Getting even more out of the box19 December 2013
Last year’s Channel in a Box (CiaB) Forum produced some interesting comments and observations. Unsurprisingly, the subject continues to generate interest within the industry. So we gathered together some of those involved with the technology to see what difference a year has made. They are (in alphabetical order) Don Ash, managing partner, director, Sales at Playbox Technology; Ed Calverley, VP Product Engineering, OASYS Automated Playout; Tom Gittins, director of Sales, Pebble Beach Systems; Ofer Lugasi, Blend product manager, Orad Hi Tec Systems; Karl Mehring, Snell’s senior product manager; Scott Rose, director of Product Management, Miranda Technologies; Mat Shell, product marketing manager at Harris; Bruce Straight, director Sales and Marketing at ToolsOnAir; Sander ten Dam, senior director Media Playout Solutions, Grass Valley; Andy Warman, senior product marketing manager for Harmonic: Jan Weigner, CTO and managing director, Cinegy GmbH; Nick Wright, CTO and co-founder of Pixel Power and Goce Zdravkoski, Managing Director, Stryme.
By Philip Stevens
TVBEurope: This technology is ever evolving. What has been the most innovative advance over the past 12 to 18 months?
Ash: (Pictured) Business wise, reaching more Tier 1 broadcasters than ever before, most didn’t even acknowledge that CiaB existed one to two years ago.
Calverley: The question of whether a standard IT server can handle playout for a fully-featured channel is behind us. The issue now is ‘how much further can we push the resources of a single box?’ Oasys has found a way to not only boost channel count without increasing the hardware needed, but also to provide additional services such as simulcast, delayed outputs and outputs with alternative content. Managing this ever increasing complexity has been a challenge, both technically and commercially but we believe that Chameleon, the integrated solution we launched at IBC, is a big step forward in addressing this challenge.
Gittins: Virtualisation is a hot topic and several vendors, including Pebble Beach Systems, are engaged in developing solutions in this space. However, one interesting advance is the pre-rendering of complex graphics. Our new Marlin automated pre-render workflow addresses this, acting as a plug in to Adobe After Effects and taking the ‘heavy lifting’ of graphics out of the CiaB box, By managing the generation of 3D graphics in advance and then reinserting them into the playlist as a clip, Marlin avoids the potential for integrated channel devices to be overloaded during playout, meaning that more channels can be handled at the same time on each box, significantly reducing costs and saving space and power.
Lugasi: Channel in box solutions have advanced quite steadily in the past year. In addition to being failure protected, CiaB supports more video clip formats and codecs and offers close caption support, audio channel shuffling, and the ability to control template graphics going to air.
Mehring: New innovations have provided many opportunities to include an ever increasing level of functionality into software solutions for inclusion in channel in a Box products, these include features such as Dolby, loudness correction, viewer measurement to name a few. Increased performance in IT equipment and connectivity offer greater power for these platforms as well as the opportunity to stream using H.264 or AVC Intra for use in more advanced systems. All of this provides compelling low cost, feature rich and future proof solutions.
Rose: Although the best channel in a box systems have already added graphics, multi-format playback, advanced audio handling and other features, the last 12 months have seen a new focus on the user experience to bring those key missing components of the channel chain into solutions. Miranda, working closely with its customers, identified that subtitling and captioning are critical components for a significant segment of the market. We acquired Softel, the market leader in captioning and subtitling, early this year and immediately began introducing their technology into our product portfolio. What is now the Miranda Softel Swift TX subtitle/caption management and delivery solution is a perfect fit with our iTX integrated playout platform. Adding a premier subtitle/captioning engine that is a critical component of the playout chain provides a level of integration within a CiaB product that has been missing from the market.
Shell: (Pictured) The innovation for CiaB products over the last 18 months has been in two areas. First, the evolution to enterprise-quality automation and workflow tools. And secondly, the maturation of graphics features plus integration with graphics presentation toolsets. CiaB products started out as something that would fit into only certain workflows, providing less functionality than ‘traditionally implemented’ complex channels. Now the best-of-breed CiaB solutions are much more comparable to the most complex channel designs. Moving forward the CiaB I/O solution of choice is going to be IP with baseband over IP. The solutions will tend to be software only and positioned for virtualisation and the cloud.
Straight: Originally CiaB solutions were dedicated proprietary hardware. The key trend in the last 18 months is acceleration in the use of standard components and a reduction in the footprint and power consumption of these components. Additionally, we (ToolsOnAir) are starting to develop our own periphery products like our flow:rage video optimised storage solution.
ten Dam: The enhancement of live control is the most significant advancement in CiaB. The platform has always had live capabilities from a playout perspective, but has been targeted to structured playout (high level of automated processes)—manual operator controls were less important. Live control is now part of CiaB – in fact it is going into a second phase.
