InBroadcast
Metadata is what makes media an asset and with software-defined workflows and exponential increases in content output maximising it is critical.
The importance of metadata cannot be stressed too highly: in a software-oriented media world it is the absolute core of the technology, the glue that makes everything work. Software-defined workflows need to know all about the content to be able to automate workflows and also make the content easier to find.
“Every single media enterprise will move to software-defined workflows so will come to depend upon metadata,” says Paul Wilkins, director - solutions & marketing, TMD. “What becomes critically important then, is that the metadata management layer – which is where the workflow orchestration resides in a well-designed system – is infinitely flexible to meet the unique needs of each media enterprise.”
Metadata is what makes media an asset and drives searching, organisation, and content identification. With more outputs and increased need to streamline production metadata is enabling more automated production within the broadcaster.
According to Ximena Araneda, evp media workflows and playout, Vizrt, it is the key that decides what an orchestration layer should do with an asset: “Without metadata, the whole process of automating production and ultimately collaborating falls apart.”
Also important is metadata exchange. The metadata does not stop in a facility, but goes via initiatives like AS-11 by Advanced Media Workflow Association to other broadcasters. “The benefit compared to the cost of metadata will only continue to be in favour of more metadata, as it is shared more effectively, and the basis for more automated production,” says Araneda.
Search and media discovery is the bread and butter of many metadata workflows. Beyond this the workflow becomes automated production typically coupled with automatic or user based task management.
A common scenario using Vizrt is for handling media ingest, quality control, multi-platform distribution and archiving of the media. As media enters the system, it will be matched against a pre-created asset placeholder, and that placeholder is connected to a user operated ingest
task. The operator can then verify that the media is associated with the correct asset and take action on that task. That will reflect a change in administrative metadata, triggering another task based workflow, for example, quality control, where the media automatically will be sent off to an external party for verification. The result of the quality control will then update additional administrative metadata, where another operator can verify the results and take action on that task, driving the workflow further.
“This type of workflow brings out better control of the lifecycle of the media, allowing for catching mistakes and errors and then act upon these in a more controlled environment,” explains Araneda.
MAM system Viz One uses descriptive, administrative and technical metadata. This type of metadata provides functionality such as resource discovery, interoperability and media lifecycle management. Metadata may also be present in the form of annotations enabling more powerful search, enhanced content description and automatic clip creation. The system also relies heavily on central metadata taxonomies such as dictionaries, thesauri, and directories.
Whereas production departments typically focus only on the next playout date, archival description of content has to serve history, cultural identity, and cultural heritage.
That’s where NOA’s mediaARC Workflows come into play. This product is designed to link production procedures to specific media or metadata results. Examples include delivering assets from an archive, validating metadata edits, or an extended QC process.
“Metadata is always changing because content is always changing,” states the Vienna-based firm. “Tomorrow’s structure of metadata description could differ significantly from today’s. Some say a non-structured, full-text database is the answer, but a strict hierarchical relationship between entities is not sufficient to describe the content of an archive. mediARC overcomes that problem by organising audio, video, and other content in media object containers. You can link media objects to multiple metadata entries regardless of whether they are complete or in segments. That means you can create logical cross-correlations even with different qualifiers describing the role of the connection of the media to its context.”
Recently NOA introduced a VideoPlayer for jobDB, the company’s system that allows users to set up workflows for the ingest, reshaping, and analysis of media for archiving or re-transcoding.
Metadata driven workflows are core to Primestream’s Dynamic Media Management system. Some metadata is automatically created or assigned when a clip or file is originally created or ingest into its MAM system, while other workflows call for additional information to be added or edited later in the process. To streamline this workflow, Primestream built specific metadata tools..For example, Primestream’s Xchange Suite integrates with the Associated Press’ metadata service to add accurate, comprehensive data to assets. For real-time logging situations like sports, Primestream’s FORK Logger turns on real-time logging along with integration with STATS for a live metadata stream into the FORK Logger module.
“If you want to find all of the goals or scores in a game for a highlight package then you are dependent on metadata to find that clip and put on air,” explains COO David Schleifer. “If ‘goals’ are logged live, an editor can have a Smart Bin where all the goals show up automatically. Primestream integration with all the leading NLE systems lets metadata and markers move between them and FORK and Xchange. Primestream also enables commenting, approval workflows and much more.”
Marquis Broadcast’s focus is on integrating incompatible systems from a workflow metadata or codec standpoint. All its solutions depend on metadata. “We often have to re-map metadata, re wrap files and transcode content ‘on the fly’ so it can move seamlessly through workflows,” says Paul Glasgow, sales & marketing director. “Our special ‘metadata sauce’ is being able to do this between fundamentally incompatible systems.”
