An Economic Analysis of DAB, DAB+, DVB-H & DRM+

This page looks at the distribution costs per station when using the DAB, DAB+, DVB-H and DRM+ digital broadcasting systems.

The analysis uses the reasonable assumption that a digital radio system that requires lower transmitter powers than another digital radio system will be cheaper to transmit. The figures presented allow you to see the trend of distribution costs for the different systems as the number of stations transmitting varies — for example, a full multiplex will have lower average distribution costs per station than a half-empty multiplex.

The main reason for writing this update page is to add the DAB+ system, which is the upgraded version of DAB, which wasn't included in the analysis of the original page.

The previous page was written in response to a highly inaccurate and, in my opinion, biased report that had been published in the EBU Technical Review. A significant part of the page consisted of discussing what the incorrect assumptions were in the EBU Tech Review article; how they should be corrected; and showing the results after they had been corrected. There is no point in repeating any of that discussion, so it won't be included on this page.

1. Choice of audio bit rate levels

In order to keep the amount of additional work needed to update this page to a minimum I will continue to use the 192 kbps MP2 bit rate used in the EBU Technical Review article. However, it must be said that I don't actually consider this bit rate level to match the audio quality on FM, and I think that digital radio stations should be provided at higher audio quality than 192 kbps MP2 provides — for example, in the figure below, MP2 at 192 kbps (the MP2 curve is labelled 'LII', because the official name of the codec is MPEG Audio Layer II — similarly MP3 is labelled 'LIII' because it is Layer III) was deemed to be "Slightly Annoying", and in the 21^st century we really should be aiming at getting much closer to CD-quality than this.

AAC/AAC+ bit rate

The figure below shows the results from a 64 kbps listening test carried out in 2003 by members of the Hydrogen Audio forums (a forum where many experts on digital audio compression discuss the subject), and AAC+ at 64 kbps (referred to as 'He AAC' in the figure) achieved an average score of 3.68 along with error bars between approximately 3.55 to 3.9:

With blind listening tests, if the error bars overlap then you have to say that there is no statistical difference in audio quality level between two bit rate/codec combinations. And as the error bars for AAC+ at 64 kbps overlap with the error bars for 192 kbps MP2 (error bars between 3.7 to 3.9 on a 1-5 scale) in the first figure, these two bit rate/codec combinations do not provide statistically different levels of audio quality. Or in other words, they provide approximately equivalent levels of audio quality.

Furthermore, AAC and in particular AAC+ encoders have improved markedly in the last couple of years since the above test was performed, so I would now expect 64 kbps AAC+ to outperform 192 kbps MP2.

I will therefore use 64 kbps AAC+ on the systems that use the AAC+ codec and 192 kbps MP2 on DAB.

2. Assumption used in the analysis

The analysis below uses the assumption put forward in the "Broadcasting to Handhelds" EBU Technical Review article that the transmitter network cost (distribution costs) follows an almost linearly proportional relationship with total transmitter power (in this document transmitter powers always refer to ERP (effective radiated power) values).

However, although I think this is a reasonable assumption for medium power levels and moderate changes in transmitter powers, I think this assumption overestimates the increase in distribution costs for large increases in transmitter powers and it overestimates the reduction in distribution costs for large decreaes in transmitter powers. Therefore, if a graph of distribution costs versus transmitter powers were drawn it would be S-shaped rather than being a straight-line. I have tried to account for this effect in the analysis below.

3. Transmission parameters' effect on transmitter powers

In order to analyse the alternative digital radio systems it is necessary to evaluate the difference in transmitter powers that result when certain transmission parameters vary.

Carrier-to-noise ratio is similar to signal-to-noise ratio (SNR), with the only difference being the location at which these parameters are measured: C/N is measured at the input of the RF section of the receiver, whereas SNR is measured at the input of the baseband section of the receiver — i.e. after the RF section.

C/N's effect on transmitter powers is simply that an increase/decrease in C/N requires an equivalent increase/decrease in the transmitter powers to compensate for the change.

The total power of a signal when measured in the frequency domain is equal to the area under the curve of the graph. For example, on a spectrum analyser the x-axis's units are Hz and the y-axis's units are power per Hz, thus the total power equals the integral of the curve, i.e. the area under the curve.

