Charles Sevior is CTO, unstructured storage division, of Dell EMC.
BY CHARLES SEVIOR
“We don’t even know what we’ve got in our archive …”
Comments such as this are all too common in the industry as media organisations have woken up to find their organically grown, sprawling content archives are missing a key ingredient: metadata.
Owners of large content libraries are bound to get overwhelmed by the sheer volume of media assets locked up on video and data tape. Without complete and accurate metadata — the information on what is stored, for instance, the producer, genre, actors, final version, and so forth — it is difficult to effectively monetise this content.
In September this year, Twitter announced 53 new premium video content collaborations with media partners across Asia-Pacific, including industry heavyweights like Bloomberg and Fox Sports Asia. As the battle to capture eyeballs intensifies with the entry of Internet companies such as Netflix and Amazon, traditional media companies are looking for innovative ways to maximise the returns of their rich content archives. Metadata — obtained using artificial intelligence (AI) — is poised to become a valuable business resource.
Capture eyeballs with data-driven content
Change is afoot in the media industry as more production companies embrace analytics and AI, fuelled partly by the digital push from governments in the region. Singapore, for instance, launched AI Singapore last year to lower the barriers to AI adoption for local companies, regardless of size and industry. These initiatives are paving the way for media companies to harness the power of data to gain visibility into their content assets and match them to specific audience profiles. This contributes to decision-making about the types of programming they should invest in.
In a recent study by Nielsen of subscription video-on-demand (SVoD) consumer viewing habits globally, 80% of the viewing time is driven by the back catalogue of content acquired by streaming services from TV networks and studios. In short, the content in media archives is contributing to the new and rapidly growing way consumers are viewing media.
Local distributors have a tremendous opportunity to leverage existing assets to provide next-generation media services. Creating an active and intelligent media archive enables them to quickly scan, harvest and enrich content metadata using AI services. The shift from a “frozen” archive — tape libraries and frozen cloud storage that lacks agility and demands a high operating cost — to an intelligent archive using on-premise or hosted scale-out object storage will also allow companies to streamline media workflow and focus valuable resources on content creation, rather than content storage and management.
Leveraging metadata for archive management and future content production
Creating metadata has typically been a manual process, where an expert “tags” or assigns metadata to a media asset and standardises its description so that search techniques can be used with confidence. But this approach does not scale. There are not enough qualified professionals and infrastructure to allow for the laborious real-time tagging of thousands of years’ worth of audio and video materials.
What if organisations could significantly speed up this task?
Shinano Mainichi Shimbun, a Japanese newspaper, is already using an AI-enabled system to automatically summarise articles. Similarly, local media and entertainment (M&E) distributors can leverage built-in AI applications within their intelligent content archive to carry out “in-place” metadata harvesting and enrichment to build up a user-referenceable database, thus reducing the amount of time and resources required. The valuable metadata obtained through this process will serve as an important guide for the development of new types of program, including providing the context for desired narratives, thus speeding-up future content production.
The platform approach to metadata
As multiple sources of data proliferate within M&E organisations, new AI and machine learning techniques are enabling the adoption of a “platform” approach to metadata collection. Applications that can analyse and correlate media assets with sources of user data can turn these static media assets into “data capital” — an asset class which will continue generating revenue throughout the life of the asset, much like how real estate continuously generates income for its owners.
When archive managers re-run an updated machine learning algorithm against an existing media asset library, correlated with more recent user data such as social media feeds or sensor data from location-based entertainment, they may tease out previously overlooked narratives, or themes that can then inform new uses for the content.
However, the sheer size of media assets can make it impossible to leverage data from a passive archive (like a tape library) to produce, say, a quick reaction to breaking news. A geo-distributed scalable object storage platform will allow multiple workloads such as archiving, disaster recovery and collaboration to be executed without requiring multiple silos for each.
M&E companies in Singapore can also reap benefits from having AI applications access media within that platform — without the need to migrate data for that purpose. The intelligent storage architecture will allow AI algorithms to perform facial recognition, object recognition, audio transcription, and even translation on media assets. Further benefits such as resolution enhancement, noise reduction and super slo-mo are also being demonstrated. These capabilities will be useful to advertisers and content owners.
Advanced IT toolsets, having evolved from raw computer science algorithms into intelligent software frameworks and applications that can modernise and activate passive archives, have ushered in a new era of revenue generation for the M&E industry. AI is poised to transform the process of media production, management and distribution — breathing new life into old content.