This file is raw output from pdftotext and may not be ideal for distribution. If you are a maintainer for Hackipedia, please sit down when you have time and clean this text version up. Source PDF: /mnt/main/jmc-storage/docs/DVB/TR 102 376 User guidelines for 2nd gen sys for Broadcasting, Interactive Serv, News Gathering other satellite app (ETSI) (DVB-S2) V1.1.1 (2005-02).pdf Like all conversions the text below should be fully readable as UTF-8 unicode text. --------------------------------------------------------------- ETSI TR 102 376 V1.1.1 (2005-02) Technical Report Digital Video Broadcasting (DVB) User guidelines for the second generation system for Broadcasting, Interactive Services, News Gathering and other broadband satellite applications (DVB-S2) European Broadcasting Union Union Européenne de Radio-Télévision EBU·UER 2 ETSI TR 102 376 V1.1.1 (2005-02) Reference DTR/JTC-DVB-166 Keywords broadband, broadcasting, digital, satellite, TV, video ETSI 650 Route des Lucioles F-06921 Sophia Antipolis Cedex - FRANCE Tel.: +33 4 92 94 42 00 Fax: +33 4 93 65 47 16 Siret N° 348 623 562 00017 - NAF 742 C Association à but non lucratif enregistrée à la Sous-Préfecture de Grasse (06) N° 7803/88 Important notice Individual copies of the present document can be downloaded from: http://www.etsi.org The present document may be made available in more than one electronic version or in print. In any case of existing or perceived difference in contents between such versions, the reference version is the Portable Document Format (PDF). In case of dispute, the reference shall be the printing on ETSI printers of the PDF version kept on a specific network drive within ETSI Secretariat. Users of the present document should be aware that the document may be subject to revision or change of status. Information on the current status of this and other ETSI documents is available at http://portal.etsi.org/tb/status/status.asp If you find errors in the present document, please send your comment to one of the following services: http://portal.etsi.org/chaircor/ETSI_support.asp Copyright Notification No part may be reproduced except as authorized by written permission. The copyright and the foregoing restriction extend to reproduction in all media. © European Telecommunications Standards Institute 2005. © European Broadcasting Union 2005. All rights reserved. TM TM TM DECT , PLUGTESTS and UMTS are Trade Marks of ETSI registered for the benefit of its Members. TM TIPHON and the TIPHON logo are Trade Marks currently being registered by ETSI for the benefit of its Members. TM 3GPP is a Trade Mark of ETSI registered for the benefit of its Members and of the 3GPP Organizational Partners. ETSI 3 ETSI TR 102 376 V1.1.1 (2005-02) Contents Intellectual Property Rights ................................................................................................................................5 Foreword.............................................................................................................................................................5 1 Scope ........................................................................................................................................................6 2 References ................................................................................................................................................6 3 Symbols and abbreviations.......................................................................................................................8 3.1 Symbols..............................................................................................................................................................8 3.2 Abbreviations .....................................................................................................................................................9 4 General description of the technical characteristics of the DVB-S2 system ..........................................11 4.1 Commercial requirements ................................................................................................................................11 4.1.1 Commercial Requirements for Broadcast Services.....................................................................................12 4.1.2 Commercial Requirements for Non-Broadcast Services ............................................................................13 4.1.3 Common Commercial Requirements..........................................................................................................14 4.2 Application scenarios .......................................................................................................................................14 4.3 System architecture ..........................................................................................................................................15 4.3.1 The system block diagram ..........................................................................................................................18 4.3.2 Reference performance ...............................................................................................................................19 4.3.2.1 Single carrier per transponder configuration.........................................................................................19 4.3.2.1.1 Sensitivity to satellite power amplifier characteristics ....................................................................23 4.3.2.1.2 Sensitivity to roll-off .......................................................................................................................25 4.3.2.2 Multiple carrier per transponder configuration .....................................................................................25 4.4 The backwards compatible modes....................................................................................................................26 4.4.1 Hierarchical modulations ............................................................................................................................27 4.5 Adaptive Coding and Modulation ....................................................................................................................29 4.5.1 ACM: the principles....................................................................................................................................29 4.5.2 Functional description of the DVB-S2 subsystem for ACM ......................................................................32 4.5.2.1 Specific subsystems for supporting ACM with MPEG-TS...................................................................34 4.5.3 DVB-S2 performance in ACM mode .........................................................................................................37 4.6 System configurations ......................................................................................................................................38 5 Broadcast applications............................................................................................................................38 5.1 SDTV broadcasting ..........................................................................................................................................39 5.2 SDTV and HDTV broadcasting with differentiated channel protection...........................................................39 5.3 Backwards Compatible services.......................................................................................................................39 5.3.1 Hierarchical modulations ............................................................................................................................40 6 Interactive applications...........................................................................................................................40 6.1 IP Unicast Services...........................................................................................................................................41 6.1.1 Single Generic Stream and ACM command...............................................................................................41 6.1.2 Multiple (Generic or Transport) Streams....................................................................................................43 6.1.3 Encapsulation efficiency of ACM modes ...................................................................................................45 6.1.4 Scheduling issues........................................................................................................................................48 6.2 Independent frames structure for Packetized streams with VCM/ACM ..........................................................50 6.2.1 Independent framing issues (applicable to MPEG-TS)...............................................................................50 6.2.2 Example slicing process..............................................................................................................................51 6.2.3 Specific cases..............................................................................................................................................51 7 Contribution services, data content distribution/trunking and other professional applications..............52 7.1 Distribution of multiple MPEG multiplexes to Digital Terrestrial TV Transmitters........................................52 7.2 DSNG and other professional applications ......................................................................................................52 7.2.1 DSNG bit rates and symbol rates................................................................................................................53 7.2.2 Phase noise recommendation......................................................................................................................53 7.2.3 Receiver filter mask ....................................................................................................................................53 7.2.4 DSNG carrier spacing.................................................................................................................................54 7.2.5 Link budget examples for DSNG ...............................................................................................................55 7.2.5.1 Generic Hypothesis ...............................................................................................................................55 ETSI 4 ETSI TR 102 376 V1.1.1 (2005-02) 7.2.5.2 DSNG Examples ...................................................................................................................................58 7.2.6 DSNG transmitting station identification ...................................................................................................61 7.2.7 DSNG Services using ACM .......................................................................................................................61 Annex A: Low Density Parity Check Codes ........................................................................................63 A.1 Structure of Parity Check Matrices of Standardized LDPC Codes ........................................................64 A.2 Description of Standardized LDPC Codes .............................................................................................65 A.3 Performance Results...............................................................................................................................66 Annex B: DVB-S2 Physical Layer Frame and pilot structure...........................................................69 B.1 Structured PLS code for Frame Synchronization...................................................................................69 B.2 Pilot Structure.........................................................................................................................................71 Annex C: Modem algorithms design and performance over typical satellite channels...................73 C.1 Modulator with Pre-Distortion ...............................................................................................................74 C.2 Clock Recovery ......................................................................................................................................76 C.3 Physical Layer Frame Synchronization..................................................................................................76 C.3.1 An algorithm for Frame Synchronization.........................................................................................................76 C.3.2 An Alternative Frame Synchronization Algorithm ..........................................................................................77 C.3.2.1 Acquisition procedure description ..............................................................................................................77 C.3.2.2 Performance Analysis .................................................................................................................................78 C.3.2.3 Acquisition parameters optimization ..........................................................................................................79 C.4 Carrier Frequency Recovery ..................................................................................................................80 C.5 Automatic Gain Control .........................................................................................................................82 C.6 Carrier Phase Recovery..........................................................................................................................82 C.6.1 Pilot-Aided Linear Interpolation ......................................................................................................................82 C.6.2 Fine Phase Recovery for High Order Modulations ..........................................................................................84 C.7 Performance Results...............................................................................................................................84 Annex D: Capacity assessment in ACM modes...................................................................................86 D.1 System Sizing Issues ..............................................................................................................................87 D.2 Methodology Description.......................................................................................................................87 D.3 Study Case Results .................................................................................................................................88 Annex E: Physical layer adaptation in ACM systems ........................................................................95 E.1 Channel estimator...................................................................................................................................95 E.2 Physical Layer Selector ..........................................................................................................................97 E.2.1 Shifted Threshold .............................................................................................................................................98 E.2.2 Hysteresis .........................................................................................................................................................99 E.3 Performance results ................................................................................................................................99 Annex F: ACM receiver implementation ..........................................................................................101 F.1 Type 1 receiver.....................................................................................................................................101 F.2 Type 2 receiver.....................................................................................................................................102 History ............................................................................................................................................................104 ETSI 5 ETSI TR 102 376 V1.1.1 (2005-02) Intellectual Property Rights IPRs essential or potentially essential to the present document may have been declared to ETSI. The information pertaining to these essential IPRs, if any, is publicly available for ETSI members and non-members, and can be found in ETSI SR 000 314: "Intellectual Property Rights (IPRs); Essential, or potentially Essential, IPRs notified to ETSI in respect of ETSI standards", which is available from the ETSI Secretariat. Latest updates are available on the ETSI Web server (http://webapp.etsi.org/IPR/home.asp). Pursuant