Sometimes I have to laugh when a client remarks on how many cables are in a system. In many cases there are only a handful and I think back to production facilities that use thousands of cables. But anyone who’s done residential AV knows how even one wire can be perceived!
Nevertheless, cable counts have fallen dramatically over time, mainly due to the advent of digital signals, multiplexing, and networking. I believe it’s necessary to think about signal transport, because there are many ways to get audio and video between devices or locations independent of the signal or the medium.
Arguably, one of the positives about AV moving onto Ethernet networks is that signal transport becomes simpler. Follow the rules for networking and whatever payload you send will get there. But not everything should travel as network data, and some things likely never will.
So transport may mean coax, audio or CATegory cable; HDMI and similar; LAN, WAN or internet; USB or Thunderbolt; or a fiber optic link. And those are just examples with physical connections. Through-the-air transport includes microwave and satellite, conventional TV and radio broadcasting, Wifi (wireless Ethernet), infrared, and other options across the RF spectrum.
Along with many of these transport methods come protocols that operate within equipment to format data for a particular medium and negotiate between sender and receiver. In some cases, the protocol is the transport mechanism, having no relationship to the carrying medium.
The sections below attempt to organize modern and “legacy” transport methods into common groups. The grouping is not perfect, some transport types may be missing, and there’s a lot of overlap and ambiguity, so terms like “format” and “connection” mostly serve to get the idea across. It’s not feasible to explain all the underlying technical details here, so that is left for the reader to research as needed.
One critical point is that similar or identical cables and connectors may carry entirely unrelated, incompatible signals!
Defining Terms
Unidirectional
This category covers transport methods that are unidirectional–signals move only from source to destination. These methods are generally the oldest and usually carry baseband signals, except as noted.
Note that “conventional” OTA broadcasting has one important attribute not shared by most other transport methods: inherent one-to-many distribution. This makes it ideal for certain types of communication, such as emergency notifications, as well as efficiently reaching a large audience.
Handshaking
This category covers analog and digital transport that includes bi-directional communications between source and sink devices, usually to establish compatible parameters.
* Note: VGA and DVI analog (DVI-I or DVI-A) are compatible with passive adapters. DVI-D and HDMI are compatible with passive adapters. HDMI and Displayport are partially compatible with passive adapters (cannot go HDMI to DP passively).
Protocol Based
This category includes transport modes that have no “native” signal format or medium. That is, the hardware or software devices use protocols to establish the nature of the connection and what happens to the data.
A typical example is streaming, where the mechanism is Ethernet (wired, wireless or the internet) and the transport is handled by a protocol like RTMP. In this case protocols may be chosen for specific reasons, but the underlying transport mechanism is essentially the same.
Using USB physical connections to bring video and audio into a conferencing app like Zoom is kind of an odd case. This approach uses UAC (USB audio device class) and UVC (USB video device class) drivers included in computer operating systems to recognize and receive content from webcams and other devices. USB, which is a generic data path, is the underlying transport mechanism, and UAV/UAC are enabling conduits/protocols.
It’s important to make the distinction between USB in UAC/UVC applications, versus using USB (or Thunderbolt) for audio and video hardware interfaces to connect with production and editing software; for example, an AJA IOX3 interface talking to Adobe Premiere. In this case different OS components may come into play, or drivers may need to be installed for the hardware to be seen. These connections are not UAC/UVC, and particular software applications may or may not support UAC/UVC. Conversely, conferencing apps may support non-UVC devices, in whole or part, but often do not. This changes constantly.
To make things more interesting, because UAC/UVC is fairly universal, manufacturers may include those capabilities in hardware or software (such as the NDI Webcam Input Tool) so that non-conferencing devices can “look like” a webcam. Or, conversely, so that webcams can be used with professional production equipment.
Of course USB-C and Thunderbolt are also used for video directly from computer graphic outputs to displays. These implementations use alternate modes that repurpose some connector pins to carry HDMI or Displayport video instead of other data.
Special Cases
In the early days of HDMI I seem to recall that cable length was supposed to top at around 20 feet, but we all know that HDMI cables can work at much longer distances. In some cases, this can be achieved simply by how the cable is designed—using larger conductors, reducing capacitance, better internal shielding, and other techniques can improve signal recovery.
Beyond that are cables that include active signal amplification, handily powered by the HDMI interface itself. Even more distance can be achieved by cables that convert the copper HDMI to fiber optic and back. Extension technologies like HDBaseT are another option.
As is always the case, when data rates go up, distances on simple copper cables go down. So passively extending USB 1.0 beyond the “official” 15-foot limit may be workable, but not so with USB 2, 3 or Thunderbolt. Which is why active copper or fiber extenders are available for USB and Thunderbolt. Ironically, the very signal formats that have reduced the number of cables also make long runs less feasible!
In a different category, external distribution amplifiers and “boosters” have been around for as long as signals have run on wires. Amplification and cable equalization have always been important for analog video to pass beyond a few dozen feet without degradation. Likewise, amplification and reclocking is often used with SDI video and AES audio.
Amplifiers inherently add noise, but this is mostly a concern with analog audio, where the noise is in the audio band. Digital signals are relatively immune to the kinds of noise created by amplifiers, or picked up on cables, because it does not become “part of” the signal payload. However, digital signals suffer loss of level, rounding of waveform edges, and smearing (jitter) which can make the data unrecoverable— hence the need for reclocking amplifiers which produce a new, clean output signal on long runs.
Fiber optic connectivity is in many ways a unique case. Fiber can, theoretically, carry any type of signal with the proper interfacing, and exhibits none of the degradation of copper. The light passing through does diminish with distance (optical loss), but that distance may be several kilometers for single-mode fiber.
Not only can fiber carry immense amounts of data, which means that many bitstreams can be multiplexed into a single light wavelength, it can also carry multiple wavelengths of light simultaneously. The science and practice of using fiber has its own rules and methods.
One small note about fiber: In Ethernet environments two strands are generally required for a complete, bi-directional circuit. Just as at least two pairs are used in an Ethernet cable (one pair in each direction between two devices) the same is generally true with fiber; Ethernet connections come in pairs. Because of this, connector polarity must be observed to maintain the correct direction of each fiber in the pair. For more see: https://blog.leviton.com/fiber-optic-polarity-101-b-polarity
