There is no doubt that today we are facing a new and exciting way to consume TV content as interest in 3DTV is growing. This technology is currently spreading within the digital production and post-production arenas, and broadcasters are becoming concerned about how to distribute future 3DTV content to consumers through the various Digital TV platforms such as satellite, cable, IP networks and terrestrial.

In a context of where there is a massive push for new digital channels and new digital content, the interest in 3D content and its delivery to consumers is growing exponentially. In parallel, the second generation (2G) of the terrestrial Digital TV standard, known as DVB-T2¹, is being rolled-out in the UK, and this is a great opportunity for broadcasters who want more transmission capacity for more value-added services based on HDTV content.

The objective of this article is to provide some key elements that explain how 3DTV can be conveyed over a DVB-T2 network. After a brief introduction to basic principles and the projected market trends, the service requirement for 3DTV to the home is identified. A generic illustration of a 3DTV transmission network is introduced and a possible implementation using the DVB-T2 standard is explained.

TEAMCAST is capitalizing on more than 15 years of experience in implementing terrestrial Digital TV solutions. Through this article, TEAMCAST is pleased to share its own market and technology vision, built on more than two years of successful and pioneering experience in the domain of DVB-T2.

Coming Fast — 3D Television
During the last couple of years, stereoscopic TV technology (so-called “3DTV”) has developed significantly within cinema and digital broadcasting ecosystems.

3D content is not new in the cinema, as the very first 3D films have been available since the 1920’s (see the illustration, The Creature from the Black Lagoon, from the 1950’s). However, digital technology has recently pushed 3D to the forefront with more than twenty 3D films released, or planned for release, by the close of this year. In 2010, the success of the 3D movie Avatar (the film grossed US$430 million; 78 percent from the 3D screen²) demonstrated viewers’ growing interests in 3D content and also the potential for significant revenues from this technology.

Consequently, many players in the Digital Broadcast ecosystem are becoming quite interested in the development of 3DTV content, the potential transmission solutions for 3D, and of course the availability of consumer 3D-ready products (3D-ready TVs, LCD monitors, Set-Top-Boxes (STBs), Mobile handsets, Notebook PCs, and so on).

In the situation where Pay TV operators are monitvoring the acceptance of 3D services closely to decide whether consumers are going to be satisfied with the 3D content experience and what kind of 3D technology would bring the best results in term of QoE (Quality Of Experience), some early adopters such as BSkyB³ in the UK and DIRECTV⁴ in the US are preparing 3D services’ delivery over satellite. Pay TV satellite broadcaster Canal+⁵ has announced its intent to launch a 3D channel before Christmas of this year and many others broadcasters, such as MediaPro in Spain for its football channel Gol TV, are investigating in 3D technology.

As usual, the question of the availability of consumer receiving and decoding products arises. Everything is moving fast and major display manufacturers such as Samsung, Sony, Philips, Panasonic, and others, are competing to be ready first for 3D TV mass production!

Recently, Samsung released a 3DTV set with prices starting at $2,500 — the set comes with two pairs of proprietary, active shutter style, 3D glasses, while Sony’s first 3D Bravia products have already arrived in Australia.

The market for 3D displays — which is encouraged by the recent success of 3D cinema — is forecast to take off in 2010, providing a welcome boost to the electronics industry. As seen in Figure 1, DisplaySearch forecasts 3D-ready TVs will grow from 0.2 million units in 2009 to 64 million units in 2018. It is predicted that 3D-ready TVs will be the largest consumer product in terms of revenue in 2018, with $17 billion of sales worldwide.

Seeing on the one hand the fast emergence of 3DTV content at the production level, and on the other hand, the increasing number of 3D displays on the consumer side, the need by broadcasters for efficient 3DTV transmission solutions will intensify during the next few years.

3D Service Broadcast Requirements
Following early experiments and trials for 3DTV, broadcasters are still wondering, precisely, how much capacity a 3D service would require, and how 2D and 3D simulcast service could be broadcast in the most efficient way, as spectrum is a scarce resource.

Today, there is no clear specification of 3D service technical requirements, as the 3D service transmission techniques are multiple and still under evaluation.

In July 2009, ZetaCast, on behalf of the UK regulator Ofcom, published an independent report⁶ Beyond HDTV: Implications for Digital Delivery. In this document, ZetaCast illustrated what would be the required bit-rate for 3D channels relative to 2D, and for the different 3D techniques. ZetaCast assumed that it would be required to deliver full 1080p/50Hz HDTV resolution video to each eye with different options. Broadcasting synchronized left and right eye views is straightforward, but is rather wasteful with bit rate (200 percent relative to 2D). Other techniques provide the opportunity to decrease the video bit-rate, but they bring some additional technical constraints on the signal processing at the decoding, which consequently might create some additional costs at the consumer site.