to the ETSI IPR Policy, no investigation, including IPR searches, has been carried out by ETSI. No guarantee can be given as to the existence of other IPRs not referenced in ETSI SR 000 314 (or the updates on the ETSI Web server) which are, or may be, or may become, essential to the present document. Foreword This Technical Report (TR) has been produced by Joint Technical Committee (JTC) Broadcast of the European Broadcasting Union (EBU), Comité Européen de Normalisation ELECtrotechnique (CENELEC) and the European Telecommunications Standards Institute (ETSI). The work of the JTC was based on the studies carried out by the European DVB Project under the auspices of the Ad Hoc Group on DVB-S2 of the DVB Technical Module. This joint group of industry, operators and broadcasters provided the necessary information on all relevant technical matters (see clause 2). NOTE: The EBU/ETSI JTC Broadcast was established in 1990 to co-ordinate the drafting of standards in the specific field of broadcasting and related fields. Since 1995 the JTC Broadcast became a tripartite body by including in the Memorandum of Understanding also CENELEC, which is responsible for the standardization of radio and television receivers. The EBU is a professional association of broadcasting organizations whose work includes the co-ordination of its members' activities in the technical, legal, programme-making and programme-exchange domains. The EBU has active members in about 60 countries in the European broadcasting area; its headquarters is in Geneva. European Broadcasting Union CH-1218 GRAND SACONNEX (Geneva) Switzerland Tel: +41 22 717 21 11 Fax: +41 22 717 24 81 Founded in September 1993, the DVB Project is a market-led consortium of public and private sector organizations in the television industry. Its aim is to establish the framework for the introduction of MPEG-2 based digital television services. Now comprising over 200 organizations from more than 25 countries around the world, DVB fosters market-led systems, which meet the real needs, and economic circumstances, of the consumer electronics and the broadcast industry. ETSI 6 ETSI TR 102 376 V1.1.1 (2005-02) 1 Scope The present document gives an overview of the technical and operational issues relevant to the system specified in EN 302 307 [2] "Digital Video Broadcasting (DVB): Second generation framing structure, channel coding and modulation systems for Broadcasting, Interactive Services, News Gathering and other broadband satellite applications", including service quality and link availability evaluation for typical DSNG and fixed contribution links, with the purpose to facilitate its interpretation. 2 References For the purposes of this Technical Report (TR), the following references apply: [1] ETSI EN 300 421: "Digital Video Broadcasting (DVB); Framing structure, channel coding and modulation for 11/12 GHz satellite services". [2] ETSI EN 302 307: "Digital Video Broadcasting (DVB); Second generation framing structure, channel coding and modulation systems for Broadcasting, Interactive Services, News Gathering and other broadband satellite applications". [3] ISO/IEC 13818 (parts 1 and 2): " Information technology - Generic coding of moving pictures and associated audio information". [4] ETSI EN 301 210: "Digital Video Broadcasting (DVB); Framing structure, channel coding and modulation for Digital Satellite News Gathering (DSNG) and other contribution applications by satellite". [5] ETSI TR 101 154: " Digital Video Broadcasting (DVB); Implementation guidelines for the use of Video and Audio Coding in Broadcasting Applications based on the MPEG-2 Transport Stream". [6] ETSI EN 300 468: "Digital Video Broadcasting (DVB); Specification for Service Information (SI) in DVB systems". [7] ETSI EN 301 192: "Digital Video Broadcasting (DVB); DVB specification for data broadcasting". [8] ETSI EN 300 429: "Digital Video Broadcasting (DVB); Framing structure, channel coding and modulation for cable systems". [9] ETSI TR 101 221: "Digital Video Broadcasting (DVB); User guideline for Digital Satellite News Gathering (DSNG) and other contribution applications by satellite". [10] U. Reimers, A. Morello: "DVB-S2, the second generation standard for satellite broadcasting and unicasting", International Journal on Satellite Communication Networks, 2004; 22. [11] R. Gallager: "Low Density Parity Check Codes", IRE Trans. on Info. Theory, January 1962. [12] M. Eroz, F.-W. Sun and L.-N. Lee: "DVB-S2 Low Density Parity Check Codes with near Shannon Limit Performance", International Journal on Satellite Communication Networks, 2004; 22. [13] E. Casini, R. De Gaudenzi, A. Ginesi: "DVB-S2 modem algorithms design and performance over typical satellite channels", International Journal on Satellite Communication Networks, 2004; 22. [14] F.-W. Sun Y. Jiang and L.-N. Lee: "Frame synchronization and pilot structure for DVB-S2" International Journal on Satellite Communication Networks, 2004; 22. [15] R. Rinaldo, M. Vazquez-Castro, A. Morello: "DVB-S2 ACM modes for IP and MPEG unicast applications", International Journal on Satellite Communication Networks, 2004; 22. [16] S. Cioni, R. De Gaudenzi, R. Rinaldo: "Adaptive Coding and Modulation for the Forward Link of Broadband Satellite Networks" in the Proc. of the Globecom 2003 Conference, San Francisco, Dec 2003. ETSI 7 ETSI TR 102 376 V1.1.1 (2005-02) [17] E. Chen, J. L. Koslov, V. Mignone, J. Santoru: "DVB-S2 Backward-compatible modes: a Bridge Between the Present and the Future", International Journal on Satellite Communication Networks, 2004; 22. [18] R. Rinaldo, R. De Gaudenzi: "Capacity analysis and system optimization for the forward link of multi-beam satellite broadband systems exploiting adaptive coding and modulation", International Journal on Satellite Communication Networks, 2004; 22. [19] U. Reimers (ed.): "Digital Video Broadcasting - The DVB Family of Standards for Digital Television", 2nd ed., 2004, Springer Publishers, New York, ISBN 3-540-43545-X. [20] M.A. Vazquez-Castro et Al.: "Scheduling issues in ACM DVB-S2 systems: performance assessment through a comprehensive OPNET simulator", under preparation. [21] C.E. Gilchriest: "Signal to Noise Monitoring" JPL Space Programs Summary, No 37-27, Vol IV, pp 169-176. [22] D. J. MacKay and R. M. Neal: "Good codes based on very sparse matrices", 5th IMA Conf. 1995, pp.100-111. [23] D. J. MacKay and R. M. Neal: "Near Shannon limit performance of low density parity check codes", Electronics Lett. Mar. 1997, vol. 33, no.6, pp. 457-458. [24] T. Richardson and R. Urbanke: "Efficient encoding of low-density parity check codes", IEEE Trans. Info. Theory, vol. 47, pp.638-656, Feb. 2001. [25] T. Richardson, A. Shokrollahi and R. Urbanke: "Design of capacity approaching irregular low density parity check codes", IEEE Trans. Inform. Theory, Feb. 2001, vol. 47, pp. 619-637. [26] U. Mengali and A.N. D'Andrea: "Synchronization Techniques for Digital Receivers", Plenum Press, New York, USA, 1997. [27] R. De Gaudenzi, A. Guillen i Fabregas, A. Martinez Vicente: "Turbo-coded APSK Modulations for Satellite Broadcasting and Multicasting- Part I: Coded Modulation Design", submitted to IEEE Trans. On Wireless Communications 2004. [28] R. De Gaudenzi, A. Guillen i Fabregas, A. Martinez Vicente: "Turbo-coded APSK Modulations for Satellite Broadcasting - Part II: End-to-End Performance" submitted to IEEE Trans. On Wireless Communications 2004. [29] G. Karam and H. Sari: "A Data Pre-distortion Technique with Memory for QAM Radio Systems", IEEE Transactions on Communications, Vol. COM-39, No. 2, pp 336- 344, February 1991. [30] R. De Gaudenzi and M. Luise: "Design and Analysis of an All-Digital Demodulator for Trellis Coded 16-QAM Transmission over a Non-linear Satellite Channel", IEEE Trans. on Comm., Vol. 43, No. 2/3/4, February/March/April 1995, part I. [31] F. M. Gardner: "A BPSK/QPSK timing-error detector for sampled receivers", IEEE Trans. On Communications, vol. COM-34, no. 5, May 1986. [32] M. Luise and R. Reggiannini: "Carrier Frequency Recovery in All Digital Modems for Burst Mode Transmissions", IEEE Transactions on Communications, COM-43, pp. 1169-1178, Feb./March/Apr. 1995. [33] M. Oerder and H. Meyr: "Digital Filter and Square Timing Recovery", IEEE Trans. Commun., COM-36, May 1988. [34] Jack K.Holmes: "Coherent spread spectrum systems", Wiley Interscience, pp 395 to 426. [35] A. Morello, V. Mignone: "DVB-S2 ready to lift-off", IBC'04 Conference, Amsterdam, 9-13 September, 2004. [36] D.R. Pauluzzi and N.C. Beaulieu: "A Comparison of SNR Estimation Techniques for the AWGN channel", IEEE Trans. Comm., vol. 48, pp. 1681-1691, Oct. 2000. ETSI 8 ETSI TR 102 376 V1.1.1 (2005-02) [37] L. Castanet et al.: "Comparison of Various Methods for Combining Propagation Effects and Predicting Loss in Low-Availability Systems in the 20-50 GHz Frequency Range", Int. Journ. on Sat. Comm., Vol. 19, pp. 317-334, 2001. [38] E. Casini: "DVB-S2 end-to-end performance with linearized and non-linearized TWT amplifiers" ESA document Ref. TEC-ETC/2004.84/EC/ec. [39] F. J. Williams, N.J.A. Sloane: "The Theory of error correction coding", Elsevier, New York, 1977. [40] ETSI EN 300 744: "Digital Video Broadcasting (DVB); Framing structure, channel coding and modulation for digital terrestrial television". [41] ETSI EN 301 790: "Digital Video Broadcasting (DVB); Interaction channel for satellite distribution systems". [42] ETSI ETS 300 801: "Digital Video Broadcasting (DVB); Interaction channel through Public Switched Telecommunications Network (PSTN)/ Integrated Services Digital Networks (ISDN)". [43] ETSI EN 301 195: "Digital Video Broadcasting (DVB); Interaction channel through the Global System for Mobile communications (GSM)". [44] ETSI ES 200 800: "Digital Video Broadcasting (DVB); DVB interaction channel for Cable TV distribution systems (CATV)". [45] ITU-R Recommendation SNG.770-1: "Uniform operational procedures for satellite news gathering (SNG)". 3 Symbols and abbreviations 3.1 Symbols For the purposes of the present document, the following symbols apply: α Roll-off factor BS Bandwidth of the frequency Slot allocated to a service c codeword C/N Carrier-to-noise power ratio (N measured in a bandwidth equal to symbol rate) C/N+I Carrier-to-(Noise + Interference) ratio DFL Data Field Length Eb/N0 Ratio between the energy per information bit and single sided noise power spectral density Es/N0 Ratio between the energy per transmitted symbol and single sided noise power spectral density Es/(N0 +I0) Ratio between the energy per transmitted symbol and single sided noise plus interference power spectral density Φ Antenna diameter ψ MPEG (L ) MPEG encapsulation efficiency ψ DVB − S 2 (L ) Overall DVB-S2 encapsulation efficiency ψ MS (η , L ) DVB-S2 Mode and Stream adaptation efficiency ψ framing (η ) DVB-S2 Physical layer framing efficiency i LDPC code information block i0 , i1 ,..., ikldpc −1 LDPC code information bits H ( N − K )× N LDPC code parity check matrix I, Q In-phase, Quadrature phase components of the modulated signal KBCH number of bits of BCH uncoded Block NBCH number of bits of BCH coded Block kldpc number of bits of LDPC uncoded Block nldpc number of bits of LDPC coded Block ETSI 9 ETSI TR 102 376 V1.1.1 (2005-02) η Spectral efficiency ηc code efficiency ηMOD number of transmitted bits per constellation symbol L IP packet length m BCH code information word m(x) BCH code message polynomial p0 , p1 ,... p nldpc − kldpc −1 LDPC code parity bits q code rate dependant constant for LDPC codes θ deviation angle in hierarchical constellations ρ roll-off rm In-band ripple (dB) Rs Symbol rate corresponding to the bilateral Nyquist bandwidth of the modulated signal Ru Useful bit rate at the DVB-S2 system input S Number of Slots in a XFECFRAME Ts Symbol period Tloop loop delay Tprop propagation time Tq waiting time UPL User Packet Length TST Threshold on SOF in Tentative state TPT Threshold on PLSCODE in Tentative state TSL Threshold on SOF in Locked state 3.2 Abbreviations For the purposes of the present document, the following abbreviations apply: ACM Adaptive Coding and Modulation AGC Automatic Gain Control AVC Advanced Video Coding AWGN Additive White Gaussian Noise BB Baseband BER Bit Error Ratio BC Backwards-Compatible NOTE: Referred to the system allowing partial stream reception by DVB-S receivers. NBC Non-Backwards-Compatible BCH Bose-Chaudhuri-Hocquenghem multiple error correction binary block code BS Broadcast Service BW Bandwidth (at -3 dB) of the transponder CBR Constant Bit Rate CCM Constant Coding and Modulation CRC Cyclic Redundancy Check D Decimal notation DAGC Digital AGC DA-VT AGC Data-Aided version of the Vector-Tracker Automatic Gain Control DD Decision Directed DNP Deleted Null Packets DSNG Digital Satellite News Gathering DTH Direct To Home DTT Digital Terrestrial Television DTV Digital TeleVision DVB Digital Video Broadcasting project DVB-S DVB System for satellite broadcasting as specified in EN 300 421 DVB-S2 DVB-S2 System as specified in EN 302 307 DEMUX DEMUltipleXer EBU European Broadcasting Union ETSI 10 ETSI TR 102 376 V1.1.1 (2005-02) EN European Norm ETS European Telecommunication Standard FDM Frequency Division Multiplex FEC Forward Error Correction FF Feedforward FIFO First In First Out GS Generic Stream HDTV High Definition Television HP High Priority HPA High Power Amplifier IBO Input Back Off IF Intermediate Frequency IMUX Input MUltipleXer - Filter IRD Integrated Receiver Decoder IS Interactive Services ISCR Input Stream Clock Reference ISI InterSymbol Interference ITU International Telecommunications Union LDPC Low Density Parity Check (codes) LG Low Gain LNB Low Noise Block LP Low Priority ML Maximum Likelihood MPEG Moving Pictures Experts Group MPE Multi-Protocol Encapsulation MUX Multiplex NA Not Applicable NCO Numerically Controlled Oscillator NG Nominal Gain OBO Output Back Off OMUX Output Multiplexer - Filter NP Null Packets PER (MPEG TS) Packet Error Rate PID Packet Identifier PFA Probability of False Alarm PND Probability of Non Detection PL Physical Layer PLL Phase-Locked Loop PLS Physical Layer Signalling PLSCODE PLS code PS Professional Services PSD Power Spectral Density PSK Phase Shift Keying QEF Quasi-Error-Free QoS Quality of Service QPSK Quaternary Phase Shift Keying RF Radio Frequency r.m.s. root mean square RX Receiver RR Round Robin SDTV Standard Definition Television SNG Satellite News Gathering SNR Signal to Noise Power Ratio SMATV Satellite Master Antenna TeleVision SOF Start of Frame SRRC Square Root Raised Cosine Filter TDM Time Division Multiplex TS Transport Stream TV Television TX Transmitter TWTA Travelling Wave Tube Amplifier UW Unique Word ETSI 11 ETSI TR 102 376 V1.1.1 (2005-02) VCM Variable Coding and Modulation 16APSK 16-ary Amplitude and Phase Shift Keying 32APSK 32-ary Amplitude and Phase Shift Keying 8PSK 8-ary Phase Shift Keying 4 General description of the technical characteristics of the DVB-S2 system DVB-S2 is the second-generation DVB specification for broadband satellite applications, developed on the success of the first generation specifications, DVB-S for broadcasting and DVB-DSNG for satellite news gathering and contribution services, benefiting from the technological achievements of the last decade. It has been designed for: • Broadcast Services for standard definition TV and HDTV. • Interactive Services including Internet Access for consumer applications. • Professional Applications, such as Digital TV contribution and News Gathering, TV distribution to terrestrial VHF/UHF transmitters, Data Content distribution and Internet Trunking. The DVB-S2 standard has been specified around three key concepts: best transmission performance, total flexibility and reasonable receiver complexity. To achieve the best performance-complexity trade-off, DVB-S2 benefits from more recent developments in channel coding (adoption of LDPC codes) and modulation (use of QPSK, 8PSK, 16APSK and 32APSK). The result is typically a 30 % capacity increase over DVB-S under the same transmission conditions. In addition, for broadcast applications, DVB-S2 is not constrained to the use of QPSK and therefore it can deliver significantly higher bit rates over high power satellites, thus still increasing capacity gain with respect to DVB-S. Furthermore, when used for interactive point-to- point applications like IP unicasting, the gain of DVB-S2 over DVB-S is even greater: Variable Coding and Modulation (VCM) functionality allows different modulations and error protection levels to be used and changed on a frame-by-frame basis. This may be combined with the use of a return channel to achieve closed-loop Adaptive Coding and Modulation (ACM), thus allowing the transmission parameters to be optimized for each individual user, dependant on its own link conditions. DVB-S2 is so flexible that it can cope with any existing satellite transponder characteristics, with a large variety of spectrum efficiencies and associated C/N requirements. Furthermore, it is not limited to MPEG-2 video and audio source coding, but it is designed to handle a variety of audio-video and data formats including formats which the DVB Project is currently defining for future applications. DVB-S2 accommodates any input stream format, including continuous bit-streams, single or multiple MPEG Transport Streams, IP as well as ATM packets. This future proofing will allow other current and future data schemes to be used without the need for a new specification. It is based on the "tool-kit" approach that allows to cover all the application areas while still keeping the single-chip decoder at reasonable complexity levels, thus enabling the use of mass market products also for professional applications. Backwards compatible modes are available, allowing existing DVB-S services and set-top-boxes to continue working during any transitional period. When there is no problem of legacy receivers, NBC modes offer the full benefits of DVB-S2. 