Warman: For Harmonic, the biggest advance has been the release of the Spectrum ChannelPort integrated channel playout system with powerful new channel in a box capabilities such as dual DVE functionality, dual live inputs, independent branding of simulcast channels, and support for external key/fill to enable the use of graphics systems at the same time as ChannelPort’s on-board graphic branding. The addition of AVC-Intra, H.264, and ProRes in addition to standard MPEG-2 and DVCPRO codecs enables exceptional flexibility, and new subtitle insertion allows for just-in-time playout of multilingual content. In combination, these innovative enhancements enable playout of complex and sophisticated programmes from a CiaB system. All of these capabilities are available as up to four independent channels in just 1RU, including storage.
Weigner: The arrival of 4K/UHD. Even if you do not care about the increase in resolution, finally interlace is dead! And of course while we are at it – everyone is going IP, so SDI must die.
Wright: [CiAB] technologies are a gradually maturing offering. Over the last 18 months or so those people operating in that business area have been continuing to make their offerings more stable and more complete. There hasn’t been any earth-shattering development in CiaB. It’s important to say that while CiaB appears to be such an all-encompassing notion it really isn’t one thing. You can’t define channel-in-a-box with a checklist and once you have all the boxes checked then you have a CiaB system. For different people solutions need to achieve different things. The challenge for us as a manufacturer is to understand how much to integrate with ChannelMaster – loudness management, subtitling, widescreen signalling? Indeed, it’s also a challenge for the customer in terms of understanding their business, where the growth is and whether CiaB technologies can assist in that growth. Obviously for us, the graphics, clip-playing and audio handling are very well rounded capabilities – you’d expect nothing less. Where we have been working hard is partnering with other manufacturing experts to integrate other leading technologies – our work with Screen on subtitling is a prime example.
Zdravkoski: (Pictured) Clearly, the most innovative advance has been the change from a channel in a box to a multi-channel in a box, including ingest, graphics, playout as well as H.264 streaming capabilities to mobile devices. At the same time, the TV business itself has changed: more and more TV stations are moving away from their own in-house playout system towards outsourcing the entire playout. This is becoming an increasingly attractive option and our fourth generation GENESIX VideoServer perfectly responds to this trend. As an all-in-one failsafe multi-channel solution, the GENESIX VideoServer offers professional HD quality and comprises both an advanced graphics engine for realtime graphic feeds as well as a streaming functionality at a reasonable price. Stryme is one of the few broadcast solutions providers that offers multi-channel streaming in the H.264 format.
TVBEurope: How practical is it to install a CiaB in a remote broadcast facility and monitor it from a central point?
Ash: It is not only practical, it is a reality. PlayBox Technology has around 2000 playout systems around the world. These are scheduled, operated and monitored remotely. PlayBox Technology has a workflow product called ‘EdgeBox’ that is designed to achieve this for Tier 1 broadcasters. PlayBox Technology also offers Multi Playout Manager which allows single and multiple channels to be remotely managed, controlled and monitored with different levels of rights management assignment from viewer to administrator. For commercials insertion, PlayBox Technology produces a product called AdBox for the remote insertion of adverts – or programmes, interstitials, etc – at a remote location. These can be controlled and monitored centrally via the internet. Insertion is triggered by cues that run a preloaded playlist.
Calverley: Very practical. So long as there is no time-critical messaging between the remote and central sites, there is no reason why the actual playout device cannot be located anywhere in the world. The beauty of systems like Oasys Chameleon that have evolved from the original concept of CiaB is that whilst they work well as part of a large multi-channel system, they can also regress to their original roots and operate simply as a standalone box receiving playlists and media with no reliance on connections to any centralised services.
Gittins: Depending on the architecture of the solution, it can be highly practical. The Pebble Beach Systems solution deploys a distributed architecture with Marina automation controlling the Dolphin integrated channel device(s). Unlike more basic solutions, it does not simply offer a confidence feed which is returned to the central site, but also incorporates the ability for operations to know in advance that all primary and secondary content is loaded. In fact exactly the same status and media validation updates are given as if the Dolphin were being controlled locally.
Lugasi: CiaB should first and foremost be designed to work in a remote facility with minimum user interaction, where possible. As most aired channels do not include ‘live’ or unpredictable programming that cannot be pre-scheduled, their media files can be prepared in advance. CiaB should handle media copying/deletion and playlist updates automatically. When a ‘live’ programme is scheduled, it is expected that the operator control the transmission playlist and react to countdowns coming from the ‘live’ production.
Mehring: There is a definite swing in the market towards broadcasters wishing to sit the CiaB on the edge in a remote facility and control and monitor from a central point. One thing we have found is that different broadcasters have different requirements for how much equipment is at each site. As a result we remain flexible in our approach so that the only consideration for our customers is what type and speed of connectivity they need between sites.
