!

!
!
!

!
!
!
!
!
!
!
!
!
!
!
White Paper on the use of DVB-S2X
for DTH applications, DSNG & Professional Services,
Broadband Interactive Services and VL-SNR applications

DVB Document A172

March 2015
 TM-S ad-hoc group

White Paper on the use of DVB-S2X for DTH applications,

application
DSNG & Professional
rofessional Services, Broadband Interactive
nteractive Services
and VL-SNR
VL applications

Scope
In October 2012 the Commercial Module (CM) requested improvements to the DVB-S2 DVB system to enhance
performance in the core markets (Direct to Home, contribution, VSAT and DSNG) and to increase the range of
applications of the standard to cover either emerging
emergi markets such as mobile (air, sea and rail) or professional
applications.. In order to allow for a rapid market launch the proposed extensions were to be an evolution of the DVB-S2
DVB
standard rather than a fundamental change to the architecture.
This document nt describes the advantages of the DVB-S2
DVB S2 extensions (herein referred to as DVB-S2X)
DVB and aims to
provide high level guidance to broadcasters and operators considering adoption of this system.
It assumes a reasonable familiarity with the original DVB-S2
DVB standard and has been intentionally kept simple in the
interests of clarity and brevity.
The DVB-S2 S2 (EN 302 307) document is now divided in two parts. parts Part 1 (EN 302 307-1)1) is the original DVB-S2
specification and Part 2 (EN 302 307-2) 307 defines the DVB-S2X extensions. Any DVB-S2X S2X receiver is backwards
compatible with the DVB-S2 S2 specifications as the Part 1 implementation is mandatory, but legacy DVB-S2 receivers
are not forward compatible with the DVB-S2X
DVB extensions. Accordingly, the legacy DVB-S2
S2 receivers will
wi not decode
transmissions using the new DVB-S2X S2X features,
features while the new DVB-S2XS2X receivers will decode both DVB-S2X and
DVB-S2 transmissions.

Summary of main technical improvements

1
DVB-S2X extends the range of operation of DVB-S2 DVB with a very-low SNR operation range (VL-SNR)(
( ), which will
extend functionality in noise compromised environments and low power applications such as VSAT networks, networks and a
very-high SNR operation range (VH-SNR SNR), which will improve throughput on high capacity trunk and contribution
contribu
links.
DVB-S2X enhances the physical layer signalling to provide a finer granularity of operative points (i.e. more modulation
and coding (MODCOD) schemes), ), enabling more flexibility with regard to optimising channel usage.
usage In addition DVB-
S2X allows the use of reduced roll-off off factors in order to decrease the occupied bandwidth,
bandwidth and to optimise satellite
transmissions in the “linear channel” (for instance in the case of multi-carrier per transponder in Ka-band).
DVB-S2XS2X also improves some features of the satellite transmission at the system level:
• the ability to configure the scrambling sequence,
sequence which was normative in DVB-S2 for all applications except
broadcast (DTH) is now made normative for DTH as well. This allows systems to cope more readily with high level
co-channel interference in multi-satellite
satellite environments
• the changes in modulation, which were normative in DVB-S2 S2 for all applications except DTH (where it was
optional), are now possible on a frame-by-frame
frame basis for DTH as well. This permits the real time optimisation of
the transmission efficiency versus the transmission robustness
• channel
hannel bonding is permitted which can provide an increase in throughput, by merging capacity across acr multiple
transponders, thus enhancing the performance of the statistical multiplexing of services
• thehe higher level protocols (i.e. GSE, GSE-lite)
GSE have been improved allowing for greater integration
in with IP based
systems
• the “all-IPIP streaming” capability has been introduced.

In summary, these technical enhancements offer the following benefits to the service provider:
provider
• for DTH applications each of the DVB-S2X features provide only a small improvement in terms of efficiency
and flexibility, but in combination they constitute a significant improvement over the original DVB-S2 DTH
system,, specifically they offer a road map to new generation services such as UHDTV
• for VSAT applications, the DVB-S2X
DVB specification opens up the possibility to support advanced techniques
for future broadband interactive networks such as intra-system
system interference mitigation, beam-hopping
beam as well
as multi-format
format transmissions. These techniques are made possible by the Super-Framing
Framing structure described

1
VL-SNR frames may be inserted in DVB-S2 S2 and DVB-S2X transmissions but require compatible receivers (Note ( that special care should be
taken when using these frames in DVB-S2
S2 transmissions in order to verify that the receiver has the capability to ignore them).

Page 1 of 14
 in Annex E of the DVB-S2X specification and when implemented they may offer significant gains in capacity
and flexibility of broadband interactive satellite networks
• for Professional & DSNG applications high efficiency modulation schemes allow spectral efficiencies
approaching 6 bps/Hz allowing for optimised satellite capacity usage.

Table 1: system configurations and application areas for DVB-S2 (Part 1) and DVB-S2X (Part 2)

Broadcast Interactive Profession

SYSTEM FEATURES DSNG VL-SNR
services services al services
Part I INPUT Types Single Transport Stream N N N N
Part I Multiple Transport Stream O O O O
Part I Single Generic Stream NA O NA O
Part I Multiple Generic Stream NA O NA O
Part II High Efficiency Mode for GSE/GSE-Lite input N N N N N
Part I INPUT Processing Input Stream Synchronizer NA O O O
Part I Null Packet Deletion NA O O O
Part I Dummy Frame insertion NA N N N
Part I TRANSMISSION Constant Coding & Modulation N N N N
Part II Variable Coding & Modulation O !N O !N O !N O !N N
Part II Adaptive Coding & Modulation NA N O O N
Part II PL PACKET PLHeader = 7+1 bits = 128 additional N N N N N
Part I PLH Wide-Band = 8+8 bits defined in Annex M O O O O
Part II PLH Wide-Band = 8+8 bits (extended for slicing) O O NA O O
Part I WAVEFORM Roll-Off range = 35%, 25%, 20% N N N N
Part II Roll-Off range extension = 15%, 10%, 5% N N N N N
Part II Channel Bonding Split of PLPackets over several transponders N NA NA O NA
Part II Super Frame Group of PLPackets constitutes a Super Frame NA O O O O
N = Normative, O = Optional, NA = Not Applicable