DAB, DVB-H and DRM+ all use the OFDM transmission scheme, which has a spectrum that is flat across the signal bandwidth (when transmitted). Therefore, the bandwidth of a signal is directly-proportional to the transmitted power, so the difference in transmitter power due to bandwidth can be written mathematically as follows:

Power difference between transmitting with bandwidths B₂ and B₁ (dB) = 10 log (B₂ / B₁)

The antenna gain for antennas with a length of up to half the wavelength of the signal is directly proportional to the length of the antenna. Therefore, the power difference due to antenna length for a dipole antenna that is shorter than half a wavelength is given as follows:

Power difference due to antenna length (dB) = 10 log (antenna length / half-wave dipole length)

The relationship between transmission frequency and required transmitter powers can be found from Friis' equation for free space propagation (the power loss resulting from transmitting over line-of-sight paths):

where P_t and P_r are the transmitter and receiver powers, G_t and G_r are the transmitter and receiver aerial gains, d is the transmitter–receiver separation, L is a loss factor, and λ is the wavelength. The relationship between wavelength and transmission frequency is given by: λ = c / f, where c is the speed of light (3 x 10⁸ m/s) and f is the transmission frequency.

Therefore, for a given location the first and third terms on the right-hand-side of the equation remain constant, and the required transmitter power will change as the transmission frequency changes according to the 2nd term only:

Power difference between transmitting at frequencies f₂ and f₁ (dB) = 20 log (f₂ / f₁)

Video reception requires much lower BER (bit error rate — the number of bit errors divided by the total number of bits transmitted) values than audio reception requires, because video is much more sensitive to bit errors than audio is. For mobile TV, a typical required BER is of the order of 10^-8 to 10^-9, whereas for audio reception the required BER is around 10^-4, and this applies to both MP2 and AAC+ audio services that are using the MPEG-4 Audio Version 2 Error Resilience tools, which DAB+ services are likely to use.

For example, the following figure shows estimated BER vs C/N curves (which have the typical 'waterfall' shape, which all BER vs C/N curves have) for DVB-H where the end points of the curves (at the bottom of the graph) use the required C/N figures that are quoted in the DVB-H Implementation Guidelines standard (ETSI TR 102 377).

The C/N figures quoted in the DVB-H Implementation Guidelines standard are for mobile TV reception, so as audio reception can tolerate higher BER levels then the C/N required for audio reception will be lower.

This translates into audio transmissions requiring lower C/N than video transmissions, which reduces the transmitter powers and hence the distribution costs accordingly.

4. DAB+

DAB+ is the name that has been dubbed for the upgrade to the DAB system. The following block diagram shows the differences between DAB and DAB+:

4.1 How it works

The DAB+ system works by first encoding the audio using an AAC+ encoder followed by applying Reed-Solomon (RS) FEC (forward error correction) coding, and then this datastream is sent over the DAB system in a data channel — i.e. from the DAB system's perspective it is simply carrying a data channel.

The DMB and DAB-IP systems used for mobile TV work in an identical way to DAB+, with the only difference in the block diagram above being that both audio and video encoders and decoders are needed instead of just an audio encoder and decoder — both DMB and DAB-IP use RS coding and are transmitted over the core DAB system in a data channel.

4.1.1 Reed-Solomon Error Correction Coding

The RS coding that has been added is a very efficient way of significantly increasing the strength of the error correction coding. The 'code rate' of the RS code is 188/204, which in this instance means that the RS encoder takes 188 bytes then adds 16 more (redundant) bytes, which are used in the decoder to try and correct any errors present. The error correction capability of this RS code is that it can correct any 8 bytes that are received in error irrespective of the number of bit errors in each of the 8 bytes.

RS coding is particularly good at correcting bursts of errors, which is precisely the kind of error patterns that a Viterbi convolutional decoder (which is what is used for the inner FEC decoder) produces when operating with signals transmitted over the mobile channel (i.e. wireless signals that can be received when stationary or on the move).

With the addition of RS coding, the error correction coding used on DAB+ will be effectively identical to that used on DVB-T, which also uses EEP (Equal Error Protection) convolutional coding for its inner layer of coding and the same RS code as DAB+ is adopting (DAB+ is borrowing this technology from the DVB systems) for its outer layer of coding.

