The bottom line

Increasing statutory requirements, viewer expectation and geographically diverse content distribution mean that subtitling is now an integral part of modern broadcasting and content delivery, writes Sam Pemberton.

Compressed production cycles, requirements for easier global distribution and high definition are driving a shift to file-based or tapeless workflows. As broadcasters, networks and content creators switch to file-based operations the traditional captioning workflow has ceased to be a viable way of handling the growing requirement for captioning, especially as much content is now delivered across multiple formats.

In file-based broadcasting, the typical subtitling workflow is to use early bound subtitles, where the subtitles are encoded into the asset stored on the server well in advance of broadcast, or may even be ingested to the server complete with subtitles. The subtitles may be created on a subtitle preparation workstation, or require extraction and editing, for example to fit a version of the content that has been squeezed to fit a time-slot. A subtitler may need to review a browse quality clip stored on the server or digitised video from tape, listening to audio, extracting relevant dialogue and sounds, then typing into a new subtitle file with in and out timing. Alternatively, the subtitles may be a translation of existing subtitles with relevant retiming and editing for inappropriate material.

Often subtitles may not be available in advance of time of air, or need late editing, and this requires a facility for late playout. While live events, such as news or sports broadcasts, require live input from a subtitling workstation via a subtitle transmission unit, either through the use of high speed input devices such as Steno or through re-speaking of dialogue or commentary, using voice recognition that is trained to the subtitler’s voice, with the generated subtitles sent to air either by a voice or keyboard command, with the ability to correct misspellings or incorrectly recognised words. These early, late and live workflows require an end-to-end subtitling solution, with seamless integration into a broad range of automation, workflow and video server systems and support for numerous file types and formats.

Often content needs to be repurposed for broadcast or for web and mobile formats, which may require that the content is also edited down or squeezed to fit a particular time-slot or to remove unsuitable material.

If subtitles have previously been created and encoded into the video, these must be extracted using Apple Final Cut Pro or similar. When this is done, an Edit Decision List (EDL) is created, which describes the edits, detailing the time code or frame reference of a cut or squeeze.

Using a file-based subtitle and caption data processor, subtitles can be extracted from the asset to a file and can then be edited as appropriate, referencing the EDL, using a subtitle preparation workstation. It’s important to select software designed to make creating and editing subtitles as efficient as possible. Look for features to quickly reformat subtitles or change in/out times, and adapt subtitles to different frame rates at the click of a button. The preparation software should allow files created for one format of playout, such as DVB, to be exported in other formats like DVD or Blu-ray bitmaps.

Also useful are pre-definable house styles that include quality control measures such as reading speed, along with the ability to review subtitles at any stage, either before committing them to the file or after, via a browse quality clip simulating burnt-in subtitles.

Once the required edits have been made, the subtitle file is reinserted into the original or edited video asset, using the file-based subtitle and caption data processing software. If content is to be repurposed for the web or mobile, these platforms can be handled by transcoding the video to the required output format then either burning in for mobile or adding a suitable subtitle file (for example DFXP or SAMI) to the asset for web playout.

Integration with video servers and workflow systems for early bound workflows

Seamless integration with video servers and workflow systems is vital for efficient file-based subtitling workflow. Integrating with the video server ensures that native file formats including key variants such as 436M tracks in MXF files are fully supported, as well as making job processing more efficient by using built-in APIs and interfaces. Subtitle processing can usually be performed against native browse quality clips, saving time, while final encoding should be faster than real-time, so for example an hour long program can be processed in a matter of minutes.

By integrating with workflow systems, subtitle processing jobs can be made not only more efficient through increased automation, but can also be closely aligned and prioritised with the broadcast schedule. The subtitling solution should offer different methods to automate job processing: where the product is used manually in a post production environment, jobs can be submitted in batches for efficiency. At a higher level, it should also feature a flexible Software Development Kit (SDK), providing an extensible interface that allows integration with virtually any automation or workflow system.

At the core of an integrated subtitling solution, a sophisticated subtitle transmission unit can play a pivotal role and become more a master caption management platform, driving off-line caption insertion (early binding), identification and playout of late and/or near live caption files and switching of live input.

With the right blend of integrated technologies and expertise, multi-lingual subtitling in file-based workflows can be made more efficient and cost-effective.

Sam Pemberton is CEO of Softel.