The channel-in-a-box device has evolved from a simple system to a sophisticated solution that can enable a TV station to go on air within a short period of time, says Andrew Warman Integrated playout systems also known as channel-in-a-box products are seen to be an important part of the broadcast future. If they […]
The channel-in-a-box device has evolved from a simple system to a sophisticated solution that can enable a TV station to go on air within a short period of time, says Andrew Warman
Integrated playout systems also known as channel-in-a-box products are seen to be an important part of the broadcast future. If they are to be a critical component of broadcasting, what are the fundamental requirements we need to consider?
When first proposed 20 years ago, playout automation was primarily a way of reducing operational expenditure in what was traditionally a people business. In the past, in order to play out a channel with programmes, commercials and trailers required significant numbers of staff to load tapes and check assets.
Once the technology became stable, and particularly when video servers took over from robotic tape libraries, two more advances were seized. First, it became practical to transmit more than one channel from a single installation, and systems were developed which broadcast tens and even hundreds of channels from one location.
Second, a new degree of sophistication was added to the transmission. Graphics and branding became more prominent to retain audiences. Automated rundowns and squeezebacks were added to the mix to create an immersive experience that viewers will not want to switch away from.
Both these trends were originally supported by the traditional playout architecture of broadcast-specific devices switchers, servers, graphics units and so on operating under the control of a real-time computer system, which, in turn, took its instructions in the form of playlists derived from scheduling, advertising sales and asset management systems. The transmission playout system depended upon powerful software with both collaborative and real-time elements.
In a multi-channel playout environment, adding or modifying a channel required, at the very least, new configuration files in the playout system, together with additional broadcast signal paths and possibly, additional capacity in areas like the server and graphics. It made for a very robust construction, capable of delivering the five nines reliability 99.999% up time that broadcasters came to expect. But it was a little inflexible, and project timescales to add channels were typically measured in months.
IT solution
The power of standard IT platforms has continued to develop, and some developers recognised that a device which is essentially an off-the-shelf PC had sufficient power to perform the basic functionality of broadcast playout. As the traditional user interface was not required, these developers put the processor board, a commodity video card and some storage into a 1U rack cabinet, creating the channel in a box.
Adopting such a simplistic approach has led to many questioning its overall reliability, and in truth there are a number of areas where such an approach can go wrong.
The top priority has to be to maintain a perfectly smooth and seamless video output, which provided the ingested content is good is a relatively trivial task for a PC and a reasonable disk drive. What can confound it is peaky processor demand for other functionality.
That can be operating system demands, for instance, to communicate with external systems downloading schedules or adding fresh content. More likely, it will be the graphics requirement, which tends to demand bursts of activity to render and deliver. Because each graphic element is likely to be different, the level of these spikes is not consistent and results in a variable load.
Within a good system design, typically limiting the amount of interaction with external systems and the sophistication of the graphics, channel-in-a-box systems have proved themselves and have become a part of the world of playout.
Typically, they are seen as suitable for smaller and niche channels, which leads to an obvious question: what happens if the small, niche channel becomes successful? Do you have to move to a full-blown automated channel, and what happens while the necessary integration work is being carried out?
More generally, can traditional and channel-in-a-box architectures co-exist? Is there an infrastructure which will allow playout centres to use all the available technology? And if there is, will it lead to a new sort of agility in building, developing and managing channels?
A new architecture
The first target for a new kind of system is that it must support four configurations. The classical channel-in-a-box has internal automation processing and content storage and so is completely self-contained. In a playout centre, though, this could lead to unnecessary file transfers and layers of content management, so a second configuration would be to have internal automation but with the device talking to a shared storage network to access the content it needs.
Just as storage should be internal or external, so too should the automation. If you have a multi-channel playout system running a large centre, it should be capable of managing channel-in-a-box devices too. Again, this could be with internal or external storage, making the four configurations.
The logical way to accomplish the requirement for internal or external automation is to make them essentially the same: to take a proven playout software structure and use the parts required to manage a single channel. The channel-in-a-box device will then perform in exactly the same way whether it is under its own control or a larger playout network.
The same logic applies to graphics. If you have a graphics workflow for the creation of templates and special sequences, you do not want a separate workflow for the channel-in-a-box services. It adds cost and complexity, and makes it hard to migrate from niche to premium channels. So the device should contain standard graphics software.
Finally, it needs to be extremely stable, so the architecture has to be more sophisticated than the simple PC in a box. The configuration could contain three separate disk systems: one, for the core automation processes and operating system; another, with very fast I/O, for graphics; and still another, with very precise latency for video playout. Rather than using a generic video card, teaming this last disk with the same hardware and software used in a full video server will also ensure predictable performance.
Such a device will fit into an existing playout centre with virtually no integration time: a channel can be on air in a matter of hours. It requires no retraining of staff to use it, and no additional resources.
As a final comment, this type of channel-in-a-box device requires a new business model attached to it. The hardware cost is relatively low, and so playout centres could be encouraged to install a number of devices in advance, only paying a licence fee when they go on air. This would mean that the theoretical claim of getting a channel to air in a matter of hours could actually be achieved in practice, giving playout businesses an incredible degree of responsiveness without significant costs and no compromise to technical and creative quality.
Andrew Warman is senior product marketing manager at Harris Broadcast Communications.
