The huge demand for content on the ever-increasing array of platforms and services is being matched by an increasing reliance on the use of metadata by broadcasters and content producers to help them find, manage and monetise their content. 

14788_aiinsight3x2_517483

Metadata: More important than ever in helping companies keep track of their content

The old media industry adage of content being all important might be in need of an update. Although ‘content is king, when fully appended with metadata’ is a bit of a mouthful, it does portray some of the increasing significance that this additional information is having for the broadcast industry. 

  • Themed Week: Content Management This article is part of IBC365’s Content Management Themed Week. For more articles click here

Metadata is being used right from the start in many media industry activities. On the production set, lens, focus, pan and tilt, and iris information can be harvested during shooting by intelligent lens and camera systems and stored as metadata, enabling workflows such as virtual production and VFX to use the metadata to accurately place CG assets in a scene, or place actors into digital environments.  

In a similar manner, on-set colour correction and digital dailies are preserving the shot data for use in post-production. In newsgathering, metadata systems, such as DPP Metadata Exchange for News, have evolved which gather metadata from every news workflow stage, from planning through to information automatically generated from camera acquisition and editing/production and on to ingest. The DPP claims use of its common metadata set promises easier search and discovery, clarity around rights information and smoother collaboration.  

While these examples are obviously a step forward, an industry awash with metadata is only useful if you can organise it, as well as monetise the relevant content. Here too systems have evolved to harvest, intelligently analyse and sort the asset information. 

The BBC has been developing Object-Based Media (OBM), a format that allows a piece of content to change according to the requirements of individual consumers. OBM productions exploit automated personalisation and production metadata to allow viewers to receive tailored content and control camera angles for example, while authoring tools such as StoryFormer can be used to create responsive stories. 

“Using GrayMeta’s Curio to automatically tag the footage has brought what had been a ‘dead’ archive very much to life and, importantly, made it available,” Dan Carew-Jones, Arrow 

14784_graymetacurio_920977

GrayMeta: Curio platform cuts down on tagging time

Arrow International Media has been using metadata and AI to discover and unlock its extensive archive of previously unused footage. It applied GrayMeta’s Curio platform to create, extract and store intelligent metadata from 70,000 minutes of content. “Using Curio to automatically tag the footage has brought what had been a ‘dead’ archive very much to life and, importantly, made it available,” says Dan Carew-Jones, post production consultant, Arrow. “We use the object recognition functions to discover what is actually contained in the footage and make that information searchable, and then use facial recognition to build more granular searches for specific individuals. We can then use all of this intelligence to inform the development and production processes for future projects.” 

“On a basic level we can search for specific shots in our archive based on specific production requirements which can reduce the load on our shooting teams,” he adds. “Beyond that we can search the archive for footage that will compellingly illustrate new ideas or stories for future programmes.”  

D Carew Jones_whitebg

Dan Carew-Jones

According to Carew-Jones, the rights management aspect of the workflow is still an additional manual process. “We can see how that aspect could be built into the system using OCR functionality to tag print documents and bring them together with their respective media,” he says. “This has unlocked an asset that we knew we had but hadn’t had the opportunity to fully develop through manual tools.” 

Distribution demands 
There’s a huge demand for content on the ever-increasing array of platforms and services and, according to Kip Welch, SVP special project at MovieLabs, that trend has been matched by the growing importance and use of accurate metadata across multiple distribution channels.  

“Broadcasters and other content producers today need consistent metadata management across linear broadcast, catch-up TV, TVOD platforms and the many new SVOD and AVOD platforms,” he says. “Marketing and consumer engagement with the content needs to be synchronised across those platforms, so good metadata systems that can deliver the right metadata with the right quality to all channels is very important.”  

Consumers need to be able to find the content on any platform where it appears, says Welch, “so delivering the right metadata to enable accurate universal search across platforms has to be part of any content producer’s mission”. 