Rose: It’s not just practical; it’s a reality as proven by hub and spoke operations already in operation in North America. iTX is running large station groups this way, sharing both control and media. Additionally, advanced monitoring with video probing and playlist-aware alerts allow these often complex networks to operate very efficiently. The only real constraint is a business model that allows exploration of local or even hyperlocal markets for advertising and events.
Shell: An IP and virtualised approach to infrastructure and services greatly facilitates the playout in dispersed geographies and is one of the early steps in true geo-dispersed origination. Remote installation and technical monitoring – realtime and near realtime – are key components for a remote, edge playout facility. From the central location it should be possible to apply updates or configuration changes, troubleshoot periodic or ongoing events through a correlated logging server and receive realtime monitoring information about the health of the remote playout facility.
Straight: (Pictured) Very well. Really this is the only practical and reliable way to do this type of installation.
ten Dam: Installing and servicing CiaB in a remote facility is just as easy as in any local facility. That’s because all elements work together in a modular way and are, in fact, architected to serve a multichannel environment—making the distance between playout nodes and the control centre irrelevant. However, one should consider the impact of operating remote channel(s). Usually remote playout is chosen because of cost reduction and thus low speed connections. Content is often drip-fed into the playout facility over that cost efficient connection. This means it’s probably slower than real time, which has an impact on channel content refresh rates. Caching content on the playout nodes is the best way to solve this problem. Various scenarios can have content loops with minimal refresh to use as a default and/or emergency scenario. It’s not uncommon that similar solutions are actually used for disaster recovery (DR) playout.
Warman: The Harmonic CiaB solution can be installed in a remote broadcast facility and monitored from a central point; this is an approach employed for conventional playout using the Spectrum media server systems on which the ChannelPort is based. Designed to be controlled over a network, the system supports remote networked status monitoring. The range of monitoring functionality available differs according to the control system managing the CiaB solution.
Weigner: Very. That’s why being IP-capable – to receive or emit streams via IP – including the confidence, preview, multi-viewer etc. – is a must. Something we do from day one. If the solution was not architected with IP-based remote operations in mind then it’s too late now.
Wright: From Pixel Power’s point of view, the entire way that ChannelMaster was architected from day one is that it’s an enterprise-grade, client/server architecture. That allows the technology to be sited where it’s most convenient for the customer’s business. Before a single line of code existed this was a decision we’d taken. So in terms of how robust and scalable ChannelMaster is, the answer is very in both cases.
Zdravkoski: It is quite simple to manage such a remote operation. Our recent project for Telecom Italia is a good example: the VideoServer is located in Milano, yet the operators are in Rome. Nowadays, the actual location of a VideoServer and its place of operation are quite irrelevant. In other projects, for example in our project for DAF, we have installed the VideoServer in Vienna but the studio and the channel operator are located in Kulmbach, Germany.
TVBEurope: It has been suggested that the systems could be dubbed channels in a box. Just how many channels could be run from one box?
Ash: Theoretically 10 to 20 channels are currently possible using IP pump from a single PlayBox Technology server, but practically the cost of the box is so small that most broadcasters still prefer to run one channel, sometimes two channels, from one box. The number of channels from one box is also limited in character generation and graphics are required. PlayBox Technology has a number of broadcast customers that use IP pump to run 12 barker channels or NVOD for movies starting at 15 minute intervals. The technology is not really the driving factor to how many channels from one box, it is really down to operational issues.
Calverley: That all depends on your definition of a channel. Some incorrectly assume CiaB systems are simply a combination of video server and graphics keyer, when in reality most channels require a whole lot more features such as subtitle and ancillary data processing and support for a variety of input and output triggers and protocols. For simple SD requirements, channel counts can be comfortably pushed up to eight, with potential for more.
Gittins: The number of channels which can be hosted on a single platform is largely determined by the CPU power each channel consumes. Factors contributing to this include the video bitrate, the complexity of graphics, additional up and down conversion, audio track handling etc. The Marina/Dolphin solution can host up to four channels in a single box.
Lugasi: The amount of playable channels from a single box depends on the box’s performance, CPU, memory, graphics performance, and PCI lane speed. Technically, a few channels can run from a single box. The question is if the broadcaster wants to risk having few channels come from a single box or would prefer a dedicated box for the channels.
Mehring: How many channels can run from one box depends on the box or even if there is a box. For a box solution that depends on how much and what type of signals are being output. For a completely software, virtualised or cloud based solution there is no box, so the number of channels comes down to how much flexible computing power you have access to which in the cloud theoretically is just about limitless.