To give you an idea of the strength of the RS coding, despite the fact that it only adds 7.8% of redundant data (which is very low in error correction coding terms — virtually 50% of the data transmitted on UK DAB multiplexes is redundant error correction data), the DVB-T specification states that the BER at the output of the Viterbi decoder must be 2x10^-4 in order for the RS decoder to reduce the BER to 10^-11 for the quasi error-free channel for MPEG-2 video. This means that the RS coding corrects 20 million bit errors for every error it fails to correct, on average. Although the RS coding won't be able to improve the BER as much on DAB+ as on DVB-T due to the fact that DVB-T is expected to be received on rooftop aerials (where you have 1 strong signal and weaker multipaths) and DAB+ is expected to be received over a true multipath channel (with numerous multipaths of similar strength), the RS coding will hugely improve the robustness of reception compared to DAB — obviously, sufficient transmitter powers must be used though.

Basically, it is fair to say that if someone can receive DAB in the UK, then even if they suffer from having dreadful reception (i.e. the signal suffers from bubbling mud and/or cuts out intermittently), if RS coding were added to protect that multiplex signal they would receive that stream almost perfectly.

One difference between radio stations using DAB+ and MP2 radio stations on DAB in the diagram above is that MP2 radio stations would use UEP (Unequal Error Protection), whereas it has been assumed in the diagram above that DAB+ will use EEP for its inner FEC coding. This may not turn out to be correct, because the DAB specification allows the use of UEP for audio streams other than MP2, but until the new DAB+ specification has been released it is impossible to know whether they have specified to allow the use of UEP with AAC+ or not — it would be better if they did, but on this page I've assumed that they haven't used it, therefore if they do use it the multiplex capacity will be higher than that shown here.

4.1.2 Higher multiplex capacity

Although I've discussed the advantage of using stronger error correction above in terms of the fact that it offers more robust reception, an intricately linked advantage is that it increases the spectral efficiency, which is the data capacity that can be transmitted at a given level of robustness. For example, the following table shows the multiplex capacities for DAB and DAB+:

System	Error Protection Level	Multiplex capacity kbps	Spectral Efficiency bits/s/Hz
DAB	PL3	1184	0.69
DAB+	PL4A	1512	0.88

This increased spectral efficiency that the RS coding allows combined with the far more efficient AAC+ codec to allow far more radio stations to transmit on a multiplex:

System	Error Protection Level	Bit Rate	Audio codec	Number of stations per multiplex
DAB	PL3	192	MP2	6
DAB+	PL4A	64	AAC+	24

4.1.3 Lower C/N

The DAB-IP mobile TV service that has recently launched on the Digital One UK national commercial DAB multiplex is using the PL4A protection level, and these channels are being transmitted alongisde normal DAB radio stations that are using PL3. So, as the receiver sees a constant C/N figure for the multiplex as a whole, all the services are received with the same C/N figure.

The DAB-IP system uses the same outer layer of RS coding as DAB+ will use, but Digital One must have concluded that DAB-IP mobile TV services can be received as robustly using PL4A as radio stations can be using PL3. Therefore, the reduction in required C/N for audio services relative to mobile TV services (i.e. the less robust video signals) described above will also apply to DAB+ stations relative to DAB-IP mobile TV channels, so the transmitter powers for a DAB+ multiplex could be reduced by approximately 2.5 dB relative to a DAB multiplex.

This assertion is also backed up by the findings given in a white paper published by Radioscape following their DMB trial (DMB's transmission scheme is also identical to DAB+'s), where they said about the RS coding (what they refer to as 'Outer Layer Coding'):

“The primary implication for DMB network design is that the use of OLC [Outer Layer Coding] effectively allows DMB networks, which require a much lower error rate than DAB, to be planned with very similar transmission requirements as DAB audio networks.”

“The results indicate that additional transmission power will not be required and indeed there appears to be 2-6dB of advantage dependent on the conditions.”

4.2 DAB vs DAB+ Distribution costs per station

5. DVB-H

In order to avoid the large reduction in transmitter power due to the change in frequency from UHF to Band III (which due to the S-shaped relationship between distribution costs and transmitter powers would overestimate the reduction in distribution costs), I will concentrate on the differences between the technologies used by the systems in order to find the differences in transmitter power levels, and hence the distribution costs per station.

5.1 Differences between the DAB+ and DVB-H systems

The best way to fairly compare two systems is to compare them on a like-for-like basis, and then account for any differences between them. So, before I discuss what the differences are between the two systems I'll say what is assumed to be the same, because when parameters are the same for both systems they can be ignored in a comparison between systems.

