HbbTV adds ATSC 3.0 watermarking for app discovery, synchronization

The European HbbTV standard and the much more recently developed North American ATSC 3.0 are often now presented as competing alternative technologies for interactive hybrid services enabling broadcasters to derive similar benefits to OTT services. Yet HbbTV has stolen some of ATSC 3.0’s clothes, or at least one of its garments, in order to incorporate audio and video watermarking both for security and to help synchronize apps accessed over a broadband connection with the broadcast signal. This is important in the light of one crucial difference between the two, which is that HbbTV uses the long-established MPEG-2 Transport Stream (TS) for its transport layer, while ATSC, having the benefit of some extra years development of streaming, went for an all-IP system based on the DASH Route (Real Time Object Delivery over Unidirectional Transport) and MPEG Media Transport (MMT) standard. This inherits features of MPEG-2 TS but optimized for delivery of streams over unmanaged and unpredictable IP networks.

DVB, however, argues that MPEG2-TS itself remains the best technical solution for unidirectional distribution of broadcast content, including terrestrial, satellite and cable. This reflects a distinction that often seems forgotten between HbbTV itself and ATSC 3.0, which is that the former has been designed for all four legacy broadcast media, that is terrestrial, satellite, cable and even IPTV following the 2017 publication by ETSI of TS 103 555. This defines how HbbTV terminals, such as set tops and smart TVs, can be used with IPTV services. By contrast, ATSC 3.0 is purely for digital terrestrial and so the focus is entirely on integrating that as harmoniously as possible with broadband services, which is best accomplished by incorporating that work on DASH and MMT.

HbbTV has borrowed from some of that work as witnessed in its recent announcement of phase 2 of application discovery over broadband (ADB). But first it is worth backtracking a little. HbbTV predates ATSC 3.0 by about 10 years, having emerged in February 2009 from the French H4TV project and the German HTML profil project. The initial focus was on interactive services and then as a way for broadcasters to incorporate OTT and benefit from the same personalization and also audience measurement possibilities which had been denied broadcasters, as opposed to cable TV and IPTV operators with their return paths. Most recently the greatest interest has been generated by the potential for broadcasters to deploy addressable advertising directly without need for cooperation with pay TV operators delivering their channels, with greater scope for aggregation.

But HbbTV had become bogged down by its own legacy, with the HbbTV 1.5 standard, which is still widely deployed, requiring the old HTML4 family of standards. The HbbTV Association updated the standard with the 2.0 version introduced in 2015 with HTML5 specifications, allowing much more dynamic UIs. This also allowed much greater commonality and reusability of code between the web and TV screen.

HbbTV was also suffering from a chronic deficiency, lack of a robust mechanism for ensuring that apps associated with a broadcast channel are discoverable in time for execution within a hybrid service, whether to insert an alternative ad, or overlay some personalized graphics such as subtitles. The point is that the service must be driven by the broadcast channels and so when a terminal is connected to both broadcast and broadband signals it must be capable of discovering the relevant app. The necessary information is contained in the HbbTV Application Information Table (AIT), but it was not guaranteed this would always reach the terminal in time, for two principal reasons. Firstly, while the TV or other device would understand HbbTV, it might be connected over HDMI to a legacy set top that does not. Alternatively, or in addition, the service provider carrying the broadcast DVB signal might omit the AIT. In either case HbbTV would not work.

The HbbTV Association therefore developed the application discovery over broadband (ADB) technology, initially as a separate component, for all kinds of set tops including IPTV and OTT as well as cable, terrestrial and satellite. This enables consumers to access the HbbTV services from their TV sets even when that is not supported by the set top, but for this to be possible it has to be able to recognize the channel quickly. Then when a compliant TV set receives the signal of a particular TV channel from the set top, it can discover via its broadband internet connection if there is an HbbTV app related to that channel. This enables it to run that app and if subsequently the user changes channel, that will be detected and may stop the HbbTV app if it is no longer required or relevant.

But for that to happen, ACR (Automatic Content Recognition) must be employed and that is where the latest phase 2 of AIT comes in by supporting audio and video watermarking. However, video watermark extraction requires native support in chipsets, while audio watermarking does not work if the sound is muted on the set top so would prove unreliable.

The HbbTV Association therefore realized that whatever approach it chose would have to be standardized and enjoy wide support. This factor led it to the ATSC 3.0 work, which was already well advanced in standardizing watermarking for ACR under ATSC 3.0. As a result, HbbTV and ATSC 3.0 are totally aligned on watermarking, which makes it easier and cheaper for makers of set tops and TV sets to comply with both.

In both cases, watermarking will play a key role enabling the relevant revenue generating applications, including addressable advertising. It can identify the content, both ads and programmes, the channel, and also the distribution network or source. It can also distinguish between different versions of the same content, for example SD versus HD, as well as between live, on-demand and catch-up TV viewing.

The potential for addressable advertising has helped revive HbbTV in some cases, as at French pubcaster France Télévisions, which in 2017 had closed down its HbbTV app France TV Zoom aimed at 18 to 34-year olds only 8 months after launch, amid a reevaluation of its digital strategy. The broadcaster at that time decided to focus on social media, video platforms and the set tops of the operators distributing its programming. Then just a few months later in January 2018, France Télévisions’ advertising subsidiary FranceTV Publicité, in conjunction with transmitter company TDF, started trialing targeted advertising in the Le Mans metropolitan region. This was extended a year later to wider testing with advertisers Sofinco and St Hubert of substituting commercials with targeted spots for DTT households with HbbTV-compliant set tops or smart TVs.

Across the border in Luxembourg, RTL Group has never at least publicly wavered in its enthusiasm for HbbTV with deployments and trials in several European countries including Germany. The group has focused on HbbTV’s role in scaling up its data capabilities as a prelude for a major push into addressable advertising now taking place. It has first-party data available from logins to its streaming services, while HbbTV TV sets provide consumption data. It sees HbbTV as a vehicle for trialing and then deploying addressable advertising targeting groups or individuals via its ability to insert IP based graphics or video into broadcast signals within a set top, smart TV or other connected device.

The appeal then of HbbTV for broadcasters, as well as ATSC 3.0, in this context, is that it frees them from reliance on pay TV distributors and allows them to aggregate their whole audience to create more appealing addressable segments for advertisers. Of course, content owners might have a different perspective again and want to be free of broadcasters as well and for them audio fingerprinting might appeal because it is agnostic to the delivery technology, but that is another story.