Rose: (Pictured) How many channels could – or should – be run is a key question. Some believe that running as many as possible from a single device is a good idea, based solely on cost considerations. However, you need to ask if committing that many channels to one device is really what you want to do. Sure, you can back up that device, but managing multiple channels to fall over in an emergency or planned maintenance is a headache most large broadcasters could do without. Plus, there is a trade-off between pure channel count and functionality. Most broadcasters demand the full channel chain inside a box, including key requirements such as Dolby E/D, Nielsen watermarking, captions and advanced graphics, not just lots of simple playback in one device.
Shell: With the continued and rapid improvements in CPU and GPU technology, the amount of functionality available on CiaB increases all the time and with that the possibility of running more than one channel in a box. The move towards software only solutions and IP I/O will allow CiaB systems to be virtualised, running on blade chassis located in private or public clouds. In this way it is going to be possible to run many channels on a single server.
Straight: It all depends on your definition of a box. If you look at the TOA Broadcast Suite, we can run two channels ingest SD on one MacMini and a single channel playout on one MacMini. So there you have two boxes.
The MacMini box can be installed in a Sonnet RackMount and you can put as
many of these as you want in your rack. Eight channels ingest and four channels play out? Not a problem, you’re just going to need a big box. The beauty of it is, that because of our distributed software module you can control as many of these channels as you want from an unlimited number of clients.
ten Dam: There is no simple answer here. It is up to the vendor to define a clear offering of what a CiaB really is. Obviously this can only be done after understanding customer needs and expectations. In a software-based solution, the question is actually how much performance is available on your playout node—then you decide how to use that in a particular situation. This means you can have a flexible, non-standard platform, which must have defined boundaries to offset scalability challenges. So a single CiaB playout node can technically house, for instance, several SD channels with minimal graphical performance or a single HD channel with a lot of graphics and PIPs, and anything in between. Since the platforms are more and more based on standard IT hardware this technical challenge will be less and less of an issue as performance capacity continues to grow.
Warman: The ChannelPort system can run in either one of two ways: integrated into a 1-RU MediaDeck 7000 chassis to function as a stand-alone solution server including storage, or integrated in a MediaPort chassis to add channels to Spectrum’s shared storage system. Both offer channel in a box playout capabilities. Both approaches to storage support four ChannelPort channels and four simulcast channels for a total of eight channels per rack unit. This gives Spectrum’s ChannelPort unparalleled channel density for a CiaB solution. Each playout channel and the associated simulcast channel are controlled by a single playlist. An added advantage is that every one of the eight channels can have unique graphic branding. The playlist controlling the channel plus simulcast pair does not have to take care of driving two sets of graphic templates; this task is handled internally by the ChannelPort to keep operations simple.
Weigner: (Pictured) Sixteen SD channels or up to eight HD channels using the currently fastest dual Xeon server (which fits in 1RU). Or when using blades we can deliver 32 HD channels from a 3RU box or 448 HD channels from a 42RU rack. If SDI is not required we can run the CiaBs as virtual machines in load distributed in a fail-over cluster and the density can massively increase.
Wright: It depends on what you mean by a channel! Again, it’s certainly as much down to the overall broadcast architecture as it is in-box technical capabilities. A single physical enclosure sooner or later has single points of failure in it and how much this is pushed is down to the user’s appetite for risk. If a customer wants eight channels from one box and a catastrophic failure happens to that box – the unthinkable happens and the dual-redundant power supply completely fails – then you are going to put all those channels off air. So you want a reasonable density, but you want to manage single points of failure from a business resiliency point of view. As the technologies progress and capacity within units grows, this will be a business decision more than a technical one.
Zdravkoski: Using our Genesix VideoServer, customers can run four entirely separate channels – each with one input and one output, including graphics, support for all codecs and H.264 streaming). Operating more channels means one gets a far more favourable price per channel.
TVBEurope: What QC capabilities are offered within CiaB systems?
Ash: PlayBox Technology offers a product called QCBox to monitor content that often comes from many different sources, encoded, transcoded and repurposed at a variety of bitrates, formats and compression standards for both SD and HD delivery. This multiplicity presents considerable challenges for AirBox playout. QCBox is designed to provide automatic monitoring of media content and delivers verification of whether legal and technical obligations are being met. Tests include checking for correct duration, continuity (no missing or overlapped frames), frozen and black frames as well as audio tests, including loudness, phase and audio levels.
Calverley: There are many people providing QC tools with a near infinite number of tests and checks that can be configured. For a CiaB system the real question is ‘will this file play?’ The tools offered by Chameleon can do a number of basic checks (e.g. black/silence/loudness), but the question of whether the file will play or not is normally the only one that really matters. Customers who want deeper QC tests should be free to choose from the best of breed solutions on the market and not be constrained by their choice of CiaB system.
Gittins: One of the biggest issues for CiaB systems is validation – the buyer needs to be sure that the system can handle all of the specific file formats and compression systems which will be delivered to it. Our solution focusses on pre-validating media as it is cached into the box to ensure compatibility with playback codecs. We also have the ability to offer a dedicated preview port or an IP stream which allows content to be viewed in baseband video using the same decoders as for on air playback.
