About Video Metadata
Current information to describe video tend to be top level tags and short text descriptions, they provide an overview of the entire piece of content. Metalabs create rich layers of timecoded information that map each scene and provide time specific keywords and tags that add value to the video.
Diagram 1: TimeTagging: How metadata is generated:
By analyzing the video signal we can determine where the breaks are located between scenes within the programme.
By analyzing the audio track we can extract a transcript of the spoken word in the video, by applying natural language processing techniques we can identify keywords and tags that identify places, people, activities and other useful handles.
New Advertising Opportunities
Content that has been TimeTagged opens up opportunities for advertisers, ads can be linked to keywords and as the content is now mapped at the scene level mid-roll video advertising messages can be streamed at key points during audience playback.
The traditional broadcast advertising model, where ads are placed at quarter hour intervals within a programme does not apply to online streaming. Where a piece of content has been TimeTagged an ad can be placed at the natural break between scenes resulting in a less fatiguing viewer experience than can be the case with pre-roll ads and less chance of viewer switch off than with a post-roll ad.
Broadcast Video - The traditional linear broadcast hour: Ads are placed at defined intervals.
Online Video - The 'on demand' hour: Ads can be placed between scenes.
TimeTagging identifies the multiple potential locations for ad insertion. Combining this information with Keywords will facilitate the delivery of the right ad at the right time during playback.