4.1 Commercial requirements (Extracted from doc. DVB-BSS17rev1 "Advanced Coding and Modulation Schemes for Broadband Satellite Services - Commercial Requirements"). The DVB-S standard was developed primarily with unidirectional broadcast applications in mind, but has been adopted for other purposes, such as point-to-point data transmission. One of the reasons for this is the availability of inexpensive receive silicon as a result of the high volume broadcast receiver market. The Commercial Module of DVB (DVB-CM) foresaw a similar process for the new system, where the volume driver is expected to remain broadcast applications. ETSI 12 ETSI TR 102 376 V1.1.1 (2005-02) To avoid confusion in the distinction between broadcast and non-broadcast applications, with increasing provision of entertainment services over IP networks, and increasing use of interactivity with TV, DVB-CM introduced the following definitions: • Broadcast services are defined as TV, radio and associated data (e.g. teletext, EPG, etc.) in contribution (e.g. DSNG), distribution (e.g. SMATV and cable feeds) and direct to home applications. In the case of interactive services, it is intended only the forward path. • Non-broadcast services are defined as point-to-point and point-to-multipoint data services. 4.1.1 Commercial Requirements for Broadcast Services Broadcast services are characterized by having a large coverage area and providing audio-visual and data services to an extensive base of similar reception systems (both antennas and receivers) with a high degree of availability. For broadcasters there are various reasons to use higher order modulation and/or advanced coding schemes, including: • Increased data throughput in a given bandwidth. • Increased availability through improved link margin. • Increased coverage area. A key factor for many established broadcasters is the issue of backwards-compatibility. Large populations of DVB-S receivers in the field must continue to provide service to customers for at least several years. This is particularly important where there is a subsidy. Backwards-compatible modulation systems that allow DVB-S receivers to continue operating, while providing additional capacity and services to new, advanced receivers, are seen as the only commercially viable way forward for some operators. Backwards-compatible systems however suffer from two disadvantages: • Compatibility will cause the overall performance to fall short of that achievable by non backwards-compatible systems. • There will be some performance penalty in the behaviour of existing QPSK receivers. Note that some operators are reluctant to accept even a slight performance penalty, as this increases service call-outs and churn. On the basis of the above considerations, DVB-CM concluded that the technical specification should provide for two approaches: • A non backwards-compatible scheme, intended for use in systems requiring the highest efficiency, and not requiring that transmissions should be receivable by existing receiver populations. • A backwards-compatible scheme that can be received by existing receiver populations, but provides additional capacity to enhanced receivers. This should have the capability of migrating to a more efficient non backwards-compatible mode once all DVB-S receivers have been replaced. Furthermore the technical specification should also take into account the following: • A bit-rate increase of at least 35 % over DVB-S should be achieved. • Service availability targets remain no more than one uncorrected error per hour. • For DVB-S2 backwards-compatible transmissions, there should be no requirement for any change to existing DVB-S receivers, including the antenna and LNB. This assumes continued use of the same transponder. • For DVB-S2 non backwards-compatible transmissions, reliable reception should be possible using receive antenna diameters in the range 0,4 m to 0,8 m. Due account shall be taken of anticipated satellite system characteristics (e.g. power, bandwidth, interference) over the next ten years. • While the technical specification is concerned with the transmission format, it is important that this does not impose an undue burden on receiver costs. New receiver silicon, enabling multi-mode reception, including DVB-S, backwards and non-backwards compatible DVB-S2, in volume should cost no more than 15 % more than current DVB-S devices. ETSI 13 ETSI TR 102 376 V1.1.1 (2005-02) 4.1.2 Commercial Requirements for Non-Broadcast Services In general, DVB-CM stated that characteristics that are desirable exclusively for non-broadcast services should only be included in the base specification if the burden of cost to the broadcast receiver is negligible. The specification should be available for consideration as an alternative forward path for DVB-RCS Ku- and Ka-band systems and other data systems currently using DVB-S. In the future non-broadcast two-way services will need to take advantage of such techniques as adaptive modulation, adaptive coding and adaptive power systems. The services provided will include: • Point-to-point services (e.g. IP-backbone). • Point-to-multipoint services (e.g. VPN services). • Two way mass market services (e.g. Internet access via satellite). The above services are differentiated from broadcast services by the following requirements and characteristics: • The possibility of targeting particular receivers with particular content within the common transport system. It is therefore not necessary to be able to decode the entire data stream at a typical user terminal. • The possibility of establishing different quality of service targets for services to different customers in different areas. • Receiver network volumes may not be as high as broadcast applications, but could benefit from inexpensive silicon arising from broadcast applications. • Satellite capacity is generally the most expensive part of a link, always more expensive compared to the equipment behind it. • Business users may be able to accept significantly larger receive antennas that residential users. In such cases it would be desirable to use this to allow greater user data rates to be transmitted. The technical specifications shall additionally take into account the following requirements: • Non-broadcast services should be receivable on antennas down to typical consumer sizes. • A wide range of system parameters should be available to address applications across consumer to business antenna sizes, and telecom to broadcast satellite powers. • As in the case of broadcast applications, it should be possible to manufacture low cost receivers. If the impact of advanced features such as adaptive coding and modulation that are not relevant to broadcast applications is to raise broadcast receiver costs excessively, then these features could exist in an enhanced profile, available as an option. • The technology shall aim to optimize the use of the transponder (keeping all other technical parameters the same, a minimum capacity increase of 100 % shall be targeted for the professional, non-broadcast market case using adaptive technologies). • The target BER shall be at least as good as for broadcast applications. • The system shall allow for power control technologies, adaptive modulation and adaptive coding. ETSI 14 ETSI TR 102 376 V1.1.1 (2005-02) 4.1.3 Common Commercial Requirements The primary goal of new broadband satellite systems is to deliver a significantly higher net data rate in a given transponder bandwidth than the current DVB-S standard. Technical specifications for such a new standard shall not prevent operators from taking into account all issues covered by local, national, and international laws, especially those related to security (protection of personal data, encryption of data and services). The new specification will cover transmit-end functions only, but take into account the consequent cost of receive silicon. The market will determine what features are actually implemented in receive silicon. For the protection of existing business, the current DVB-S standard shall not be modified, nor shall changes to other standards cause any existing feature to become invalid. The new specification will contain a form of modulation and coding that is compatible with existing DVB-S receivers ("backwards compatible"), and if it offers greater efficiency, a non-backwards compatible form also. The new specification will contain a range of options for coding that are appropriate to a wide range of applications. The new specifications must be application neutral and media content independent. The specifications must be transmission frequency neutral, or contain the elements allowing for an adaptation to the specifics of certain frequency ranges (e.g. Ka Band). Specifications for broadcast and non-broadcast applications shall be provided by the definition of a limited number of profiles. The specification shall not prevent the use of any kind of scrambling and security system at the transport layer. 4.2 Application scenarios The DVB-S2 system has been optimized for the following broadband satellite application scenarios. • Broadcast Services (BS): digital multi-programme Television (TV)/High Definition Television (HDTV) broadcasting services. DVB-S2 is intended to provide Direct-To-Home (DTH) services for consumer Integrated Receiver Decoder (IRD), as well as collective antenna systems (Satellite Master Antenna Television - SMATV) and cable television head-end stations (possibly with remodulation, see [8]). DVB-S2 may be considered a successor to the current DVB-S standard [1], and may be introduced for new services and allow for a long-term migration. BS services are transported in MPEG Transport Stream format. VCM may be applied on multiple transport stream to achieve a differentiated error protection for different services (TV, HDTV, audio, multimedia). Two modes are available: - Non Backwards Compatible Broadcast Services (NBC-BS), not backwards-compatible with [1]. - Backwards-Compatible Broadcast Services (BC-BS), backwards-compatible with [1]. In fact, with a large number of DVB-S receivers already installed, backwards compatibility may be required for a period of time, where old receivers continue to receive the same capacity as before, while the new DVB-S2 receivers could receive additional capacity broadcasts. When the complete receiver population has migrated to DVB-S2, the transmitted signal can be modified to a non-backward compatible mode, thus exploiting the full potential of DVB-S2. To facilitate the reception of DVB-S services by DVB-S2 receivers, implementation of DVB-S in DVB-S2 chips is highly recommended. • Interactive Services (IS): interactive data services including internet access DVB-S2 is intended to provide interactive services to consumer IRDs and to personal computers, where DVB-S2's forward path supersedes the current DVB-S standard [1] for interactive systems. No recommendation is included in the DVB-S and DVB-S2 standards as far as the return path is concerned. Therefore, interactivity can be established either via terrestrial connection through telephone lines, or via satellite. DVB offers a variety of return link specifications, such as for example DVB-RCS (EN 301 790 [41]), DVB-RCP (ETS 300 801 [42]), DVB-RCG (EN 301 195 [43]), DVB-RCC (ES 200 800 [44]). Data services are transported in (single or multiple) Transport Stream format according to [7] (e.g. using Multiprotocol Encapsulation), or in (single or multiple) generic stream format. DVB-S2 can provide Constant Coding and Modulation (CCM), or Adaptive Coding and Modulation (ACM), where each individual satellite receiving station controls the protection mode of the traffic addressed to it. ETSI 15 ETSI TR 102 376 V1.1.1 (2005-02) • Digital TV Contribution and Satellite News Gathering (DTVC/DSNG) Digital television contribution applications by satellite consist of point-to-point or point-to-multipoint transmissions, connecting fixed or transportable uplink and receiving stations. They are not intended for reception by the general public. According to ITU-R Recommendation SNG.770-1 [45], SNG is defined as "Temporary and occasional transmission with short notice of television or sound for broadcasting purposes, using highly portable or transportable uplink earth stations, etc.". Services are transported in single (or multiple) MPEG Transport Stream format. DVB-S2 can provide Constant Coding and Modulation (CCM), or Adaptive Coding and Modulation (ACM). In this latter case, a single satellite receiving station typically controls the protection mode of the full multiplex. • Data content distribution/trunking and other professional applications (PS) These services are mainly point-to-point or point-to-multipoint, including interactive services to professional head-ends, which re-distribute services over other media. Services may be transported in (single or multiple) generic stream format. The system can provide Constant Coding and Modulation (CCM), Variable Coding and Modulation (VCM) or Adaptive Coding and Modulation (ACM). In this latter case, a single satellite receiving station typically controls the protection mode of the full TDM multiplex, or multiple receiving stations control the protection mode of the traffic addressed to each one. In either case, interactive or non-interactive, the present document is only concerned with the forward broadband channel. For all these applications, DVB-S2 benefits from more recent developments in channel coding and modulation, achieving typically a 30 % capacity increase over DVB-S [1]. When used for point-to-point applications like IP unicasting or DSNG, the gain of DVB-S2 is even greater. Adaptive Coding and Modulation (ACM) functionality allows different modulation formats and error protection levels (i.e. coding rates) to be used and changed on a frame-by-frame basis within the transmitted data stream. By means of a return channel, informing the transmitter of the actual receiving condition, the transmission parameters may be optimized for each individual user, dependant on path conditions. Furthermore DVB-S2 is compatible with Moving Pictures Experts Group (MPEG-2 and MPEG-4) coded TV services [3], with a Transport Stream packet multiplex. Multiplex flexibility allows the use of the transmission capacity for a variety of TV service configurations, including sound and data services. All service components are Time Division Multiplexed (TDM) on a single digital carrier. 