Lugasi: Orad’s Blend should be considered like any other file-based video server and graphics template. That means media files should QC similar to a standalone video server, graphics template should be uploaded to the graphics engine and updated with relevant data. Putting both the video server and graphics engine in the same box does not change how the QC works. One of the safety features implemented in Orad’s Blend is the ability to validate all playlist events, and check media availability in the Blend storage when the clip is introduced to the playlist or when the playlist is updated, giving the operator time to prepare and locate missing media files prior to playout.
Mehring: (Pictured) Many systems don’t include QC because it is seen as an upstream process. The approach we take is that material needs to be checked to make sure that it does not cause a decoder failure when it comes to playout. For this reason we use a validation service to play in faster than realtime as soon as material arrives into the system. This allows rogue material to be quarantined and operational staff alerted to this exception. We also check the loaded schedule and alarm on media errors for problems such as incorrect material start points, missing captions or subtitles, missing audio files and so on.
Rose: Quality control within playout is present in two key areas – checking files and media upon arrival, and ensuring error-free playout. Quality control of media delivered to playout is assisted by file analysis tools, a number of which are in daily use with broadcasters. What’s crucial to the QC process is to enable operators to take the reports generated, interpret the results and see the media at the point of the suspected fault. Quality control of the output has advanced dramatically. Integrated monitoring enables operators to manage more channels and, more recently, playlist-aware monitoring using playlist data reduces the number of false-positive alarms by activating monitoring systems only when events that need to be checked are actually on air.
Shell: The processing power required for file based QC is high and will impact the performance of CiaB, for which the primary function is for realtime channel playout. Broadcasters should consider verifying and checking the technical compliance of content around its channels rather than from individual playout devices, for example during the ingest process or directly on archive or NAS. Most CiaB systems do have a baseband review port that can be used for the manual QC of content which can be especially useful in smaller systems where separated ingest capability does not exist. Tools that enable content to reviewed over IP as opposed to baseband video will become important with the transition to all IP I/IO.
Straight: I can only speak for the TOA broadcast suite. We currently do not offer file level QC. There are multiple third party vendors that handle this well. Video and audio QC will be added as option at NAB 2014.
ten Dam: QC within CiaB can be done through a low resolution proxy, high resolution from the playout device, or high resolution from a dedicated device. The best choice depends on the existing and desired infrastructure and workflow.
QC functionality is always on the requirements list but in real life it isn’t used that often. QC is expensive in operations so there is a real drive to eliminate this as much as possible. Also, the automated processes around content management are very reliable, so there is no real reason to check if the right clip will be played or not. And finally, reliable technical checks can be automated by integration with external dedicated devices, and limit operator intervention even further.
Weigner: Doing QC at the point of playout is normally a very bad idea. If you still encounter problems this late in the workflow chain you have a problem. That said, in the real world there is always that last minute piece arriving that needs to go to air. In our playout UI the operator can clearly see whether the file did load and whether the content decodes, the AFD, closed captions, parental rating and also other metadata. These can also be modified if necessary. Each clip can be previewed and inspected, then all the audio channels checked, the in/out points verified or changed or items split to insert commercials. Graphics can be previewed and changed, secondary events checked and changed. A horizontal timeline view improves the overview of what is going on especially for multi-channel playout scenarios.
Wright: That varies based on what you mean by QC. If you want to QC the assets that you are going to play, then we’d suggest that the way to do that is to integrate your favourite QC tool as part of a media management workflow. For example, if you’re moving material from the archive to playout, then do it via an active agent that QCs that material. The fundamental role of a CiaB solution is to playout your material, but you have to tell it that the material is fit for playout. It is possible to include this QC capability in a CiaB solution, but is it really desirable?
Zdravkoski: Experience has shown that video clips are the most critical links in the chain. We have therefore come up with a smart solution: a completely isolated process within the system. Our InProcess QualityCheckService virtually plays all files frame by frame, thus preventing problems or failures when video footage goes on air.
TVBEurope: One major concern among some broadcasters is the ability to take over manual control if, say, a sporting event overruns. How easy is to handle such situations?
Ash: Very easy and very simple. The automation in PlayBox Technology allows for manual override of automated playout. So for live productions, any changes to the playlist during on-air sessions are possible! There are no queued or locked clips. Every clip in the playlist, except the one which is currently playing, can be trimmed, edited or repositioned. Moreover, playlist order can be changed on-the-fly with commands like skip-to-next or jump. Such order changes are performed seamlessly without stopping current playout session. Live productions are facilitated by the powerful Live Show Clipboard which allows insertion and/or execution of various events or live streams.
