Industry Insights: What you need to know about media asset management

In this installment of our Industry Insights roundtable, our experts from the field of media assessment management (MAM) technology examine costs, the march to the cloud, broadcast workflows and metadata. 

Industry Insights: Machine learning, AI and software-first workflows coming to production switchers

What is the current path to hybrid or cloud MAM?

“We envision both hybrid and cloud MAMs being deployed based on the workflow needed. When a customer has on-prem requirements for either signal based or ultra-high resolutions workflows, a hybrid model works very well to reduce cost of cloud. When a customer on the other hand requires highly flexible, geo-dispersed and burstable access, that’s where cloud MAM can deliver the most value,” said Alan Dabul, product manager of MAM & media I/O for Ross Video.

“We understand that a lot of media companies want the benefits of cloud but also want a gradual path with a flexible deployment architecture which can evolve over time. That’s exactly what AMPP offers, together with a cloud-native SaaS control and orchestration layer, as well as migration strategies for customers with existing Grass Valley on-premise installations so that they can leverage their previous investments,” said Camilla Powell, a product manager for Grass Valley.

“Ateliere utilizes several paths for migrating our customers’ media to the cloud. For large amounts of media, we provide direct transfer, a bucket-level feature that enables fast, easy, and secure transfers of files over long distances to an Amazon AWS S3 bucket. When bandwidth is limited, or the volume of media makes transfers over the internet impractical, we often utilize the AWS Snow family of secure edge computing devices, which allow users to move terabytes or even petabytes of media into the cloud. Another option is using Aspera’s high-speed secure data transfer service for customers who have smaller volumes of media to move into the cloud,” said Bill Admans, COO at Ateliere.

“Initially, the first step for MAM workflows transitioning to the cloud was simply relying on the cloud for long-term storage with the remainder of functionality on premise. Then, certain transformation and customization workflows began moving to the cloud. The path has now shifted to those MAM and production workflows that are real-time or live in nature to the cloud. The last 18 months have been transformational in terms of the ability to support remote real time video workflows,” explained Rick Young, head of global products at LTN Global.

Column: Even in the cloud, content is never at rest

“Hybrid storage environments are gaining traction, although a large piece of the market will continue with On-Prem storage for some time. With DIVA and Kumulate Content Management the storage migration is completely abstracted from MAMs, thereby ensuring that the MAM can be hosted in any location, including the public cloud, while continuing to operate seamlessly,” said Savva Mueller, director of business development for Telestream.

Advertisement

“The scope involved depends very much on the technology within the solution. Questions around infrastructure, third party integration needs and live workflows need to be explored too. At Dalet, we appreciate that each organization will be faced with some challenges in moving to the cloud, which is why we offer solutions that accommodate several deployment configurations. Whether on-premises, fully in the cloud or a hybrid operation between the two,” said Mathieu Zarouk, director of product strategy, Media Workflows at Dalet.

“Hybrid infrastructure no longer requires convoluted and complex solutions – simple agents co-located with assets or systems keep everything securely connected. For some components, there is still a cost advantage to on-prem, but with the pace of change in the industry and consumer demand, the flexibility offered by a PaaS solution is often a better match to business requirements. Where your assets are stored may be an exception to this for a number of reasons, but good media supply chain solutions should have broad support for different storage systems and vendors, as well as being operationally agnostic,” said Karsten Schragmann, the head of product management for Vidispine.

“The best path to a modern media supply chain is to start with a cloud-native platform that can manage workflows across both on-premise and cloud-based infrastructures. It’s also important to realize that you don’t have to transform the entire supply chain all at once. Start with a step where it’s most practical or most needed, such as content receipt, migrate that first, and then build on that success in subsequent stages of your supply chain,” answered Geoff Stedman, CMO for SDVI.

What are some important MAM cost considerations?

“Does the content need to be accessed on-prem? Does the content need to be in its original codec and resolution? Does the content need to be archived off-site or in the cloud?,” suggested Dabul.

