SMPTE: Standards for communicating the world

Interview with Bruce Devlin, a.k.a Mr. MXF
Standards help turn disagreements into compromises and then into industries. These words from Bruce Devlin, outgoing Vice President of Standards at SMPTE, exemplify well the need to create rules in a world that increasingly needs understanding. Rules allow communication between devices from different manufacturers in a world where the end result, the content, needs communication to exist. The technology is already available and the rules are already written.
However, the transition to IP in broadcast is taking time. And, as a general rule, the advocates of this technology cite lack of education as the root of the problem. Here the work of SMPTE (Society of Motion Picture and Television Engineers) stands out. Education through the exchange of information and the creation of standards so that everything that is between manufacturers and end-users works correctly are their main actions.
How does SMPTE work and what are its main functions? What are the advantages of embracing the IP transition? What is SMPTE 2110 capable of doing compared to NDI or other standards? What are its next steps? Let us discover these answers together.
What is SMPTE? Could you give us an overview?
The SMPTE is the Society of Motion Picture and Television Engineers. We’re just over 100 years old. We were founded, because back in 1916, there were lots of different formats for the size of motion picture film. This variety caused a lot of interoperability problems. Then, someone thought it would be a good idea to write down a standard. The idea was to have all the film with the same size, same thickness and same number of sprocket holes, so that we could distribute film around the country, America in that case.
If you fast forward to today, you look around, and there are people putting different sorts of codecs over IP. There are people putting different resolutions over IP, people have different synchronization mechanisms for IP, etc. If you are, for example, communicating with me through a Zoom call and you want to put it on air on a live broadcast, somehow you need to check the compatibility of the audio-visual signal at each one of those little steps in your chain, and you’d like to do it quickly.
It’s certainly the same interoperability problem that we had in 1916, except nowadays, the boxes are full of electronics and not full of bits of plastic with bright lights, and occasionally, smoking things melting. Actually, no. Today, we still have boxes of things that occasionally melt, so that aspect of developing new systems has stayed the same. [Laughs].
Basically, SMPTE started as a membership organization and continues to be a membership organization with sections around Europe, around America, Japan, India and everywhere. Yes, we’ve been making standards for 100 years, and we’re still doing them, but one of the main things that we do nowadays is education. We try to teach people about the fundamentals of audio-visual technology.
How do these standards come about?
What will happen if somebody has a bright idea? That somebody will probably try to exploit that idea and make it successful in the market. A standard helps companies get more users of that technology more quickly. Let’s imagine that they come up with some brand new way of distributing sound. It’s likely that many people around the world are doing similar things. For example, with Immersive audio there might be several, similar ways of creating a bitstream. They all have the same fundamental goal of providing multiple channels that can be combined into a great audio experience, but we have different interested bodies who are all trying to produce a similar result, with different technologies.
Imagine that you’re a cinema owner. Representatives of the individual solutions might come to your cinema and try to convince you to install their solution in your theatre. Additionally, movie distributors might show up at your theatre and tell you that their movies are made with different bitstreams. And, of course, you want to show all the movies. What do you do about it?
That’s really where SMPTE comes in. We say: “Well, it is possible to build an infrastructure that delivers a common bitstream and therefore simplify the distribution if we get all the proponents around a table and agree on the compromises necessary”. SMPTE creates a practical level of standards that allows operators to show more movies to more people and allows TV stations to show more TV to more people. We just want to make the world of cinema, television and streaming bigger and easier for all.
You perform like a middle stage between the manufacturer and the customer. How do you do that?
I’m going to use the analogy of an app store to answer you. You, as an end-user, want to record your voice with your phone. You search through the app store for audio recording, and you find tools that are for voice only, tools that are for music, tools that allow you to plug in an external array of microphones, tools that want you to pay €10 or €1 per year, and others that have in-app purchases, that enable recordings longer than 30 seconds. Eventually, you find one that’s just perfect. Just the right price and with just the right features for you.
Let’s think about what the app store is doing. The app store is providing metadata that allows you to make a choice. Some engineer thought that in order to make this market exist we have to give the end-user choice. They have to be able to compare apps against each other. We’re going to force the app designers to write down some key features of their apps. We also know there’s going to be two, three or four different ways you’re going to buy it. The person who’s designing the App Store is trying to figure out what things the app designer wants to do, and what things the user wants to do. And you find the level of technology in the middle.
In a way, what SMPTE does, is to act as an intermediary between what the manufacturers want to build and what the users want to do. SMPTE’s members provide the technology that has to be in the middle, between these two end-points of the market so that both can understand each other. A standard does not tell the people how to run the business, a standard merely provides interoperability solutions that allow tool providers to enable better business. That’s what we care about.
Going back to SMPTE’s areas of action, you were talking about standards, membership and education. Could you describe these areas in more depth, please?
We will start with membership. SMPTE has around 8,000 members all around the world. We tend to split our membership into sections that are usually located in a geographical region to encourage members to meet and network. Sections often arrange meetings where people who are leaders in the industry talk about what they do. It’s a very simple formula. We get real people with real experience talking what they do in the sections. Because of COVID, we’ve been doing a lot more of this online.
The next thing is education that comes in different forms. The first is actually oriented around the section events where local expertise is given a bigger audience in section event. For example, the SMPTE UK section recently had experts from the BBC research and development department explaining how are they building IP networks, and how are they doing high dynamic range. Then we have a range of formal education courses where you might learn about IP video, or how colour works, how IP networking works, how the components D-cinema work. This is SMPTE education: we educate the rest of the world about what we do, but we are also educated by other industries about what they do.