5.1.1 Similarities between the systems

Therefore, all that remains are the differences due to the architectures of the two systems.

5.1.2 DVB-H uses far stronger error correction coding than DAB+

The following block diagrams shows the error correction coding schemes used by DAB, DAB+, DVB-T and DVB-H:

As you can see in the figure, both DAB+ and DVB-H use RS and convolutional coding, which are the green and light blue blocks in the figure above, respectively. The RS coding (green block) is actually identical on both systems, and the convolutional coding is close enough to being identical that it will be treated here as being identical — it's just standard Equal Error Protection (EEP) convolutional coding, which isn't very strong, and there will only be tiny differences in strength between the convolutional coding used on the two systems.

The only difference in the error correction coding schemes used by the two systems is that DVB-H uses a third layer of coding, which is absent from DAB+. This third layer is called the MPE-FEC, which stands for Multi-Protocol Encapsulation Forward Error Correction. The MPE-FEC is also an RS code, but it is far more powerful than the RS code used on DAB+.

One of the characteristics of RS codes is that you know from the structure of the code how many errors can be corrected, and the following table shows the error correcting capabilities of the MPE-FEC and RS codes used on DVB-H and DAB+ (I've used the same colour-coding as in the above diagram):

FEC code	RS codeword length	Error correcting capability	Percentage of bytes correctable	Code rate	Systems used on
MPE-FEC	255 bytes	64 bytes	25%	0.75	DVB-H
RS	204 bytes	8 bytes	4%	0.92	DVB-H & DAB+

That is, the MPE-FEC can correct any 64 bytes out of a 255-byte 'codeword' (think of it as a packet of data) irrespective of the number of bit errors in each errored byte. The weaker RS code that both systems use can correct any 8 bytes out of a 204-byte codeword.

The following two graphs show the error correcting capabilities of these two RS codes when the bytes that are in error are uniformly randomly distributed. This overestimates the performance of the RS codes in comparison to what you would actually get with transmission over the mobile channel where errors tend to occur in bursts rather than being uniformly randomly distributed, although interleaving is used to try and spread error bursts out over time to try and emulate this ideal situation. However, the figures do graphically show the enormous difference in strength between these two RS codes:

Error correcting performance of the MPE-FEC	Error correcting performance of the RS code

The two graphs above show the input/output relationship when the percentage of bytes that are in error is high, because as RS codes can correct a certain number of bytes, when the respective number of errored bytes is exceeded the RS code fails and the high number of errors are passed through to the output, and the user is likely to experience poor reception quality.

Error correction code	Percentage of bytes in error at input	Percentage of bytes in error at output
MPE-FEC	10%	0.0000000001%
RS code	10%	10%
RS code	1%	0.0005%

From the figures in the table, you can see that when 10% of bytes are in error the weaker RS code fails to correct virtually any bytes, because 10% is well above the 4% limit of errored bytes that it can correct. In contrast, the MPE-FEC corrects virtually all of the bytes when 10% of all bytes are in error at the input, because 10% is well below the 25% of bytes that it is capable of correcting. This also means that, when the weaker RS code fails when the BER is relatively high, as it would on DAB+ because there is no outer-layer of error correction to catch the errors, the MPE-FEC on DVB-H can still correct virtually all errored bytes until the percentage of errored bytes is very high.

As I've said, the figures do overestimate the performance of these codes, but hopefully you get the picture about just how strong the MPE-FEC is in comparison to the weaker RS code used on DAB+ (and which is also used on DVB-H).

However, the point of using stronger error correction on a digital communication system is not to provide extraordinarily low BER levels, because audio typically requires a BER of 10^-4or 10^-5. The point of using stronger error correction is to provide higher capacity (higher spectral efficiency) and/or to reduce the transmitter powers. This situation is summed up by the following fundamental three-way trade-off for all wireless digital communication systems:

For example, stronger error correction increases the robustness of the signal, and this allows the use of a less robust transmission mode, such as changing from using QPSK to the higher capacity 16-QAM modulation, or it allows the transmitter powers to be reduced, or it allows a combination of both higher capacity and lower transmitter powers.