“A key aim of our GV AMPP is to allow broadcasters to no longer provision and budget for peak demand throughout the year, and instead enable them to scale up and down for key events,” said Powell.

“Benjamin Franklin said only two things are certain in life: death and taxes. With the cloud, it’s ingress and egress; the cost of moving media into and out of cloud storage environments. Other important considerations are managing the cost of computing and your storage footprint, which can quickly get out of control. But, you can manage these costs through policies controlling how your application scales the compute environment and automating the media life cycle between on-demand and archive cloud storage environments. Another challenge that racks up costs is replication and the number of versions you are storing. To solve this problem, we utilize the Interoperable Master Format (IMF) to componentize all the versions of a content title and save only the differences, often achieving over a 75% reduction in storage footprint. The savings can add up to millions across large media libraries,” Admans noted.

“For hosted or cloud scenarios, storage and data egress are drivers. It’s important for companies to run scenario-driven estimates in order to understand their real cost before moving forward with hosted solutions. Vendors should also provide simplified pricing to allow customers to benefit from cost transparency and predictability that would allow them to plan and manage their budgets over time,” said Young.

Column: Media asset management evolution in broadcast production

“It’s important to match a MAM and content management system to the needs of the workflow. Far too many expensive MAM solutions have been shelved because staff refused to change their working habits to accommodate them. Some workflows will benefit from a simple, cost-effective content management solution,” Mueller added.

“I’d say three mostly. Web-friendly preview creation, transcode, and media processing to generate the content previews. Secondly, egress of previewing content and media movement. Additionally, storage optimization and management to help move infrequently-accessed content to cheaper storage,” Zarouk said.

“The most important cost consideration when managing media supply chains is to ensure that the platform and the services it enables are based on a consumption model where you only pay for what you use. This model encourages agility because you can switch easily between tools depending on the requirements of the media processing job, and it promotes efficiency because it eliminates over-provisioning and underutilization. It also allows you to align your costs to the value of the project you’re undertaking, ensuring you make good business decisions,” Stedman said.

What about broadcast workflow considerations?

“Does the customer require baseband signals to perform their workflows? Does the customer require graphics workflows with baseband output signals?,” Dabul responded.

“With remote working now being so fundamental, it is important for users to be able to execute their broadcast workflows regardless of where they are working from, and regardless of where the media is located. Our larger customers, with multiple production sites spread nationally and globally, benefit from the federated view of all assets and the ability to use any content, wherever it was ingested or created,” Powell said.

Column: Preparing your content for machine learning

“New and upcoming ultra-fast networks like 5G and Wi-Fi 6E don’t just mean more data streaming from the cloud; they mean faster access to the cloud from the point of content creation. Broadcasters are looking to these new technologies as readily available gateways to the cloud, and reporters are already starting to stream video live to air over 5G mobile networks. Recently, several technology innovators came together at the Masters’ tennis tournament in Indian Wells, California, to test streaming matches live to the cloud over 5G UltraWideband and the results were impressive. Wi-Fi 6E promises more bandwidth and speed than previous standards and is anticipated to be a viable replacement for video over IP. In dense environments like stadiums, Wi-Fi 6E means people will not compete for bandwidth like they do today, which means new opportunities for streaming content and delivering experiences during sports matches and other events,” said Admans.

“When building MAM-based broadcast workflows, it is critical to understand a technology stack’s ability to support high end, professional video specifications especially as the demand for 4K and HDR grows across consumer platforms. In addition, I see the coming together of non-live, file-based workflows alongside real-time and live collaboration use cases. Both are critical in today’s media landscape,” said Young explained.

Advertisement

“Today, important considerations are not only around broadcast workflows, but how to optimize overall production and distribution for broadcast and OTT, streaming, syndication, and all the multiple platforms where content is consumed. Designing workflows to create once but deliver many is ideal. Orchestrating and automating steps so that work does not need to be manually duplicated, will help streamline operations to serve workflows that are no longer ‘just for broadcast,'” Zarouk said.

