Multiple choices – technology choices for multiscreen TV

The technologies used to deliver content to multiple screens is still relatively new and non-standardised. Content and service providers will need to weigh their options carefully to deliver appealing services that make business sense. Stuart Thomson looks at some of the issues.

TV Everywhere has been the topic du jour at pay TV industry events for some time now, but as ever, there is a gap between the insider buzz amongst industry executives and acceptance amongst the public at large. While that highly sought after figure the early adopter has embraced multiscreen services with enthusiasm, this group is arguably more willing, or at least prepared, to tolerate glitches with the technology, a limited range of choices and a less than perfect viewing experience than the wider population.

And these glitches have yet to be fully overcome. The technology is still being rolled out and is still relatively immature, while content rights-owners have yet to fully take the plunge and deliver universal rights to single distribution partners.

Content preparation

For Steve Plunkett, chief technology officer at media management specialist Red Bee Media, the main challenge facing broadcasters seeking to deliver content to devices other than the TV is rights – the lack of availability of content is the one factor that is most likely to hinder the development of services and their widescale adoption.

“The most successful commercial and public broadcasters are outsourcing TV playout and also ingest and transcoding.” Matt Vidmar, Vision 247

On the technology side, the fragmented nature of delivery to devices beyond the TV also presents multiple challenges, says Plunkett. “Video needs to be prepared in a different way. You have multiple renditions of video assets which means changes to workflow,” he says. “You are trying to create the best available quality for the device, without too much bandwidth consumption. There is much less control over the distribution chain and even with a single device like the iPad you can make assumptions but you don’t have control of the end user’s experience.”

As broadcasters face an expanding range of devices and capabilities they also have to make choices between maintaining a degree of consistency and trying to provide services that make maximum use of the capabilities of each device. “Each of these devices comes with different user experience expectations,” says Plunkett. “Broadcasters need to avoid the lowest common denominator. That means building the experience and the people to create really good user experiences for each of those platforms.”[icitspot id=”28417″ template=”box-story”]

Lack of standardisation in workflows management and metadata that can be used to track and discover content is one key issue that could hinder the devlopment of services, according to Jon Folland, CEO of Nativ. “There are no standardised workflows at the video level, at the metadata level and at the best practice level,” says Foland. “The big challenge is fulfillment – getting content to all these platforms at the right time and in the right format.” Ultimately, the most powerful players in the market – such as Apple – will define how content is presented and formatted.

Metadata is an area of difficulty because practices that are common in the production environment are very different from requirements further downstream towards the consumer. “It’s a very hard problem to solve. It’s about the kind of data you use. What would make life easier is a single standard for managing metadata for the lifecycle of the content,” says Folland. While there are initiatives under way to simplify workflow practices – Folland points to the example of the Advanced Media Workflow Association (AMWA), whose activities in this area have been supported by both software providers and content owners – the likelihood of a comprehensive metadata standard is remote.

For Matt Vidmar, chairman and chief technology officer at UK-based playout and content aggregation specialist Vision247, the sensible approach to the barrier-to-entry problem for broadcasters is to outsource ingest and playout to third-party providers, even at a relatively early stage. “The most successful commercial and public broadcasters are outsourcing TV playout and also ingest and transcoding. One could argue that the factor of scale would not justify this approach. However the reality is that one can offload technical operations to external service provider at a fraction of the cost of what payroll costs alone would be in-house,” says Vidmar. “If a network operator elects to develop bespoke apps for a number of channels and for all popular devices, mobile and wired, this can turn into real nightmare time-wise and cost-wise. Again, I would say that the answer…is to outsource to a specialist company that can do it fast and very cost-effectively, typically repurposing already developed and well proven code which is simply adapted and rebranded.”

Adaptive evolution

Once the problem of ingesting and managing content is solved, operators face a bewildering range of choices about which distribution platforms to support.