4.3 System architecture To achieve the best performance, DVB-S2 is based on LDPC (Low Density Parity Check) codes, simple block codes with very limited algebraic structure, discovered by R. Gallager in 1962 [11]. LDPC codes have an easily parallelizable decoding algorithm which consists of simple operations such as addition, comparison and table look-up [22], [23]; moreover the degree of parallelism is "adjustable" which makes it easy to trade-off throughput and complexity (see note 1). NOTE 1: The maximum decoder complexity was set to correspond to 14 mm2 of silicon using a 0,13 µm technology, and the reference symbol rate was 55 Mbaud. Their key characteristics, allowing quasi-error free operation at only 0,6 to 1,2 dB from the Shannon limit [12], are: • the very large LDPC code block length (64 800 bits for the normal frame, and 16 200 bits for the short frame); • the large number of decoding iterations (around 50 SISO iterations); • the presence of a concatenated BCH outer code (without any interleaver), defined by the designers as a "cheap insurance against unwanted error floors at high C/N ratios". In comparison, the DVB-S and DVB-DSNG soft-decision Viterbi decoder takes decisions on blocks of only 100 symbols, without iterations, and the RS code over blocks of about 1 600 bits (interleaving factor 12), offering already quite good performance (see figure 3), around 3 dB from the Shannon limit. ETSI 16 ETSI TR 102 376 V1.1.1 (2005-02) Digital transmissions via satellite are affected by power and bandwidth limitations. Therefore DVB-S2 provides for many transmission modes (FEC coding and modulations), giving different trade-offs between power and spectrum efficiency. Code rates of 1/4, 1/3, 2/5, 1/2, 3/5, 2/3, 3/4, 4/5, 5/6, 8/9 and 9/10 are available depending on the selected modulation and the system requirements. Coding rates 1/4, 1/3 and 2/5 have been introduced to operate, in combination with QPSK, under exceptionally poor link conditions, where the signal level is below the noise level. Computer simulations demonstrated the superiority of such modes over BPSK modulation combined with code rates 1/2, 2/3 and 4/5. The introduction of two FEC code block length (64 800 and 16 200) was dictated by two opposite needs: the C/N performance improves for long block lengths, but the end-to-end modem latency increases as well. Therefore for applications not critical for delays (such as for example broadcasting) the long frames are the best solution, while for interactive applications a shorter frame may be more efficient when a short information packet has to be forwarded immediately by the transmitting station. Four modulation modes can be selected for the transmitted payload (see figure 1). Q Q I=MSB Q=LSB 100 110 10 00 000 ρ= ρ=1 I 010 φ=π/4 I 001 11 011 01 101 111 (a) QPSK (b) 8PSK Q Q 01101 11101 01001 1010 1000 01100 11001 R3 00101 00001 0010 0000 MSB R2 11100 00100 00000 01000 LSB R2 0110 1110 R1 1100 0100 10100 10101 R1 10001 10000 11110 11000 text I text I 0111 1111 10110 10111 10010 1101 0101 10011 01110 00110 00010 11010 0011 0001 11111 00111 00011 01010 1011 1001 01111 11011 01011 (c) 16APSK (d) 32APSK Figure 1: The four possible DVB-S2 constellations before physical layer scrambling QPSK and 8PSK are typically proposed for broadcast applications, since they are virtually constant envelope modulations and can be used in non-linear satellite transponders driven near saturation. For some specific broadcasting applications (i.e. regional spot beams) and interactive application operating with multi-beam satellites, 16APSK provides extra spectral efficiency with very limited linearity requirements if proper pre-distortion schemes are employed. 32APSK modes, mainly targeted to professional applications, can also be used for broadcasting, but these require a higher level of available C/N and the adoption of advanced pre-distortion methods in the up-link station to minimize the effect of transponder non-linearity. Whilst these modes are not as power efficient as the other modes, the data throughput is much greater. 16APSK and 32APSK constellations have been optimized to operate over a non-linear transponder by placing the points on circles. Nevertheless their performances on a linear channel are comparable with those of 16QAM and 32QAM respectively. All the modes are also appropriate for operation in quasi-linear satellite channels, in multi-carrier Frequency Division Multiplex (FDM) type applications. By selecting the modulation constellation and code rates, spectrum efficiencies from 0,5 to 4,5 bit/second/Hz are available and can be chosen dependant on the capabilities and restrictions of the satellite transponder used. ETSI 17 ETSI TR 102 376 V1.1.1 (2005-02) DVB-S2 also features the presence of a Physical Layer (PL) scrambler that is X-oring the I-Q modulator symbols (inclusive of the optional pilot symbols but excluding the PL header) with a complex binary randomization sequence of length truncated to the current PLFRAME duration. The complex randomization sequence has an original period of 262 143 symbols and is the one used in the terrestrial UMTS standard. It provides good auto and cross-correlation properties and allows to simply generate up to 262 142 distinct complex sequences. The main advantages of the physical layer randomization presence in DVB-S2 are: • capability to uniquely "sign" individual carriers present in a multi-carrier multi channel transponder; • randomization of periodic pilot symbols pattern when pilot is time interleaved in the carrier; • randomization of other satellite or same satellite other beams interference. It should be remarked that in case of interfering signals with lower baud rate than the useful carrier the physical layer descrambling present in the DVB-S2 demodulator will make interferer appearing as wideband interference thus reducing their degradation effect; • the physical layer randomization also allows the application of repetition coding at the DVB-S2 modulator to further increase the C/(N+I) operating range of the system. DVB-S2 has three roll-off factor choices to determine spectrum shape. These are α=0,35 as in DVB-S and two others, namely α=0,25, α=0,20 for tighter bandwidth shape restriction. DVB-S2 is suitable for use on different satellite transponder bandwidths and frequency bands. The symbol rate is matched to given transponder characteristics, and, in the case of multiple carriers per transponder (FDM), to the frequency plan adopted. Two levels of framing structures have been designed: • the first at physical level, carrying few highly-protected signalling bits; • the second at base-band level, carrying a variety of signalling bits, to allow the maximum flexibility on the input signal adaptation. Physical Level framing The first level of framing structure has been designed to provide robust synchronization and signalling at physical layer [14]. Thus a receiver may synchronize (carrier and phase recovery, frame synchronization) and detect the modulation and coding parameters before demodulation and FEC decoding. The DVB-S2 physical layer "train" is composed of a regular sequence of periodic "wagons" (physical layer frames, PL Frame): within a wagon, the modulation and coding scheme is homogeneous, but may change (Variable Coding and Modulation) in adjacent wagons. The PL framing structure is application independent (Constant Coding and Modulation or Variable Coding and Modulation). Every PL Frame is composed of: • a payload of 64 800 bits (normal FEC frame) or 16 200 bits (short FEC frame), generated by encoding the user bits according to the selected FEC scheme; thus the payload corresponds to a code block of the concatenated LDPC/BCH FEC; • a PL Header, containing synchronization and signalling information: type of modulation and FEC rate, frame length, presence/absence of pilot symbols to facilitate synchronization. The PL-Header is always composed of 90 symbols (using a fixed Π/2 binary modulation), and the payload is always composed of an integer multiple of 90 symbols (excluding pilot symbols). Since the PL Header is the first entity to be decoded by the receiver, it could not be protected by the powerful LDPC/BCH FEC scheme. On the other hand, it had to be perfectly decodable under the worst-case link conditions. Therefore designers selected a very low-rate 7/64 block code, suitable for soft-decision correlation decoding, and minimized the number of signalling bits to reduce decoding complexity and global efficiency loss. For example, assuming a 64 800 bit frame, the worst case efficiency of the PL Frame is 99,3 % (excluding pilot symbols). ETSI 18 ETSI TR 102 376 V1.1.1 (2005-02) Base-band Level framing Another level of framing structure, the "baseband frame", allows a more complete signalling functionality to configure the receiver according to the application scenarios: single or multiple input streams, generic or transport stream, CCM (Constant Coding and Modulation) or ACM (Adaptive Coding and Modulation). Thanks to the LDPC/BCH protection and the wide length of the FEC frame, the Baseband (BB) Header may contain many signalling bits (80) without loosing neither transmission efficiency nor ruggedness against noise. This BB-Header carries other important signalling information, such as: labelling the modulator input streams, describing the position and characteristics of user packets, indicating the presence of padding bits in the transmitted "baseband frame", signalling the activation of specific tools (Null-packet deletion function, Input Stream Synchronization function, as described in [14]), signalling the adopted modulation roll-off (see note 2). NOTE 2: The roll-off factor needs not be signalled at physical layer, since (sub-optimum) reception is possible even assuming unknown roll-off. 4.3.1 The system block diagram The DVB-S2 System is composed of a sequence of functional blocks as described in figure 2. The block identified as "Mode Adaptation" is application dependent. Input sequences may be single or multiple Transport Streams (TS), single or multiple Generic Streams (packetized or continuous): the block provides input stream interfacing, optional tools required for ACM (e.g. synchronization (see note 1) and null-packet deletion for Transport Streams (see note 2), described in clause 4.5.2.1), CRC coding (and replacement of SYNC bytes) for error detection in the receiver for packetized input streams. Furthermore, for multiple inputs, it provides merging of input streams in a single transmission signal and slicing in FEC code blocks, Data Fields. These latter are composed of DFL bits, where KBCH -80 ≥ DFL ≥0, taken from a single input port, to be transmitted in a homogeneous transmission mode (FEC code and modulation). KBCH is the BCH uncoded block length, which is dependent on the FECFRAME length (normal or short) and on the coding rate, and 80 bits is the BBHEADER length. The Base-Band Header is appended in front of the Data Field, to notify the receiver of the input stream format and Mode Adaptation type. NOTE 1: Data processing in DVB-S2 may produce variable transmission delay. This block allows to guarantee constant-bit-rate and constant end-to-end transmission delay for packetized input stream [15]. NOTE 2: To reduce the information rate and increase the error protection in the modulator. The process allows null-packets re-insertion in the receiver in the exact place where they originally were [15]. MODE ADAPTATION BB DATA Single Signalling Input Input Input Stream Null-packet Dotted sub-systems are CRC-8 Stream interface Synchroniser Deletion Buffer Encoder not relevant for ACM (ACM, TS) COMMAND single transport stream Merger broadcasting Slicer Multiple applications Input Input Input Stream Null-packet Buffer CRC-8 Streams interface Synchroniser Deletion Encoder (ACM, TS) QPSK, PL Signalling & α=0,35, 0,25, 8PSK, 0,20 rates 1/4,1/3,2/5 16APSK, Pilot insertion 1/2, 3/5, 2/3, 3/4, 4/5, 32APSK 5/6, 8/9, 9/10 I PL Bit mapper SCRAM BB Filter BB BCH LDPC Bit PADDER into Q and SCRAM Encoder Encoder Inter- BLER constel- Quadrature (NBCH,KBCH) (nldpc,kldpc) leaver Dummy Modulation BLER lations PLFRAME Insertion STREAM ADAPTATION FEC ENCODING MAPPING PL FRAMING MODULATION LP stream for to the RF BBHEADER BC modes satellite DATAFIELD BBFRAME FECFRAME PLFRAME channel Figure 2: Functional block diagram of the DVB-S2 system ETSI 19 ETSI TR 102 376 V1.1.1 (2005-02) In case the user data available for transmission are not sufficient to completely fill a BBFRAME, padding is provided by the "Stream Adaptation" block to complete it. Base-band scrambling is also provided. "FEC Encoding" carries out the concatenation of BCH outer code and LDPC inner codes. Depending on the application area, the FEC coded blocks (FEC frames) can have length 64 800 (normal frame) or 16 200 (short frame) bits. When VCM or ACM are used, FEC and modulation mode are constant within a frame but may be changed in different frames; furthermore, the transmitted signal can contain a mix of normal and short code blocks. Bit interleaving is then applied to FEC coded bits for 8PSK, 16APSK and 32APSK to separate bits mapped onto the same transmission signal. "Mapping" into QPSK, 8PSK, 16APSK and 32APSK constellations is then applied to get a complex XFECFRAME, composed of 64 800/ηMOD or 16 200/ηMOD modulated symbols (ηMOD being the number of bits carried by a constellation symbol). "Physical Layer Framing", synchronous with the FEC frames, provides optional dummy PL frame insertion (when no useful data is ready to be sent on the channel), PL header and optional pilot symbols insertion (2,4 % capacity loss) and scrambling for energy dispersal. Pilot symbols insertion occurs at regular intervals (36 pilot symbols each 1 440 data symbols), starting after each PLHEADER. This allows achieving the high channel estimation accuracy indicated by the standard and needed to track channel variation when ACM is utilized. The modulated symbols are inserted in a regular physical layer frame structure, composed of fixed length slots of 90 symbols. As the XFECFRAME length is dependent on both the frame type (short or normal) and the modulation order, it occupies a variable integer number of slots, which is larger the lower is the modulation order (see table 11 in [2]). The PLFRAME is obtained by adding the PLHEADER, which occupies one extra slot and carries the information related to the frame type and to the physical layer mode. After decoding the PLHEADER, the receiver can derive, through the knowledge of the transmission parameters, the current frame length and thus the start of the following frame, even if the status of the channel does not allow for successful data decoding in the current frame. Finally, "Modulation" applies Base-Band Filtering and Quadrature Modulation, to shape the signal spectrum and to generate the RF signal. Square-root raised cosine filtering is used at the transmit side, with choice on three roll-off factors: 0,35, 0,25 and 0,20. 4.3.2 Reference performance The DVB-S2 system may be used in "single carrier per transponder" or in "multi-carriers per transponder" (FDM) configurations. In single carrier per transponder configurations, the transmission symbol rate Rs can be matched to given transponder bandwidth BW (at -3 dB), to achieve the maximum transmission capacity compatible with the acceptable signal degradation due to transponder bandwidth limitations. To take into account possible thermal and ageing instabilities, reference can be made to the frequency response mask of the transponder. Group delay equalization at the transmitter may be used to increase the transmission capacity or to reduce degradation. In the multi-carrier FDM configuration, Rs can be matched to the frequency slot BS allocated to the service by the frequency plan, to optimize the transmission capacity while keeping the mutual interference between adjacent carriers at an acceptable level. 4.3.2.1 Single carrier per transponder configuration Dependant on the selected code rate and modulation constellation the system can operate at carrier to noise ratios from -2,4 dB using QPSK 1/4 to 16 dB using 32APSK 9/10 (assuming AWGN channel and ideal demodulator) (figure 3). These results have been obtained by computer simulations for a Packet Error Rate of 10-7, both for DVB-S2 and DVB-S/DVB-DSNG, and correspond about to one erroneous Transport Stream Packet per transmission hour in a 5 Mbit/s video service (see note 1). NOTE 1: It should be noted that this definition is slightly different from the Quasi Error Free target adopted in EN 300 421 [1]. Furthermore modem implementation margins reported in [1] and [4] are not included in figure 3. On AWGN, the result is typically a 20 %to 35 % capacity increase over DVB-S and DVB-DSNG under the same transmission conditions and 2 dB to 2,5 dB more robust reception for the same spectrum efficiency. ETSI 20 ETSI TR 102 376 V1.1.1 (2005-02) 5 4,5 Dotted lines = modulation constrained Shannon Limit 32APSK 4 3,5 16APSK Ru [Mbit/sec] per unit Symbol rate Rs DVB-S2 3 6QAM 2,5 8PSK 2 8PSK DVB-DSNG QPSK 1,5 1 DVB-S 0,5 0 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 C/N [dB] in BW=Rs Figure 3: Required C/N versus spectrum efficiency, obtained by computer simulations on the AWGN channel (ideal demodulation) (C/N refers to average power) Figure 4 shows instead the DVB-S2 performance for constant satellite bandwidth BW = Rs(1+ρ) on the AWGN channel assuming ideal demodulation. The figure does not take into account the performance degradation which is expected on the satellite channel due to the signal envelope, which increases with decreasing roll-off. For DVB-DSNG only the normative roll-off 0,35 is considered, even if DVB-DSNG also includes, optionally, 0,25. ETSI 21 ETSI TR 102 376 V1.1.1 (2005-02) 4 0,20 0,25 3,5 32APSK 0,35 3 16APSK ) ◊ Ru [Mbit/sec] per unit BW=Rs(1+ 2,5 DVB-S2 0,35 16QAM 2 8PSK DVB-DSNG 1,5 8PSK QPSK 1 DVB-S 0,5 0 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 ρ C/N [dB] in BW=Rs(1+ρ) ρ Figure 4: Required C/N versus spectrum efficiency for constant satellite bandwidth BW = Rs(1+ρ) on the AWGN channel (ideal demodulation) (C/N refers to average power) Figure 5 gives examples of the useful bit rate capacity Ru achievable by the system versus the LDPC code rate, assuming unit symbol rate RS. The symbol rate RS corresponds to the -3 dB bandwidth of the modulated signal. RS(1+α) corresponds to the theoretical total signal bandwidth after the modulator, with α representing the roll-off factor of the modulation. The figures refer to Constant Coding and Modulation, normal FEC frame length (64 800 bits), no padding field, no pilots (the pilots would reduce the efficiency by about 2,4 %). Typical BW/RS or BS/RS ratio is 1+α = 1,35: this choice allows to obtain a negligible ES/No degradation due to transponder bandwidth limitations, and also to adjacent channel interference on a linear channel. The use of the narrower roll-off α = 0,25 and α = 0,20 may allow a transmission capacity increase but may also produce larger non-linear degradations by satellite for single carrier operation. BW/RS factors less than 1+α may also be adopted, but careful studies should be carried-out on a case-by-case basis to avoid unacceptable interference and distortion levels. ETSI 22 ETSI TR 102 376 V1.1.1 (2005-02) 1/4 1/3 2/5 1/2 3/5 2/3 3/4 4/5 5/6 8/9 9/10 4,5 4,0 32APSK 3,5 3,0 16APSK RU 2,5 2,0 8PSK 1,5 QPSK 1,0 0,5 0,0 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 LDPC code rate Figure 5: examples of useful bit rates RU versus LDPC code rate per unit symbol rate RS When DVB-S2 is transmitted by satellite, quasi-constant envelope modulations, such as QPSK and 8PSK, are power efficient in single carrier per transponder configuration, since they can operate on transponders driven near saturation. 