Defining Metadata Schemas for Media Workflows

Cutting to an ad-break in the middle of a crucial scene is never a good idea. Ad-breaks during episodic or feature length content must make sense to the viewer and if the timing is right, they are more willing to accept the interruption and watch the ads. Commercial content providers need to keep viewers engaged, while also keeping advertisers happy.

Identifying optimal moments to cut to an ad-break can be a time-consuming task. Their positioning must avoid distracting the viewer from the flow of the program, and instead capture their attention during a natural pause in the action. Media organisations need to offer advertisers the best possible opportunity to reach their target demographic without alienating them. So how can they strike a balance between their viewers needs and advertisers requirements?



Finding the Right Spot

A typical workflow in commercial broadcasting is to identify black frames in content, to determine suitable locations for ad-breaks. This involves the content processing team manually scrubbing through, looking for shot and scene changes and placing ad-break markers into a programme timeline. There are a few technical tools and AI services that can automatically detect black frames. However, simply processing your content through a service isnt enough. To make informed decisions, operators need to visually inspect and confirm that detected segments are suitable.



Keeping up the Pace

By using AI services for black frame detection, then processing media content for final QA/QC, operators can drastically reduce this traditionally manual and labour-intensive validation process. AI will detect video intervals that are (almost) completely black, meaning there needs to be several subsequent black frames in order to be flagged as a black interval. Audio is also taken into account. AI detects the combination of silence and black frames to identify segments for removal. The duration and the minimum pixel and frame thresholds can be customised via settings within the black frame detection filter. Speed-wise, black frame detection is a relatively quick process, executing at 17x speed, meaning a 17-minute-long video will take roughly 1 minute to analyse.



Striking a Balance

Black frame detection can either run automatically as part of the ingest process, or on-demand. Media companies can leverage AI generated metadata, but there may still be instances where skilled operators need to step in and manually create markers. Accurate.Video visualises time-based metadata and helps operators make faster editing decisions, whilst keeping them in control of the process. This offers the ultimate flexibility, automating where it makes sense, and connecting seamlessly with existing tools and workflows.



Black frame detection is a common use case for QC and Validation, when operators need to process content for ad-break insertion. Accurate.Video will help highlight and suggest chapter transitions, commercials, or invalid recordings, so media companies can deliver with confidence. Find out more about Accurate.Video

LINK:	https://www.codemill.se/post/metadata-schemas-for-media-workflows...
	See more stories from cantemo

Most recent headlines