That brings us into another activity of SMPTE, which is liaisons. As Head of Standards, one of my jobs is to communicate with other organizations to make certain that we’re not duplicating our effort but we’re working cooperatively and collaboratively. For example, we work with DVB because DVB deals with satellite and terrestrial links and SMPTE deals with contributing to those links. We work with the EBU because the EBU is a group of end-users, whereas SMPTE is more about the vendor community. And we work with many others including the W3C, CTA, ATS. The list is very long.
That’s where SMPTE really thrives and we end up developing standards in specific areas that are contributed by our membership, we provide education on those standards that bring in new members who have new ideas. Those new ideas go back to the members and we make new standards. This help drive the industry forwards.
Let’s talk about IP facilities, what are the main advantages of migrating to IP in a broadcaster?
Flexibility. That is the one-word answer. With SDI each cable carried one signal in one direction. Imagine a flexible environment where, in the morning, you could be recording a soap opera in a studio, but in the afternoon you turn the set around and have a game show and in the evening you have a comedy program. Physically, it is a real challenge to move all the sets around, but also, in terms of the way that you might move the cameras or the signals, you’re fixed with SDI.
As soon as you go to IP the signals flowing around your facility no longer depend on where the wires are. You can be bidirectional on a single cable. If suddenly, you were shooting a 4K drama in the morning but in the afternoon you want to have some kids’ game show where every child has a GoPro on their head and you want a hundred cameras, you can do it with IP, and an appropriate budget.
One of the biggest, most flexible IP facilities is being built right now (Eurosport) and is able to take 1,000s of signals and route them from anywhere to anywhere and bring them together and split them apart. They can make Italian football, Spanish basketball, British snooker, German motorcycle racing, or competitions across thirty different countries in the morning and, in the afternoon, they can broadcast World Cup matches, for example. The morning could be in HD and the evening in 4K. It could be SDR or HDR. They are using the same kit, configured differently. And that is the flexibility that IP brings.
And why is taking so long and so much then?
I’m going to start with a bit of history. In 1988, SDI was first demonstrated at IBC. That was the big thing, but really SDI only worked between two videotape recorders, two cameras, and one switcher. It took around ten years for SDI fully replaces analog circuit switching.
If all you want to do is replace an SDI workflow with an IP workflow you just lift out the SDI, replace it with IP, and you can do it that in a few weeks. You go to the shops. You buy the kit. You plug the kit in. You put some converters where there is kit that will never support IP and it just works. But you didn’t buy the IP kit just to replace SDI. You bought the IP kit to, fundamentally, change your workflows in order to be super-flexible.
Recently, we had a SMPTE+ event where we got lots of people who’ve built IP facilities to just talk about their experiences. They all said that if all you’re trying to do is replace SDI, just go and buy it. You won’t notice the difference. Then, if you don’t notice the difference, why did you spend all that money on a replacement? The reality is that IP enables flexibility. Flexibility means changing your working practices. Changing the working practices means changing the way people work. Changing the way people work … that’s hard.
I should also say it’s not just SMPTE who does IP standards. There’s an open specification called NDI. You can go out and buy it and it’s integrated into lots of tools and it works. You play with it and you discover you’re doing things in a different way to the way that you did before. Once you’ve trained your people you discover that there are differences between the way ST 2110 works and the way that NDI works. If all you’re trying to do is a self-contained small production with compressed feeds then ST 2110 is probably overkill. If you’re trying to do an Olympic Games where you want 8K and 4K, and high dynamic range, and you’ve got 1,000 microphones all the way around the opening stadium, and they all have to come back in sync to the one mixing station, and they’re all 24-bit audio, and all the cameras that would do 12-bit RGB because you want the perfect quality; NDI can’t do that. ST 2110 is your only solution. That’s the thing about IP. You’ve got the flexibility that you want depending on what sort of program you’re trying to make and what your budgets are like.
What will be the next step?
Let me tell you about the way I think technology will change in the coming years. When I joined the TV industry in the 1980s, we were fighting physics. You could not build the chips fast enough at a reasonable price to do 4K TV. In the 1990s, we could build chips fast enough to do SD and HD. In the 2000s, we could write low-level software that could do SD and HD. In the 2010s, we could write software to do SD and HD and 4K, and we could build chips to do any resolution that you like. We’re now in the 2020s, I can go to the shop and spend €400 on a phone that will do 4K.
Now, in 2022, we have the unlimited compute potential of the cloud, and with my credit card, I can swipe and buy enough computing to do 8K in real-time if I want to. There are no longer any barriers to processing video. It’s literally, how much do I want to spend and how clever am I? Those are the only two barriers that exist.
The standards and specifications we need nowadays are the opposite of what we needed in the 1980s. In the 1980s, you had to design the whole system before you actually started putting anything on silicon. Now you build the whole system and figure out what it takes to make it stable. Then you build the platform out of stable components. For example, how do I connect a Unity games engine to a big LED wall with some metadata capture on the back of a camera? And, more complex, how do I interconnect workflows such as remotely controlling multi-vendor virtual cameras with a joystick? Or even more important, how do I could get my director photography in Los Angeles with an Oculus headset to walk around the set, and wherever they looks, they control the virtual camera(s) in Spain? The answer is, today we’ll build it, figure out what standards we need to make it repeatable, and standardize them, which is exactly the opposite order of how we did it in the 1980s.
How is it going to change? Everything is going to be defined by software. The standards we will need will have to do with the vocabularies that the different computer programs need to communicate with each other. In this way, we will be able to have two vendors making different devices but communicating in the same way in the same ecosystem. This will reduce the risk of miscommunication or misunderstandings.