Formatting content remains very complex. To solve the issue of limited and unpredictable bandwidth in unmanaged networks, content providers have taken to a ready-made solution in the form of adaptive bit-rate streaming – where multiple resolutions of a video asset are made available and the stream adapts dynamically to changes in the availability of bandwidth to give a smooth video experience to users – but this brings its own complexity as multiple versions of the same piece of content need to be created and managed.

In terms of the complexity caused by multiple bit-rate formats, Vidmar says the best approach is “to stream live and VOD as multi-bitrate RTSP and also as adaptive HLS, with the default preference to serve only HLS to all Android devices above version 3”. He says this is a standard feature in most Content Delivery Networks (CDNs).

Whether the ecosystem for multiscreen could be simplified by the introduction of standards for streaming video remains to be seen. Much work has been done on, MPEG-DASH, a proposed specification for adaptive streaming that could theoretically supplant HLS and Microsoft Smooth Streaming. However, it is more likely that the standard will effectively be one more choice for video service providers rather than a replacement, particularly given the existing base of devices that support legacy technologies.

“A couple of years back the challenge was to create all these different profiles but we are beyond that now,” says Boris Felts, vice-president of product marketing at Envivio. “The difficulty now is around value added services and how to create experiences that are the same or superior to those on the main TV screen – things like picture quality, bit-rate, sub-titling and multiple audio channels.”

Thierry Fautier, senior director of convergence solutions at Harmonic, believes that the economics of delivering content to multiple screens could be boosted by the deployment  in the near future of two standards-based bandwidth-saving technologies – the MPEG-DASH initiative to provide an open standardized way of delivering adaptive bit-rate video, and HEVC, the next-generation video codec that offers a step change in efficiency from the current H.164 standard. Fautier believes OTT video providers will migrate to MPEG-DASH to unify their streaming services around a single format and will be the first to adopt HEVC as the most efficient way to deliver video to multiple devices. “The first big step will be during the Olympics with connected TV manufacturers adopting DASH in Europe – that will show there is an ecosystem that works,” says Fautier. While Apple’s HLS (HTTP Live Streaming) technology has been widely adopted thanks to Apple’s grip on the tablet market, Fautier believes service providers will not want to be hostage to Apple, which has the power to make unilateral changes to the technology or to change the terms under which it is licensed.

One of the other key choices that has to be made, according to Fautier, is between deploying consumer premises gateway devices that can transcode content in the home and managing transcoding and formatting from a centralised location in the network. While doing it in the home could save bandwidth, Fautier does not believe that deploying expensive CPE with video processing capability makes any kind of economic sense. “People are not watching video on iPads 24 hours a day so we think it makes sense to do [the video processing] from the headend,” he says.

Fautier’s view is informed by his belief that out-of-home consumption of streaming video will be limited. Out of the home, connectivity, or the lack of it, will be a bigger issue than bandwidth, he believes.

Range of devices

The range of devices that are now capable of consuming video continues to grow, but in this universe some devices are more equal than others. While Samsung TVs and iOS devices including the iPad are essential, the ever expanding range of Android smartphones and tablets presents content providers with a greater dilemma, especially as many of these devices are not optimised for video. The number of devices multiplied by the number of adaptive bit-rate formats multiplied by the number of profiles required for each format equals a tremendous amount of video processing that needs to be undertaken to support the delivery of services.

Whereas previously service providers only needed to support single devices with multiple users – the TV screen – now they have to support single users with multiple devices. Beyond the first flush of enthusiasm from early adopters, winning mass adoption for these services will require them to offer something new, such as network-based DVR that enables users to access the content wherever they are, in the same quality that they are used to on the main TV screen.

While the iOS ecosystem – typically the first port of call for operators and broadcasters looking to deliver second screen distribution of their services – is relatively straightforward, being controlled by Apple, the Android world remains much more fragmented.