Broadcast Interactive Profession

FECFRAME (64800 bits) DSNG VL-SNR
services services al services
Part I QPSK 1/4,1/3, 2/5 O !N N N N
Part I 1/2, 3/5, 2/3, 3/4, 4/5, 5/6, 8/9, 9/10 N N N N
Part II 13/45 N N N N N
Part II 9/20, 11/20 N N N N N
Part I 8 PSK 3/5, 2/3, 3/4, 5/6, 8/9, 9/10 N N N N
Part II 23/36, 25/36, 13/18 N N N N N
Part II 8 APSK-L 5/9, 26/45 N N N N N
Part I 16 APSK 2/3, 3/4, 4/5, 5/6, 8/9, 9/10 O !N N N N
Part II 26/45, 3/5, 28/45, 23/36, 25/36, 13/18, 7/9, 77/90 N N N N N
Part II 16 APSK-L 5/9, 8/15, 1/2, 3/5, 2/3 N N N N N
Part I 32 APSK 3/4, 4/5, 5/6, 8/9, 9/10 O !N N N N
Part II 32/45, 11/15, 7/9 N N N N N
Part II 32 APSK-L 2/3 N N N N N
Part II 64 APSK 11/15, 7/9, 4/5, 5/6, O N N N O
Part II 64 APSK-L 32/45 O N N N O
Part II 128 APSK 3/4, 7/9 NA O O N NA
Part II 256 APSK 32/45, 3/4 NA O O N NA
Part II 256 APSK-L 29/45, 2/3, 31/45, 11/15 NA O O N NA
N = Normative, O = Optional, NA = Not Applicable

Broadcast Interactive Profession

FECFRAME (16200 bits) DSNG VL-SNR
services services al services
Part I QPSK 1/4,1/3, 2/5 NA N O N
Part I 1/2, 3/5, 2/3, 3/4, 4/5, 5/6, 8/9 NA N O N
Part II 11/45, 4/15, 14/45, 7/15, 8/15, 32/45 NA N O N N
Part I 8 PSK 3/5, 2/3, 3/4, 5/6, 8/9 NA N O N
Part II 7/15, 8/15, 26/45, 32/45 NA N O N N
Part I 16 APSK 2/3, 3/4, 4/5, 5/6, 8/9 NA N O N
Part II 7/15, 8/15, 26/45, 3/5, 32/45 NA N O N N
Part II 32 APSK 2/3, 32/45 NA N O N N
N = Normative, O = Optional, NA = Not Applicable

Broadcast Interactive Profession

Very Low - SNR (Part II's novelty for VL-SNR reception conditions) DSNG VL-SNR
services services al services
Part II BPSK 1/5, 4/15, 1/3 for short FEC blocks (16200 bits) NA O O NA N
Part II 1/5, 11/45, 1/3 for medium FEC blocks (32400 NA O O NA N
Part II BPSK-Spread2 1/5, 11/45 for short FEC blocks (16200 bits) NA O O NA N
Part II QPSK 2/9 for long FEC blocks (64800 bits) NA O O NA N
N = Normative, O = Optional, NA = Not Applicable

Page 2 of 14
Table 1 summarises (2) the elementary features of both the DVB-S2 system (in grey, EN 302 307-1) and the DVB-S2X
system (in orange, EN 302 307-2), and indicates the status of each feature (i.e. Normative, Optional and Not-
Applicable) for each targeted application area (i.e. Broadcast Services (DTH), Interactive Services, DSNG, Professional
Services and Very-Low SNR delivery).

Modifications to the Physical Layer

Header Modification
The DVB-S2 specification defines a satellite transmission system using a single carrier signal transmitting a series of
physical layer packets. Each packet carries a payload protected by a forward error correction (FEC) encoding scheme
(in this case a concatenation of LDPC and BCH). This is preceded by a physical layer header (PL-Header) allowing
DVB-S2 demodulators to achieve synchronisation with the transmitted signal. Additionally, the PL-Header signalling
provides an indication of the MODCOD scheme used in the payload, the presence of pilot signals and the FEC frame
length.
A new physical layer header (based on the original DVB-S2 header) has been designed extending the signalling from 7
bits to 8 bits. This has enabled the definition of more MODCOD schemes, allowing a finer SNR granularity as well as
the introduction of additional MODCOD modes optimised for linear channels. The extended physical layer header also
offers improved error-correction capabilities for very low SNR (VL-SNR) operation down to SNR levels as low as
-10 dB.
Scrambling Sequences
It is not un-common for satellite services to have to contend with a high level of co-channel interference (CCI).
Furthermore it is anticipated that, going forward, multi-spot beam satellite payloads, operating in high CCI
environments, will become much more common place (especially, but not exclusively, in Ka band).
To avoid performance degradations where channel estimation is based solely upon pilots, DVB-S2X proposes a
mechanism to mitigate CCI between DVB-S2 and DVB-S2X carriers, by invoking the use of a defined set of
scrambling sequences for the broadcast (DTH) profile as well as the professional profiles, that already previously
supported this feature in DVB-S2.
The default DVB-S2 Gold code (scrambling sequence number 0), and six other Gold codes have been defined for use in
these scenarios. A DVB-S2X receiver will try to decode using the default Gold code first and then cycle through the
others as required until the correct sequence is found.
New MODCODs and finer SNR granularity
The following section shall consider the implications and advantages of the improvement in MODCOD granularity with
respect to:
1. Broadcast services (DTH applications)
2. Interactive services (broadband applications)
3. Professional services (DSNG and VLSNR applications)

