One of the diagrams that I created back in 2015 has been popular for illustrating the differences between the transport systems used by SMPTE ST 2022-6 and ST 2110. SInce the old version referenced RFC 4175, I thought it would be good to update it for the latest version of ST 2110. Here it is:
I call this my "submarine" diagram, because it very vaguely looks like a sub with a conning tower, if you really squint your eyes hard. If you want to use this diagram in something your are writing, please go ahead, but please give me some credit and link to this website wherever possible.
Here are a few things that this diagram is meant to illustrate:
1) Each of the types of media is encapsulated as a separate packet stream, as opposed to ST 2022-6 (see below).
2) There are more (and frequently larger) packets for video than audio (the actual ratio is closer to 500:1 for HD video, but that is hard to show in a diagram). There are even fewer and smaller packets for ancillary data, which should be reduced by another two or three orders of magnitude from the audio stream.
3) For a device that is just processing audio data, only the audio packets need to be sent to or from that device. This greatly simplifies the audio device in relation to the ST 2022-6 audio device shown below.
4) The bandwidth of the video signal is lower as a stream of ST 2110-20 packets than it is as a native SDI signal (hence the packets are narrower on this diagram as compared to the input and output signals). This happens because only the video pixels are sent over ST 2110-20, leaving out the HANC space (used for audio) and the VANC space (used for data) from a standard SDI video signal.
5) This version of the diagram has the flows labeled with their SMPTE standard numbers, as well as showing separate SDP files for each flow. This is an important change from VSF TR-03, which used a single SDP file to represent all of the packet streams created from a single media source.
For comparison, here is the same type of diagram for SMPTE ST 2022-6:
Here are a few notes about this diagram:
1) In this case, the ENTIRE SDI signal is being encapsulated intact, including the HANC and the VANC. As a result, the bandwidth of the packet stream is slightly more than the bandwidth of the SDI video stream, as shown.
2) In this case, the audio and data signals are embedded (into the HANC and VANC respectively) into the SDI video signal before packetization occurs.
3) The audio processing device shown in green at the top of the diagram has a more complicated job as compared to the ST 2110-30 audio processor. In this case, the entire ST 2022-6 video stream needs to be sent to the audio device, where it needs to de-packetize the video and then de-embed the audio signals before processing them. It also needs to re-embed and re-packetize the entire stream at its output.
4) One advantage that ST 2022-6 streams have is in the area of synchronization. Using this format, the audio, video and data signals all travel together, and therefore remain synchronized with each other. In the case of ST 2110 systems, each type of media packet stream can undergo different amounts of delay from source to destination, and thus need to be resynchronized at the output. This is done use packet timestamps and IEEE 1588 Precision Time Protocol in SMPTE ST 2110 systems.
5) SMPTE ST 2022-8 is an enhancement for ST 2022-6 that specifies how timestamps can be assigned within the packet streams to allow for simpler interoperation with ST 2110 systems. It also provides info on how to create an SDP file for the stream.
For historical reference purposes, here is my original diagram for VSF TR-03 video/audio/data systems:
As you can see, in TR-03, multiple media types used a common SDP file. This was changed in ST 2110 because it was felt that each type of media should be individually specified. For example, an AES67 audio device would only need to read the SDP file for an ST 2110-30 audio stream, and not have to parse through all kinds of other info about video and data streams that would be superfluous to that device.
Thanks for reading!