Calverley: (Pictured) Nothing will ever rival the true flexibility of having a fully equipped system that might be typical of those found in any studio gallery. However, that is the domain of live production and very distant from the reality of most presentation/master-control systems that drive the majority of channels on-air today. Traditional automation systems rely on the passing of commands ahead of time to prepare server ports and graphics events to be rolled frame accurately with the switching of a mixer. This can make manual operation tricky, limiting it to simple hot cutting between sources or simple triggering of pre-prepared sequences of clips/graphics. If designed right, a CiaB system can provide far more flexibility for manual operation as the time between a user selecting a new video or graphic event and it being taken to air is now a matter of frames rather than seconds. With the correct templates pre-defined, even very complex manual presentations including multiple live sources, DVE effects and dynamic graphics can all be controlled via the click of a mouse or the press of a button. Yes – CiaB systems can work with external hard panels, too.
Gittins: The key is to deploy a system which is designed to deal with late breaking changes and live events, offering the flexibility to edit the playlist close to air if events overrun. The new SmartPanel for Marina gives the playout operator a higher level of manual control of automated and external devices, enabling them to control graphics and secondaries, even in a live environment.
Lugasi: (Pictured) Orad’s Blend can handle two additional HD SDI inputs to be used for any live feed. Operators can schedule ‘live’ events with an estimated duration and then schedule ‘manual’ events that can be played when required. Similarly, operator can override the playlist and switch to a live source without the need to schedule it in the playlist. This gives greater flexibility to choose the scheduled method: manual event, auto-advance, fix clock event, or a mix between the three.
Mehring: It is true to say that not all CiaB systems were designed with manual control in mind, however with Snell’s ICE product responsiveness and ease of intervention was included from day one. Manual control can be taken at any time if required using software or hardware panels with standard as well as highly configurable functions. We include commercial hotlist, hot start and news flash functionality in the standard package to allow for any style from fully automated to highly live and manual.
Rose: Managing live TV is not something that technology magically solves. The key to managing live events is to ensure that operators have the tools they need – when they need them – in a way that feels natural and intuitive. Some systems have removed physical control panels, or replaced them with a simple button panel. Many operators who use iTX to run highly demanding channels for national broadcasters choose it because it offers the choice of full panel control or the same capabilities via the iTX desktop.
Shell: At any time it should be possible for an operator to intervene and take manual control of the channel such as a live breakaway to an important news flash and with the ability to join the previous event where it left off, in progress or at a defined start point. Or be able to schedule a live event where the exact duration is not known and which will play until the operator manually takes the next event. During these manual events the integrity of the as-run log also has to be maintained. These scenarios are well known to CiaB vendors that have a depth of experience in traditional automation and whilst CiaB provides architectural simplicities and efficiencies it should not impair the ability to interact, control or manage the operations. Indeed, from the experiences gained from traditional automation and the aggressive investment in these technologies we should expect the operations to be simpler and more performant.
Straight: With TOAs just:play and just:live it couldn’t be easier. Using the DSK
capabilities that our hardware partners offer we can switch automatically
between video and any live signal.
ten Dam: This is definitely a concern and is being addressing through engineering efforts, and in fact there have been some significant improvements on the platform. But the challenge is with highly automated processes in the architecture, making it such a good fit for Disaster Recovery, there aren’t the conventional controls one is used to. However this is changing at a rapid pace. For example, improvements have been made for easy content search-and-replace functionality, jump/insert events and JIP.
Warman: The ChannelPort system makes it easy to take over manual control in three ways. The system supports hardware control panels from multiple vendors, so users can control switching, clip playout, and graphics manually. Harmonic also offers its OPC (On-board Playout Control) application, which allows users to break away from the schedule, override the playlist, and stay with a live event for as long as necessary. And finally, operators can use their choice of automation system to take direct control over the system.
Weigner: For live events, such as sports, manual control is the default mode as the end time is undetermined. The end is triggered by the operator and depending on the selected type of event following the live event items are either skipped, bumped or joined-in-progress which is a particular North American feature where items are assumed to have started as scheduled and the overrun of the live event eats away at this item – or something that Europeans will never understand.
Wright: (Pictured) Because of our extensive experience of master control and traditional branding playout, we know how important master control is with manual intervention. We also believe that having the option of a hardware panel – our Video Switch Panel (VSP) – is very important in this regard as it provides a truly tactile interface for critical manual activity.
Zdravkoski: This can be handled quite simply and comfortably with our Playout Automation Module. It allows the user to interrupt the current programme at any time, switch to the so-called JustInLive mode and later return to the scheduled programme. Besides that, Cueton and/or GPI frame enable precise switching between a currently scheduled programme and the live mode.
TVBEurope: How easily can CiaB handle distribution to other media platforms – tablets, smartphones etc?
