Broadcast scheduling is critical yet extremely manual task today. The complexity of the scheduling workflow increases with more number of feeds, and integration with ad schedules for each feed. Thus, a traditional scheduling process may need additional number of people, who need to be supported by additional infrastructure. This inflates the costs, and still does not limit the possibility of human errors.
How rule-based scheduling works
An automated approach to scheduling could help address both these issues. Broadcast scheduling could be completely reinvented, by defining rules for scheduling, and deploying deep-learning based systems. Such systems could be programmed to intelligently identify content segments and use the pre-defined rules to create schedules using those segments.
How automated scheduling can help TV networks
- Simplified workflow with automated validation of assets as per schedule
An automated system can notify errors such as mismatch between schedules and actual assets. It can also help in detection of errors such as varying audio levels, audio silence, black frames etc.
- Ability to create multiple schedules for multiple feeds with minimum human intervention, using pre-defined rules
As TV networks look to scale up rapidly, managing multiple feeds manually could become extremely complex, and resource-heavy. Using the rule-based method, broadcasters could simply tell the scheduling system to intelligently create schedules based on pre-defined criteria. For example, matinee movies could be a defined genre, assigned a slot in afternoon. Then, all the movies that the broadcasters wish to show in afternoon could be tagged as ‘matinee’, helping the system identify these assets. Once identified an automated system could build a schedule for entire week by itself, with minimum intervention.
- Automation of complex processes such as content rights management and graphics generation
Another area where a rule-based system could help is management of content rights etc across the globe. The system could automatically make an asset available or unavailable based on the content distribution right status of the asset. Traditionally, this would require manually checking status of each asset, and still the chances of playing out wrong asset cannot be ruled out. Automating this process could prevent any accidents, making broadcast more efficient.
There are many more ways in which automation, and machine-learning can reinvent broadcast. As the automations systems evolve, we could see more interesting applications of the technology in all areas of broadcast.