From the data provided by ZetaCast, we can assume the ratio between 3D and 2D required bit-rate is in the range of 160 percent, for a given resolution and considering a 2D plus Depth or Difference method for transmission.

In order to estimate what would be the actual bit rate for 3DTV service, the next step is to characterize the data bandwidth and bit rates currently in use for HDTV (2D) services. The following table illustrates the different data rates, the video resolutions and the transmission standards currently in use for different HDTV channels over Terrestrial Networks (in France and in UK). Assuming the 160 percent ratio, it is then possible to derive what would be the equivalent service data rate required for 3D.

In France, the HDTV services are distributed onto the DVB-T network, and a complete 8 MHz multiplexed channel referenced as R5 is used to carry three channels: TF1HD, France2HD and M6HD, using a total bandwidth of 24.88 Mbps (FFT 8K, 64QAM, GI 1/8, FEC &frac3/4;).

These three HDTV services are statically multiplexed to reach the best transmission efficiency and each service data rate is variable from 4 to 14 Mbps, providing an average total bit rate around 8 Mbps.

Regarding the picture resolution, the 1440x1080 anamorphic configuration is in use today, rather than full HD resolution (1920x1080). This is mainly for three reasons: first of all, it reduces the required transmission data rate compared to the full resolution (-30 percent). Secondly, it is the same picture resolution as content supplied from popular consumer sources such as HDV/HDCAM. Thirdly, a good percentage of HD acquisition is done at this resolution.

In the UK, recently, the Freeview platform started to be upgraded to DVB-T2 to distribute HDTV content. Today, the Freeview HD multiplex carries three services, although the channel data rate will, in fact, allow an additional service in the future. The BBC HD total data rate per service is currently around 13 Mbps on average, but this might be reduced to 9 Mbps to enable four services to be multiplexed together.

The technical characterization of current HDTV services derives the data rate necessary for a future 3DTV service. Applying a 160 percent 2D to 3D ratio, a base figure of 13 Mbps is required to broadcast a 3DTV service (assuming a 1440x1080 resolution and the 2D Plus Difference/Depth format).

Frame Compatible Format
Before the final adoption of any 3DTV optimized encoding techniques, such as 2D Plus Difference/Depth, or a new format likely based on MPEG-4 Multiview Video Coding (i.e., MPEG-4 AVC with the MVC extension⁹), broadcasters plan to start broadcasting in 3DTV using what is called a Frame Compatible format.

Indeed, during the first phase of 3DTV service broadcast, the objective is to reuse as much of the existing HDTV delivery system for early 3DTV content delivery, and if possible, the full backward compatibility of the transmitted 3DTV service with 2D displays is also targeted.

Several ‘Frame Compatible’ formats have been defined such as Side-By-Side, Top/Bottom, and so on, and these are now supported by the latest HDMI 1.4 release¹⁰.

The basis of the Frame Compatible format is to halve the resolution of the left and right images in order for the two signals to be squeezed together into one combined frame — the whole thing looks like a 2D HDTV signal to a normal STB. The two pictures are then unravelled into the left and right pictures in the TV set and displayed separately on the screen according to the 3D viewing system being used (anaglyph, polarization, shutter, etc.). This enables broadcasters to operate using exactly the same transmission system as currently in use for standard HDTV content broadcast with H.264 video encoders.

Taking the example of the SKY satellite 3DTV channel, the video encoder delivers the total video service at around 14 Mbps, with a combined pixel resolution of 1920x1080 per frame. The actual horizontal resolution for each eye is half of this — the viewer experiences an equivalent resolution of just 960 pixels by 1080 lines for 3DTV content. Note this 14 Mbps bit rate is not far from the previous projection of 13 Mbps.

Generic 3D Transmission Architecture
A generic 3D broadcast transmission architecture is illustrated in Figure 4. 3D content can be generated from either a live signal source or from content servers. In the case of live capture, a typical 3D camera provides a composite stereoscopic 3D signal, for left and right eye viewing. At the service provision site, real-time encoders generate the compressed signals to be conveyed over the broadcast network. The compressed signal includes one video service and one, or several, audio components.