1) DTH Broadcast services

When comparing performance, consideration must be given to the Channel Model in which the transmission is
operating. For the purposes of this document, two models are considered, known as models A and B.
Channel model A represents an ideal channel consisting of an ideal linearized High Power Amplifier (HPA) acting as a
hard limiter to amplitude excursions. This is driven at the optimum operating point for any MODCOD. SNR is
expressed as Psat/N, where Psat is the saturated amplifier power and N is the noise contribution within the usable
channel bandwidth Bu, where Bu=Rs(1+roll-off), which is set to 36 MHz. In this model noise and spectral efficiency
are measured in the total usable bandwidth Bu. This model assumes there is no adjacent channel interference and that
no linear distortion is imposed upon the transmission.
Channel model B represents a more realistic channel and includes IMUX and OMUX satellite filters (corresponding to
a 38 MHz -3dB bandwidth, and channel spacing of 40 MHz), and a more faithful TWTA non-linear model (including
both amplitude and phase distortions). This model also incorporates an element of Adjacent Channel Interference (ACI)
in both the uplink and downlink paths, with the same symbol rate and roll-off factors as the wanted signal. It is assumed
that the TWTA Output Back Off (OBO) is optimised for each MODCOD. The usable signal bandwidth
Bu = Rs x (1+roll-off) can be freely optimised, or it can be limited by satellite operator specifications.

2
Being a summary of « table 1 » specified in Part 1 & Part 2 of the ETSI EN 302 307 document, for the purposes of clarity, the notes explaining
the possible misinterpretation of some specific features have not been reproduced. Please refer to the original standard in case of ambiguity.

Page 3 of 14
 Figure 1: DVB-S2 and DVB-S2X spectral
pectral efficiency, at the same roll-off and symbol-rate
rate (channel model A),
A) in the
SNR range typical for broadcasting.

Figure 1 illustrates the

he simulated performance of channel model A in the SNR range of interest for broadcast services.
DVB-S2 is indicated by the green plot lot and DVB-S2X by the blue plot. The same roll-off off is assumed (3), to better clarify
the granularity effect. Figure 1 shows that DVB-S2X and DVB-S2 S2 share a large number of MODCODs, therefore their
points fall on the same ideal dotted lines.
DVB-S2XS2X adds new 8PSK MODCODs in the SNR range from 6 to 8.5 dB covered in DVB--S2 with lower granularity.
In addition, whilst 16APSK is optional in DVB-S2 for DTH legacy receivers (see violet triangles in Figure 1), DVB-
S2X mandates the use of 16APSK (adding new MODCODs) in the SNR region 9 to 12 dB. The mandatory
implementation of 16APSK in DVB-S2X S2X represents a significant capacity increase of around 5% over the original
mandatory DVB-S2 S2 specification in the SNR region around 10 dB (which is a challenging operating area for current
satellite platforms, typically characterised by a C/I of about 13 dB from adjacent satellites). Even further increases are
available when operating in channels with SNR levels better than 10 dB.
In the range of SNR appropriate for broadcast applications, the DVB-S2X
DVB curve faithfully follows the slope of ideal
Shannon capacity curve within a dB or so (4).
The performance of a satellite broadcast service is characterised by a link analysis calculation (the link budget). This
predicts the performance of the service in terms of receive margin (in dB) above the required threshold and the
performance is often described in terms of an an availability figure (% availability per annum). The performance is
determined by the SNR characteristics of the transmission channel to the specific downlink installation and location.
The available SNR is not a constant over any a given region and is dependent upon the satellite uplink and downlink
contour maps and the characteristics of the transmit and receive earth stations. In addition there is a temporal variation
caused by the prevailing weather conditions in the uplink and downlink regions at the the time of the transmission.
A broadcaster computes the minimum available SNR in the service area for a target service availability (e.g. 99.9% 99 of
the year) using appropriate statistical modelling techniques (for example, see ITU-R R Recommendations P.618 and P.837
to P.841 as a guide,, although other modelling techniques also exist).
exist
Due to its increased MODCOD granularity, DVB-S2X DVB S2X offers greater flexibility to the broadcaster with regard to
optimising system performance for specific SNR levels.
As shown in Figure 1, any SNR can be matched by DVB-S2 and DVB-S2X (as as indicated by the blue and green
continuous lines with the stair-case
case shape connecting the simulated
sim points (5)), by fine-tuning
tuning the transmitted symbol-
symbol
rate. The figure shows that DVB-S2X,S2X, with its finer MODCOD granularity, closely matches the ideal blue dotted line,
while DVB-S2 with its coarser er granularity suffers from a larger “ripple” compared to the ideal dotted green line.

3
Simulated Roll-off
off 5%, even though this is not available in DVB-S2
DVB
4
It should be noted that in the SNR range of interest for broadcasting, each tenth of dB corresponds approximately to 1% of spectral
sp efficiency
5
The continuous lines are constructed as follows: starting from a simulated point at the maximum allowed symbol-rate,
symbol rate, the symbol-rate
symbol and spectral
efficiency are progressively lowered, and the required SNR reduces by 10Log[Rs/Rs(max)]; when the lower efficient eff MODCOD becomes
preferable in terms of spectral efficiency and SNR, it is selected (the flat part of the staircase). When the symbol rate cannot
can be finely tuned, the
steps become vertical and the maximum granularity gain approaches 14%.