“The workflows of a full media supply can be complex, but AI can provide insights into any inefficiencies in the chain,” Schragmann noted.

“It’s important that the platform you are using to manage your workflows is able to integrate both automated functions (such as transcoding, file analysis, automated QC, etc.) and manual work (such as compliance editing, manual QC verification, etc.) so that the entire supply chain can be monitored and optimized. The best management platform won’t just manage assets, but will also manage the work required to prepare those assets for any distribution platform,” Stedman said.

Where does metadata fit in?

“Metadata as we all know fits into most if not all the key workflows that are driven by the MAM and orchestration layers. It can facilitate the exploration of content by producers to create new compelling content just as much as driving long term archiving workflows to reduce on premise storage needs,” Dabul said.

“Media without metadata is close to useless. Rich metadata enables rapid location of assets in the short term and maximizes utility of that asset over time. Our customers are demanding AI solutions to automate metadata creation and enrichment, whether it is logging metadata, transcription services, facial recognition, scene detection… the list is endless. These are rapidly becoming must-have core feature expectations in a product as opposed to product delighters, and that is something we certainly recognize at Grass Valley,” Powell answered.

Column: Simplifying live broadcast operations using AI and machine learning

“To operate efficiently and control costs, you need access to cloud environment metadata describing the performance, configuration, operations, security, usage, and billing data. In the same way, to fully monetize and manage your media, you need metadata describing every aspect of the files stored in your cloud environment. This metadata includes top-level data relating to the physical aspects of your files, descriptive metadata about items like synopsis or rights management, or essence-level metadata that describes what is happening within your content,” Admans said.

“Metadata enables memorialization of media assets, and for that reason, it is essential to ensure structured metadata is preserved. Technical, speech-based, or video-based, all have specific use cases today and are widely used by AI engines for modeling and recommending. This makes it essential for MAM systems to have a robust way of generating, preserving, and retrieving content based on metadata tags,” said Russell Vijayan, head of business development for Digital Nirvana.

“Metadata is key throughout the process. Rich metadata capture for raw materials — either through integration with upstream systems (cameras, production system, transmission devices) or through AI sources — is critical for content discovery and management. Detailed essence-specific metadata is also important to understand and ensure interoperability between solutions. And, of course, rich finished content metadata is a must-have when looking to distribute or syndicate content to partners or for EPGs and similar consumer discoverability mechanisms,” Young noted.

“Without extensive metadata tagged to the content, a MAM system would be largely unusable. Automatic metadata extraction should be done with machine learning algorithms that more than supplement any human logging that may have been done during ingest,” Mueller said.

“Metadata has to be a fundamental pillar for any top-grade MAM. It provides meaning to the content, especially when stored across different locations and on elastic storage solutions. Metadata also provides rules for triggering workflows in as automated fashion as possible driving the content supply chain efficiently. Metadata is critical for effective multi-platform distribution strategies. Getting content to audiences in the right format, specification and at the right time can directly impact business success,” Zarouk said.

“It is important to have structured metadata, as this enables us to automate processes and workflows. A really simple example would be knowing the codec and resolution of a file; based on this metadata we can make run-time decisions in automated workflows about how to handle that asset,” Schragman explained.

“Metadata is the fuel for the entire media supply chain engine. It must be captured right at the point of ingest, augmented as additional workflow steps happen, and used by the management platform to guide the actions and next steps in the supply chain. The best management platforms have a very flexible metadata model and utilize that metadata extensively to optimize supply chain activity,” Stedman said.

Contributors

Alan Dabul, Ross Video
Camilla Powell, Grass Valley
Bill Admans, Ateliere
Rick Young, LTN Global
Savva Mueller, Telestream
Mathieu Zarouk, Dalet
Karsten Schragmann, Vidispine
Russell Vijayan, Digital Nirvana
Geoff Stedman, SDVI