Ash: PlayBox Technology offers a product called AirBox Multi Parallel Output or AirBox MPO which enables the running of two or more outputs so that broadcasters can easily provide parallel outputs in any combination needed to deliver the content. HD SDI, SD SDI (with realtime rescaling) and IP streaming or, for example, output in H.264 (MPEG-4) and MPEG-2 at the same time to feed different media platforms simultaneously.
Calverley: CiaB is a name for products relating to the generation of linear channels using software to replace traditional hardware devices, whether it is for traditional television broadcast or online platforms. With an ever changing range of distribution formats in the IP worlds we see that the role of the playout system is to generate the mezzanine stream, whether that is SDI or an MPEG2/H.264 transport stream that can then be processed for transmission/distribution downstream in dedicated systems. Where CiaB systems can add significant value is in the cost efficiencies in enabling output streams with different graphics/branding or alternative commercial content to be handled in parallel from the same origination system. This can provide access to new revenue generation models that previously would have been cost prohibitive to achieve. Typically, content owners have realised there is value in making content available online ahead of transmission, so it is less common now for the playout system to be the source of content for the catch-up/VoD platforms. However, in the case of live programming, the ability to automatically capture a live event to file as it plays out – clean/semi-clean of graphics – can be an added benefit to reduce the complexity of turning around live programmes for online platforms.
Gittins: In many cases clip-based content which is to be delivered in advance is handled as part of the upstream workflow before arriving at the CiaB for linear playout. But CiaB solutions which can provide a suitable IP stream do have a role to play in providing the live streaming of content or fast turnaround VOD workflows that relate to live events.
Mehring: Revenues available from other media platforms are typically less than those of more traditional platforms, so channel-in-a-box lends itself to these with their cost dynamic. The use of new streaming formats such as H.264 and AVC Intra direct from these solutions allows even more of the broadcast chain for these platforms to be included.
Rose: CiaB doesn’t in itself help distribution to other platforms, unless there is a framework around the CiaB devices to provide comprehensive media management. This is a fundamental principle of a well-architected solution. Media is the centre of a broadcaster’s world and being able to view, manage and control it is something that systems that only have a ’box’ struggle to do. A system that can output channels in baseband and IP, and help master media for distribution to online and VoD platforms is a compelling differentiator when choosing between vendors.
Shell: It will be possible with CiaB solutions to output uncompressed video over IP using the SMPT 2022 5/6 specification. This output can be used to feed external encoders, especially so when the range of target media platforms are diverse. CiaB products can also produce time-locked, scaled bundles of streaming video over IP from currently available technology. With continued improvements in CPU/GPU power they will be able to output multiple low bitrate streams. When the range of target media platforms is narrower greater economics can be found by using the internal encoding capabilities from within the CiaB.
Straight: The future is definitely in IP streaming. We offer an RTMP streaming option, which allows us to send a H264 compressed stream to any CDN for final distribution.
ten Dam: For a software-based solution it’s not a big step to deliver in different formats. Today, it’s primarily streaming in baseband domain, but IP out at various bitrates is quickly becoming the standard and as is file-based delivery for on-demand systems. The architecture is ready to support all of this with relatively limited engineering investment, and as such can be brought to market quickly and without impact on workflow or user interfaces.
Warman: (Pictured) CiaB solutions were not conceived with this objective in mind. However, one benefit of the simulcast output on the ChannelPort is that it can drive downstream devices, with graphics already tailored for web or mobile distribution.
Weigner: Very easily. Just define the output resolution(s) and stream to each of these platforms with its own resolution specific branding – feed into your own Wowza streaming server or to the CDN of your choice.
Wright: CiaB is traditionally conceived as technology that has video I/O, be that baseband or ASI or something else. You can use all of the same upstream workflow management and share content between channels that require traditional output as for a CiaB solution that delivers files or a stream. The latter output is obviously more appropriate for tablets/smartphones. We see no difference except for whether the CiaB device has a BNC on it for output or some other form. Everything that’s upstream of that final decision remains the same.
Zdravkoski: We have presented these opportunities at the IBC 2013. Tablets and smartphones are ideal for the mobile operation of the VideoServer, and for remote monitoring. As far as viewers are concerned, the H.264 streaming function, flash capabilities and various connections to leading streaming platforms enable viewers to watch video clips and TV programmes via tablets and/or smartphones.
TVBEurope: Someone has suggested that channel in a box will be replaced by ‘channel on a chip’. Is that feasible – and, if so, when will we see it?
Ash: I am not so sure this will ever become a reality because every leap forward in technology that can allow this to happen, sees a leap forward in requirements from consumers – or more to the point manufacturers of consumer products – promoting higher quality and definition, such as 3D, 4K. That will no doubt be driven to 4D, 5D, 8K, 16K and so on. When technology does allow for channel on a chip, I am sure the PSU to drive the chip and the fans to cool it down will still equate to a 1U box that is the same as today.