Page 4 of 14
As previously described, in order to ensure that the simulated results were complete and representative, the TM-S2
group introduced the more complete channel model B together with the concept of an enhanced receiver (ER) with
linear equalisation. Similar results to those in Figure 1 have been achieved with this model, but in this case the
performance “ripple” due to symbol-rate fine-tuning is dependent upon the allocated bandwidth Bu. In particular, when
the transmitted symbol-rate approaches the OMUX -3dB bandwidth, the ripple amplitude reduces to around 3% of
spectral efficiency (SE) because of the equaliser effects.
In summary, when specific SNR targets are to be met, DVB-S2X reduces the DVB-S2 losses due to its finer MODCOD
granularity. The DVB-S2X capacity advantage ranges from 0% (6) to 10% (for Rs<<-3dB bandwidth) or to 3% (for Rs
approaching -3dB bandwidth).

2) Interactive Services
Frequency re-use techniques over multi-spot beam networks allow for the optimal delivery of broadband interactive
services over large geographic regions. Indeed, an increased number of High Throughput Satellites (HTS) are being
launched to support this technology. In this configuration, due to on-board satellite accommodation limits, the High
Power Amplifiers (HPA’s) are typically operated in multi-carrier mode with appropriate back-off applied. Thus the
overall end-to-end link can be considered quasi-linear (this assumption is applicable for more than three carriers per
HPA, the dual carriers per HPA case being more borderline). The new DVB-S2X specification introduces additional
MODCOD features which can take advantage of this mode of transponder operation: Linear MODCODs (indicated by a
'-L' suffix in the MODCOD name), which have been optimised for a linear channel in the presence of phase-noise (7).
These bring mixed values of gain, in some cases up to about 1 dB better SNR for the same spectral efficiency.
In addition, a finer granularity of MODCODs has been introduced, reducing the SNR threshold distances from the 1 to
1.5 dB of DVB-S2 to an average of about 0.4 to 0.5 dB for DVB-S2X. This reduces the system margin required in
ACM-based systems.
As a normative feature, DVB-S2X mandates the support of a higher modulation order of 64APSK, as opposed to the
optional 32APSK of DVB-S2. This results in an expected spectral efficiency gain of up to 10%. Higher order
modulations up to 256 points are also optionally available, thus further increasing the potential spectral efficiency for
systems where the link budget allows operating at a corresponding higher level of SNR.
Very Low SNR (VL-SNR) MODCODs can also be optionally used in DVB-S2X allowing broadband networks to
operate in regions affected by heavy atmospheric fading, at higher frequency bands (e.g. Q/V-band) or with small
antenna (even mobile) terminals.

Figure 2: Efficiency versus C/Nref (noise integrated over Rs x (1+roll-off))

for the same occupied bandwidth (AWGN linear channel).

6
0% in the case that the selected DVB-S2X MODCOD is also available in DVB-S2
7
See Table H.3 in Annex H.8 of EN 302 307-2, “Phase Noise Masks for Simulation”

Page 5 of 14
Figure 2 shows the simulated performance over a linear channel (AWGN) of DVB-S2 (blue markers connected with a
blue solid line) and DVB-S2X (red markers connected with a red solid line). The diagram illustrates normal
FECFRAME size and a 5% roll-off for DVB-S2X and a 20% roll-off for DVB-S2.
Operations in single-carrier per HPA mode is also foreseen for broadband interactive services in DVB-S2X thanks to
the inclusion of MODCODS which have been specifically optimised for this configuration. This system configuration
typically represents legacy single beam networks as well as future multi-beam networks where on-board limitations do
not prevent such operations. Further to the MODCODs specified for DTH, 64APSK constellations optimised over the
non-linear channel are also normative in broadband interactive services. Optionally, 128APSK and 256APSK
constellations optimised over the non-linear channel can also be used.

3) Professional Services (DSNG and VL-SNR)

Professional services such as trunking and digital satellite news gathering (DSNG) contribution links typically use
receive stations with larger antennas when compared to DTH terminals. Moreover, thanks to ACM and/or uplink power
control these services can operate at very low SNR margins. DSNG services typically operate with SNR values in the
range 10 dB to 14 dB and backhaul/trunking services in the SNR range of 12 dB to 18 dB. Both of these applications
stand to have significant gains (i.e. more throughput or more margin) from the new MODCODs offered by DVB-S2X
for high SNR channels. This region of SNR operation is highlighted by Figure 3.
Professional services (such as professional distribution or IP trunking) can be operated in single-carrier per transponder
mode (in C-band and Ku-band). However, these and other professional services, including DSNG contribution are
typically operated in multiple-carrier per transponder scenarios (in C-band, Ku-band and Ka-band).

Figure 3: Identical to Figure 2, but zoomed in on the professional SNR operating region
(AWGN linear channel).

