Cloud TV and OTT are coming of age as we move into 2014 with two parallel standards developments that will lead towards a universal video player for IP connected devices and an architecture for the ‘virtual living room’ to share content and services among friends and family.
These two developments send contrasting messages about the fate of the set top box or home gateway, with the universal player likely to stimulate box-less OTT services, while the virtual living room can complement local storage by enabling sharing of recorded and user generated content over the Internet. Both developments are embedded with a little controversy, in the case of the universal player over the associated Encrypted Media Extensions (EME) and the royalties payable for it. For the virtual living room the main hurdles concern management of rights when households are sharing access to their TVs or content recorded on their DVRs.
The universal player standards are being driven by the World Wide Web Consortium (W3C) with key support from DLNA (Digital Living Network Alliance), UPnP (Universal Plug and Play), CableLabs and the DASH Industry Forum. The universal player is based on DASH streaming and has been strongly backed by Netflix, Google and Microsoft, which are the three principle editors for the standard inside W3C. All three have been incorporating the universal player standard within their platforms and demonstrating parts of it during 2013. The motivation is strongest of all for Netflix, which currently has to maintain apps for many different platforms and stands to save a fortune if the core functions can be compressed into a single web application, with web browsers delegated to deal with the individual device platforms, without the need for plug-ins.
Netflix has been using Microsoft’s Smooth Streaming rather than Dash, which was only ratified in December 2011, and the Silverlight plug-in rather than web standards for the run time environment. However its latest player is web based, using DASH and EME. Netflix first demonstrated playback with DASH streaming in Microsoft’s Internet Explorer 11 (IE11) using Windows 8.1 without browser plugins in June 2013 at a developers’ conference called Build (Building Media Streaming Apps and Sites without plug-ins using MPEG Dash). This made use of code contributed by Microsoft and represented a significant step forward in cross platform content streaming by removing the need to develop proprietary plugins for each platform.
Using the Dash code written in JavaScript (dash.js) developers can instead build browser based MPEG-Dash players that comply with the DASH-AVC/264 Implementation Guidelines. This provides the foundation for the W3C’s universal player initiative, with the latest version of dash.js supporting the W3CF’s Media Source Extensions (MSE), which allow JavaScript to generate media streams for playback.
While the browser now takes care of incompatibilities at the device hardware level, the Dash player must in turn be able to cope with differences between browsers. A core part of the W3C work leading towards the universal player therefore involved changing how core video streaming operations such as seek, play and pause are implemented, in order to address potential incompatibilities between different browser implementations. Alongside this redevelopment the W3C has put in place an improved test infrastructure to help third parties make contributions to enhance the player.
Google has also demonstrated DASH working with Microsoft’s PlayReady as well as its own Widevine DRM and has been incorporating the W3C Media Source Extensions into its Chrome browser, just as Microsoft has done with IE11.
The issue of encryption around the universal player has proved more controversial, although not among the three main editors. This came to a boil in May 2013, when W3C announced it was going ahead with its Encrypted Media Extensions technology to enable use of third party encryption mechanisms and DRMs for content protection, in the face of opposition from some Internet rights groups. The Free Software Foundation and Electronic Frontier Foundation ganged up to collect 25,600 signatures for a petition, arguing that EME was a backwards step, endorsing and enabling business models that ‘unethically restrict users, and making subjugation to particular media companies a precondition for full Web citizenship.’ The two campaigning groups noted the irony that just as proprietary technologies such as Flash and Silverlight were finally being killed off, they were being replaced by a ‘control fantasy’ promoted by three equally dominant media giants.
These complaints might seem surprising given that EME does not specify either the encryption or DRM used to control access to or copying of video. Instead, it provides a standard mechanism for a browser to call a plug-in that handles the work. It is an enabling mechanism that can be standardized across all browsers so that a universal player can exploit whatever content protection is required without having to support it.
But it was the support for DRM plug-ins that attracted the complaints with the argument that this would hamper the very interoperability W3C was aiming to achieve with a universal player and also provide a back door for media companies to insist on an element of proprietary software. Even though no DRM is specified this could sustain the status of a few existing dominant providers and shut out alternatives, so the argument went. It also flies in the face of the W3C’s guiding principle that all technologies it supports should be royalty free.
W3C’s CEO Jeff Jaffe has refuted the idea that EME was standardizing the status quo, which he admitted was not ideal. He agreed that existing DRMs were proprietary and encumbered by patents. He argued that a universal player has to make use of content protection technology that is available, even if the situation is not ideal. Some standardization and openness was better than none, he argued, given that the alternative was continued reliance on proprietary plug-ins such as Adobe Systems’ Flash Player or apps that bypass web technologies altogether. At least the W3C standard converges on DASH and on HTML5, and for content that does not need to be protected, this makes for a single method of delivering content.
The DRM is clearly here to stay, since the alternative would be to rely on native applications instead, with the same problems of proliferation that W3C is trying to get away from. The only way to silence critics such as the Free Software Foundation would be for the primary DRMs being promoted for the universal player to be made available open source. Microsoft might argue that PlayReady can already be ported to devices that use non-Microsoft OSs, codecs or media players, but it has yet to release a version for open Linux distribution.