To demonstrate the capacity increases and lower transmitter powers (in the form of lower C/N figures) that using stronger error correction allows, the following figure shows the difference in required C/N between DVB-H (pink curve), which uses the MPE-FEC, and DVB-T (red curve — and DVB-T is identical to DVB-H apart from the MPE-FEC in terms of the transmission scheme (modulation and FEC coding)), which doesn't use it.

The horizontal axis is Doppler frequency, which is simply another way of expressing the speed at which the mobile receiver is travelling.

As you can see, the difference between using the MPE-FEC and not using it translates into a reduction in required C/N of approximately 3.5 dB at low speed up to around 9 dB at high speed, so it can be assumed that DVB-H can use 3.5 dB lower C/N than DAB+ due to the use of the MPE-FEC, because DAB+ uses the same error correction coding as DVB-T.

The figure also shows that DVB-H can use the higher capacity 16-QAM modulation at the same 15 dB C/N level as DAB uses with its differential QPSK modulation fo radio stations, so the capacity also increases significantly — in this example the spectral efficiency for DVB-H is 50% higher than for DAB (i.e. DVB-H can transmit 50% more data per MHz than DAB can).

5.1.3 Differential vs synchronous modulation

The other significant difference between DAB+ and DVB-H is that DAB+ uses differential modulation whereas DVB-H uses synchronous modulation, and it is well-known that using differential modulation incurs a 3 dB C/N penalty relative to using synchronous modulation.

Therefore, DVB-H can use 3 dB lower C/N than DAB+ due to the use of synchronous rather than differential modulation.

5.1.4 Multiplex bandwidth

Band III is organised to allow either 4 x 1.71 MHz-wide DAB/DAB+/DMB channels or 1 x 7 MHz-wide DVB-T/H channel, and as mentioned earlier, the spectrum of an OFDM signal is flat when transmitted, and as the units of a spectral graph are in dB/Hz (power per unit bandwidth), the area under the C/N vs frequency curve gives units of power, so the difference in transmitter power due to bandwidth is simply:

The overall difference in transmitter powers between DVB-H and DAB+ is simply the sum of the above values:

In other words, a DVB-H multiplex requires 0.4 dB lower transmitter powers than a DAB+ multiplex. Or in terms of linear units this is:

5.2 DAB vs DAB+ vs DVB-H Distribution costs per station

The following graph shows the distribution costs per station for DAB, DAB+ and DVB-H (Band III):

6. DRM+

DRM+ will be specified to use frequencies up to 120 MHz, and the two possible bands are between around 45 - 70 MHz and the FM band.. Therefore, calculations will be carried out at two transmission frequencies: 60 MHz and 100 MHz.

Band I antennas are very long (a half-wave dipole at 60 MHz would be 2.5m in length), so it will be assumed that FM aerials will be used, and half-wave dipoles are 1.5m in length (the material-dependent velocity factor cancels when calculating relative antenna loss, so will be ignored).

The assumptions I made were that the maximum speed of the receiver was 120 mph, transmission frequency was 120 MHz, DRM Mode A and (typical) FEC code rate of 0.6 were used. As a starting point the figures used the capacity values quoted in the DRM specification for 20 kHz channels with the appropriate transmission mode parameters (modulation order and FEC code rate). Appropriate OFDM subcarrier separation were used to allow for the increased Doppler spread at higher frequencies and receiver speeds, and a 300 μs guard interval was used to allow for very large SFN cell sizes.

The following table shows the estimated capacity values produced by the program:

There are no published required C/N figures for DRM+ yet, but the general consensus is that with 20 dB C/N you're assured robust reception.

However, 64 kbps is a lot lower than the 104 kbps capacity possible using 16-QAM CR 0.6 on a 50 kHz channel, therefore the code rate can be substantially reduced (i.e. error protection increased), which will reduce the required C/N:

A reduction in CR from 0.6 to 0.37 equates to a large increase in protection strength — equivalent to, say, around 5 dB.

However, DRM+'s 50 kHz channels are narrowband channels, so they will be prone to suffering from flat-fading. On the other hand, the weird and wonderful propagation that happens at HF and for international transmissions will require a higher C/N than DRM+ transmissions will at VHF. Lastly, DRM+'s multi-level coding error correction coding scheme in conjunction with its use of UEP (Unequal Error Protection) is relatively strong as single-layer FEC coding schemes go, so overall I'll estimate that the C/N should be 17 dB in order to give a few dB's worth of protection against flat-fading.

