In the early days of television, telecommunications regulators had abundance of RF spectrum to play with. Such freedom, coupled with limited interference-prediction tools, led to a conservative spectrum planning approach, which in turn yielded UHF and VHF broadcast bands with a large number of unutilized channels due to guard band requirements.

It is increasingly important to start with the end in mind. Today's broadcast industry is undergoing a major transition'from an environment that has been 100 percent analog, to a new era that comprises several different "flavors" of digital. Regulators have the opportunity to develop a roadmap for optimizing spectrum efficiency and interference control; to look at spectrum as a whole, rather than as bands of free frequencies between occupied channels. But while trying to predict and plan for the future, spectrum planners must accommodate the present, including likely simulcast periods. The key will be to select the path of least inconvenience.

Spectrum restacking: Now or later?

Most countries are assigning spectrum for DTT services in the same bands as analog to make the best use of existing infrastructure. This sees the guard bands between analog television channels first in line for appropriation.

But if the end-goal is to rationalize and optimize the use of spectrum, regulators and spectrum planners need to consider what will be the outcome of liberated spectrum, once analog services are switched off. Since digital broadcasting is spectrally very efficient, portions of some bands may be reassigned or sold for other wireless communication purposes. This will inevitably lead to some form of spectrum restacking. The question is whether to rearrange the spectrum before or after deploying digital services.

Many countries have already inserted DTT services into analog guard bands prior to restacking, resulting in adjacent analog and digital channels (Figure 1'Option A). This scenario is only really viable for shared transmission networks, where adjacent-channel services are broadcast using a common antenna system and interference can be more easily controlled.

In environments where transmission networks are not shared, the presence of multiple high-power sites leads to a higher probability of interference between adjacent (n±1) and even semi-adjacent (n±2) channels. It is more difficult to maintain the required protection ratios'the relative signal strength between "wanted" and "unwanted" services'to achieve an acceptable level of interference.

Even in shared transmission networks, the use of adjacent channels in one coverage region often means the same channels are used in neighboring regions, leading to the potential for co-channel interference. Care must be taken to ensure the specified protection ratios are met.

The approach of deploying digital services prior to spectrum restacking has usually been taken to minimize disruption to existing audiences of analog television, and allows a simulcast period. DTT services may be compromised in the short term as radiated power levels are restrained and the number of channels are potentially limited. Moreover, once analog television is switched off, the issue of spectrum restacking re-emerges, in order to avoid a "Swiss-cheese" effect of unutilized channels.

It can be far more cost-effective and practical to rearrange the analog services prior to deploying digital services (Figure 1'Option B). In either scenario, viewing is potentially disrupted, but restacking first is more likely to be the path of least cost. Regulators need to evaluate early what spectrum needs to be retained for digital services, and what spectrum will become available for re-assignment or sale for other applications. Once they have the vision of where they want to go, they can develop the roadmap of how to get there.

What's in the mix?

The anticipated mix of digital services is a critical issue for spectrum planning'particularly when the environment is to encompass both fixed-reception and mobile TV networks. This is because of the mix of high, low, and medium-power transmission sites that will inevitably be required. Interference is made more likely by the resulting differences in signal levels between the analog and/or digital fixed-reception and mobile TV networks (with the relative field strength of the mobile network generally much higher).

The situation is particularly bad if adjacent, or even semi-adjacent, channels are involved, especially if analog TV is in the mix (Figure 2). An ideal scenario would be to allocate all spectrum for mobile TV services at one end of the band, with fixed digital services at the other. Although this will not have any benefit if remains scattered throughout the whole band, it is a practical solution where spectrum is restacked prior to digital service deployment. It also highlights how important it is to pre-empt spectrum usage before planning is carried out.

An additional complicating factor is the potential introduction of mobile data devices that transmit, as well as receive, in traditional broadcast spectrum. These so-called "white space" devices could become prevalent in countries that allocate 700MHz spectrum for fourth-generation (4G) services such as "worldwide interoperability for microwave access" (WiMAX). Not only will such networks take on a distributed topology as for mobile TV networks, but the transmitters in the mobile/portable devices will almost certainly be powerful enough to cause interference with home DTT reception.

