any Talker and Listener .

A Stream can be viewed as a container for audio data. In addition to the audio content, it also includes Ethernet information, such as the source and destination MAC addresses. Ethernet information is present in every packet. This mandatory overhead for each packet should make it evident that using a larger number of audio channels per Stream is more efficient than using a smaller number.
Stream format diagram

It is worth noting that a Stream is configured with a fixed PTO before it starts streaming. The value range for the PTO is from 0.25ms to 2ms. Due to the fact that all devices in the network have a shared understanding of time, it is possible with Milan to guarantee the latency that has been configured for the Stream . Details on how this is possible are described in Traffic Shaping Section.

Media Clock Data

The previous section Network Synchronization explains the distinction between Network Time and Media Time in Milan .

In summary:

  • Network Time is the shared global time base provided to all devices in the network by gPTP .
  • Media Time is the timing domain used specifically to synchronize audio recording and playback clocks in Endstations.

From a transport perspective, a Stream can carry either audio date or Media Time clocking data.