16APSK and 32APSK, which are inherently more sensitive to non-linear distortions and would require quasi-linear transponders (i.e. with larger Output-Back-Off, OBO) may be greatly improved in terms of power efficiency by using non-linear compensation techniques in the up-link station [13]. In FDM configurations, where multiple carriers occupy the same transponder, this latter must be kept in the quasi-linear operating region (i.e. with large OBO) to avoid excessive inter-modulation interference between signals. In this case the AWGN performance figures may be adopted for link budget computations. Table 1 shows, for the single carrier per transponder configuration, the simulated C/N degradation using the satellite channel models and phase noise mask given in [2] (non linearized TWTA, IMUX and OMUX filters, phase noise relevant to consumer LNBs), at the optimum operating TWTA point (see note 2). CSAT is the un-modulated carrier power at HPA saturation, OBO is the measured power ratio (dB) between the un-modulated carrier at saturation and the modulated carrier (after OMUX). Phase noise degradation figures refer to a pilot-based carrier recovery system [13]. The figures show the large advantage offered by the use of dynamic pre-distortion for 16APSK and 32APSK. The large phase noise degradations quoted for APSK, and in particular for 32APSK, can be considered as pessimistic, since they refer to consumer-type LNBs, while for professional applications better front-ends may be adopted at negligible additional cost. NOTE 2: The following parameters have been simulated [13]: Rs = 27,5 Mbaud, roll-off = 30 % (not available in DVB-S2, but giving performance between roll-off 0,35 and 0,25). Table 1: CSAT/N loss [dB] on the satellite channel (simulation results, Single Carrier per Transponder, optimum TWTA operating point) Transmission CSAT/N loss [dB] CSAT/N loss [dB] CSAT/N loss [dB] Mode without predistortion with dynamic predistortion with dynamic predistortion without Phase Noise without Phase Noise with Phase Noise QPSK 1/2 0,62 (IBO = 0; OBO = 0,33) 0,5 (IBO = 0 dB; OBO = 0,38) 0,63 8PSK 2/3 0,95 (IBO = 0,5; OBO = 0,35) 0,6 (IBO = 0;OBO = 0,42) 0,85 16APSK 3/4 3,2 (IBO = 5; OBO = 1,7) 1,5 (IBO = 1; OBO = 1,1) 1,8 32APSK 4/5 6,2 (IBO = 9; OBO = 3,7) 2,8 (IBO = 3,6; OBO = 2,0) 3,5 ETSI 23 ETSI TR 102 376 V1.1.1 (2005-02) 4,5 32APSK 4,0 3,5 16APSK s Ru[bit/s] per unit R 3,0 8PSK 1,9 dB 2,5 2,0 QPSK 1,5 DVB-DSNG 0,3 1,0 2,2 dB 0,5 DVB-S 0,0 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 C/N [dB] in Rs Figure 6: Examples of RU versus required C/N by satellite, in single carrier per transponder configuration Figure 6 shows, in the C/N - Spectral Efficiency plane, the overall performance of DVB-S2 by satellite, compared to DVB-S and DVB-DSNG (see note 3). The C/N gain of DVB-S2 versus DVB-S and DVB-DSNG, at a given spectral efficiency, remains substantially constant around 2 dB to 2,5 dB, confirming the results on AWGN. Similarly, the capacity gain at a given available C/N confirms to be in the range 0,3 to 0,4 bit/s/Hz (2,6 % pilot symbol loss is not indicated, since pilots are optional). Compared to AWGN simulations, satellite simulation curves for 16APSK and 32APSK are more aligned with QPSK and 8PSK curves, due to the amplitude limitation of the non-linear TWTA characteristics. NOTE 3: The transmission modes indicated by circles are fully simulated [13], while the other configurations are extrapolated. The degradations of table 1 are added to the AWGN simulated figures, for the relevant constellation order, neglecting the effect of coding rates on degradations; M-QAM degradations is assimilated to M-APSK. 4.3.2.1.1 Sensitivity to satellite power amplifier characteristics The sensitivity to different satellite power amplifiers has been investigated by means of software simulations [38]. The same channel model and phase noise mask as in [2] has been used. Three different amplifiers have been considered: the two models defined in [2], linearized TWTA (LTWTA) and non-linearized TWTA (NL-TWTA), and a second linearized-TWTA (LTWTA2), whose characteristics are reported respectively in figures 7 and 8. Only the 16ASPK modulation scheme with LDPC rate ¾ is considered for a 27,5 Mbaud link and roll-off factor 0,3. ETSI 24 ETSI TR 102 376 V1.1.1 (2005-02) 0 TWTAs Model AM-AM Characteristic 60 DVB-S2 Non-Linearized TWTA Intelsat DVB-S2 Linearized TWTA -5 50 TWTAs Model AM-PM Characteristic Output Amplitude [dB] Output Phase [deg] -10 40 30 -15 20 -20 10 DVB-S2 Non-Linearized TWTA Intelsat DVB-S2 Linearized TWTA 0 -25 -30 -25 -20 -15 -10 -5 0 5 10 -30 -25 -20 -15 -10 -5 0 5 Input Amplitude [dB] Input Amplitude [dB] Figure 7: AM/AM and AM/PM characteristic of DVB-S2 linearized (LTWTA) and non linearized TWTA (NL-TWTA) models 0 0 -10 -5 -20 Output Amplitude [dB] Output phase [deg] -30 -10 -40 Linearized TWT vs DVB-S2 TWT Linearized TWT vs DVB-S2 TWT -50 AM-PM Characteristic AM-AM Characteristic -15 Linearized Linearized DVB-S2 DVB-S2 -60 -20 -16 -12 -8 -4 0 4 8 -20 -16 -12 -8 -4 0 4 8 Input Amplitude [dB] Input Amplitude [dB] Figure 8: AM/AM and AM/PM Characteristic of linearized TWTA LTWTA2 Tables 2 and 3 report results for this configuration; in these tables, D represents the demodulator loss due to the non-linearity. Simulations findings indicate that linearized TWTAs work better than the non-linearized one. This is true both for the case when pre-distortion algorithm is used and for the case when the original constellation is transmitted. LTWTA2 provides the smallest demodulation loss (D) at the price of higher OBO. The total loss is slightly higher than for LTWTA, that has a less linear characteristic but lower OBO. Furthermore, using simple static constellation pre- distortion the L-TWTA can bring about 0,3-0,4 dB gain over classical TWTA, whereas without pre-distortion the gain goes up to 0,8 dB. The performance of NL-TWTA, when a pre-distorted constellation is used, is comparable to the one of the linearized TWTA without pre-distorted constellation. Since the AM/AM characteristic of the two amplifiers is very similar in the region of interest when IBO>12 dB, the advantage appears to be related to the flatter AM/PM characteristic of the linearized amplifier. Table 2: Summary of results for 1 carrier/HPA with no pre-distortion algorithm, 16APSK 3/4 CSAT/N loss HPA Optimum IBO [dB] OBO [dB] D [dB] [dB] NL-TWTA 4,5 1,5 1,7 3,2 LTWTA 4 1,3 1,1 2,4 LTWTA2 3 1,6 0,85 2,45 ETSI 25 ETSI TR 102 376 V1.1.1 (2005-02) Table 3: Summary of results for 1 carrier/HPA with Static pre-distortion algorithm, 16APSK 3/4 CSAT/N loss HPA Optimum IBO [dB] OBO [dB] D [dB] [dB] NL-TWTA 2 1 1,4 2,4 LTWTA 2 1,0 1,0 2,0 LTWTA2 2 1,35 0,8 2,15 4.3.2.1.2 Sensitivity to roll-off Decreasing the roll-off factor causes an increase of the loss over non-linear channel due to the fact that the ISI becomes stronger. Simulations carried on with the same channel model and phase noise as for the previous case have shown however that the resulting effect is only marginal. A roll-off factor of 0,2 produces a loss of about 0,1 dB w.r.t. the case of roll-off 0,3 for all the modulation schemes, in case of adoption of static pre-distortion techniques. Similar results are expected when using dynamic pre-distortion. 4.3.2.2 Multiple carrier per transponder configuration In [38] simulations of multi-carrier systems have been carried out on the same satellite channel model as in clause 4.3.2.1. 8PSK rate 2/3 and 16APSK rate 3/4 have been tested in this configuration. Results for 2 carrier/HPA are summarized in tables 4 to 7. From the tables is immediately clear that there is no great advantage in using a pre-distortion technique for none of the modulation schemes when 2 carriers travel along the same amplifier. As expected the conventional TWTA (NL-TWTA) behaves worse than the others. Recalling the characteristics of the two linearized tubes defined in clause 4.3.2.1 (figures 7 and 8), LTWTA2 is expected to behave better than LTWTA, since its AM/AM characteristic is more linear. Such a thesis does not match the results. Indeed, despite a slightly less demodulation loss (D column), the OBO of LTWTA2 is higher and makes the total loss higher. Therefore, once more, this indicates the importance of OBO factor for the total degradation of the whole system. From the same results it appears that the use of pre-distortion techniques does no mitigate the total loss. It is then clear that when two or more carriers go through the same tube, the intermodulation products due to the non linearities represent the main contribution to the global ISI. Therefore, neither a static pre-distortion technique nor one with memory can attenuate this contribution. A positive outcome of using pre-distortion techniques is the reduced OBO that results in a slightly higher DC to RF conversion efficiency (see tables 6 and 7). Table 4: Summary of results for 2 carriers/HPA with no pre-distortion algorithm for 8PSK 2/3 CSAT/N loss HPA Optimum IBO [dB] OBO [dB] D [dB] [dB] NL TWT 1,0 1,0 1,0 2,0 LTWTA 1,5 1,05 0,7 1,75 LTWTA2 0,0 1,25 0,8 2,05 Table 5: Summary of results for 2 carriers/HPA with Static pre-distortion algorithm for 8PSK 2/3 CSAT/N loss HPA Optimum IBO [dB] OBO [dB] D [dB] [dB] NL TWTA 1,0 1,0 1,0 2,0 LTWTA 1,5 1,05 0,7 1,75 LTWTA2 0,0 1,25 0,8 2,05 Table 6: Summary of results for 2 carriers/HPA with no pre-distortion algorithm for 16APSK 3/4 CSAT/N loss HPA Optimum IBO [dB] OBO [dB] D [dB] [dB] NL-TWTA 5,0 2,01 2,1 4,1 LTWTA 4,5 1,75 1,35 3,1 LTWTA2 3,0 2,0 1,5 3,5 ETSI 26 ETSI TR 102 376 V1.1.1 (2005-02) Table 7: Summary of results for 2 carriers/HPA with Static pre-distortion algorithm for 16APSK 3/4 CSAT/N loss HPA Optimum IBO [dB] OBO [dB] D [dB] [dB] NL-TWTA 5,0 2,0 2,05 4,05 LTWTA 4,0 1,6 1,5 3,1 LTWTA2 3,0 2,0 1,45 3,45 Simulations for 5 carriers/HPA have been carried only for the LTWTA being the one that performs better in case of 2 carriers/HPA. Moreover since data pre-distortion does not provide any noticeable advantage, only the original constellation of 8PSK and 16APSK schemes has been used. Results are shown in table 8. Table 8: Summary of results for 5 carriers/HPA on LTWTA with no pre-distortion algorithm for 8PSK 2/3 and 16APSK 3/4 CSAT/N loss Modulation Optimum IBO [dB] OBO [dB] D [dB] [dB] 8PSK 2/3 4,0 1.70 1,4 3,1 16APSK 3/4 6,5 2,8 1,6 4,4 4.4 The backwards compatible modes The large number of DVB-S receivers already installed makes it very difficult to think of an abrupt change of technology in favour of DVB-S2 for many established broadcasters, especially where there is a receiver subsidy, and for free to air public services. In such scenarios, backwards-compatibility may be required in the migration period, allowing legacy DVB-S receivers to continue operating, while providing additional capacity and services to new, advanced receivers. At the end of the migration process, when the complete receiver population has migrated to DVB-S2, the transmitted signal could be modified to the non-backward compatible mode, thus exploiting the full potential of DVB-S2. Optional backwards-compatible (BC) modes have therefore been defined in DVB-S2, intended to send, on a single satellite channel, two Transport Streams, the first (High Priority, HP) being compatible with DVB-S receivers (according to EN 300 421 [1]), as well as with DVB-S2 receivers, the second (Low Priority, LP) being compatible with DVB-S2 receivers only. Backwards compatibility can be implemented according to two approaches [17]: • layered modulations, where a DVB-S2 and a DVB-S signals are asynchronously combined on the radio-frequency channel (therefore this operational mode does not require any specific tool in the DVB-S2 specification), the DVB-S signal being transmitted at significantly higher power level than DVB-S2. Since the resulting signal shows large envelope variations, it must be transmitted on a quasi-linear transponder, far from saturation. As an alternative, to better exploit the satellite power resources, HP and LP signals can be independently transmitted on the up-link, and amplified each by an independent satellite amplifier (HPA), driven near saturation; the resulting signals are then combined on the down-link channel. This requires the design and launch of new generation satellites. • hierarchical modulation, where the two HP and LP Transport Streams are synchronously combined at modulation symbol level on a non-uniform 8PSK constellation (note that hierarchical modes are also used in EN 300 744 [40]). Since the resulting signal has quasi-constant envelope, it can be transmitted on a single transponder driven near saturation. This solution is included in the DVB-S2 standard as an option. ETSI 27 ETSI TR 102 376 V1.1.1 (2005-02) 4.4.1 Hierarchical modulations Figure 10 schematically represents the hierarchical transmission system structure: it consists of two branches, the first compliant with DVB-S standard for the high priority layer, the second increasing the constellation dimensionality to a non-uniform 8PSK (figure 9) for the low priority layer. I=HP Q=HP 10 Q 00 MSB LSB QUADRANT QUADRANT 1 1 LP bit 0 ρ=1 0 2θ I 0 0 1 1 11 01 QUADRANT QUADRANT Figure 9: Non-uniform 8PSK constellation EN 300421 HP TS Is,Qs MUX DVB-S bits outer & inner coding Hierarchical Spectrum Quadrature mapper shaping modulation LP TS (α=0,35) MUX DVB-S2 PL Header outer & inner insertion coding 2θ EN 302 307 Figure 10: Functional block diagram of hierarchical backward compatible DVB-S2 system The LP DVB-S2 compliant signal is BCH and LDPC encoded, with LDPC code rates 1/4, 1/3, 1/2 or 3/5, according to DVB-S2 transmission standard [2] and extracted after LDPC encoder (figure 10). Of the possible configurations of the DVB-S2 stream, only the normal FEC frame configuration is allowed, with 64 800 bits exiting from the LDPC encoder, divided into 720 slots of 90 bits. The Physical Layer (PL) header sequence (90 bits) is then inserted, to indicate the LP code rate being transmitted. The hierarchical mapper maps three bits per symbol as in figure 9: one bit from the DVB-S2 LDPC encoded signal, following PL header insertion; two bits from the HP DVB-S encoder. The deviation angle θ is not transmitted and may vary according to user requirements (see note 1): large θ figures improve LP ruggedness against noise and interference, but penalize HP. NOTE 1: The half angle θ may be constrained by legacy receiver performance. ETSI 28 ETSI TR 102 376 V1.1.1 (2005-02) Table 9 gives the bit-rate ratio (%) of the LP stream versus the HP stream. Table 9: (LP bit-rate/HP bit-rate)x100 [%] DVB-S2 Coding (LP) DVB-S Coding (HP) 1/4 1/3 1/2 3/5 ½ 26,6 35,7 53,7 64,6 2/3 20,0 26,7 40,3 48,4 ¾ 17,8 23,8 35,8 43,0 5/6 16,0 21,4 32,2 38,7 7/8 15,2 20,4 30,7 36,9 BC Hierarchical modulation schemes performance is given in the following [17]. Simulations through FEC decoding of the HP layer were performed to determine the impact of the added LP layer upon the legacy HP layer. The simulation model included the convolutional encoder, Viterbi decoder and QPSK mapper, addition of lower-layer signal, imposition of phase noise according to the aggregate phase noise mask of [2], clause H.8 and carrier recovery according to a QPSK-directed second-order loop [13] with