“Broadcasters, content providers and distributors, I come not to bury SDI, but to praise it.”
Unfortunately, the technology transition from SDI to IP has for the most part played out like a Shakespearean drama, with IP being cast as the villain. Over the past few years, slings and arrows from all directions have been flung at the idea of moving media operations to commercial off-the-shelf (COTS) equipment.
The shame of the situation is that it need not have been so. IP is not SDI’s adversary, but simply its successor. Just as analog gave way to digital and SD to HD, the venerable serial digital interface is passing the technology torch to a faster and more versatile way of producing, playing out and distributing content.
And let’s be clear, nobody is talking about a burial. While SDI will eventually find its way to that undiscovered country, it will remain a vibrant and vital part of the broadcast workflow for decades to come, living harmoniously alongside IP-based systems and equipment. In addition to being misleading, the framing of the SDI-to-IP transition as a winner-take-all, fight-to-the-finish death match has only delayed the inevitable – as well as the industry’s advancement to a more agile technology foundation.
The good news is that the finger-pointing and the mudslinging, for the most part, appear to be over. Seeing is believing and the availability of dozens of real deployments in which broadcast-quality video is being produced, played out and delivered using standard IT equipment has done wonders in recent months to change attitudes and opinions about the suitability of IP as the technology foundation of the future.
With the animosity toward IP dissipating, more and more media professionals are beginning to recognize that as venerable as SDI has been, there are actually a few things that IP is pretty good at. In fact, some of the attributes of IP that had once sounded the loudest alarm bells with broadcast engineers are turning out be the technology’s best features:
Enhancing the reliability of IP-based networks is SMPTE 2022-7, a standard introduced in 2013 that brings broadcast-specific behavior characteristics to common, IT-based networks. SMPTE 2022-7 provides “hitless” or “seamless” switching — meaning the restoration of service is imperceptible to viewers — by instructing sending devices in the network to issue two separate streams to a destination device over different paths in the network. Should errors occur in the primary stream, due to a fiber cut or a more subtle disruption, the receiving device automatically switches to the secondary path. As long as both paths do not fail simultaneously, SMPTE 2022-7 provides broadcast facilities with service-level redundancy that is not available in the SDI domain.
In general, signal processing in an IP realm can be considerably more computationally efficient than in an SDI-based plant. In many instances, processing an audio stream or metadata from an SDI-based signal requires a de-embedding process, which eats up significant amounts of computing capacity. Standards that define practices and procedures for transporting media over IP, such as the forthcoming SMPTE 2110, do not require audio or data streams to be de-embedded, sparing media companies the expense of associated computational overhead.
Ideally, when the industry looks back on the SDI to IP transition in a few years, the general consensus will be that it was Much Ado about Nothing.
In reality, the broadcast industry has been working diligently over the past few years to condition COTS environments to deliver the same, if not better, performance that media professionals expect from SDI-based facilities. The SMPTE 2059-2 standard for precision timing is a great example of these forces at work. SMPTE 2059-2 is a broadcast-specific profile of the IEEE 1588 Precision Time Protocol (PTP) standard. It defines a broadcast-specific procedure for delivering accurate network timing services in an IP-based production facility, as well as a hybrid facility that contains a mix of SDI and IP-based components. The standard provides a common clock for all components in the signal path that allows for the sub-microsecond timing of media flows through the network, enabling the alignment of all signals in the network with the same precision found in traditional broadcast environments.
With the unprecedented success of the Alliance for IP Media Solutions (AIMS) rapidly eliminating interoperability concerns related to running media operations on COTS-based equipment, the only substantial obstacle in the way of clearing the path for a smooth and thoughtful transition from SDI to IP is working through lingering misperception and misinformation — and getting familiar with the new technology and how to use it.
Blogs and opinion pieces are still appearing that warn media companies, for example, that latency can be an issue in IP networks and that engineers will need to be careful to design their IP-based broadcast facilities in a way that will mitigate latency issues. That’s just common sense, as well as a mandatory requirement of all network design projects. Nobody ever claimed that moving to IP would eliminate the need to apply sound design principles. Designing an IP-based video facility with latency issues and then blaming the technology is the equivalent of boarding a local instead of an express train and then cursing out the transit authority for making you late for work.
The transition from designing traditional broadcast networks to ones running on standard computing equipment will not be completely seamless. Broadcast engineers will need to hone a few new skills and acquaint themselves with new design principles – just as they do for nearly any transition. The good news is that there are plenty of examples of IP-based networks in action, as well as plenty of places for media professionals to acquire hands-on experience with IP.
The bottom line is that the move to IP need not be a tragedy — or A Comedy of Errors, for that matter. Ideally, when the industry looks back on the SDI to IP transition in a few years, the general consensus will be that it was Much Ado about Nothing.