As a normative feature for professional applications, DVB-S2X mandates the support of a higher modulation orders up
to 256APSK, as opposed to the 32APSK of DVB-S2. This results in spectral efficiency gains up to 51% as depicted in
Figure 3, which zooms in on the typical operational region for professional applications. Due to the lower roll-offs, finer
granularity and improved MODCODs, significant gains are possible when compared to DVB-S2. For example up to
2.5 dB SNR gain for the same spectral efficiency or up to 21% in the sub 12 dB SNR region and 51% in the sub 20 dB
SNR region efficiency improvements are possible for the same SNR value.
The VL- SNR MODCODs can also be optionally used in DVB-S2X allowing the operation of professional services in
high atmospheric fading areas.
Transmission of VL-SNR Frames Mixed with Regular Frames
The original DVB-S2 PLS header was designed to operate reliably at symbol-energy per noise levels of slightly lower
than Es/No = -2.5 dB. However this performance is not adequate to support the additional VL-SNR MODCODs of
DVB-S2X which are required to operate at Es/No levels as low as -10 dB. A straightforward revision to the DVB-S2
PLS header design to accommodate these additional lower Es/No levels would have required a significantly higher

Page 6 of 14
transmission overhead to all receive terminals. Considering the fact that mobile and VL-SNR terminals are likely to be
only a small minority of the population in any given network, another less onerous solution was deemed to be required
in the implementation of the DVB-S2X header architecture in order to economically support the additional modes.
An alternative approach which allows regular terminals to use the original DVB-S2 90-symbol PLS Header with only
minimal modification to support the larger number of DVB-S2X MODCODs has therefore been adopted. As shown in
Figure 4, terminals with higher Es/No thresholds will continue to use the 90-symbol PLS header (i.e. the regular frame
sync), and are therefore not penalised by accepting mobile or VL-SNR terminals into the network. Each code frame
intended for a mobile or VL-SNR terminal also contains the original 90-symbol PLS header which, in this case, is used
to signal to the normal terminals that the following is a frame for a VL-SNR terminal, and specifies the length of the
code frame (only two lengths are allowed even though nine VL-SNR MODCODs are defined). A regular terminal can
then skip this code frame and look for the next PLS header at the corresponding time with no disruption of frame
synchronisation.

Regular Frame Regular Frame Mobile Frame Regular Frame

Key:

Standard PLS Header Mobile frame Sync

Figure 4: Time-multiplex format of a DVB-S2 extension carrier consisting of both regular code frames and code frames
for VL-SNR terminals

In Figure 4, the additional VL-SNR frame synchronisation serves two purposes:

•to provide burst mode synchronisation of the VL-SNR frames, and
•to identify the specific VL-SNR MODCOD for the mobile terminals.
This is accomplished by a set of specifically designed unique words (UW), each identifying one of the nine
MODCODs, plus a dummy frame. These unique words are generated from a single mother sequence of 900 symbols,
such that an efficient correlation algorithm can be used to detect both the presence the signal and, at the same time,
identify the MODCOD. The length of the UW is selected to provide more than adequate margin to achieve
synchronisation at the lowest operating SNR range without introducing excessive overhead.
For most of the VL-SNR range, π/2-BPSK modulation is used. The two lowest MODCODs are further spread by a
factor of 2. To interoperate with legacy DVB-S2 receivers in VCM modes, the length of LDPC codes are modified so
that the frame size of the VL-SNR MODCODs, including the 900-symbol frame sync and pilots, is the same as either a
QPSK or 16APSK modulated DVB-S2 MODCOD with pilots.
Sharper roll-off
The gains achieved from sharper roll-off depend upon the flexibility the operator has to set the optimum symbol rate
and the constraints imposed by the satellite operator in terms of uplink spectral mask.
Assuming that the operator is totally free to optimise the symbol-rate to increase the spectral efficiency (as illustrated in
Figure 7), and using the DTH satellite channel model B and an enhanced receiver, the TM-S2 has calculated that there
is no clear benefit to use sharper roll-offs (see Figure 5). Although sharper roll-offs allow higher symbol rates (37
Mbaud) the additional capacity gain is below 2% compared with a 20% roll-off and symbol-rate of 34 Mbaud.
Nevertheless, increasing symbol rate may generate interference in adjacent channels, which can be better controlled by
sharp roll-offs.
New simulation results based on the satellite channel model described in the User Guidelines document TR 102376-2
which includes an additional increase of +8 dB in the uplink ACI and with a symbol rate of 34 Mbaud, have shown that
a roll-off of 20% can provide 0.1 - 0.2 dB better performance when compares with 5% roll-off using a fractionally
spaced equaliser in the receiver. Note that these results did not cover higher symbol rates.
In cases where the satellite operators limit the usable bandwidth (Bu) by invoking the (1+roll-off) (8) rule, then
maximum transmittable symbol-rate becomes Bu/(1+roll-off). In this scenario, assuming that within the usable
bandwidth, transponder bandwidth limitations are not generating significant distortions (i.e. we are operating as channel
model A), then the sharper roll-offs will allow for a proportional increase of the transmitted symbol rate. In this case the
spectral efficiency gains roughly account for half of the roll-off difference. This is because the other half of the gain is

8
The transmitted symbol rate Rs must be lower to BW/(1+roll-off) , where BW is the usable bandwidth.

Page 7 of 14
“masked” by the increased noise bandwidth at the receiver. For example, comparing DVB-S2 with roll-off 20% to
DVB-S2X with roll-off 5%, the symbol-rate increase is about 14%, however the receiver noise power increases by
around 0.6 dB and, overall, the net spectral efficiency gain is reduced to around 7%.

Figure 5: Effect of sharper roll-offs on channel model B, “free symbol-rate optimisation”: DVB-S2X (red and green
curves) compared with DVB-S2 at roll-off 20%

Figure 6. Effect of sharper roll-offs on channel model B, “1+roll-off rule” with Bu=36 MHz: DVB-S2X (red and green
curves) compared with DVB-S2 at roll-off 20%

To validate such results on the more complete channel model B, and to compare the available capacity with that of
Figure 5 (free symbol-rate optimisation), the spectral efficiency with roll-off’s of 5%, 10% and 20% have been
compared imposing the (1+roll-off) rule over Bu=36 MHz. For a 5% roll-off the measured gain when compared to a
20% roll-off approached 6.5%, only slightly less than the gain on the simpler channel model A, without bandwidth
limitation distortions (9).

