Many of the latest AI-enhanced broadcast solutions focus on delivering efficiencies to different facets of sports and news production, writes David Davies.
The long-term implications of artificial intelligence are complex and potentially very significant, so it’s hardly surprising that questions of an ethical and existential nature have dominated recent industry discourse. But as critical as those topics undoubtedly are, it’s also important to remember that this phenomenal technological movement is also yielding a remarkable amount of innovation, including in broadcast production, distribution and delivery.
‘Deep and advanced’ conversations
As might be expected of such an established mainstay of live production solutions, AI is not a new topic for Chyron. From an early stage the company has been working on sports-related AI applications such as player-tracking and camera calibration, the latter involving the “tracking of the movement of the camera relative to the pitch without the actual tracking information coming from the camera itself,” explains Mathieu Yerle, Senior VP of Strategy, Product and Marketing at Chyron. “More recently, we have been looking at a different application of AI, still within sports, but based around training models to detect typical specific plays in different sports. There are various ways that can be used to accelerate user workflows, for example in producing automated or semi-automated instant replays where previously you would have a replay operator sitting and finding the best in-point, the best out-point, the best angle, and so on.”
Read more Training models responsibly: AI and creator rights
As of version 1.4 of the Chyron LIVE cloud-native live production platform, launched last July, a new AI-based instant replay tool supports production crews – especially those in single-operator scenarios – by automatically detecting and clipping major plays so they can be quickly and easily inserted into the live programme. Initially only available for football, the tool is intended to provide a “technological foundation” for other sports in future LIVE releases.
Yerle agrees that replay remains a hotbed of AI-related media innovation, along with the compilation of standout moments from matches. “An extension of [AI usage in replays] is to do highlights re-creation,” he says. “We know that many customers in sports end up doing a lot of clip selection through the game so they can use those [highlights packages] during the breaks. That’s obviously a relatively fast-paced workflow, and one that can be automated or partly automated through scene detection.”
Whilst it can sometimes feel that even the language surrounding AI is still developing, the signs are that the conversations around its use in broadcast are increasingly mature. “We’re having a lot of very deep and advanced conversations with customers [about AI]. So that means we’re probably spending less time doing evangelisation than we might have done in the past. There aren’t many days now where there isn’t a reference made to some kind of AI project,” says Yerle.
Accordingly, he implies that we have now entered into a fascinating phase in which “AI has opened up a realm of possibility around a lot of use cases, and customers are looking at them closely and trying to find the efficiencies. There is a [distinct trend] towards automating elements of workflow so that broadcasters can keep talent focused on the high-value-added parts, and I think we are going to see more of that.”
Ad verification with AI
The use of AI in ad verification processes is the focus of a newly announced innovation from Skandha Media Services, which specialises in scalable playout services that leverage cloud technology, automation, AI and machine learning. Now it has introduced Evince, an AI-based verification tool designed to bring “total peace of mind” to live sports streaming platforms and advertising agencies.
By automating the detection and capture of primary and secondary ad elements on live sports streams, Evince allows agencies and advertisers to discern exactly “how an ad appeared on a specific channel via mobile, digital or CTV, and at which point during a live event,” says Neelesh Dalvi, Head - Digital Operation, Skandha Media Services.
Video spot ads (including those carrying QR codes) are saved as files in MP4 format, and L-Band, Aston and Bug Ads are saved as PNG, JPG and MOV files. The files are instantaneously and securely stored in the cloud so that each ad insertion record can be accessed reliably in real time.
Read more AI and DAI: Machine learning and the quest for even greater dynamism in advertising
Significantly, Evince also addresses a growing “bottleneck” in the OTT era that sees brands and agencies potentially having to verify advertisements on an ever-increasing number of platforms and outlets. “The advertisers want a proof of delivery, and the conventional method is that the broadcasters or OTT operators [have a log] and tell them that it was run during so-and-so period,” says Dalvi. “But they really need a [visual authentication] of that, and with Evince we have trained the model in such a way that it can identify the particular advertising event, take screenshots according to the time frame, and make a document of it that which is given to the customer.”
“This technology goes far beyond providing traditional logs for a live sports stream that may be firing hundreds and even thousands of ads over the full length of a live game or competition,” says Yogesh Salian, Director of Client Services at Skandha Media Services. “The advertising agency not only has the ability to check that a particular ad appeared visually correctly during an event on a specific stream or platform, but also that it was delivered at a specific time, or in a particular spot, for the expected duration.”
AI for social media highlights
The streamlined generation of sports highlights for social media channels is among the primary targets of attention for Magnifi by Videoverse. Providing an AI-powered platform that enables the creation of “instant key moments” and auto short-sized videos, Magnifi works with rights holders – including the Indian Premier League, Women’s Premier League Professional Squash Association and leading sports broadcasters – to “increase their presence and viewership” on socials and digital platforms.
Current projects include the creation of highlights and web stories for several broadcasters – based in the Indian Sub-Continent and the Middle East – covering the T20 World international cricket tournament, and assisting with social media solutions for sports marketing agency SPORTFIVE’s coverage of two premier tournaments in the ASEAN Football calendar: the Mitsubishi Electric Cup 2024 and the Shopee Cup 2024/2025.
According to Meghna Krishna, Chief Revenue Officer of Magnifi, the solution is essentially “providing an AI tool for real-time content creation. [Our platform enables the] creation of short-form content in real-time to publish to social media or digital platforms, including personalised or localised content specifically for sports.”
Taking the hypothetical example of a five-day cricket match, she explains that “if you had to create a digital synopsis of that game, it might take you up to 24 hours. [With Magnifi] it can be possible to publish highlights within six minutes – and that’s not just the general ‘five-star’ highlights, but also highlights from [different perspectives] such as the home team or the visiting team. A lot more content can be created faster and which is personalised to the viewer’s liking; that’s what we can do with AI today.”
Krishna says there is no ceiling to the amount of material that the platform can sift through in order to generate those highlights packages. “That is the best part,” she says. “There is no limitation [so] the sky is the limit because everything is tagged and then how you choose to use that content is up to you.”
Like Chyron’s Yerle, Krishna feels that the increased awareness of AI-based solutions – a trend to which she suggests the advent of ChatGPT has been integral – is ushering in a distinct new phase of industry-wide development. Similarly, she also sees sports as a starting point for the application of AI-based techniques to other areas of media production.
“As we grow in the industry, we realise what the other use cases are,” she says. Ultimately, “the idea is not only to provide software for sports, but for entertainment and news as well, even for security, infrastructure and manufacturing.” Certainly, the potential applications in news – such as the distillation of complex news sources into concise summaries – would appear to be plentiful.
There is likely to be much more of this in the coming years. AI-powered technologies put through their paces in the intensely scrutinised areas of live sports will go on to find natural outlets elsewhere in media production and delivery. In the best possible scenario, the resulting efficiencies will free up operatives to concentrate on other, potentially more involving tasks. Only time will tell as to whether some human-led roles are also reduced or even lost altogether.
Read more Adobe’s Deepa Subramaniam: How AI video will shake-up post production
No comments yet