Your browser is out-of-date!

Update your browser to view this website correctly. Update my browser now

×

‘AI is fast, but is it fast enough for live?’

EVS CTO Alex Redfern discusses the impact AI could have on live sport production, stating that it needs to "speed up" if it is to take over processes currently done by humans

Alex Redfern, chief technology officer at EVS, has been discussing the impact of artificial intelligence on the media industry – and live production in particular.

Speaking at an event for journalists, Redfern considered the impact AI could have on the live production side of the broadcast industry.

He said that because speed is so important to live production, AI needs to “speed up” if it is to take over processes currently done by humans. “It’s fast, but is it fast enough for live?” mused Redfern. “We will get there, I’m absolutely convinced, but we’re not there yet.”

Redfern compared AI’s capabilities to that of EVS operators, who, he said, “see every input of every server. They’re watching it. They are understanding what’s happening. They can make split-second decisions that I’m not sure a machine can, and I’m not sure AI will be able to replace humans fully.”

Alex Redfern

Where AI will have an impact, he believes, is around automation, and that will ultimately lead to fewer humans interacting with equipment. “I don’t think the equipment won’t be necessary, because it will be needed to do things like make replays, or route signals, or whatever it may be. But I strongly believe that there’ll be a reduction in the human workforce in the long term as we go down that automation route, and that will probably have something of a negative effect on the industry.”

Redfern went on to discuss where EVS is considering putting generative AI into its products and how that will benefit its customers. “XtraMotion is one of the main products that we put the energy and effort into gen the AI part because we can start to generate frames of video for super slow motion. We generate video that wasn’t there before.

“The crucial thing for us is, AI often comes down to creativity versus automation. At EVS we’ve always focused on the creativity aspect, that’s what’s always been the most interesting thing for us, how can we make the image crisper? How can we slow it down? How can we give the customer more visibility? It’s not necessarily about automating, but putting those tools at the fingers of operators, giving the operator the choice, giving the director the choice of whether to use it or not, whether it makes sense, making it faster. So anything that can be a single button press, anytime where we can put, for example, a super slow motion replay, a shallow depth of field replay, at the hands of an operator with a single button push within a couple of seconds, that’s where we really start to benefit from Gen AI.”

Another key element of introducing AI into EVS’ products is that it will help to reduce costs, said Redfern. “Traditionally features, like super slow-mo, shallow depth of field, perhaps even sharpening, required really dedicated pieces of hardware. [With AI] we can start to do that from any camera, any piece of footage that comes in from around the venue, onboard cameras in motorsports, or cameras in goal posts, really small cameras can suddenly be shallow depth of field, and you start to increase the creativity, but you decrease the cost, because it means you can deploy perhaps cheaper cameras, or you don’t need to deploy as many of the expensive camera channels.”

Sustainability: hardware v software

As part of the event, Redfern also discussed how EVS is addressing sustainability issues within the industry. He described reducing the energy consumption of EVS’ product is key to how the company can help its customers with their sustainability challenges.

“If we reduce our carbon footprint, the customer does too,” he added. “Continuous improvement optimisation at CPU level, let’s say, is really important. EVS is obviously well-known as being a hardware manufacturer. Hardware can bring real benefits to ESG, it can be far less power-hungry.

“Neuron [EVS’ Network Attached Processor] as an example, for the density, for the size, for the amount of hardware used, the COTS equivalent doing it in CPU and perhaps GPU, is about four to five times as power hungry as the Neuron. Also, hardware generally lasts longer. The lifecycle generally is 7-10 years, as opposed to 4-5 years for COTS. So thinking of that balance between hardware and software, that marriage between the two, where you have more control over the software that runs on the hardware is really, really important to us.”