bandwidth 60 kHz and damping factor 0,707. The symbol rate was assumed to be 27,5 Mbaud. For each combination of legacy rate and half angle θ, CNR was varied until the threshold for DVB-S QEF (Quasi Error Free) was determined, corresponding to BER of 2 × 10-4 after Viterbi decoding: specifically, for each of these combinations. For the performance evaluation of the incompatible Low Priority Stream, the following approximated method has been applied: if the chosen low-priority DVB-S2 scheme requires a given CNR at a target BER on BPSK constellation, its performance degradation ∆L when applied to the non-uniform 8PSK constellation may be estimated as: ∆L = -20 × log10(sin(θ)). CNR can be derived from the DVB-S2 specification: in [2], table 13, QPSK performance at packet error rate equal to 10-7 is indicated. Therefore: CNRLP = CNRQPSK-3+ ∆L+M where M represents an additional margin to take into account the lower target performance of DVB-S with respect to DVB-S2 and implementation losses. In figure 11, the required CNR of the HP (compatible) and LP (incompatible) bit-streams is represented, versus the non uniform 8PSK constellation angle θ, to achieve DVB-S target QEF (see note 2). For the DVB-S branch, CNR values are derived by simulations, adding 0,8 dB for implementation losses (see note 3); for the DVB-S2 branch, they are extracted from [2], adding 0,8 dB to take into account for the lower target performance and implementation losses. With increasing θ angles, the HP C/N goes up, while the LP C/N goes down. The throughput in bits/s/Hz for each code rate of the HP and LP streams is indicated. NOTE 2: To be consistent with EN 300 421 [1], C/N figures for HP and LP corresponds to DVB-S QEF target (which is more severe than DVB-S2 QEF target [2]). An implementation margin of 0,8 dB is also included in both branches. NOTE 3: The 0,8 dB value allow to approximately match simulation results for θ equal to 0° to values indicated in the DVB-S standard [1]. The figure also indicates the cycle slip limit to the BC mode performance. In fact, as the value of θ is increased, the constellation becomes more 8PSK-like, and the carrier recovery in the legacy receiver becomes prone to cycle slips. Reported results refer to simulations at a symbol rate of 20 Mbaud, for as system roll-off factor equal to 0,2 (worst case). The phase noise and satellite models are those reported in clauses H.7 and H.8 of [2].Carrier recovery is based on a second order hard decision directed loop, damping factor is 1,0, natural frequency is equal to the symbol rate multiplied by 0,00114. ETSI 29 ETSI TR 102 376 V1.1.1 (2005-02) 16 3/5 1/3 1/4 1/2 14 LP 12 C/N[dB] HP 10 DVB-S 7/8 5/6 8 3/4 2/3 Sat 6 Lin 1/2 4 4 6 8 10 12 14 16 18 θ [°] Figure 11: required CNR of the HP and LP bit-stream versus the non uniform 8PSK constellation angle θ on the AWGN channel, for the different code rates of DVB-S on the HP stream and DVB-S2 on the LP stream. Grey dashed lines indicate the cycle slip limits 4.5 Adaptive Coding and Modulation The DVB-S2 standard has been conceived for a wide range of satellite broadband applications, including point-to-point applications like IP unicasting or DSNG, with the adoption of Adaptive Coding and Modulation (ACM), allowing different modulation formats and error protection levels (i.e. coding rates) to be used and changed on a frame-by-frame basis within the transmitted data stream. By means of a return channel, informing the transmitter of the actual receiving condition, the transmission parameters may be optimized for each individual user, dependant on path conditions. ACM has been considered as a powerful tool to further increase system capacity, allowing for better utilization of transponder resources. As a consequence, in DVB-S2 standard ACM is included as normative for the interactive application area and optional for DSNG and professional services. As IP traffic is driving interactive services and applications offered by broadband unicast systems, the new standard is intended to be IP friendly, by improving the efficiency of carriage of IP data. This goal has been coupled with the legacy induced by the current wide utilization of MPEG Transport Stream (MPEG-TS) packets for encapsulating IP datagrams [7]. The result is a highly flexible standard, which supports for interactive applications both generic packet streams and MPEG-TS. One of the implications of such a high flexibility is the multiplicity of solutions allowed by DVB-S2 for implementing ACM in interactive systems. Finally, support of individual Quality of Service targets has been recommended for interactive applications. 4.5.1 ACM: the principles Current point-to-point multi-beam satellite systems based on the DVB-S standard are designed for link closure in the worst-case propagation and location conditions. The standard, envisaged for broadcasting applications, considers a fixed coding rate and modulation format, which are selected according to the assumed coverage and availability requirements. This approach implies the occurrence of high margins in the majority of the cases, when interference and propagation conditions allow for higher Signal to Interference plus Noise Ratio (SNIR). Typical Ku-band broadcasting links are designed with a clear-sky margin of 4 dB to 6 dB and a service availability target of about 99 % of the worst month (or 99,6 % of the average year). Since the rain attenuation curves are very steep in the region 99 % to 99,9 % of the time, many dBs of the transmitted satellite power are useful, in a given receiving location, for only some ten minutes per year. Unfortunately, this waste of satellite power/capacity cannot be easily avoided for broadcasting services, where millions of users, spread over very large geographical areas, receive the same content at the same time. ETSI 30 ETSI TR 102 376 V1.1.1 (2005-02) However, this design methodology devised for broadcasting systems is not optimal for unicast networks. In fact, the point-to-point nature of link connections allows exploiting spatial and temporal variability of end-user channel conditions for increasing average system throughput. This is achieved by adapting coding rate and modulation format (ACM) to best match the user SNIR, thus making the received data rate location and time dependent. Fixed link margins are therefore avoided and a remarkable improvement in system capacity is obtained thanks to better utilization of system power resources. Assuming a fixed beam power allocation, the key parameters responsible for SNIR variability within the satellite coverage are the fading attenuation, the satellite antenna gain and the antenna C/I. The fading attenuation variation is due to both geographical dependency of rain statistics and to propagation channel time variations. Examples of fading attenuations statistics and Satellite Terminal (ST) received power variations during heavy fading events can be found in [15]. In [18], figure 5, some examples of satellite antenna C/I distributions are shown, which have been obtained in a study case assuming European coverage. Because of the spread of the distributions, which is covering 15-20 dB range depending on the selected frequency reuse factor, antenna C/I is an important source of SNIR variations within the coverage. In annex D a methodology [18] is illustrated for deriving ACM systems capacity and assessing the advantage with respect to current DVB-S multimedia systems. Capacity and link availability figures have been derived following the proposed methodology for a study case of an ACM multi-beam Ka-band system covering Europe, showing an important gain with respect to non-adaptive systems. One of the key factors, which can heavily affect capacity, is the granularity of physical layer schemes supported by the system. It is important to ensure a reasonably small step in Es/N0 and spectral efficiency between consecutive schemes in order to maximize the data rate provided to STs. In particular, it is of the highest importance for capacity purposes utilizing a sufficiently fine granularity in the SNIR range experienced in the coverage during clear sky conditions. On the contrary, larger SNIR intervals can be assumed between the more robust modes without major capacity penalty, due to their lower occurrence probability. This is confirmed by simulation results shown in annex D. In the DVB-S2 standard code rates have been designed in order to obtain a granularity of about 1 dB in C/N (see figure 3), covering the wide range from -2 dB up to +16 dB on the AWGN channel. Figure 12 shows the scheme of an ACM satellite link, composed by the Gateway (GW), which includes the ACM DVB-S2 modulator, the Satellite, the Satellite receiving Terminal (ST) connected to the GW via a return channel. Although DVB-S2 standard is also applicable to regenerative satellite systems, here a transparent satellite system architecture is considered. yawetaG ACM DVB-S2 High bit-rate MODULATOR forward-link Modulation & coding SNIR/MODCOD selection signalling source rate control (*) Satellite Terminal Info Return channel SNIR measurement SOURCE(s) (*) Source rate control may be directly applied to source(s) or locally at the GTW input or via network traffic control Figure 12: Block diagram of a DVB-S2 ACM link ETSI 31 ETSI TR 102 376 V1.1.1 (2005-02) The DVB-S2 ACM modulator operates at constant symbol rate, since the available transponder bandwidth is assumed constant. ACM is implemented by the DVB-S2 modulator by transmitting a TDM sequence of physical layer frames, each transporting a coded block and adopting a uniform modulation format. However, when ACM is implemented, coding scheme and modulation format may change frame-by-frame. Therefore service continuity is achieved, during rain fades, by reducing user bits while increasing at the same time FEC redundancy and/or modulation ruggedness. Physical layer adaptivity is achieved as follows: • Via a return channel, individual STs provide to the Gateway (GW) information on the channel status, by signalling the SNIR and the most efficient modulation and coding scheme the ST can support. As periodic reports increase the signalling overhead on the return link, a preferable approach would be the ST sending a message whenever channel variations imply a change in the spectral efficiency. • The ST indications are taken into account by the GW in coding and modulating the data packets addressed to each ST. Therefore service continuity is achieved, during rain fades, by reducing user bits while increasing at the same time FEC redundancy and/or modulation ruggedness. On the contrary, during the more frequent clear sky periods a much higher information rate can be delivered to the ST thanks to a higher spectral efficient physical layer mode. • In order to avoid information overflow during fades, a source bit rate control mechanism has to be implemented, adapting the offered traffic to the available channel capacity. This general principle can be implemented in different ways according to service requirements, system architecture and traffic statistics [15]. For example, when the source is co-located with the GTW, its bit-rate may be directly controlled according to the SNIR reports. Instead, when information providers are remote, network traffic control policies may be implemented (e.g. TCP/IP protocol automatically reduces bit-rate when large reception delays are experienced), while low priority packet dropping may be an acceptable strategy to manage short traffic peaks under "best effort" service level requirement. When the aggregate channel capacity variations are small (e.g. because of the averaging effect on a large number of users, with different C/N+I conditions), and when capacity margins are allowed on the channel exploitation, user bit-rate control might even be avoided. The introduction of transmission buffers can be useful to absorb short traffic peaks, at the cost of an increase of the delays on data and, in some cases, on the ACM control loop. The GW can impose error protection applied to a given portion of user data according to two methods [15]: • via the ACM command (see the system block diagram in figure 2); • by splitting user data into various streams (one per required protection level), and feeding each of them to a different DVB-S2 modulator input. The modulator will apply a constant and suitable protection level to each input stream. A crucial issue in ACM systems is the physical layer adaptation loop delay, as it is strictly linked to the system capability of tracking channel variations. The adaptation loop can be defined as the set of operations, which occur starting from channel estimation at the ST premises ending with reception of the information encoded/modulated according to the reported channel status. Whenever the user link is degraded in such a way that it requires a protection level higher than the estimated one, the received packet cannot be correctly decoded. On the contrary, if the channel conditions have improved so that a higher efficient physical layer mode could be supported, the result is a loss in efficiency due to the utilization of a not optimum physical layer scheme (i.e. carrying too much code redundancy). Therefore, the longer the loop delay the higher the impact physical layer adaptation has on user QoS, because of propagation channel dynamic. Maximum rain variation rates at Ka band have been estimated to be of about 0,5 dB to 1 dB per second, while faster SNIR variations are due to fading components with a high dynamic behaviour. The loop delay is a sum of several contributions, including: the delay for transmission of the channel status reports, the return link propagation delay (depending on the selected return channel), the delay introduced within the GW (buffering, processing and transmission of data packets), and finally the fixed 250 ms propagation delay on the forward link. Different system architectures assuming different strategies for information processing and buffering at the GW side will lead in general to different QoS performance for the same channel variation rate. ETSI 32 ETSI TR 102 376 V1.1.1 (2005-02) A possible algorithm for channel estimation is the DA-SNORE [21], which, relying on pilot symbols, provides unbiased and accurate SNIR estimation for Es/N0 higher than 0 dB. With the frame format assumptions of the DVB-S2 standard it can be shown that such algorithm leads to small channel estimation delays (in the order of 10 ms) for an error standard deviation of about 0,2 dB. Together with loop delay reduction, accurate channel estimations are critical for efficient physical layer adaptation. In case of DVB-S2, as the Es/N0 distance between physical layer schemes is in the order of 1 dB, an error standard deviation better than 0,3 dB is suggested to fully exploit system capabilities. Depending on the channel estimation accuracy and on the system dynamicity, the number of physical layer configurations to be used by a certain ST can be optimized in order to ensure the desired QoS. However, as mentioned before, a reduction in the physical layer granularity has a direct impact on the throughput. In [16] the key parameters and trade-offs for efficient physical layer adaptation are shown. Besides, a methodology is illustrated, which allows, through physical layer thresholds optimization, meeting user Packet Error Rate (PER) requirements with a negligible impact on link efficiency. 