“Cable operators want a way to deliver the same channel line-up to existing clients on multiple devices,” says Felts. US operators have managed to secure the same rights for multiscreen distribution and have moved quickly to deploy services, he says. “The Android system is a bit cumbersome because not all releases are supported by all devices,” adds Felts. While newer Android devices, including tablets, can play back HTTP Live Streaming, legacy phones require 3GPP implementations or operators need to hire third-party software companies to develop dedicated apps that will play back video.

For Vision247’s Vidmar, the solution to the problem of having to format user interfaces for multiple screen sizes within the Android universe is to focus on the market leading devices – namely Samsung and HTC devices.

Operators will continue to deliver separate services for cable and satellite, IPTV and OTT, but Vidmar says solutions are available to simplify things, citing the Xtreme IPTV solution used by Vision247. The transition to OTT can be managed, but industry participants will wait in vain for universal standards. “I believe the only way to achieve the transition to OTT is to replace the back office with a system that can support all broadcasting formats across all the networks. The bottom line is, as in the old days of tape, when we had many different tape formats, and then switched to file-based broadcast but ended up with many different file formats, now the reality is that we have many streaming formats,” he says. “We might be lucky sometime in the future, and might end up with truly consolidated industry where a single VOD file format and a single streaming format will be all that it is needed to reach 100% audiences. However this dream is for the distant future.”

The need to prioritise the more popular devices nothwithstanding, Nativ’s Folland argues that the relentless pace of innovation in consumer electronics means that broadcasters really can’t afford to pick and choose. In addition to improvements in the processing power of devices enhancing their ability to play back high quality video, there are also ongoing improvements in the user experience driven by social networks and software providers that means broadcasters need to have access to rich metadata. “At the moment it’s a great big mess – it’s an exciting mess but people need to understand the challenges for content owners,” says Folland.

Quality assurance

Given the heterogeneity of the environment in which those content owners are now expected to operate, one of those challenges is measuring and monitoring the consumer experience and gathering data on consumer expectations and how users are behaving.

According to Johan Görsjö, strategic product manager at quality assurance specialist Agama Technologies, extending quality assurance solutions to the OTT environment is not particularly problematic, with software embedded in the player via which the video is consumed rather than on the set-top box. Multiscreen means a more complex delivery chain, however, says Görsjö. “You have more devices in the home and a more complex delivery chain with more entities involved,” he says. “Having a monitoring system that can tie all the parts together is an important benefit for the operator, because TV everywhere is more complex than cable.”

TV Everywhere services could involve video being handled by multiple CDN providers that do not provide comparable data sets, meaning that it is difficult to see how the customer experience is affected, he says (see sidebar for more on CDNs). Only one part – the service provider – is going to be held responsible by the customer for anything that goes wrong, on the other hand. “Many CDNs offer information but you may not be able to compare it and decide what it means for the customer experience. It’s going to be a challenge for operators and something we want to provide to give them and idea of what it all means for service quality,” says Görsjö. Embedding monitoring software in the player at the point of consumption can enable information to be provided to the service provider without the need to invest in expensive hardware, he says.

Görsjö also points out that adaptive bit-rate encoding, which also adds complexity to monitoring services, should not be seen as a silver bullet by service providers. “You could give customers inconsistent quality because bit-rates keep toggling between bit-rates or customers keep getting driven down to the lowest bit-rate – you don’t want Game of Thrones pushed down so it looks like a YouTube clip,” he says.

Agama’s CEO Mikael Dahlgren says that operators typically are not looking to target every Android device in the market. “It’s not possible to have a high-end service delivered to so many devices – you create one for a limited number,” says Dahlgren.

Dahlgren says that providing a full system for monitoring multiscreen deployments can make the operator’s life easier in that customer care departments will be more easily able to trace problems. Very large amounts of information can be extracted from the monitoring system and it is the job of companies like Agama to present that in a way that is useful to the service provider.

“You want to extract data that’s relevant for each part of the organisation, so you need power to process that in real time. We can take a lot of data from viewing devices and use that to back-track to find out where problems originate,” says Dahlgren.

