Advancing through AI-based Closed Captioning in the era of FCC regulations


One of the largest US TV networks that deal with major news and sports channels. 

Over the past few decades, news and sports broadcasting has evolved rapidly and the challenge to meet and comply with FCC regulations on closed captions, that evolved in 2017, has been ever-growing. In addition to the existing responsibility of adding captions to repurposed broadcast content, news and sports broadcasters, under a new mandate, were also faced with the challenge of inserting highly accurate captions to their assets within 12 hours from the time of the actual broadcast. Considering the increased user consumption of highlights and montages over the Internet, this became an onerous task.


Due to its seasonality, the unpredictable flow of volumes would be difficult to manage right from the stage of job allocation, caption editing, role assignment, report generation, and status monitoring for improved project management. This, coupled with the fact that the captions have to be turned around within 12 hours from the actual broadcast time and further integrated into existing workflows would require the solution to be highly reliable and easily available.


As a PaaS (Platform as a Service) provider, Digital Nirvana (DN) provides closed captioning services as a part of its media services offerings. Upon provision of access to the media content by the customer, DN picks it up for processing, creates a closed captioning file, and delivers the output back to the customer. The company will host the storage for the platform in its own instances. The solution provided had the STT (speech-to-text) engine at its core and seamless API-based integration to the customer’s workflow. Soon after a clip is generated, it is transferred to DN’s cloud for STT generation, and the resultant output, along with the proxy video, is transferred to DN’s processing centers for manual editing to ensure 100% accuracy. The presence of STT data makes the captioning process efficient, and in turn, returns captions within a much shorter time frame. Language model adaptation with text data and the ability to train STT engines specifically with relevant sports data lets DN improve the quality of STT output, thereby considerably reducing manual effort. This, in turn, helps DN turnaround requests at a faster pace while ensuring regulatory compliance.
In light of the FCC’s increased focus on enforcement, the company is required to be in compliance with all applicable requirements, which includes the returning of captions in a shorter time frame in spite of a volume increase. Violation of regulatory requirements finds the company liable for a penalty. High volume hiring of full- time captioning employees could be expensive and would be ineffective in terms of employee utilization as the requirement is seasonal. Recruiting freelancers could be a viable option in terms of cost, but could at times have a negative effect on output quality and service reliability. By leveraging artificial intelligence to automate the process, industry players are benefited in terms of cost, time, and resources.
Efficiency gains leading to a significant increase in output per individual captioner.
• Ability to scale up production at short notice.
• Cost containment. • APIs for easy integration.
Among the many players, what makes Digital Nirvana’s captioning service different? While turnaround, cost, quality, service reliability, etc. are all key factors, the major differentiator would be the ability to convert the process as a solution than just a part of a service. Asset retrieval closed caption generation, and the provision of timely and quality output is a part of the Digital Nirvana solution. The ability to precisely understand customer requirements, recommend and work with the customer to ensure that the solution best fits their needs made Digital Nirvana the preferred solution provider for this customer.



Contact Us

Schedule Meeting