9
When comparing the 20% roll-off in Figure 5 with the 5% roll-off in Figure 6, the (1+roll-off) rule with a Bu=36 MHz limitation produces a
negligible performance penalty. In the case that Bu is increased (e.g. from -1 dB transponder bandwidth to the channel spacing), then the roll-off
gain progressively reduces and approaches the case with free symbol-rate optimisation. To compare results of Figure 6 with Figure 1 (both at roll-
off 5%), SNRs in Figure 1 are to be reduced by 10*log(38/36)=0.2 dB, and SE divided by 38/36=1.0555. After such scaling, it can be deduced for
example that 8PSK rate 5/6 on channel A performs 0.7 dB better than on channel B.

Page 8 of 14
Similarly, in broadband and professional applications, the gains from sharper roll-off roll off depend upon the broadband
network configuration Particularly, whether the broadband network is operated as an ‘open system’ or as a ‘closed
system’ and whether the on-boardoard HPAs are operated in single carrier or multi-carrier
multi carrier mode. In multi-carrier
multi mode, it is
expected that the (1+roll-off) rule applies, as near optimum performance is achieved when the carriers within the
transponder do not overlap in frequency. In single
single carrier mode, when the network is shared by different operators using
different uplink stations, the satellite operator may impose a maximum occupied bandwidth so that again the (1+roll-
off) rule applies. When the (1+roll-off)
off) rule applies, and assuming that withinin the transponder bandwidth limitations we
are not generating significant distortions, sharper roll-offs
roll offs allow a proportional increase of the transmitted symbol rate.
In this case the spectral efficiency gain depends upon the SNR region. For example, ple, comparing DVB-S2 with 20% roll-
off with DVB-S2X with 5% roll-off, off, the symbol-rate
symbol rate increase is 14%, the receiver noise power increases by 0.6 0 dB, and
the net gain in the professional region varies from 8% (SNR=10 dB) to 11% (SNR=20 dB).
In ’closed systems’,, where the network is fully utilised by one operator, and the satellite payload is operated in single
carrier per HPA mode, the operator itself may decide on how to optimise the inter-transponder
transponder interference in order to
optimise the overall system capacity. Inn this case, the same conclusions as for the corresponding DTH case apply.
Use of Enhanced Receivers
Basic equalisation technologies are implemented in most DVB-S2 chipsets. DVB-S2X, S2X, through the use of channel
model B, encourages manufacturerss to include even more advanced equalisers by default. As such, when using the new
receivers, broadcasters will often be able to optimise to a higher symbol rate within the correction capability offered
of by
these equalisers.
Computer simulations on channel modelodel B, using free symbol-rate optimisation with adaptive equalisation (in enhanced
receivers, ERs) demonstrated spectral efficiency
e gains of around 7% to 9% over conventional receivers (CR), as shown
in Figure 7.
Nevertheless, the standard does not specify the receiver's implementation, hence the upper symbol rate limit still
depends upon the precise receiver implementation, the channel distortions
distortions and the satellite operator's operating rules.
Thus we can say that the introduction of the advanced equaliser concept will bring benefits but that it is up to the
implementers to realise those benefits through careful optimisation. Of course, the equaliser ser is less critical when the
transmit spectral mask bandwidth is significantly less than the transponder separation (e.g. following Ru=Rs(1+rolloff)
rule imposed by some satellite operators).
operators)

Figure 7: Spectral efficiency vs symbol rate for receivers

rece vers with and without linear equalisation (with = enhanced
receiver ER, without = conventional receiver, CR).
receiver,

Variable Coding and Modulation

DVB-S2X makes variable coding and modulation (VCM) mandatory for DTH applications, thus enablingenab the change of
MODCODs on a frame by frame basis. This allows operators to adjust transmission robustness versus efficiency
according to the prevailing atmospheric conditions or even to tailor services according to quality of service
requirements (e.g. a highly robust standard definition TV channel simulcast with an ultra-high
high definition TV channel).

Page 9 of 14
By sacrificing the picture quality during heavy rain fading (e.g.
( SDTV instead of UHDTV), it is possible to significantly
increase the overall system spectral efficiency
fficiency whilst at the same time improving the perceived availability of the
service to the end user. Two examples of comparative performance gains are presented in Table below.
below
Ku-Band Ka
Ka-Band
Study Case 95 °E geostationary orbit 10 °E geostationary orbit
Coverage over India 8--Beams
Parameter Coverage over Europe
Transponder Spacing (MHz) 40 40 60 60
Transponder Bandwidth (MHz) 36 36 54 54
Total Number of Transponders 24 24 40 40
Aggregate bandwidth (MHz) 960 960 2400 2400
Air Interface (DTH Profile) DVB-S2 DVB-S2X DVB-S2 DVB-S2X
Transmission Mode CCM VCM CCM VCM
Average bit Rate per stream (Mbits/sec) 20 20 (HQ) 20 20 (HQ)
HEVC 2 (LQ) 2 (LQ)
Availability 99,9% 99,0%(HQ) 99,9% 99,0% (HQ)
99,9%(LQ) 99,9%(LQ)
Symbol Rate (MBaud) 34 34 51 51
MODCODs QPSK 2/3 8PSK 3/5 (HQ) QPSK 2/3 8PSK 5/6 (HQ)
QPSK 2/3 (LQ) QPSK 2/3 (LQ)
Total number of UHD video streams
48 72 120 240
(TV channels) delivered by the satellite