Laying the groundwork

Spectrum planners need to have a thorough understanding of the existing analog services before embarking on a digital spectrum plan. However, the records of decades-old networks are not always comprehensive. Moreover, it is not only channel allocations that are required, but also site information such as antenna coverage patterns, effective radiated power (ERP) and height, plus geographical data such as land use, elevation, and demography. Such information is essential to understand how far the RF signal will propagate and determine the most viable coverage areas.

It is also important to establish protection ratios for all digital and analog services, including those of international neighbors. This is the minimum requirement for cross-border relations: coordination and cooperation between neighboring countries will facilitate spectrum planning overall. However, with many countries working to different digital deployment timelines, cross-border issues can be a challenge, especially where co-channel operation is being considered.

The broad digital coverage objectives'both area and data payload'must also be considered. The payload in particular has a significant impact on the required signal strength, which in turn influences the ERP. While higher data rates for either multi-channeling or HDTV may promote DTT take-up, the use of higher order modulations translates to a greater susceptibility to interference.

Another key issue affecting DTT coverage is the "cliff effect," so-called owing to the abrupt deterioration in signal quality over a short distance (equating to a reduction in signal level). Below certain signal strength the digital picture becomes unwatchable, in contrast with the analog signal which degrades gradually. As a result, the digital signal field strength needs to be higher for a greater distance away from the transmission site in order to match analog coverage. The situation is even more challenging for mobile TV reception, where the signal level needs to be 30 or 40dB higher than fixed-reception signals at the "cliff edge." Both scenarios need to be taken into account from a frequency planning point of view.

Single-frequency networks

Single-frequency networks (SFNs) allow the reuse of spectrum across a broad area, and as such are an invaluable spectrum planning tool. SFNs can involve many low-power sites within a coverage region (as in a mobile TV network), or they can span multiple neighboring coverage regions.

The most fundamental limitation of an SFN is one of content. All transmitters in an SFN must broadcast exactly the same content, disallowing any regional variation. So, while a single SFN could technically span thousands of kilometers across an entire country, a desire for regional programming tends to promote a more conservative application.

It is accepted that SFNs will be essential for mobile TV networks'from both a spectrum availability and signal availability point-of-view. They are also immensely practical for fixed reception networks. The distance between transmission sites is determined by the modulation parameters used, the characteristics of the transmission network, and population distribution.

It is also important to recognize that the ideal transmission site locations for an SFN may differ from those used in multi-frequency networks (MFNs). Where it is not possible to establish the "ideal" network'such that all receivers operate within the required signal level differentials and/or within the guard interval'SFNs can become subject to "intra-SFN" interference, which can limit the overall SFN area. Moreover, the potential introduction of adjacent or co-channels may determine the size-limit of the SFN that is laid over an MFN.

The big picture

Ultimately, the regulator needs to look at the "big picture" on behalf of all spectrum users, as well as adhere to government policies regarding the number of channels and coverage requirements. This includes establishing the framework for how much interference is acceptable, and ensuring that the spectrum allocations are workable. Once the spectrum plan is defined, the service planners take over where broadcast network operators determine how to cover a particular area within the defined boundaries.

Here, the manner of specifying the spectrum plan is significant. There are two basic approaches. One more traditional method is to set a nominal site for a coverage region, then define transmit powers in specific directions to achieve the required levels of coverage and ensure protection to neighboring regions. However, this is rather prescriptive and borders on service planning. Alternatively, the regulator can define the percentage of an area to be covered, along with limitations on signal level required to maintain the required protection ratios and minimize over-spill. This allows the broadcaster to determine what they can install to achieve the required coverage within the required field strength limits.

Needless to say, applying the theory in practice demands a great deal of experience and understanding, with a vast number of parameters that need to be taken into account. Here, sophisticated software tools can make a wealth of difference, although should not be employed without experienced planners to drive them and interpret the outcomes.

Spectrum planning is an exercise that needs to be carried out systematically and carefully to ensure the expanding mix of next-generation wireless services can flourish and be commercially viable, while at the same time ensuring the needs of the terrestrial broadcasters are appropriately preserved.