Welch also observes a broader adoption of common supply chain specifications across producers and across platforms. “For example, major Hollywood studios have repurposed the common metadata and media delivery specs used for data exchange with their TVOD distribution partners for use with their own direct-to-consumer services,” he says. “We think this trend of unifying workflows in the distribution supply chain to support different kinds of content services (such as TVOD and SVOD) will continue. We are also seeing metadata management and standardisation moving upstream in the supply chain.”  

 “Delivering the right metadata to enable accurate universal search across platforms has to be part of any content producer’s mission,” Kip Welch, MovieLabs 

“The industry is adopting more cloud-based production workflows and is moving to support a variety of content release window strategies across distribution channels [Welch points out the growth of simultaneous theatrical and digital releases],” he continues. “Metadata management [therefore] becomes a more important part of early production workflows, along with early registration of data in registries such as EIDR.” 

EIDR, or the Entertainment Identifier Registry, is a universal unique identifier system for movie and television assets, described by Welch as “a lingua franca for title identification and version management”. EIDR was founded by Movielabs, CableLabs, Comcast and Rovi on the principle of open participation and operates on a non-profit, cost-recovery basis.  

“EIDR provides an industry-run, low-cost infrastructure for universal identification,” says Welch. “It delivers a consensus data structure for titles and versions that eliminates massive duplication of effort across content workflows. 

“That is absolutely necessary in today’s content ecosystem where virtually all content moves across facilities, systems, channels and platforms to reach consumers. From a practical perspective, there is no such thing as a content walled garden anymore. Content producers and distributors cannot afford to rely solely on bespoke identification systems that slow down workflows with partners and require constant rematching and reintegration.” 

Kip_Welch_2018-1001-e1524163135853

Kip Welch

Talking the same language 
Welch says there are new challenges facing broadcasters, production companies, post facilities and OTT service providers looking at working with metadata for content management.  

“Some of the challenges are very familiar, such as metadata quality and preservation of metadata from production forward into distribution,” he says. “Other metadata challenges are new and arise directly from the trends we are seeing: it is more important than ever to link different metadata systems, which means a greater need for common metadata formats and mappings between them. Similarly, the use of entirely different bespoke workflows for different channels and platforms is just no longer feasible in a world where distribution occurs simultaneously across platforms and regions.” 

Thankfully there are existing and emerging common standards available to help the broadcast/film industry manage content. “There are standards and specifications that help a lot,” says Welch. “We are obviously big proponents of the MovieLabs Digital Distribution Framework (MDDF), which provides standardised approaches for metadata and media delivery, as well as licensing and availability data necessary to launch products on digital platforms. Registration of titles and versions in EIDR also provides a standardised method for precise identification of content and versions of content that all producers, platforms and service providers can share. It is eye-opening to see what a difference just that one step can make in helping partners work together across systems and platforms to deliver content to consumers.” 

Full auto 
“The holy grail of digital production and distribution,” is how Welch describes the goal of adding editorial, rights and technical metadata to content fully automatically. 

“We like to think it is possible, because new automation specifications are available right now to help make it happen,” he says. 

Welch adds that the industry is actively developing new production workflows, especially cloud-based, to help. “These enable greater automation around many common production tasks and more linkage across production workflows and on into distribution.” 

But human intervention is still required today for some of the steps. “Although there are automated tools that can help with identifying sensitive content, oversight and application of different standards across different territories often involves subjective judgment, so human involvement remains necessary,” Welch admits.

“However, accuracy is something that automated systems can help with a lot. When content is registered in EIDR, group curation by the EIDR community of metadata experts has the potential to improve data accuracy, and then those improvements can be shared and propagated across companies and systems for quality assurance purposes. There are also machine learning tools available that can go a long way towards finding and correcting common data capture mistakes. We believe there is significant upside potential for more accurate data capture as the industry adopts new and better standards and tools.” 

The ‘Holy Grail’ may take time to materialise, but Welch believes it can happen with enough industry commitment and collaboration, “especially given the economic imperatives driving it to happen”.