4.5.2 Functional description of the DVB-S2 subsystem for ACM With reference to the DVB-S2 modulator of figure 2, it may be noted that, unlike DVB-S, the second generation of the standard allows for several input stream formats, thus enhancing system flexibility. In addition to the widely used MPEG transport stream, generic streams, of constant or variable length packets, are encompassed by the standard. When this second configuration is selected, TS rules do not apply. Moreover, different encapsulation protocols with improved efficiency can be used as an alternative to the Multi Protocol Encapsulation (MPE) [7]. The data stream packets of both MPEG-TS and generic streams are called User Packets (UPs). The input interface accepts both single and multiple streams. One additional input signal available in the standard is the "ACM command". This shall be utilized in ACM systems in conjunction with a single input stream. It allows setting, by an external control unit, of the transmission parameters to be adopted by the DVB-S2 modulator for a specific portion of input data. The utilization of the ACM command interface allows for a system configuration, which is completely transparent to the physical layer scheme selection. This functionality is indeed performed by a unit external to the DVB-S2 modulator, which signals through the ACM command the transmission parameters associated to the data packets. In interactive unicast networks, ACM functionality is normative and can be implemented either through a single generic stream with the utilization of the ACM command, or through multiple transport or generic streams. In order to better exemplify the DVB-S2 functionalities involved in the three different system configurations, in figure 13-a, -b and -c the mode adaptation block diagram is shown respectively for the single GS, multiple TSs and multiple GSs case. In DSNG type of systems ACM can be implemented with a single transport stream or optionally with multiple transport streams. The different possible system configurations and their features will be illustrated in clause 6.1. ETSI 33 ETSI TR 102 376 V1.1.1 (2005-02) MODE ADAPTATION BB Signalling CRC BBHEADER Slicer encoder DATA FIELD ACM Command a) MODE ADAPTATION TS1 BB Input Stream Null-packet CRC Signalling Buffer Synchronizer Deletion encoder ..... ..... ..... ..... ..... Merger BBHEADER Slicer DATA FIELD TSK Input Stream Null-packet CRC Buffer Synchronizer Deletion encoder b) MODE ADAPTATION GS1 BB Signalling CRC Buffer encoder ..... ..... ..... Merger BBHEADER Slicer DATA FIELD CRC Buffer encoder c) Figure 13: Mode adaptation subsystem in the system configuration: a) Single GS b) Multiple TS c) Multiple GS The input interface is followed by two optional blocks, the Input Stream Synchronizer and the Null Packet Deletion. These two sub-systems are meant as tools to be utilized for implementing ACM over MPEG transport streams. The first aims at guaranteeing constant bit rate and end-to-end transmission delay, despite the variability introduced by data processing in the modulator and demodulator. The second identifies and removes null packets introduced by the transport stream multiplexer in order to reduce the transmitted information rate, by allowing for null packets reinsertion at the demodulator side. The two sub-systems will be further analysed and detailed in clause 4.5.2.1. CRC-8 encoding is applied to transport streams or generic streams with fixed length packets. ETSI 34 ETSI TR 102 376 V1.1.1 (2005-02) The input streams are then buffered, allowing a merger/slicer to read frame by frame the information necessary to fill the data field. In the case of a single stream, only the slicing functionality is required, while if multiple streams are present the merger/slicer is responsible for composing the data field by reading information bits from one of the input buffers. Hence, the merging of input streams does not take place within a single data field, but through reading different data fields from different inputs. Each data field needs to be homogeneous with respect to the physical layer mode, as it is indeed transmitted in one frame. The merging policy is application dependent. For unicast systems with multiple input streams the standard considers the possibility of performing a round robin polling with a time out for the user packets in each buffer. However, additional different policies can be implemented. When a data field is not available at any of the input buffers, a dummy frame is inserted and transmitted. The fixed length BBHEADER (80 bits) is finally inserted in front of the data field, describing its format. While for broadcasting and DSNG applications the data field can be filled to his maximum capacity (KBCH -80 bits), for unicast applications the data field may include an integer number of user packets. This allows for correct recovery of the user information when adaptive coding and modulation is utilized. As a consequence, padding is required to complete the constant length (KBCH) BBFRAME. This also happens whenever UP available data are not sufficient to fill the BBFRAME. The stream adaptation subsystem is responsible for providing padding in case DFL < KBCH -80, and scrambling the information at the encoder input. In clause 6.1.3 the impact of the introduced overhead on the overall encapsulation efficiency will be analysed. Next, the BBFRAME is sent as input to the FEC encoder: in unicast systems, the output FECFRAME can have short (ηLDPC = 16 200 bits) or normal length (ηLDPC = 64 800 bits). Mapping is then applied to get a compleXFECFRAME: when ACM is used, coding rate and modulation format may be changed frame-by-frame. PLFraming is then applied. It is worth noting that, as the PLFRAME length is dependent on both the frame type (short or normal) and the modulation order, it occupies a variable integer number of slots, which is larger the lower is the modulation order. After decoding the PLHEADER, the receiver can derive, through the knowledge of the physical layer mode, the current frame length and thus the start of the following frame, even if the status of the channel does not allow for successful data decoding in the current frame. When ACM systems are considered, pilot symbols can be inserted in the physical layer frame structure for carrier synchronization and channel estimation purposes. In fact, phase recovery for 8PSK and higher modulation orders with the specified phase noise appears very difficult without any pilot. Besides, in ACM system, the key issue is that the receiver is in general able to decode only a part of the entire stream, and precisely only the frames where transmission parameters are compatible with user channel conditions. In this context, pilot symbols both allow for carrier recovery without knowledge of the frame data even in cases when some PLHEADERs are not correctly decoded. Pilot symbol insertion in DVB-S2 signal is optional, with possibility to carry out pilot switching on a frame-by-frame basis. Finally, physical layer scrambling, baseband shaping and I-Q modulation are performed. 4.5.2.1 Specific subsystems for supporting ACM with MPEG-TS Two subsystems have been introduced in the DVB-S2 standard in order to guarantee, even in ACM mode, constant transport stream bit rate and end-to-end delay, as required by MPEG-2 (see note). NOTE: The PCR (Presentation Clock Reference) mechanism in MPEG exploits the recovered transport stream clock to reconstruct the clock for the decoded video signal. The first problem encountered by DVB-S2 designers was that a transport stream is characterized by constant bit-rate, while ACM is by definition a variable bit-rate transmission, trading-off user bit-rate with FEC redundancy during rain fades. The introduction of the "null-packet deletion" block allowed to overcome this. The second problem was that, during rate adaptation, delay and rate variations may take place in the modem. This is taken into account by the "input stream synchronizer" block. ETSI 35 ETSI TR 102 376 V1.1.1 (2005-02) Null-packet Deletion To understand MPEG transport over a DVB-S2 ACM physical layer, it is useful to remind some rules of MPEG Transport Streams: • The TS packet length is fixed (188 bytes). The packet Header includes a PID field, with limited addressing capability. • IP services are mapped into TS by gateways (GTW), using the DSM-CC multi-protocol encapsulation (MPE) scheme. An IP packet may be split into different TS packets. The TS PID does not include the IP MAC address. • Services multiplexed within a TS may be characterized by constant bit-rate (CBR) or variable bit-rate (VBR). • Input services mapping into TS packets is random (i.e. there is no cyclic fixed allocation), but the packet position in the stream cannot be modified along the transmission chain, unless precise re-multiplexing rules are satisfied. • The TS bit-rate (RTS) must be constant (rate variations require a remultiplexer, updating PCR time stamps). • The end-to-end transmission delay must be constant. Therefore, should a packet jitter be introduced on the channel, it must be smoothed before the demultiplexer by a reception buffer of appropriate dimension. Video Null-packet insertion TS Audio DVB-S2 TS Modulator Multimedia MUX ACM physical GTW layer IP RTS CBR and VBR inputs CBR output Figure 14: MPEG TS multiplexing scheme To avoid data overflow in ACM, the useful bit rate generated by data sources (e.g. a video encoder) must be controlled by a feed-back mechanism using the SNIR report from STs. When the user bit rate is modified at the input of a conventional MPEG-2 Transport Stream multiplexer, the output bit-rate RTS is kept constant by adding "MPEG null-packets" (PID = 8 191D) (see figure 14), which do not carry any useful information. To map one/many constant bit-rate Transport-Stream(s) into a variable bit-rate ACM physical layer, the DVB-S2 modulator activates the subsystem called "Null-Packet deletion" (see figure 15). The function performed by this block is to discard all (or the majority of) the null packets, so that the output bit-rate is no longer constant, but corresponds to the useful source bit-rate. A signalling information (DNP, counting the deleted NPs) allows reinsertion, at the receiver side, of the deleted null-packets in the exact position where they were after the transmit multiplexer. Figure 15 describes the "null-packet deletion" process. For multiple transport stream operation, the null-packet deletion function operates independently over each TS, since the receiver is able to recognize packets belonging to each TS. ETSI 36 ETSI TR 102 376 V1.1.1 (2005-02) Reset after DNP DNP insertion Null-packet deletion Counter Useful- DNP (1 byte) packets Insertion after Output Input Next Useful Null- Packet packets Input Optional I I I I S I S UP S UP S Null-packet Null-packet UP S Y N S S Y Y N S S Y Y N S S Y S S S Y Y N C S Y C C C DNP=0 DNP=0 DNP=1 DNP=2 S I D S I D Y UP S N Y UP S N Output N C S Y P N C S Y P Figure 15: Null-packet deletion and DNP field (1 byte) insertion Input Stream Synchronizer In an ACM system, end-to-end Transport Stream delay variations ∆DTOT may occur under dynamic bit-rate variations. For example, the following sources of delay variations may be listed: • The ACM modulator and demodulator contain fixed length buffers (of typical capacity M [bit] of some LDPC code blocks). When a useful bit-rate variation ∆R = R1 - R2 occurs (R in bit/s), these buffers are crossed at different speed, thus the delay dynamically varies in time, in a range of some tens milliseconds: ∆D = D1-D2 = M [(1/R1)-(1/R2)] = (M ∆R)/(R1R2). • In case of multiple input streams, the buffers in front of the merger (see figure 2) produce random delay jitter, depending on the merger priority strategy. Typical figures of such delay jitters are in the range of some tens to some hundreds milliseconds. • Using DVB-S2 for video contribution purposes, the ACM bit-rate control-loop may drive the source bit-rate (e.g. VBR video encoder), but this latter may show a significant delay DS (e.g. hundreds of milliseconds) in executing rate variation commands (see also clause 7.2.6 and figure 40). Therefore it may happen that the total control loop delay is too large to allow real time compensation of the fading variation. To increase the control speed, the rate control loop may be closed in parallel on the video encoder and on the DVB-S2 modulator (which may immediately react to the rate variation command). In this configuration, a large buffer of many Mbytes (M = ∆R DS) must be inserted in the modulator after null-packet deletion, to avoid data overflow. This buffer generates transmission delay variations ∆D = (∆R/R1) DS during rate adaptation, which can be as large as some seconds. In annex F two receiver schemes are proposed to regenerate the Transport Stream clock R'TS under dynamic rate variations. ETSI 37 ETSI TR 102 376 V1.1.1 (2005-02) 4.5.3 DVB-S2 performance in ACM mode In annex D a method is described for capacity assessment in ACM systems. ACM technique allows to significantly increase the average system throughput and availability of satellite networks, thus making the system economically more attractive for interactive applications. The increase is mainly dependant on the adopted frequency band, target link and service area availability requirements and related system sizing options. Compared to CCM, an ACM-based system can provide: • higher system throughput than the one supported by a CCM system, as an ACM system can take benefit from better link propagation and beam C/I conditions than the worst case link on which CCM physical layer is sized; • higher availability (time-link and/or spatial) than the one supported by the CCM system as when deeper fading occurs the ACM system can use a more robust modulation and coding scheme. This is obtained through a very small reduction of the total system throughput as the number of users affected by these deep fade events is very limited. The highest link availability and service area supported by the system depends on the lowest modulation and coding scheme supported by the system; • more optimization dimensions in the system design to cope with more pushed frequency reuse and more complex satellite antennas. Examples reported in annex D indicate that ACM could allow a capacity increase up to 200 % with respect to CCM, dependent on the link parameters and system configurations. ETSI 38 ETSI TR 102 376 V1.1.1 (2005-02) 4.6 System configurations The DVB-S2 standard defines four application areas and profiles: Broadcast, Interactive, DSNG and professional. Table 10 associates them to the system configurations and mechanisms specified in [2], either defined as "Normative" or "Optional" or "Not Applicable". At least "Normative" subsystems and functionalities shall be implemented in the transmitting and receiving equipment to comply with [2]. Configurations and mechanisms explicitly indicated as "Optional" within [2] for a given application area, need not be implemented in the equipment to comply with [2]. Nevertheless, when an "Optional" mode or mechanism is implemented, it shall comply with the specification as given in [2]. Table 10: System Configurations and Application Areas System configurations Broadcast Interactive DSNG Professional profile profile profile profile QPSK 1/4,1/3, 2/5 O N N N 1/2, 3/5, 2/3, 3/4, 4/5, 5/6, 8/9, 9/10 N N N N 8PSK 3/5, 2/3, 3/4, 5/6, 8/9, 9/10 N N N N 16APSK 2/3, 3/4, 4/5, 5/6, 8/9, 9/10 O N N N 32APSK 3/4, 4/5, 5/6, 8/9, 9/10 O N N N CCM N N (see note 1) N N VCM O O O O ACM NA N (see note 2) O O FECFRAME (normal) 64 800 (bits) N N N N FECFRAME (short) 16 200 (bits) NA N O N Single Transport Stream N N (see note 1) N N Multiple Transport Streams O O (see note 2) O O Single Generic Stream NA O (see note 2) NA O Multiple Generic Streams NA O (see note 2) NA O Roll-off 0,35, 0,25 and N N N N 0,20 Input Stream Synchronizer NA except O (see note 3) O (see note 3) O (see note 3) (see note 3) Null Packet Deletion NA O (see note 3) O (see note 3) O (see note 3) Dummy Frame insertion NA except N N N (see note 3) N = normative, O = optional, NA = not applicable NOTE 1: Interactive service receivers shall implement CCM and Single Transport Stream. NOTE 2: Interactive Service Receivers shall implement ACM at least in one of the two options: Multiple Transport Streams or Generic Stream (single/multiple input). NOTE 3: Normative for single/multiple TS input stream(s) combined with ACM/VCM or for multiple TS input streams combined with CCM. 5 Broadcast applications In figure 16 a simplified block diagram of the DVB-S2 system for the broadcast profile is given for the single transport stream and CCM configuration, derived from figure 2. Baseband PL signalling & Signalling pilot insertion Input CRC-8 BB FEC PL interface Encoder Mapper Modulator Scrambler Encoder Scrambler Figure 16: Simplified block diagram of the DVB-S2 system for the broadcast profile, single transport stream and CCM configuration ETSI 39 ETSI TR 102 376 V1.1.1 (2005-02) 5.1 SDTV broadcasting Table 11 shows comparisons between DVB-S2 and DVB-S broadcasting services via 36 MHz satellite transponders in Europe, using a 60 cm receiving antenna diameters. The example video coding bit-rates are: 4,4 Mbit/s using traditional MPEG-2 coding, or 2,2 Mbit/s using advanced video coding (AVC) systems the DVB Project is currently defining for future applications. The required C/N of the two systems, DVB-S and DVB-S2, have been balanced by exploiting different transmission modes and by fine tuning the DVB-S2 roll-off factor and symbol-rate. The results confirm the capacity gain of DVB-S2 versus DVB-S, exceeding 30 %. Furthermore, by combining DVB-S2 and AVC coding, an impressive number of 21 to 26 SDTV channels per transponder are obtained, thus dramatically reducing the per-channel cost of the satellite capacity. Table 11: Example comparison between DVB-S and DVB-S2 for TV broadcasting Satellite EIRP (dBW) 51 53,7 System DVB-S DVB-S2 DVB-S DVB-S2 Modulation and coding QPSK 2/3 QPSK 3/4 QPSK 7/8 8PSK 2/3 Symbol-rate (Mbaud) 27,5 (α = 0,5) 30,9 (α = 0,20) 27,5 (α = 0,35) 29,7 (α = 0,25) C/N (in 27,5 MHz) (dB) 5,1 5,1 7,8 7,8 Useful bit-rate (Mbit/s) 33,8 46 (gain = 36 %) 44,4 58,8 (gain = 32 %) Number of SDTV programmes 7 MPEG-2 10 MPEG-2 10 MPEG-2 13 MPEG-2 15 AVC 