It is possible in fact that the technologies required to support multiscreen distribution could themselves be used to much greater effect to provide data on customer behaviour and usage that service providers could use in a variety of ways. Delivering one of the recent SCTE Summer Lecture series in the UK, Stewart Newton, managing director, EMEA, for quality assurance specialist IneoQuest said that said that working with adaptive bit-rate technologies provided a tremendous flow of analytics. “To pay for content you need good quality even though people will watch good content in poor quality,” said Newton. Network operators are under pressure to support complex Service Level Agreements, which is made more complex by the use of third-party CDNs, he said.

“There are huge blind spots out there, especially when you are using third-party CDNs to get to the end device,” said Newton. “Adaptive techniques are very different to existing video delivery and CDNs can be a black hole for SLAs.”

Adaptive bit-rate technologies are challenging, said Newton, with HLS and Microsoft Smooth Streaming demonstrating very different characteristics in terms of how aggressively they ate up available bandwidth, for example. However, Newton said that the development of solutions to monitor adaptive streams provided extremely useful insights into user behaviour. “You need to analyse user behaviour, such as what the tolerance limit is for poor quality video,” he said . “Post-caching multiscreen analytics can provide a huge amount of data about how channels are performing, what’s happening with SLA thresholds and why people are watching assets for only a short period of time.”

Range of functions

Those usage patterns are also likely to be much more variegated given the range of functionalities and use-cases supported by different devices. In addition to offering a complete line-up across multiple device types, service providers and content owners are increasingly expected to support a full range of functions that viewers are already used to on the TV – as well as an ever-growing range of new ones that can be supported by new devices such as the iPad.

“We’ve noted that customers that were doing limited trials are going beyond that to offer an extended channel line-up but also services that are standard on TV but new in the multiscreen world like VOD, nDVR, time-shifting, multiple languages and so on,” says Nabil Kanaan, senior director, product marketing at video processing specialist RGB Networks.

Service providers are keeping the cost of deployment manageable by shying away from live deployments and by limiting the number of channels they offer to multiple devices. “They are offering 10-15 linear broadcast channels delivered OTT and they are starting with a limited library of on-demand assets and them beyond that, time-shifting,” says Kanaan. Operators are also increasingly interested in finding ways to make money from the services they have already launched, such as via targeted advertising.

For Kanaan’s colleague Yuval Fisher, chief technology officer at RGB Networks, the level of complexity that service providers need to address increases significantly when operators deliver content over the open internet to devices over which they have no control and which are used to consume content on the go. Typically operators are taking a step-by-step move towards the deployment of services, he says. “It’s an evolutionary approach. We are seeing the iPad as the first device and once that’s in place they add to it,” says Fisher Few operators are currently supporting all adaptive bit-rate video formats, he says. Moreover, operators remain hamstrung thanks to legal issues over content rights when it comes to providing on-the-go services, at least in the US.

More sophisticated services that require investment of time and money, such as so-called ‘follow me’ services that enable users to stop watching a show on one screen and immediately start on another, are still largely confined to trade show demonstrations, according to Fisher. Operators are still – of necessity – delivering their multiscreen offerings in separate silos rather than via a completely integrated headend. So far, he says, scalability issues have been manageable. However, a lack of forward-looking bandwidth planning by ISPs could have consequences if it is not addressed, he says.

Delivering content on the go also incurs bandwidth costs to the consumer. If the cost passed onto the consumer of viewing content on 3G networks remains high, services are unlikely to meet with success. This at least has been the experience to date of services in the US. In Europe, services have been deployed more successfully by integrated fixed and mobile operators such as Swisscom.

Monetising services remains a challenge for service providers and broadcasters. Operators need to market value-added services to provide additional revenue on top of free video-on-demand and catch up TV. Broadcasters are typically moving ahead to deliver additional advertising revenues from catch-up services while operators are still looking at services primarily as useful to combat churn. Given all the costs involved in deploying services in the first place, this is perhaps the greatest challenge of all.