Correction factor for non-linear relationship between transmitter power & distribution costs

As mentioned earlier, it is likely that rather than the distribution costs having a linearly proportional relationship with transmitter powers a curve of one against the other would be S-shaped, where very low transmitter powers such as would be used for DRM+ stations would need to have their distribution costs increased as a result. Therefore, I will use a correction factor to revise the distribution costs upwards.

A simple way to estimate the distribution costs for groups of DRM+ stations is to group the same number of stations as are grouped on a DAB multiplex (six stations). Then it is possible to apply an overall factor to the DAB distribution costs per station curve, because this includes the cost savings enabled by a medium to large number of stations sharing resources on a number of multiplexes.

The following table calculates the overall factor to be applied to the DAB distribution costs curve for transmission at 60 MHz:

Parameter	DRM+ Value (value 2)	Relative Value (value 1)	Equation	Difference in power at 60 MHz
Transmission frequency	60 MHz	220 MHz (Band III)	20 log (f₂ / f₁)	-11.3 dB
Antenna loss	2.5m	1.5m (FM)	-10 log (l₂ / l₁)	+2.2 dB
Signal bandwidth	6 stations x 50 kHz	1,710 kHz (DAB)	10 log (B₂ / B₁)	-7.6 dB
C/N	17 dB	15.0 (DAB)	C/N₂ - C/N₁	+2.0 dB
Correction factor				+3 dB
Total (dB)				-11.7 dB
Total (linear factor)				0.068

The following table calculates the overall factor to be applied to the DAB distribution costs curve for transmission at 100 MHz:

Parameter	DRM+ Value (value 2)	Relative Value (value 1)	Equation	Difference in power at 100 MHz
Transmission frequency	100 MHz	220 MHz (Band III)	20 log (f₂ / f₁)	-6.8 dB
Antenna loss	1.5m	1.5m (FM)	-10 log (l₂ / l₁)	+0.0 dB
Signal bandwidth	6 stations x 50 kHz	1,710 kHz (DAB)	10 log (B₂ / B₁)	-7.6 dB
C/N	17 dB	15.0 (DAB)	C/N₂ - C/N₁	+2.0 dB
Correction factor				+3 dB
Total (dB)				-9.4 dB
Total (linear factor)				0.1148

7. Distribution costs per station for all systems

The following graph shows the distribution costs per station for all of the systems:

8. Discussion of the results

8.1 DAB

As expected, DAB is far more expensive to transmit than all of the other systems.

Also, ignoring the additional spectrum required due to multi-frequency network (MFN) planning in order to avoid co-channel interference, DAB consumes far more spectrum than the other systems:

In order to carry the 56 stations used in the analysis, DAB needs to use 10 multiplexes which are each 1.71 MHz wide, and due to multi-frequency network (MFN) planning, in order to avoid co-channel interference a much larger amount of spectrum would be consumed than 10 x 1.71 MHz. Furthermore, this number of multiplexes is already more than the typical number of DAB channels that have been allocated to areas in GE-06 (typically around 6 channels have been allocated to a certain area), and DAB has to share Band III with DMB/DAB-IP mobile TV, which would exacerbate the problem with there being insufficient spectrum.

Overall, due to the very high transmission costs and insufficient spectrum, it is fair to say that it is inevitable that radio stations would be provided at low audio quality if the old DAB system were used.

The DAB system is a product of the 1980s and it should never have been launched in any country. It has never been up to the requirements that a digital radio system must meet in the 21^st century, and the sooner it is swithched off completely the better it will be for everybody concerned.

8.2 DAB+

As expected, DAB+ is much cheaper to transmit than DAB due to the far higher number of stations that can transmit on a multiplex and the lower transmitter powers required. However, DAB+ multiplexes are still limited to carrying 24 x 64 kbps AAC+ stations, so extra multiplexes are needed once the number of stations exceeds 24 and 48 stations, so unlike with DVB-H, the distribution costs do not decrease monotonically (i.e. reduce consistently without increasing again).

However, one advantage that DAB+ has over DVB-H is that it consumes less spectrum, because when there are up to 24 stations only one 1.7 MHz-wide DAB+ multiplex is required, then 3.4 MHz of spectrum is required for up to 48 stations, whereas DVB-H always requires 7 MHz of spectrum. Add to this the fact that extra spectrum is required to avoid co-channel interference and DVB-H's wide bandwidth is problematic when there is a limited amount of spectrum available.