The company maintains a large legacy and contemporary library of broadcast and media systems from which it extracts and returns the richest metadata sets possible. While rushes from cameras are very simple to manage, the heavy ‘metadata lifting’ comes from dealing with metadata associated with ‘in-production’ assets. As an example, Marquis manage metadata from complex timeline sequences, sequence translations and also manage the genealogy and tracking metadata.
For example, Primestream’s Xchange Suite integrates with the Associated Press’ metadata service to add accurate, comprehensive data to assets. For real-time logging situations like sports, Primestream’s FORK Logger turns on real-time logging along with integration with STATS for a live metadata stream into the FORK Logger module.
“If you want to find all of the goals or scores in a game for a highlight package then you are dependent on metadata to find that clip and put on air,” explains COO David Schleifer. “If ‘goals’ are logged live, an editor can have a Smart Bin where all the goals show up automatically. Primestream integration with all the leading NLE systems lets metadata and markers move between them and FORK and Xchange. Primestream also enables commenting, approval workflows and much more.”
Marquis Broadcast’s focus is on integrating incompatible systems from a workflow metadata or codec standpoint. All its solutions depend on metadata. “We often have to re-map metadata, re wrap files and transcode content ‘on the fly’ so it can move seamlessly through workflows,” says Paul Glasgow, sales & marketing director. “Our special ‘metadata sauce’ is being able to do this between fundamentally incompatible systems.”
The company maintains a large legacy and contemporary library of broadcast and media systems from which it extracts and returns the richest metadata sets possible. While rushes from cameras are very simple to manage, the heavy ‘metadata lifting’ comes from dealing with metadata associated with ‘in-production’ assets. As an example, Marquis manage metadata from complex timeline sequences, sequence translations and also manage the genealogy and tracking metadata.
“We may also extract metadata from media files and aggregate and re wrap these with other metadata sources such as from PAM, MAM and users e.g. for AS-11 compliance,” says Glasgow.
“For example, an Avid Media Composer editor within in an Interplay environment may wish to send an asset to another production system playout server, automation system DAM, archive etc. From a user perspective it’s a simple <right click<> send to>< an incompatible system>. We deal with all the complexity. In a large enterprise we may be servicing hundreds of concurrent processes interoperating between many different systems. To do this we maintain the largest independent library of legacy and contemporary third party integrations. This means customers can deploy the latest and best of breed systems, confident that they can also bi-directionally integrate with their legacy infrastructure and systems.”
Telestream’s Vantage products can process metadata in sidecar xml files, or embedded in MXF files. It can also process what it calls ‘work orders’ - metadata contained in CSV files. More than that, Vantage can create metadata by automatically analysing files, and it can convert metadata between different stylesheets so that the metadata can be used by different systems.
“In a very simple example, but one that is common in news ingest, Vantage can examine metadata on files that come into a news organisation, examine this to decide the resolution, aspect ratio and even orientation of the shot, and then apply the necessary processing rules to get this story into the news production system as quickly as possible,” explains Paul Turner, vp of enterprise product management. “This allows broadcast news companies to benefit from footage that comes from the massive number of video cameras in phones that are now first on the scene at almost every incident.”
The two main types of metadata in Editshare’s Flow production asset management system are asset-level metadata which refer to an entire media file (e.g. master clip metadata) and log level/subclip metadata, which refers to a particular section of a clip (i.e. subclip).
Metadata fields can be entered by text, pick lists, timecodes, dates, booleans (true/false), as well as some special purpose log list metadata fields called groups and categories, which are used in the Flow Logger application to perform point-and-click logging in reality or sports productions of contestants, players or activities. Extensive metadata can also originate from tapeless ingest or other third party systems and is preserved throughout Flow workflows.
“Increasingly, customers will see that good editorial metadata doesn’t just magically create itself— that the data you find is only as valuable as the time you put into entering it on the front end,” says Jeff Herzog, product manager for Flow asset management and video products. “An investment in metadata input pays dividends in the editorial process as well as to future monetisation possibilities for content.”
The ability to reuse and repurpose content has become critical in order to control the burgeoning costs of programme creation. Organisations need to understand what content they own or have the rights to use, and where the content is located.
“They also need to manage where the content is stored: across fast edit storage, slower commodity disk, tape, optical, or cloud services. Metadata lets teams take control of all of these issues and more,” says Dave Clack, CEO, Square Box Systems. Given all of this, companies are finding that their old methods of managing content are failing.”