Main improvements at the Upper Protocol Layers

Channel bonding
The DVB-S2 S2 system was designed to carry single or multiple MPEG Transport Streams (or generic continuous streams)
over a single satellite transponder. DVB-S2
DVB was developed to address the he main DTH application in 2003 which was
HDTV multi-programme deliveryelivery using advanced video coding (AVC H.264).. Typically this would require about 10
Mbit/s per program in the case of constant bit-rate
bit rate coding (CBR). Therefore in a single 36 MHz transponder, delivering
about 60 Mbit/s in DVB-S2 S2 format, it was possible to broadcast six HDTV CBR programs, which could be increased to
seven by exploiting the 20% gain of statistical
tatistical multiplexing (see Figure 8).
The DVB-S2X S2X system was developed in 2013, a time when HEVC video coding and UHDTV (four times the definition
of HDTV) technologies are beginning to be embraced by the broadcast sector. sector. Assuming that a UHDTV signal requires
four times the transmission capacity of HDTV for the same compression system, and that HEVC doubles the
compression efficiency versus AVC, we can estimate a required bit-rate
bit rate per UHDTV CBR program of 20Mbit/s. Under
such circumstances, a 60 Mbit/s DVB-S2 S2 signal could provide
pro three UHDTV CBR programs per transponder, and the
statistical multiplexing gain would be significantly reduced (e.g. 12% instead of 20%, assuming Figure 8 is applicable
to UHDTV and HEVC), thus not allowing the transmission of an additional program. This makes the prospect of
satellite carriage for UHDTV services using conventional techniques rather uneconomic.

Figure 8:: Example of Statistical Multiplexing Gain for H264 coding (10)

In order to increase the statistical multiplexing gain for UHDTV, DVB-S2X introduces the concept of “channel
bonding”. This allows the capacity of two or three transponders to be merged in order
er to transport a single “big

10
Ken McCann: “Review of DTT HD Capacity Issues, An Independent Report from ZetaCast Ltd Commissioned by Ofcom

Page 10 of 14
transport stream” (typically a large statistically multiplexed MPTS). It should be noted that this functionality is only
available if the receiver is equipped with multiple tuners to enable the simultaneous acquisition of data streams from
different transponders. These types of receivers are already becoming more commonplace with the advent of
functionalities such as ‘picture-in-picture’ and ‘watch one program, record another’.
To illustrate the potential gains of channel bonding, consider a typical 36MHz transponder carrying three UHDTV
services. If the transponder is bonded with a second transponder then the channel capacity increases to six services.
Reference to Figure 8 shows that the statistical multiplexing gain rises from around 12% to 19%, an improvement of
7%. Bonding a third transponder could take the gain to 24%, an overall gain of 12% when compared to the original
single transponder.
A further potential application of channel bonding is in its ability to gather spare capacity across transponders. Often,
due to the placement of carriers on a shared transponder, areas of spectrum become sterilised between carriers and
cannot be usefully deployed. Channel bonding offers the potential of accumulating this space and reconstituting it into
usable capacity, thus further improving transponder usage. It should be noted that the DVB-S2X supports the channel
bonding only in conjunction with Constant Coding & Modulation (CCM).
Generic Stream Encapsulation (GSE)
DVB-S2X includes the possibility to carry audio-visual and data services in full-IP format, using the GSE or GSE-lite
adaptation protocol layer instead of a traditional transport stream. This can increase the system flexibility in the future,
when the services from various networks (such as terrestrial broadband networks and satellites) converge to the IP
format. For example the broadcasting services from the satellite could be injected in a full-IP home network, and
seamlessly routed to the user together with the interactive broadband services (x-DSL or optical-fibre). The support of
GSE or GSE-lite also significantly reduces the layer-2 encapsulation overhead thus bringing additional gains in terms of
net system capacity.

Optional Super-Framing
An optional super-framing structure is available in DVB-S2X as described fully in Annex E of the standard.
The use of the super-framing structure has the following benefits:
• Increased resilience to co-channel interference caused by other beams, due to super-frame-wide scrambling
and regular (and orthogonal) pilot symbol insertion that can be aligned in time on different beams
• Support of synchronisation algorithms by the regular insertion of reference data fields, which leads to
enhanced receiver performance under severe channel conditions such as VL-SNR or link interruptions
• Future proof frame design with signalling of the super-frame content format, which is able to accommodate
and support:
• Interference mitigation techniques
• Beam hopping operations
• Single-format or multi-format transmission per carrier
• Introduction of different profiles where receivers can clearly identify supported and unsupported
super-frame content formats
The following profiles are currently specified, while the remaining 11 profiles are reserved for future use or proprietary
content:
Profile Description
0 DVB-S2X including a modified VL-SNR burst mode frame with larger pilot fields
1 DVB-S2 legacy format
2 Format with Bundled PL-FRAMEs (64800 payload size) of constant size and location
3 Format with Bundled PL-FRAMEs (16200 payload size) of constant size and location
4 Flexible format with VL-SNR PL-Header tracking and different PL-Header protection levels

The super-frame structure corresponding to the first profile introduces an extra overhead of only 0.12% when compared
to common DVB-S2X transmission. For future applications, the super-frame structure supports orthogonal start of
super-frame (SOSF) and pilot fields by using Walsh-Hadamard sequences. A set of unique orthogonal sequences can be
assigned to co-channel carriers within a multi-spot beam network (a unique sequence per beam).
Profiles 2 and 3 allow constant and aligned PL-FRAME sizes to be maintained over co-channel carriers. This enables
interference mitigation techniques to be employed at the receiver terminal (e.g. Multi-User Detection, MUD) as well as
at the Gateway (GW pre-coding techniques) as presented in Figure 9. Both techniques, depending upon the system
configurations, have the potential to provide capacity gains between 20% and 100%, with pre-coding offering the
greatest advantage. In order to benefit from these gains they should be used in combination with more aggressive
frequency re-use schemes, in particular, with full frequency re-use and two colour schemes.
Page 11 of 14
 User
The GW computes that 1 terminals
erminals estimate the
channel precoding matrix end
end-to-end channel
given the reported channel frequency response for C,
responses reported by each I1 and I2, i.e. Hi1,Hi2
terminal in the coverage and Hi3-
Hi3 up to 10~15
3 C
interference channel
responses for full
I1 I2 frequency –re-use
systems