Digital broadcast architectures can be separated typically into two groups: the IP based architectures and the Transport Stream based ones. For IP based delivery scenarios, IP packetized services are conveyed from the different service encoders to consumers via IP networks. In the case of Transport Stream architecture, a service multiplexer gathers all the difference-encoded service stream into a global Multi-Program Transport Stream, the so called MPTS. This MPTS is generally statistically multiplexed to optimize the total useful bandwidth and is then sent over a digital broadcast platform such as Satellite DTH (Direct To Home), Cable, or Digital terrestrial network.

At the consumer site, the technical architecture would vary according to the different 3D technology being used. For instance, display manufacturers have already demonstrated 3D displays that work with either polarized glasses, shuttered glasses, or no glasses at all (in the case of auto-stereoscopic — multiple viewing position or look around) solutions. In the case of Pay-TV service, the digital terminal located at the consumer site (STB or PVR) is likely to operate transparently to 3D content. Pay-TV operators who want to develop 3D service are investigating technical scenarios that would limit the impact on the consumer terminal, so existing network use would be straightforward.

Emergent DVB-T2 Market
Since its introduction in 2008, the 2nd generation of the terrestrial Digital TV standard DVB-T2 has been embraced by broadcasters who wish to launch a new generation of terrestrial Digital TV services. From the technical point of view, compared to its predecessor, DVB-T2 brings several key advantages, as illustrated in Figure 5.

The key advantage of DVB-T2 is the capability to transmit a higher data rate, which is especially important for transmitting HDTV and 3DTV services.

Today, several countries have expressed their interest in DVB-T2. Early adopters, such as BBC and Arqiva¹¹ in the UK, have decided to deploy a DVB-T2 service and network infrastructure — half of the population in the UK was able to access terrestrial HDTV service during the FIFA World Cup!

In Finland, two licences for HD service multiplexes have been awarded to DNA. The Mobile Phone Operator is currently planning to roll-out a DVB-T2 network that will cover 60 percent of the population by 2011.

Other countries are also interested in DVB-T2 technology. A massive deployment of DVB-T2 networks is likely to occur after the ASO (Analog Switch Off) plans in Europe are implemented over the next few years.

There are several key drivers for DVB-T2 technology development, as illustrated in Figure 6. Some drivers are related to the terrestrial TV service penetration. For example, countries such as the UK, Spain, France, and Italy with a significant number of terrestrial TV households are more likely to be interested in the transition to DVB-T2. The future success of DVB-T2 can, therefore, be related to today’s increasing consumer demand for HDTV content and the future need for 3DTV. Indeed, as has been shown, such content will require more channel data capacity and the transition from DVB-T to DVB-T2 will be essential.

The question of terrestrial DTV competitiveness is also crucial. In the context of active competition between different network platforms (satellite, cable, IPTV, terrestrial), the race for more capacity is endless and this is pushing the 2G of transmission technologies.

For satellite and cable platforms, DVB-S2 and DVB-C2 are being rolled-out to permit significantly more data rate compared to their predecessors, DVB-S and DVB-C. DVB-T2 can, therefore, be foreseen as a logical and certain evolution to ensure the terrestrial broadcasting platforms remain in the most competitive position when compared to the alternative networks.

DVB-T2 Transmission Capacity
In line with DVB’s aim to provide a coherent and compatible family of standards, DVB-T2 uses OFDM (orthogonal frequency division multiplex) modulation to deliver a robust signal and to offer a range of different modes, making it highly flexible. DVB-T2 employs the same LDPC (Low Density Parity Check) error correcting codes as used in DVB-S2 for excellent performance in the presence of high noise levels and interference. A significant number of highly innovative features such as Physical Layer Pipes (PLP), support of Multiple-Input-Single-Output (MISO) and Rotated Constellations are also included to improve the reliability and efficiency of the system.

Rotated Constellation
In order to allow greater capacity in a given channel bandwidth, DVB-T2 implement signal constellations of up to 256-QAM per carrier. Additionally, guard interval ratios down to 1/128 reinforce the opportunity to maximize the useful bandwidth especially in good receiving conditions (fixed service with roof top antenna).


There are a large number of possible combinations between the different modulation parameters.

This typical configuration is currently used in different countries where DVB-T2 is under trial and is referred to as a SISO (Single Input Single Output) implementation. However, in DVB-T2, other network topologies can be introduced, such as MISO (Multiple Input Single Output) and MIMO (Multiple Input Multiple Output). In one example of MIMO, a pair of antennas is used at the transmitting and receiving ends — one of each pair using vertical polarization; the other using horizontal. The effect of this can be to double the transmitted data capacity, in theory.

Although MIMO technology might be highly attractive for broadcasters who wish to increase data capacity, this approach falls foul of one of the commercial requirements of T2, which stipulates the T2 broadcast will be compatible with existing DVB-T (and, by implication, analog TV) domestic installations, i.e., single antenna, single down lead.

