Canada’s CPAC Taps Dalet for Bilingual Closed-Captioning Workflow
Dalet Digital Media Systems has supplied and integrated an end-to-end workflow for CPAC’s (Cable Public Affairs Channel) English and French closed captioning. The solution combines the strength of Dalet AmberFin and Dalet Galaxy to provide an advanced automated workflow that maintains genealogy and captioning data for the 24/7 operation.
Based in Ottawa, CPAC is Canada’s only privately-owned, not-for-profit, bilingual television service providing a commercial-free window on Parliament, politics, and public affairs. Through its unique style of balanced, long-form coverage of the people, politics, and events shaping public policy, CPAC facilitates and fosters participation in the Canadian democratic process. CPAC is delivered by cable, satellite, and wireless distributors to more than 11 million homes in Canada and worldwide via 24/7 web streaming on the CPAC TV 2 GO mobile app and the www.cpac.ca website. The organization ingests up to 80 hours of new content daily and has made over 30,000 hours of archived content available on these digital platforms.
“We urgently needed a closed captioning solution that could manage a variety of content from several different types of production workflows: news and magazine-style shows, live and studio production, and long-form programming. For each of these workflows we needed to be able to manage, and more importantly maintain caption data, especially after editing has been completed or the content is passing through another system,” says Eitan Weisz, senior manager of broadcast operations, CPAC. “Those are all facets that we felt Dalet could deliver for us in one package.”
The solution devised by Dalet in concert with CPAC deploys Dalet AmberFin as an advanced media and processing platform for insertion and extraction of caption data, leveraging the existing Dalet Media Asset Management (MAM) deployment to manage transformation and distribution in an end-to-end media workflow with a key emphasis on maintaining captions as metadata.
During live shows or events, CPAC content is transcribed by operators using keyboard or stenography methods to generate live captions. Live feeds with captions are ingested into Harmonic and Dalet Brio video servers and written to an MXF file including a standard SMPTE ST436 data track. For preproduced content, a proxy file is sent to an external captioning house for transcribing and generation of a separate caption file.
Where the Dalet MAM detects valid ST436 data, Dalet AmberFin automatically extracts the caption information to a standard SCC caption file, associated with the relevant video file in the MAM. Also automatically generated is TTML (Timed Text Markup Language), a standardized XML structure that represents all the captioning information as time-coded metadata that can easily be catalogued, searched and modified.
“The TTML pivot format is vital. Instead of keeping the information at file level, we expose it as timed metadata to the user so that they can directly view and search it as well as perform basic editing and correction of the caption information,” explains Dalet Sales Director Frederic Roux. “CPAC has rich automated metadata created in the Dalet system and tagged to all its content through this captioning process.”
In the other production scenario at CPAC, when an SCC file is received from the captioning house, Dalet will instantly match it with the video file and then perform the same workflow: transforming the SCC file into searchable, viewable, useable data.
“A common and expensive error in closed captioning workflows is the risk of losing captioning data during the production process,” explains Roux. “Editing can destroy the captioning information at file level and require the user to recapture or recreate the original material, which is time-consuming and costly.”
Retaining caption data during the production process is integral to the Dalet end-to-end workflow. As Dalet maintains the genealogy of all produced content, the Dalet MAM automatically associates the corresponding caption metadata from the various pieces of production content to the final package created by the NLE. The final new asset then inherits the captions of the original material.
Distribution is also managed by Dalet at CPAC. For linear playout, Dalet uses the metadata caption information or the SCC file with associated video to re-embed the captions in the media file within the ST436 data track. Options exist for managing the captioning separately as a side car file. Should there not be an SCC file, Dalet will automatically regenerate it based on the timed text metadata.
For distribution to VOD, Web or mobile platforms, the text will be rendered automatically into a TTML file, the most suitable web format.
“With this automated system we can prepare a video file for broadcast and for all non-linear distribution complete with captioning information in both French and English languages and in full confidence that metadata will be preserved,” says Weisz.
The closed-caption system is the latest evolution of CPAC’s relationship with Dalet dating back to 2009. The partnership began when CPAC began digitizing its legacy tape content and selected Dalet as a front end MAM to catalogue its new online archive. A year later, CPAC and Dalet built a full MAM workflow to manage and control the client’s long-form programming all the way to air. The next step, in 2011, was to design, install and execute a newsroom computer system for all CPAC’s news output based on the comprehensive set of Dalet tools for news production.
“The longstanding relationship we have enjoyed with CPAC is a testimony not only to the confidence they have in our technologies and services, but also to the unique way in which Dalet can grow with a customer as its needs change and grow,” says Roux. “We congratulate CPAC on its successful mission to date and look forward to many years of close partnership together.”