The pace of technology change in the broadcasting business, which accelerated rapidly during the COVID-19 pandemic, hasn’t slowed much in 2023. The continued growth in streaming and the resultant pressure on legacy linear businesses has pushed both networks and stations to make their infrastructures more efficient and cost-effective.
Broadcasters are continuing to adopt cloud technology and IP networking while maintaining a strong defense against cybersecurity threats. And they are also exploring the potential benefits — and hazards — of artificial intelligence (AI) to their operations.
Those were some of the key takeaways from Technology Leaders on Trends in 2024, a TVNewsCheck webinar moderated by this reporter that gathered top broadcasters and vendors to examine the current state of media workflows and forecast key technology trends for the new year and beyond. “Hybrid” was a theme that came up repeatedly, both in the context of business models and broadcasters’ technology architectures. (To watch the video of the full webinar, click here.)
Tackling Hybridity
Paul Cheesbrough, CEO of Tubi Media Group for Fox Corp., noted the “key milestone” when Nielsen Gauge data in July 2023 showed broadcast and cable TV linear viewing fell below 50% of overall TV consumption for the first time as streaming viewing continued to grow (to 38.7% over the same month).
“That key tipping point is pretty material, and it kind of summarizes nicely where we are as a business, which in terms of consumer distribution of our content is very hybrid with the traditional platforms but also the new emerging platforms like streaming,” Cheesbrough said. “As a technology organization, supporting that hybrid of platforms where the consumer exists is really central to how we’ve invested and how we structure ourselves. Now, the big focus on ’24 is obviously to keep those platforms running, but really lean in to making sure streaming is a good long-term business for us, that it monetizes well, and that the economics of it really start to pull through.”
Fox has already created a cloud-first technology infrastructure based around its Tempe, Ariz., content hub to support this hybrid distribution model of linear TV and streaming, and there are other broadcasters like Sinclair that are also aggressively shifting their workflows to the public cloud. But Bryan Bedford, head of IoT and verticals for Cisco Systems, said that most customers are looking to implement hybrid technology infrastructures that combine cloud compute with a healthy dose of private cloud or on-premise hardware. That is largely due to the cost of running broadcast workflows on public cloud platforms.
“Now people are coming to us and asking, how can we look at a hybrid environment?” Bedford said. “Not to discredit that [the cloud], there are a lot of great uses cases for public cloud, it’s awesome. But they’re looking at cost innovation and what can we do in our own data center, to maybe alleviate some of that opex sourcing.”
Bedford added that investment in 2110 IP routing infrastructures continues to grow. He said the industry has moved out of the “bleeding-edge adoption phase” as workflows have solidified and integration expertise has grown, and that Cisco has seen a “steady increase across all major areas” for 2110 including studios, production control rooms, master control rooms, stadiums and mobile production units. The company now has several hundred live 2110 deployments around the world.
“We don’t see it slowing down for some time,” Bedford said.
Lawo is also seeing interest in hybrid architectures, said Jeremy Courtney, senior director in the company’s CTO office. He estimated that most Lawo customers are still using on-premise hardware for 70%-80% of their technical infrastructures.
“There are some customers that have absolutely embraced the cloud and have gone all in, but we’re seeing more who favor the hybrid approach,” Courtney said. “If you are going to use the cloud 24/7, there are some costs that come with that. And I think it’s not unreasonable to say if you are using functionality on a 24/7 basis there’s no advantage to having that all in the cloud, by egress costs, ingress costs, etc. Then to be honest, run it on premise.”
Courtney said many broadcasters used to equate “agility” with working in the cloud, but that the industry has learned that is not necessarily true. While public cloud can help at peak times, he said, “you can absolutely have a level of agility in your data center on premise.”
As it looks to help its customers support live production, Lawo is busy developing a three-layered “dynamic media facility” with IP as the key enabler, Courtney said. The bottom layer is connectivity and edge compute, with the ability to take in and pump out signals from legacy standards like SDI, as well as “tactile” surface like pushbuttons and consoles that operators “can feel and be creative with.”
The middle layer is the local network, where broadcast technology is transitioning from “FPGA purpose-built devices with fixed functionality” to standard servers with “agile functionality, the ability to spin things up and spin them down as and when you need it.” The top layer is the public cloud, used in combination with low-bandwidth transport technologies like SRT to help manage the unpredictable nature of the internet.
“We’re bringing this all together in one package, and underpinning this package you need a cloud-native software platform that allows you to discover devices, to control those devices, and coming back to that agile piece, that platform needs to enable you to spin things up and spin them down as you see fit,” Courtney said. “And we very much believe that as our customers learn to leverage this, they’ll absolutely see much better utilization in how they consume products and services.”