3DTV Over DVB-T2
DVB-T2 would be definitively the preferred technology to carry 3DTV services over terrestrial networks, compared to its predecessor DVB-T, as the data capacity is about 40-50 percent greater. The exact number of 3DTV services that could be conveyed over DVB-T2 cannot yet be defined, as these are the early days of 3D and DVB-T2 technologies and many parameters are not yet completely defined and verified. Nevertheless, there are some assumptions that will lead to a good estimate for the number of 3D channels in a DVB-T2 multiplex, as seen in Table 9.

We end up with three 3DTV services within a DVB-T2 multiplex inside a terrestrial 8 MHz bandwidth channel. Just to illustrate, today in France the multiplex referenced as R5 is dedicated to carry three HDTV services via DVB-T (TF1 HD, M6 HD and FRANCE 2 HD). In this context, the conclusion is that the number of 3DTV services would be equivalent to the number of HDTV services, considering the technology transition from DVB-T to DVB-T2.

It can be expected that some of the key assumptions listed here will change as DVB-T2 experience develops, such as: the video coding rate, the statistical multiplexing efficiency, and so on. The remaining key question revolves around the picture resolution necessary to achieve the viewer-expected quality for a good 3D experience: is this adequate to ensure the success of 3DTV in the consumer marketplace?

Regarding the transmission network architecture, DVB-T2 networks would be quite transparent to the 3DTV service content. As shown in Figure 10, the service multiplexer is interfaced to the T2 Gateway, which is used to generate the T2-MI output stream and controls the correct stream timing and time stamping to operate in Single Frequency Network (SFN). The T2-MI can be send through an IP based or a MPEG-TS/ASI based network, to feed the different transmitter sites.

Each of the transmitter sites pictured in Figure 10 consists of a T2 modulator that then feeds the UHF or VHF transmitter power amplifiers. The signal is radiated into the air via the mast and antenna, to be received by consumers in their homes or even on the move.

Will Dreams Become Reality?
Although 3DTV technology is still undergoing trials, and many technical parameters are subject to change, it is possible to predict what would be the required service bit rate for a 3DTV service, taking into account several assumptions. In this article, the data rate has been estimated to be around 13 Mbps, which leads to the conclusion that the number of 3DTV services that can be broadcast within a single multiplex in a DVB-T2 terrestrial network is three.

There is also the advantage that the transition from 2D to 3D content will remain transparent in terms of channel bandwidth usage when the transmitter technology is upgraded from DVB-T to DVB-T2.

Regarding the transmission architecture, the transition to 3D content involves mainly new service encoders and compliant 2D/3D displays. The transmission chain through the multiplexer, the distribution network and the terrestrial transmission will operate transparently for either 2D or 3D content. DVB-T or DVB-T2 networks, already in use or under deployment, can be easily used to carry 3D content to consumers. The dream of 3DTV will, indeed, become a reality.

Footnotes
¹ DVB is a registered trademark of the DVB Project
² Source: CRN
³ Source: SKY
⁴ Source: DIRECTV
⁵ Source: PCInpact
⁶ Source: ZetaCast Report
⁷ Source: digitalbitrate.com, reghardware.co.uk
⁸ Assumptions: 2D Plus Difference. Most probable 3D/2D bitrate ratio at 160 percent
⁹ ITU-T Rec. H.264 & ISO/IEC 14996-10 Advanced Video Coding (AVC) standard on Multiview Video Coding
¹⁰ See HDMI 1.4: Specification
¹¹ Source: DigitalSpy

About the author
E. Pinson is the Business Unit Manager at TEAMCAST

About the company
TeamCast is a highly active member of the TV Broadcasting ecosystem worldwide, with innovative technology offerings based on a solid expertise in both Digital TV and Mobile TV transmission. TeamCast is a company with strong technical credentials who supplies product technologies to infrastructure equipment providers for Digital television broadcast networks. TeamCast’s strategy consists in developing Terrestrial DTV and Mobile TV solutions as soon as new standards and new technologies are released.

In the specific framework of DVB-T2, TEAMCAST has been involved right from the beginning as an active member of DVB. Its active participation within several projects such as Celtic Project B21C, SME42 project, and Celtic Project Engines has been a key opportunity to build a concrete experience concerning this emerging technology. Furthermore, TEAMCAST has developed a close relationship with early DVB-T2 adopters such as Arqiva in UK, Teracom in Sweden and DNA in Finland, having delivered DVB-T2 professional products and solutions.

More information is available at
www.teamcast.com