This includes inconsistencies in file and folder naming, a reliance on ‘hero’ individuals who are the go-tos for performing video search and retrieval, and manual methods such as spreadsheets and documents. These methods simply can’t keep pace with the exploding amount of content being created, in addition to the huge physical file sizes that resulting from emerging technologies such as 4K/8K, HDR, HFR, and VR.
Clack explains that CatDV is extremely effective for capturing and logging all types of content and resource metadata. This ranges from asset-based, technical metadata for items such as cameras, exposures, video and audio formats, and containers to custom metadata that reflects the needs of an organisation.
“CatDV captures a huge variety of data types including text, numbers, lists, multiple selections, and predictive text entry – anything to make logging simple, quick, and useful,” he says. “It also captures time-based and temporal metadata. Examples include markers for the best moments in the content, such as the goals in a soccer match, or markers for QC or bad language, as well as integration points for automating image detection and speech transcripts.”
TMD prefers to divide metadata into two broad categories: technical metadata, which aids automation; and descriptive, which aids discovery.
“One of the key issues we’ve found, particularly during data take-on from legacy systems, is that there is sometimes confusion between the two,” reports Wilkins. “To quote an obviously disastrous example, one major broadcaster had made the number of audio channels a descriptive text field not a prescribed technical one. That resulted in hundreds of different descriptions for a stereo soundtrack, a nightmare to translate to a modern, rigorous database structure.”
Content creation and/or preparation with multi-versioning & IMF packaging is among the most discussed workflows currently. For this type of workflows, the Dalet AmberFin media processing platform, together with the Dalet Workflow Engine, allows sophisticated technical decisions to be controlled by business rules that enter the system as simple XML files, watch folder triggers, or API calls.
Multi-platform delivery has caused the industry to move away from the simplicity of ‘one profile per customer,’” explains Kevin Savina, Director of Product Strategy. “But with the help of a powerful workflow engine, we can create an environment where a single workflow can produce all the desired outputs just by changing the metadata that initiates a particular process.”
SMPTE’s standardized mastering format IMF – the Interoperable Master Format has a wealth of facilities for identification, auditing and tracking of media, titles and metadata. Although based on human-readable XML, it is fundamentally optimised for machine processing.
“With the huge variation of starting points for IMF creation, it is imperative that a workflow engine is versatile and able to use the right tools at the right time to form a valid and verifiable IMF bundle at reasonable speed and for reasonable cost,” says Savina. “Dalet Workflow Engine has been optimized for these kinds of workflows where the number of input files is not known until the job starts, and the workflow proceeds without losing or changing vital information. The icing on the cake is the ability to see the performance of the jobs in a data analytics engine that is able to spot trends in operation so that continual optimization of the tools can be performed.”
Metadata evolution
There are already plenty of examples of technical metadata being used to seamlessly link content from creation to consumption, setting the optimum format at each stage of the process. The next stage will be for an extension of descriptive metadata. More information will be gathered and created, some of it perhaps automatically.
“Intelligent systems might ‘listen’ to the script and ‘view’ the content to build a comprehensive description and audience rating of a programme,” suggests Wilkins. “That rich metadata could then be exposed to consumers to help them identify their sort of content. Ultimately this, too, could be automated. Intelligent set-top boxes and online clients could build a profile of the user, learning what sort of content they like.”
Perhaps in the near future, Artificial Intelligence systems will take advantage of metadata by using metadata along with complex search algorithms to find new relationships between media and audience interest,” suggests Schleifer. “Technology such as speech-to-text, voice-search systems like Siri, Cortana and Alexa are all maturing into powerful tools that will improve how we search and find what we are looking for. Properly catalogued media gives the broadcaster more opportunities to leverage their libraries to the public.”
Vizrt’s Araneda confirms that metadata will increasingly be populated with a higher degree of automation enabled by new technologies such as image analysis for face recognition, speech-to-text, and object recognition, and many more.
“The end result will be a higher degree of metadata and more automated rules and tools that use that data,” she says.
“Finding footage of a particular person, or past comments on a particular topic, will be just a few clicks away for the producer. We also believe that with the social media explosion, keeping track of use data will be more and more important in the planning stage of producing a story, in order to take decisions on how the story creation should be done, and on what platform to target similar content.”
In conclusion, the success of online video relies, in part, on metadata. Publishers need to manage the creation of metadata as a part of the production of the video. Video is a complex medium that requires both automatically created metadata and human authored, which is more flexible and accurate, thus providing a fuller experience to the consumer.
“Higher quality metadata leads to a more engaged consumer, which means monetized video assets,” emphasises Dalet’s Savina. “Publishers need to incorporate media asset management systems that allow authoring and managing asset metadata in their production workflow in order to build engaged audiences and maximize the value of their internet advertising.”
|
No comments:
Post a Comment