The fact that DAB+ multiplexes are 4-times narrower than DVB-H also makes DAB+ much more flexible than DVB-H in terms of frequency planning.

DAB+ is suited to carrying national, regional or local stations, and the only kind of station that it is unsuited to carrying is small local stations, such as those with a small coverage area. It is also unsuited for use in areas where there aren't enough stations to fill a multiplex. In these circumstances DRM+ is a superior alternative.

8.3 DVB-H

Despite the DVB-H multiplex being 4-times wider than the DAB+ multiplex, DVB-H still manages to use slightly lower transmitter powers than DAB+ due to the stronger error correction coding and use of synchronous as opposed to differential modulation.

In contrast to DAB+, all of the 56 stations in this analysis could be carried on the lowest capacity transmission mode on one DVB-H multiplex, thus the distribution costs per station fall monotonically.

However, as alluded to above, due to the wide channel bandwidths, DVB-H is rather inflexible in terms of frequency planning, and it consumes a lot of bandwidth.

8.4 DRM+

Due to DRM+'s lower transmission frequency and narrower bandwidth per radio station than the other systems, DRM+ has the lowest distribution costs per station.

And as the following diagram shows, DRM+ is far cheaper than the other systems when the number of stations transmitting in an area is low:

DRM+ is the only system which isn't expected to have stations multiplexed together — i.e. each DRM+ station is expected to transmit in its own 50 kHz channel. (The only reason I grouped 6 stations together in the above analysis was to allow comparison with DAB multiplex costs.) So, DRM+ is the only suitable system for individual stations that want to transmit, such as small stations with a small coverage area.

However, because DRM+ channels are so narrow it would be possible to transmit a large number of stations from the same transmitter. For example, a transmitter with a 3 MHz bandwidth could transmit up to 60 DRM+ (non-multiplexed) stations, so the distribution costs per station would be even lower than given above due to resource sharing, because the 1/number of stations relationship would allow the distribution costs to fall monotonically over a larger number of stations as they do for DAB+ and DVB-H multiplexes.

DRM+ is also by far the most flexible of the systems in terms of frequency planning, because stations can transmit individually on their own channel, and as they use just 50 kHz-wide channels the frequency planning can be 'fine-grained' as opposed to planning for the far wider DAB+ and DVB-H multiplexes. Coverage areas can also be tailored to suit the station's analogue coverage area, which isn't typically possible on DAB+.

Personally, I think the DRM+ system is very underrated. It does have one drawback of flat-fading due to using narrowband channels that the other systems don't have, but its far lower transmission powers per station allows this to be compensated for, and of course the error correction and time-interleaving mitigate this problem anyway.

DRM+ is also expected to have the option of longer guard interval durations (DRM has very long guard interval durations now) than DAB+ and DVB-H, which will allow larger SFN cell-sizes, thus making DRM+ the best option for covering very large coverage areas, in particular sparsely-populated coverage areas.

I think broadcasters and regulators should give DRM+ more consideration than they are currently doing, because DRM+ is a much better solution for certain situations than DAB+ is, and for some countries, such as Sweden, where the country has a very large and relatively sparsely-populated surface area, it would seem to me to provide a better solution than using DAB+ would.

WorldDAB and the DRM Consortium have agreed to cooperate with one another, so it is expected that future digital radio receivers will support both DAB+ and DRM+, so this will give broadcasters the option of choosing the most appropriate system for their requirements.

An Ecomonic Analysis of DAB, DAB+, DRM+ & DVB-H

Introduction

1. Choice of audio bit rate levels

AAC/AAC+ bit rate

2. Assumption used in the analysis

3. Transmission parameters' effect on transmitter powers

4. DAB+

4.1 How it works

4.1.1 Reed-Solomon Error Correction Coding

4.1.2 Higher multiplex capacity

4.1.3 Lower C/N

4.2 DAB vs DAB+ Distribution costs per station

5. DVB-H

5.1 Differences between the DAB+ and DVB-H systems

5.1.1 Similarities between the systems

5.1.2 DVB-H uses far stronger error correction coding than DAB+

5.1.3 Differential vs synchronous modulation

5.1.4 Multiplex bandwidth

5.2 DAB vs DAB+ vs DVB-H Distribution costs per station