As content providers acquire rights to distribute media worldwide, the demand to produce and play out content in multiple languages is increasing. This necessitates the acquisition, management and control of multiple-language versions of each piece of core content. But how can this media be generated and managed efficiently while keeping international playout centers cost-effective, and what are the pitfalls of multilingual playout operations?
Historically, a combined video-audio asset was produced and held — typically on tape — for each language version of any program. The creation of a complete new combined asset was necessary as the requirement to add a new language arose. To produce new tapes and physically transport them to and from the dubbing house was costly and time consuming. It also required a high degree of human intervention before finished versions were ready for transmission.
With the advent of file-based workflows, it has become possible to assign multiple audio tracks to a single video file. This caters to the proliferation of language variants within a single asset and enables more efficient media management. Now, the number of audio tracks capable of being stored is restricted only by the server capacity and any limitation of the server product being used. The number of languages that can be played out simultaneously is also an important consideration when implementing multilingual services.
The required languages are usually listed in the traffic and rights management systems, and these requirements may change over a period of time. Although the addition of new languages may be handled immediately upon arrival of the new program, subsequent updates may be necessary as distribution rights change, or new channels are launched. It is not uncommon for new languages to be added to a digital asset long after it is first created.
Furthermore, new language tracks could be required either for assets that already exist within a given system or for media that is yet to arrive. It is vital that all language requirements are scheduled ahead of time. Advanced notification could come from traffic. Then, the media asset management (MAM) or automation system could initiate processes accordingly, depending on whether an asset exists, or if an asset arrives without the full complement of languages attached.
However, not all traffic systems have access to information relating to which languages or audio formats are required for broadcast. Or they may lack the ability to deliver this information. That is where today's sophisticated automation and MAM systems offer significant operational advantages.
Automating language provision
If audio-track assignments are captured at ingest time, an advanced automation system can find and select the correct languages for playout based on which broadcast channels are using the media. And if the asset is only partially complete, lacking one or more of the languages required, an automatic workflow can be triggered to address the problem.
Having established the audio-track requirements, the automation system can alert operators if a specific audio track is missing, so that it can be added in time for playout if it is available within the facility. It is also possible to automatically generate a low-resolution proxy version of the program. This small, easily distributed file can be sent, along with original audio, to the facility where the additional audio tracks will be created.
An automatic workflow enables the number of outstanding jobs with the dubbing facility to be tracked and easily managed. The smart automation system will listen for a response from the dubbing house to confirm that the new audio track has been created and sent.
Upon receipt of the dubbed, discrete, audio files, the system can stitch all the elements together, automatically associating the new audio tracks with the original high-resolution video asset, and updating the media metadata. This normally uses standardized track tagging information following one of the ISO standards.
It is important early in the project to standardize on the language codes that will be used (e.g., two-letter or three-letter language codes) and also on the tags that will be used to differentiate between mono and stereo (e.g., ENG L, ENG R).
However, post-production systems and commercial distribution companies do not always deliver multiple-language media in file formats that contain the correct metadata (e.g., ISO tags) required for reliable control at playout. The automation system needs to provide tools that allow for manual or automated tagging of files before they enter the core system. This way, assets arrive in the playout center carrying the correct metadata within them and can be fully utilized without delay. If implemented as stand-alone applications, these tools offer the advantage of being usable before or after the file is delivered, without needing to access the main automation database.
In many workflows, adding a new language will result in the need to QC the program again before moving the content to an archive or directly to the playout environment. This is particularly true as many countries now have strict rules concerning loudness of program material. Where the automation system offers content management and configurable operational rules, this process can be managed automatically. QC software is available that can be integrated within the system to perform essential checks automatically.
When it comes to playout, the asset may be broadcast simultaneously on separate audio tracks on the same channel, allowing the viewer to select the preferred delivery language via his or her set-top box. The automation system needs to ensure that audio is present on all of the output tracks in the broadcast chain, a task that often requires complex rules to be configured. If, for example, one of the required languages does not exist, which language should be substituted in its place?
Where programming requirements dictate that the automation system handles playout of the single video asset at different times with different associated audio tracks, the video server may not support that kind of complexity. For example, a movie may be broadcast with original sound and subtitles at one time, but may be required to be broadcast with a dubbed soundtrack at another time. In order to make this happen, the automation system needs to be capable of event-by-event audio track control. A system capable of using the video server's API, which allows audio track selection within the server, avoids the need for external audio processing, thereby saving costs.
Not all video servers have the ability to shuffle audio tracks, so an automation system that can offer frame-accurate control of external devices such as master control switchers and external audio shuffling devices ensures that the correct combination of audio tracks is implemented on the final output of the system.
If audio track management is available on a per-event basis, the system can deliver more efficient use of the available resources. However, key to this is the ability of the automation system database to be language track aware. This can be accomplished either by the language code being entered at the ingest stage manually, or by subsequent interrogation of the file header to extract the audio track data and combine it with the audio requirements determined by the traffic interface.
At playout, the automation system should prevent the operator from running an event that has incorrect or missing audio, by including audio-track availability in the media validation checks. If the availability of the audio track is set as one of the key criteria for content being ready for transmission, the operator will be warned if it is not present, thereby allowing him or her to select an alternate asset for playout. This is particularly important for fast-paced channels such as multilingual news services.
Finally, the complexity of scheduling commercial breaks in multilingual programs should not be underestimated. The placement of breaks needs to factor in the various cadences of different languages to ensure that the dubbed tracks are not interrupted while speech is still occurring. If the automation system can offer the ability to browse a low-resolution proxy video along with multiple audio tracks and to move smoothly between them, the operator will be able to select accurate timecode points for commercial breaks for each language. Each language version, although sharing the same program and media ID, can then have a unique pattern of breaks.
In the new file-based workflow environment of multilingual playout, it is the automation system that is uniquely placed to assess media files and to control devices accordingly. Playout management is significantly streamlined, with the ability to automatically track assets and changes without manual intervention.
Modern automation systems enable more efficient use of storage, as only one instance of the video asset is required, regardless of how many audio tracks are added. Also, server decoders can be allocated flexibly, taking into account other programming requirements. Further savings are made thanks to the ability to distribute low-resolution copies of files to dubbing houses: no more waiting around for tape deliveries. The system is also easier to scale: New countries and languages can simply be added automatically into the existing workflow.
Alison Pavitt is marketing manager at Pebble Beach Systems.