The issue of royalties has also dogged the codec movement, with H.264 bogged down just as heavily by patents. There again W3C turned a blind eye to that in the development of HTML5 in its desire to get the standard out of the door, in the absence of a royalty free alternative to H.264. Far from being resolved it looks like the situation may be even worse for the latest H.265 codec, otherwise known as HEVC (High Efficiency Video Coding), which halves the bit rate for a given video quality compared with H.264. H.265 is also riddled with patents and the main issue there is that, unlike with H.264, it looks like some of the patent holders are intent on pursuing their own royalties rather than acting on a group basis through a patent pooling arrangement. This could effectively strangle the standard at birth in patents claims, or certainly hamper its deployment. This in turn could have knock on effects for OTT and IPTV services and also for roll out of UHD. Having said that, we have the same furor over royalties every time a new codec is standardized upon, as more companies want to get in on the bandwagon. The truth is that they all know that without a patent pool, the system will not be adopted, and it is all gamesmanship.
The other impending development for OTT and cloud TV of the virtual living room, also has some issues around security, but in terms of reconciling differences between individual subscribers’ rights rather than over technologies and royalties. Here UPnP is the architect with its cloud TV architecture due for ratification around the end of the year, according to its Principal Architect of Advanced Platforms and Services, Clarke Stevens. This will allow devices to be used and shared in principle from any Internet connected viewing devices in a secure manner. It will also allow multiple web services to be combined into a single complete solution for users, so has close synergy with the W3C universal player development.
The UPnP Forum has been pushing this as part of its more advanced UPnP+ standard first announced in August 2012, extending the remit of UPnP beyond the living room with more robust security and remote control. ‘This takes everything we’ve done in UPnP, where we can control the TV, play things off the Internet or local storage, and share it remotely,’ said Stevens.
As Stevens pointed out the essential challenge of secure messaging or information exchange within closed user groups had already been solved by the chat room standard XMPP (Extensible Messaging and Presence Protocol), which is used for accessing Facebook Chat for example. ‘That allows you to take a logical model of a chat room and map that naturally onto UPnP,’ said Stevens. ‘So you can set up a virtual living room and invite people to share devices. My mother can share my TV and I can share home movies with her and play them on her TV.’ This of course is where rights issues may arise, but at least the same content security mechanisms supported by the DLNA protocols are available. DLNA is based on UPnP for device discovery and automatic configuration.
The UPnP Forum has also been active on another front, management of sensors under the banner of the Internet of Things (IoT). An interesting aspect of this is that these sensor management specifications were developed as part of UPnP+, which the forum is promoting as a platform to unify media and automation within the home as a facilitator of multiplay services. For this reason the UPnP Forum timed the launch of its sensor management standards to coincide with IBC 2013, rather than shows aimed at home automation. ‘It is defined as an architecture to bridge from a UPnP command and control out to other networks like Zigbee or Zwave or Bluetooth,’ said Stevens.
It is aimed at broadband and pay TV operators seeking to attach sensors and actuators to their home networks, with the ability to monitor and control the devices remotely. Here there is overlap with the cloud architecture within UPnP+, given the aim to interoperate where required with non-UPnP networks. A key part of this strategy lies in full integration of the IPv6 protocol with full backwards compatibility with IPv4. IPv6 will gradually take over from IPv4 for interoperability among IP devices and will be essential for communicating with many emerging sensors and actuators. UPnP+ also brings improved support for low-power and mobile devices as well as access to UPnP devices and services from web browsers.
The particular focus of the UPnP sensor management specifications is in connecting islands of home automation or service such as energy management, automotive and health/wellness ecosystems and bridging them seamlessly into the IP home network that IT and AV devices currently connect to. The result would be a single integrated solution spanning all underlying network technologies, including Wi-Fi as well as the likes of Zigbee and Zwave.
The specifications also support retention and distribution of sensor data once it is delivered, with the ability to coordinate processing of multiple sensor data streams. This can provide valuable insights over time into trends, such as household power consumption, measured month-to-month and year-over-year, or personal health trends such as weight, blood pressure and glucose levels over time. It also provides the foundation for coordination between these islands of automation, rather than merely their connectivity. It could, for example, turn up the heating in a given room if an infra-red sensor indicated occupants’ body temperatures were falling just a fraction of a degree. That may be not be a realistic example, but gives an idea of where the UPnP Forum is heading. It now has a group called the E-Health & Sensors Working Committee that is defining interfaces for devices in the home network to communicate with each other for distributed sharing of sensor information from personal health devices. This is based on the IEEE-11073 or ANT+ standards to provide real-time and/or historical processing of personal health data for display on existing home network AV (Audio/Video) devices.
The forum anticipates growing synergy between television and healthcare in the home. At the same time it has prepared the ground for a proliferation in remote extensions to home services, whether TV through its cloud architecture, or automation/monitoring/control through Sensor Management.