Migrating to high definition services

While the past decade has seen high definition (HD) services taking root across the delivery chain from production to consumer displays, this decade will see the acceleration of growth in HD in a bid to future proof content and to prepare for delivery with stereoscopic 3D.

No doubt consumers have been anticipating HD for some sometime now, many have already purchased HD ready flat screens. Key technical issues concerning bandwidth, distribution picture quality, method of transmission and content availability as part of a live or catch up service are challenging both the broadcasters, service providers and their accountants.

As it stands, competition is heating up between cable, satellite, telco, terrestrial and even ‘over-the-top’ Internet based video services. HD is no longer a differentiator but rather an expected base service offering.

To many broadcasters decisions have been already made in the production environment and the push to deliver HD content to the viewer is the major hurdle. HD is bandwidth hungry. The combination of channel expansion and high definition far outstrip any capacity created by gains at the RF level. In MPEG-2, high-definition requires roughly seven times the bandwidth of a modern standard definition encoder.

Fortunately, the introduction of AVC coding technology helps solve some of this problem by requiring only 4x the capacity of a typical SD MPEG-2 service.

The single greatest advancement for improving or maintaining a high level of video quality across a number of channels in a multiplex is variable bit rate processing or statistical multiplexing. Unlike Telco IPTV and Internet Streaming architectures where services are delivered individually, traditional broadcast networks (cable, satellite and terrestrial) leverage a larger transport containing many services.

In a statistical multiplex, encoders perform a ‘look ahead’ process to generate a quality value or metric based on a characterisation of the video input. The look-ahead metrics are influenced by the input video quality and any pre-processor settings. With the intent to free up bandwidth and all other channels being roughly equal in complexity, the ideal settings will minimise the effect of any one channel skewing the pool average.

In some cases the filter tool is adaptive and so setting a nominal value across all encoders is an acceptable practice and reduces the possibility one or more channels gouges bandwidth from the adaptively.

The use of statistical multiplexing has become a commonplace practice in the delivery of standard definition services, but in many instances high definition services are only transmitted as constant bit rate (CBR) streams. The use of CBR appears to result from the limited number of HD channels being broadcast by the operator. In cases where statistical multiplexing is used for HD services the services are typically segmented off into their own transports, separate from the MPEG-2 standard definition services.

This segmentation by format and codec is useful, especially if different transponders have divergent modulation schemes, but it does limit the operator’s ability to leverage all the possible Statistical Multiplex options.

The complexity of effectively sharing bandwidth between mixed formats and codecs is no small task, but the benefits can be noticeable. Digital terrestrial broadcasts in particular are well suited for mixed Statistical Multiplex pools given the fact that there is often only on transport available.

Mezzanine audio formats like Dolby E attempt to balance the quality versus compression equation in the distribution channel between programmers and service providers, offering an intermediate format that enables a multitude of format options for the final distribution.

With an incoming feed of Dolby E, a service provider is relatively unconstrained in what can be provided to the end user. Both 160Kbps stereo and 384kbps 7.1 multi-channels services can be derived from the Dolby E source.

For the most part so-called mezzanine options are not available today for video services. Especially in secondary markets there is no notion where the distribution feed has already been through two or three satellite links and decode re-encode steps, often with bit rate reductions.

Such a tier would require a 2-3x increase in satellite distribution capacity and would be cost prohibitive in many cases. As a result, programmers are constrained by the cost implications of the distribution network, which in turn drives codec and bit rate requirements. Programmers understand this reality and go to great lengths to find the ideal compromise between video quality, bandwidth and the constraints of the downstream networks.

As operators look to expand or launch high definition services in an ever more competitive PayTV environment it is critical to use all the tools available to balance the need to add as many services as possible and maintain expected levels of video quality. The most significant improvements to video quality can be achieved by acquiring content in the best quality possible. While this is the most difficult area to influence and often out of the direct control of the operator, it is critical that the focus be given to codec selection and possible resolution reduction.