GWY

2 Hi1, Hi2, ….HiN (N complex numbers):

reported on the return channels when significant
changes are detected – worst case every ~ 1 sec
in deep fading conditions similarly to the ACM
signaling

Figure 9: Functional description of GW-based

GW Pre-coding
coding techniques

In addition to interference mitigation, the new super-frame

super structure supportss the idea of satellite beam hopping. This
technique is currently being considered for use in broadband networks (through proprietary solutions) with the scope to
increase the flexibility of resource allocation over both geographic coverage
verage and over time. The concept consists of
illuminating the different beams in a time division fashion with bursts of data, such that at any given time only a set of
beams are simultaneously illuminated. The complete set of beams within the overall coverage rage will obtain illumination
within the so called “beam-hopping” window. Initial system
ystem simulation results show that beam hopping can better track

Page 12 of 14
the time-variant nature of user demand within the coverage when compared to other techniques (such as flexible HPAs).
From the air interface point of view, beam hopping requires the receivers to work in a burst mode, as each terminal will
only see the desired carrier for a fraction of the beam hopping window. This is made possible with super-framing due to
the structure of the payload data container and the data gaps which terminate the super-frame with dummy symbols (in
formats 2 and 3) or dummy frames of arbitrary size (in format 4). Thus, consecutive containers act as bursts plus guard
time, which allow the satellite payload to re-configure according to the beam to be illuminated.

Conclusions
The 2003 DVB-S2 standard already offers excellent spectral efficiency performance with respect to the Shannon-limit,
particularly for DTH applications, and provided significant efficiency gains (of around 30%) when compared to the
previous DVB-S system. DVB-S2X represents an evolution and refinement of the DVB-S2 standard rather than a
fundamental step change in the technology. Therefore, in general, DVB-S2X cannot provide such large advances in
performance as offered by the DVB-S to DVB-S2 transition and, indeed, that is not its intention.
However, the changes presented in DVB-S2X do offer significant advances and opportunities, to service providers,
particularly in the areas of:
• New generation DTH broadcast services (e.g. VCM with simulcasting or channel bonding in support of
UHDTV)
• Broadband services (e.g. multi-beam hopping applications supported by new super-framing structure)
• Professional services (e.g. finer MODCOD granularity offering more flexible SNR usage and VLSNR
applications)
Finally, we shall summarise the advantages and potential gains for each application area in turn.

DTH
DVB-S2X fine-tunes both the physical and the upper protocol layers of DVB-S2, producing a highly attractive package
for those service operators intending to launch new generation services requiring new receivers. The typical application
is the launch of UHDTV services based on the new HEVC coding algorithms. The table below summarises the spectral
efficiency gains of DVB-S2X versus DVB-S2 for UHDTV broadcasting.

DVB-S2X Spectral Efficiency gain vs DVB-S2 (channel model B, enhanced receiver for DVB-S2 and DVB-S2X)
DVB-S2X Characteristic (1+roll-off ) rule; Bu =36 MHz Free symbol-rate
optimisation
Physical Better MODCODs 0% (for SNR<10 dB) 0% (for SNR<10 dB)
Layer 5% or larger (SNR>10 dB), (11) 5% or larger (SNR>10 dB),
(12)
Finer granularity 0% (8) to 10% 0% (8) to 3%
Sharper roll-off 6% to 7% 2%
(5% vs. 20%)
System Channel bonding StatMux gain (UHDTV-1): 7% to 12%
Layers (two or three Plus the possibility to cumulate spare capacity for additional
transponders) programs
VCM simulcast Up to 50% in geographical areas with a high level of atmospheric
(UHDTV-1 and SDTV) attenuation in Ku frequency band (A higher gain is expected in Ka
frequency band)

Broadband
For legacy (single beam) broadband networks, DVB-S2X fine-tunes both the physical and the upper protocol layers of
DVB-S2. However, the best expected performance gains are foreseen for future multi-spot beam systems especially in
the Ka-band multi-spot environment where aggressive spectrum re-use schemes together with advanced interference
mitigation techniques can significantly boost the capacity (up to 100%) when compared to more conventional systems.
This is made possible due to the new super-framing structure presented in DVB-S2X. In addition, by utilising the beam
hopping technology supported by the new super framing structure, it is envisioned that dynamic resource allocation
over the coverage area will be able to be greatly improved.

11
Reference DVB-S2 without (optional) 16APSK. In case of available SNR=12 dB, the DVB-S2X better MODCODs gain may increase 10% to 15%
(Figure 1).

Page 13 of 14
Professional Services
Professional services will benefit from finer MODCOD granularity allowing much more flexibility in system design
with regard to trading throughput, availability and SNR. This will allow for more the efficient deployment of DSNG
and contribution links and the more flexible usage of transponder capacity. Similarly, the lower roll-off filters will give
the flexibility to space carriers closer together, again, potentially increasing transponder usage.
VLSNR MODCODS and higher order constellation schemes extend both the low and high end SNR operating region
providing potential new application areas, particularly for VSAT (VLSNR) and high throughput trunk/contribution
links (high order constellation schemes).

Page 14 of 14