Calverley: Early CiaB systems were limited, but have now proven how powerful they can be. It is perfectly feasible to build a channel playout device on a card or chip and whilst initial models may be limited, like their predecessor they will no doubt improve. The important question is ‘what is the driver behind them?’ The traditional broadcast hardware vendors have an obvious motivation for selling custom hardware platforms – be it channel on a chip or some other non-standard platform – as it locks the customer in to only ever using that hardware with their platform. This ensures that customers have limited choice when it comes to expanding and evolving in future. The key value in most CiaB systems is that they are built upon relatively low cost standard IT hardware, something that, due to the scale of the overall IT industry, we know will continue to grow in power and reduce in cost. This means that, over time, channel density will increase and the per-channel hardware cost will decrease. Ultimately the real value in most CiaB systems is the software. Software can be upgraded over time, perhaps adding features that go beyond the capabilities of the original hardware platform, resulting in the need for new hardware. Alternatively, licences could be combined together to reconfigure systems as requirements change. The argument that broadcast hardware is ‘more reliable’ or ‘more robust’ than IT hardware is 10-15 years out of date. Add to that the fact that recruiting engineers who are confident to support broadcast equipment is getting harder and there is a clear argument towards moving away from custom broadcast hardware where there are better software-based alternatives.
Gittins: (Pictured) Whilst it may ultimately be feasible, we are far more likely to see a channel hosted purely in software, and ultimately in the cloud.
Lugasi: Aired channels are still relying on SDI equipment and feeds, so CiaB cannot get much smaller if equipped with graphics and I/O board. Once SDI becomes obsolete, new solutions will come powered by chips producing Channel on Ethernet stream.
Mehring: Channel on a chip would imply the use of bespoke hardware, our vision is to take this segment into a new way of working in a fully virtualised environment to allow fast deployment and new levels of operational efficiencies only available with elastic computing.
Rose: As the primary mover in the playout space, Miranda is constantly challenging perceptions on how to streamline operations and enable new business models. What we do is as much about providing business agility as it is about groundbreaking technology. The playout form factor – be it in a box, on a chip or in the cloud – is driven by how viewers want to consume content, not just by the form factor flavour of the month. When will we see the next innovation? Very soon I suspect, but it won’t just be playout in a smaller box or virtualisation for the sake of it. Whatever form it takes, it has to be an innovation that demonstrably delivers tangible benefits to the broadcast business.
Shell: There is no reason to believe that a chip could not host all the elements required, but today we see the drive for channel playout as being in an IP infrastructure, software only and running in a generic IT data centre. Some of the advantages to this approach are that software codecs provide greater flexibility, commodity platforms are economical and virtualisation can provide better resilience.
Straight: Interesting thought. It is not something we are currently working on nor do we see happening in the immediate future.
ten Dam: (Pictured) It’s always a challenge to predict, but definitely could be the evolution of CiaB. We will see significant changes over the next five years. Today, propriety hardware is primarily needed for high-level performance and very demanding baseband I/Os. With the increased processing power of standard IT hardware and the move towards IP output, these two are eliminated, freeing the way to ‘think outside the box’.
Warman: The industry is constantly refining the channel in a box concept and its feature set. This segment has moved very quickly in the past few years and will continue to do so. More and more broadcasters, teleports, and distribution facilities are buying into CiaB – many are no longer on the fence. CiaB solutions have been proving themselves in high-value implementations, and at this point most companies consider them a real playout option for their operations. How this translates to a channel on a chip has yet to be determined, but anything is possible.
Weigner: Never. It’s not about hardware but software. A traditional linear or on-demand channel is a software service running on a commodity IT compute platform – either locally in a ‘box’, or remote, or remote hosted in a datacentre which we can then also call cloud, if we want. Hardware-centric CiaB vendors are doomed just like traditional playout vendors.
Wright: This comes back down essentially to risk. Even if you can have a channel on a chip it doesn’t ultimately help you with a lot of the other costs. It’s not an enabler for the industry and, in many regards, plays against the drive to UltraHD and higher frame rates. Yes, if we stuck a stake in the ground and said, no, we’re never as an industry going to go beyond 1080 50i HD then maybe. But the landscape is changing so quickly that whether it’s a chip or a card or a box or a piece of software isn’t really the question. The questions are what you want to do, how you want to operate and how you get cost-effective, day-to-day operations, now and going forwards?
Zdravkoski I cannot yet imagine channel on a chip. But, I am convinced that we will see Multi-channel in a Blade Server as a new standard in three to five years from now. We are already working on such a solution. It is our aim to equip the Multi-channel in a Blade Server with capabilities for 32 or 64 completely independent channels, including ingest, graphics, playout and streaming, using just a single 4HE machine.