The media storage paradox with AI
Jonathan Morgan
Issue: May/June 2021

The media storage paradox with AI

Up until very recently, it was widely accepted that the main focus of content storage was simply to archive and protect assets. Nowadays, in order to truly maximise content within vast storage archives, it’s absolutely key that assets can be located quickly and efficiently. In some large archives, it can take days to manually search for a particular piece of content, and that’s assuming producers even know it exists within the archive in the first place. AI metadata tagging offers a solution that could enable media libraries to be searched at speed, drastically reducing this process into something that can be performed in seconds. It offers organizations the ability to fully utilize the extent of their archive, therefore maximizing its value. Let’s explore why this is, and how exactly companies can utilize this AI capability to make the most of their content. 

Knowing the worth of your archive
In the early days of media storage, companies were moving from linear tape open (LTO) format on a physical shelf, to putting digital files on a virtual shelf. Now, over time, these digital storage archives have bloated immeasurably as more and more assets of larger sizes have been added. Quickly, there became a paradox within the industry that the more content you were storing, the more difficult it was to find the content you needed, and thus the less valuable your archived content became. 

Your media producers may think they have footage of Princess Diana from the 1980s, but if this footage is unsearchable, or takes valuable time to manually find, then its value is certainly not being maximized. Being able to find content equates to being able to monetize it. If this is not the case, then much of the content within an archive is wasted. 
AI has become somewhat of a ubiquitous term. It’s a term that’s bandied about in almost every sector, but where AI really becomes exciting (or at least it does for us) is where it can enable advanced image recognition. AI metadata tagging is one form of this, offering organizations the ability to utilize the full extent of their archives and, crucially, extract value from old/legacy projects. 
By running analytics across archives and media sets, AI algorithms can be used to build up databases of metadata that enable searches. With this technology, video archives that once sat dormant can be queried for all types of things – even terms which are quite sophisticated. Semantic video analysis is another term for the uber-advanced type of video tagging, which could even accurately return content to highly specific searches, such as, “Find me footage of Sean Connery as James Bond”. 
A big challenge with current storage archives (on top of searchability) is simply having the video data in the necessary place when it needs to be processed. This involves storage systems knowing where to move specific assets and when, and AI can help here. When combined with accurate image and video tagging, AI-based automation of whole workflows can substantially improve efficiencies. 
Resurrection – is it a myth?   
AI video analysis can help to unlock the value of existing media archives, as well as those that are newly ingested. Metadata can be extracted and embedded within existing content using AI capabilities, which essentially enable the resurrection of legacy content. The value of legacy footage and the ability to reuse content has never been more obvious than during the ongoing Coronavirus pandemic. Production has become so much more difficult logistically, especially at the beginning of the global quarantine periods. Broadcasters in particular have regularly been calling upon existing footage to pad out broadcast schedules and entertain new viewers. 

In the sports world, where much of the content is of course live footage of games and analysis, fans were left with all scheduled games and matches cancelled, and no live content to entertain them. Tennis’ Wimbledon was cancelled in the height of the pandemic, and in the US, the cancellation of the NBA was predicted to lead to a loss in revenue for Disney (the owner of ESPN) of $306 million. What’s rather unique about the sports production industry, however, is the insatiable appetite of the fans for archived content. Fans of tennis, for example, would be very likely to enjoy a program that explores the “10 Greatest Wimbledon Champions”, or basketball fans would be sure to watch content that features footage from “The Best NBA Games of the Last Decade”.  Many broadcasters took advantage of this, creating programs and content that  featured legacy footage they already had available to them. The value of easily accessible legacy content has never been clearer. 
Those media companies who also had access to their content in the cloud were much more agile and able to react quicker, as their employees were able to remotely access and produce content from the start. 
Ideally, automatic semantic tagging of assets with metadata should be done upon entry of content into storage. But with new AI capabilities, automatic tagging can be put in place for all future ingests, with older archives analyzed and tagged now for searching in the future. 
A maturing technology 
The ability to extract, index and tag video and image content with metadata enables companies to truly make the most of the content they have already invested considerable money and time in developing. The paradox of ‘the more content you have, the less valuable it becomes’ should shift to a thing of the past, and we predict it will as AI grows in its sophistication. Storage management has come a long way and continues to develop. Essentially, the technology can only get better. 

In this new unpredictable way of life, COVID-19 restrictions look set to continue for some time. For producers and media companies, it’s essential that content can continue to be generated from archive footage and this can only be done if this footage is searchable and accessible; this is where AI comes into its own. Once AI tagging and indexing is built into the media industry workflow, it can provide value over the long-term as we recover from economic uncertainty by increasing the value of all content within storage archives.

Jonathan Morgan is the CEO, Object Matrix (, an award winning software company that has pioneered object storage and the modernization of media archives.