NextGen TV Developments
Along with its cloud initiatives, Sinclair continues to march ahead with the rollout of the ATSC 3.0, or NextGen TV, standard, said Mark Aitken, president of ONE Media and SVP of advanced technology for Sinclair. Close to 75% of the country is now reached by a 3.0 signal, he said, and an important development in 2024 will be adding second and even third 3.0 stations in markets where 3.0 is already on-air to enable new businesses like data distribution.
“[ATSC] 3.0 is headed into the picking-upsteam phase,” Aitken said. “We’re doubling up in major markets we’re already deployed in. What’s important about that is having additional capacity. The ability to bring more stations in the same market leads to additional capacity to do things other than simply television programming.”
Given the capacity crunch, 3.0 stations are also looking to use guide data broadcast in the over-the-air signal to steer consumers to “virtual channels” and interactive applications like betting and gaming that can be delivered to smart TVs through a broadband connection. Aitken said that functionality will be demonstrated at CES in Las Vegas next month.
“It’s enabling the broadcast equivalent on the internet, that’s tied directly into the ‘store’ that a station represents,” he said.
AI’s Prospects And Pitfalls
Not surprisingly, the opportunities and pitfalls of AI dominated much of the discussion. Cisco has just published an “AI readiness index” polling 8,000 businesses across 30 different markets, said Bedford, which indicated that plans are already in place for many companies, but actual rollouts are lagging behind.
“AI is the number one topic in both C-suites right now,” Bedford said. “Our data says that 95% of CEOs have an AI strategy documented, and they’re trying to figure out how they implement it.”
Tegna is still in the exploratory phase with most AI applications, said Kurt Rao, Tegna SVP and CTO. However, he noted that Tegna, like many other large broadcasters, has already been using AI-based closed captioning systems for years.
“What’s really new in the last 12 months is generative AI, the large language models,” Rao said. “We’re very bullish on what we think it could do on a number of fronts. On the content side of the house, we think there are opportunities there to create better products and/or improve the production workflow internally. That’s one vertical use case.”
Gen AI could also be used to automate back-office functions like invoice processing or T&E, he said, as well as in IT applications like monitoring network or device logs for cybersecurity purposes.
“We’re at the very beginnings of experimenting with this, so I don’t want to say we’ve proven anything out,” Rao said. “But in pilot mode or experimentation mode, we’re certainly seeing signals that it could be very helpful.”
Aitken said that Sinclair is collaborating with Korean broadcasters on using AI to add sign language functionality as an assistive service in ATSC 3.0 broadcasts. He also thinks AI could be helpful in optimizing data flows for data distribution applications that use 3.0 in combination with other wireless delivery networks like 4G or 5G, such as delivering software updates or infotainment to automobiles.
Fox appears to be farther ahead than most broadcasters in addressing AI. Cheesbrough described how Fox has already trained AI algorithms to produce catch-up highlights for streaming viewers who are joining a game already in progress, something it first did for FIFA World Cup coverage last year and is now providing with NFL coverage this fall.
Fox has worked with OpenAI to develop search and discovery functionality for its Tubi streaming app, which allows a viewer to type in a generic theme like “romantic Christmas movies from the ’80s” instead of having to search for a specific title. And Cheesbrough said Fox is also experimenting with “caution” on various production applications for news and sports coverage, including assistive tools for journalists and AI-generated match commentary.
“In the last year we’ve seen a lot of organic adoption, which is both a blessing and a curse, given some of the security rails you need to put around some of the public large language models,” said Cheesbrough of the rapid rise of AI in general.
On that note, Fox has created a method to protect its content as it is being “crunched up and spat out on the other end” by large language models from AI companies, Cheesbrough said. He described the system, which Fox aims to open-source in 2024, as “modern-day digital rights management.” Every piece of content that Fox publishes digitally is now fingerprinted using blockchain technology, with commercial terms of trade attached to it. That becomes the integration point for companies like OpenAI and Google to start working with Fox’s content.
“It’s very early days on that, but one of the biggest things it does is it allows attribution,” Cheesbrough said. “So, if you’re in ChatGPT and you search for something and some of our content has helped produce that result, we get attribution back for the generation of that result, and ultimately, we’d like to get paid for that as well. So those rails that we’re starting to build are super-strategic and critical from a distribution point of view.
“We’re pretty cautious about [AI],” he continued. “You know, the first wave of the internet, a lot of IP and content was harvested, and media companies and publishers didn’t get fair payment for that. We’re making sure that we go into this with optimism, but also some level of caution and control.”
To watch the video of the full webinar, click here.
16 May 2024
23 May 2024
Comments (0)