21 AVC 20 AVC 26 AVC 5.2 SDTV and HDTV broadcasting with differentiated channel protection The DVB-S2 system may deliver broadcasting services over multiple Transport Streams, providing differentiated error protection per multiplex (VCM mode) (see note). A typical application is broadcasting of a highly protected multiplex for SDTV, and of a less protected multiplex for HDTV. Figure 17 shows an example configuration at the transmitting side. Assuming to transmit 27,5 Mbaud and to use 8PSK 3/4 and QPSK 2/3, 40 Mbit/s would be available for two HDTV programmes and 12 Mbit/s for two-three SDTV programmes. The difference in C/N requirements would be around 5 dB. NOTE: It should be noted that the DVB-S2 system is unable to differentiate error protection within the same TS MUX. DVB-S2 Modulator QPSK rate SDTV 3/4 coder 1 Mode adapter Stream Multiple Transport Streams MUX Input Interface Adapter VCM 1 & adaptation tools M Roll-off=0,25 SDTV E FEC Padding: not present coder R G Coder E FECFRAME: 64800 HDTV R Pilots: on coder Mod 2 Input Interface MUX & adaptation tools 2 HDTV 16APSK coder rate 3/4 Figure 17: Example DVB-S2 configuration for TV and HDTV broadcasting using VCM 5.3 Backwards Compatible services Backwards compatible modes may be used to extend the services delivered by a transponder, without disturbing the legacy DVB-S receivers [17]. ETSI 40 ETSI TR 102 376 V1.1.1 (2005-02) 5.3.1 Hierarchical modulations Different possible application scenarios can be envisaged. The first application scenario here analysed guarantees the same service availability (e.g. with respect to rain attenuation) for the two priority levels, as it may happen when HP and LP both carry video services. Figure 18 shows, in addition to the DVB-S-only curve, the overall hierarchical DVB-S2 (HP+LP) throughput per unit bandwidth, versus the available C/N (see note), for each combination of HP and LP code rates. Points are only indicated if the intersection half-angle is less than 18°. It can be seen that additional capacity may be obtained via hierarchical modes only for C/N ratios above 7 dB, while for lower C/N ratios the use of DVB-S alone is preferable (in terms of capacity as well as of full compatibility). In particular, substantial throughput gain is achievable by means of the hierarchical modes when the available C/N is larger than required by DVB-S rate 7/8. For example, let us assume that the link budget ensures 10,8 dB C/N for the target availability (e.g. 99,9 % a.y.). For a symbol rate equal to 27,5 MHz, the use of DVB-S at rate 7/8 would offer 44 Mbit/s only, while, introducing hierarchical modulation (LP rate 3/5), an increase in capacity of 16 Mbit/s for the DVB-S2 users would be available, and still guaranteeing the target 99,9 % a.y. service availability for all users. NOTE: Figure 18 is derived from figure 11, taking the intersection of the HP and LP curves (θ angles corresponding to balanced C/N). On each curve, the five possible HP code rates are drawn from right to left, starting from 7/8 to 5/6, 3/4, 2/3and 1/2; the sequence is interrupted on left side when the θ angle exceeds 18°. 2.5 3/5 1/2 7/8 5/6 1/3 5/6 7/8 2.0 3/4 5/6 7/8 1/4 7/8 η [bit/s/Hz] 3/4 5/6 2/3 3/4 7/8 1.5 2/3 5/6 3/4 2/3 1/2 1/2 1.0 DVB-S only 0.5 4 5 6 7 8 9 10 11 C/N [dB] NOTE: Lines are drawn connecting points with the same configuration on the LP branch. HP configuration is also indicated. Figure 18: Normalized User Data Throughput versus C/N If the availability targets are relaxed (e.g. 99 % a.y.) for the Low Priority stream (this could represent the situation in which LP stream does not contain TV programmes but additional data services), the LP branch could benefit from an extra 2 to 4 dB C/N - depending on the climatic zones - that could be used to increase its code rate and transmission capacity. Making reference to figure 11, and assuming an available C/N of 7 dB for HP (99,9 % a.y. availability) and 10,5 dB for LP (99 % a.y. availability), rate 2/3 could be used for HP and rate 1/2 for LP (θ = 15°). For a symbol rate of 27,5 Mbaud, the additional HP capacity would be of about 13,8 Mbit/s (to be compared with the additional capacity of 4,1 Mbit/s, achievable by using DVB-S at rate 3/4 instead of 2/3, 99,9 % availability). 6 Interactive applications Interactive data services may take advantage of the possibility offered by DVB-S2 to change the modulation format and error protection level, by using the ACM functionality, thus allowing to differentiate service levels (priority in the delivery queues, minimum bit-rate, etc.). By means of a return channel informing the transmitter of the actual receiving conditions, the transmission parameters may be optimized for each individual user, dependant on path conditions. ETSI 41 ETSI TR 102 376 V1.1.1 (2005-02) 6.1 IP Unicast Services Figure 19 (derived from figure 12) shows a possible exchange of information (info request and info response) between the user, the Satellite Gateway and one of the information providers during an Internet navigation session by satellite (forward high capacity link) [15]. IP unicast links using DVB-S2 ACM must adapt error protection on a user-per-user basis, where the number of users may be very large (e.g. up to hundreds of thousands). According to the negotiation between the Satellite Terminal (ST) and the "ACM routing manager", an "ACM router" may in principle separate IP packets per user, per required error protection and per service level. The aggregate input traffic on the various protection levels shall not overload the available channel capacity; this applies to the average input traffic, while the peak traffic may temporarily exceed it, compatibly with the input buffering capacity and the service requirements on maximum delays. To fulfil this constraint when the total offered traffic becomes larger than the channel capacity, various strategies may be implemented: for example lower priority IP packets may be delayed (or even dropped) in favour of high priority packets, or the bit-rate delivered to users under poor reception conditions may be reduced. If the control-loop delays (including routing manager and ACM router) are too large to allow error free reception under fast-fading conditions, real time services (e.g. video/audio streaming) may be permanently allocated to a high protection branch, while lower priority services (e.g. best effort) may exploit the higher efficiency branches (i.e. lower cost) provided by ACM. In the ACM router, the polling strategy of the input buffers may be statically or dynamically profiled according to the traffic statistics, the propagation characteristics, and the traffic prioritization policy of the service operator. ACM DVB-S2 SYSTEM ACM Router MCA Buffers per: • Protection level ACM ACM routing etilletaS • user Command High bit-rate manager yawetaG • service level forward-link BUF BUF BUF Info Response Info C/N+I Info Response signalling Response Router Interaction Info channel Satellite Provider GW Terminal Return channel Info Request Figure 19: Example of IP services using a DVB-S2 ACM link As illustrated in clause 4.5.2, the ACM router may interface with the DVB-S2 modulator via a Single Generic Stream input and the ACM Command input, or via Multiple (Transport or Generic) Stream inputs. The choice between the different options has a significant impact on the definition of the system architecture (intended as data processing, routing, buffering and transmission strategy) and consequently on the overall system performance in terms of efficiency, dynamicity, legacy constraints, user packet format, complexity. In the following two clauses, some examples of possible architectures for DVB-S2 unicast systems supporting ACM are shown and some considerations regarding their performance carried out. 6.1.1 Single Generic Stream and ACM command According to this system configuration, the DVB-S2 ACM modulator receives two input signals. The first is the data stream, continuous or packetized. The second is the ACM command, carrying the MODCOD information used by the modulator for encoding and mapping each specific portion of the input data stream. This approach has the important implication that the scheduling function, which performs the selection and aggregation of the information to be transmitted in each frame, is located outside the DVB-S2 modulator subsystem. This strategy leads to a complete transparency of the modulator to layer 2 functions and therefore has the important advantage of an absolute flexibility in the choice of the scheduling algorithm. ETSI 42 ETSI TR 102 376 V1.1.1 (2005-02) It is worth noting that in ACM systems the choice of the physical layer mode to be used in each frame is necessarily linked to the scheduling process, as all the UPs included in one frame are transmitted with the same physical layer parameters. From a different perspective, the data for filling a specific frame need to be selected taking into account the physical layer mode requested by the STs to whom they are addressed. This is the reason why the MODCOD information is generated at the same time of the user data selection and sent as an input to the DVB-S2 ACM modulator. Figure 20 shows the block diagram of a possible architecture for buffering and processing the data prior to conveying them to the ACM modulator. The input data stream, composed of a sequence of User Packets, is routed according to the addressed user and his QoS requirements. Therefore, if L is the number of active users and N the possible QoS levels in the network, L*N is the maximum number of buffers needed to appropriately discriminate UPs. As active users we mean those STs, which have an open session. When no traffic is sent to an individual ST for a certain period of time, a time-out is exceeded and the user is not considered active any more. As a result, the associated buffers are de-allocated. In general, a user can simultaneously support applications requiring different QoS levels. However, as most of the users will probably support less than N simultaneous different applications with different QoS requirements, the number of allocated buffers is consequently reduced. In the simplified case where the service level agreement defines only one QoS level for each user, the number of buffers equals the number L of active users. Figure 20: Block diagram of a possible system architecture with single generic stream input to the ACM DVB-S2 modulator For each frame the merger selects from the input queues a number of packets, and combines them for building a set of information bits. Frame by frame successive data sets composed in this way are sent to the ACM modulator, together with the associated transmission parameters. When the number of bits in one set is not sufficient to completely fill the BBFRAME, the modulator will provide padding by automatically choosing the most suitable type of FECFRAME, with short or normal length. The merger selection is driven by an ACM Routing Manager, which is responsible for packet scheduling. The scheduling policy is application dependent and needs to be designed for maximizing system efficiency while meeting QoS requirements. In order to achieve these goals, the ACM Routing Manager can take advantage of the channel status information reported by the STs, of the different priority levels and QoS requirements of the input queues, and finally of the information concerning the buffer occupation. In fact, the first type of information is needed in order to combine in one frame packets with the same transmission parameters; the second allows for meeting QoS requirements (maximum delay, minimum rate, etc.); the third can be used e.g. for satisfying QoS requirements without sacrificing efficiency (see clause 6.1.3) in presence of scarce traffic associated to a certain physical layer mode. In this case indeed, a smart scheduler policy could decide of merging these few packets with others more numerous requiring a lower efficient physical layer mode. It is worth to mention that the high level of discrimination of the input flow (packets are routed according both to their QoS level and to the addressed user) increases the degree of choice in the packet selection process, allowing for a scheduler policy very flexible and effective. The above described queue organization thus makes easier support of individual ST QoS targets. ETSI 43 ETSI TR 102 376 V1.1.1 (2005-02) Furthermore, in the analysed system architecture the loop delay is minimized, thanks to the fact that the choice of the transmission parameters is made immediately before the encoding and mapping functions. The waiting time in the buffers before the scheduling process is not contributing to the loop delay, as a channel variation occurring when UPs are waiting in the queue does not lead to a wrong physical layer mode selection. On the other hand, no queues are present within the DVB-S2 modulator and the time interval comprised after the decision on the physical layer mode and before signal transmission is minimized. The complete separation between scheduler and modulator allows for a robust system, whose dynamicity is not affected by data buffering delay within the GW. The system good performances in terms of flexibility, channel tracking capabilities and QoS satisfaction need nevertheless to be traded off against system architecture complexity. The quite large number of buffers, though of limited memory size requirements, and their dynamic allocation dependent on user traffic variations make implementation a challenging task. However, a functionally equivalent solution can be implemented with only N input buffers, in which the user packets are separated per QoS levels. As for each frame the ACM Routing Manager shall aggregate packets with the same requirements in terms of transmission parameters, there is a need for accessing all the buffers memory locations, instead of the first one only. Thus, the L*N FIFO buffers of figure 20 can be in principle replaced by N buffers of larger dimension, where all the packets in the buffers are made available to the merger for building the data field. This prevents the need for dynamic buffer allocation, thus simplifying implementation. Other simplified yet high-performance scheduler implementations can be envisaged in the system implementation. What shall be avoided is to devise architectures whereby the STs affected by the less favourable link conditions are significantly impacting the delay in packet delivery to other users [15]. 6.1.2 Multiple (Generic or Transport) Streams Figure 21 shows the block diagram of the DVB-S2 ACM system according to the system configuration where the DVB-S2 modulator interfaces with a number of input data streams. As regarding the data stream format, two solutions are possible: • IP datagrams can be encapsulated in Transport Streams (Multi-Protocol Encapsulation - MPE), according to EN 301 192 [7]; • IP datagrams can be fragmented and encapsulated in variable or fixed length layer-2 packets, or directly mapped in the transmitted TDM stream. MPE or other encapsulation protocols can be assumed. As explained in clause 4.5.2, the merger within the DVB-S2 modulator reads the data fields from one of its inputs. Since physical layer mode homogeneity is required for the data field, each data stream within the modulator needs to be associated with a certain physical layer mode. For this reason, the ACM router splits the users' packets per service level (priority) and per required protection level, and sends them to the multiple DVB-S2 input interfaces, each stream being permanently associated to a given protection level. Therefore, each input stream merges the traffic of all the users needing a specific protection level, and its useful bit-rate may (slowly) change in time according to the traffic characteristics. ETSI 44 ETSI TR 102 376 V1.1.1 (2005-02) DVB-S2 Modulator DVB-S2 Demodulator MPEG-TS only Protection Framing decoding VBR level 1 NP TS /Protection level Source 1 deletion selection TS & MUX Buffer GTW IP (TS only) NP IP TS 2 M ACM Re- E ACM services GTW MUX Mod insertion ACM R Dem (TS only) Router G & E Buffer R TS CBR NP CBR Source K TS deletion IP & GTW MUX Buffer SNIR RTS (TS only) measure recovery Protection ACM Protection level K Routing Level control Manager Return channel NOTE: For Generic input Streams, GTWs, TS Muxes and null-packet deletion are not required. Figure 21: IP Unicasting and ACM: Multiple input streams - uniform protection per stream Dotted boxes in figure 21 address the specific case of IP services encapsulated in Transport Streams (Multi-Protocol Encapsulation - MPE), according to EN 301 192 [7]. In this case, K MPE gateways (GTWi) are associated to K TS multiplexers, to feed K DVB-S2 input streams (one per active protection level). Null-packet deletion, applied to each branch, reduces the transmitted bit-rate. The decoded TS, after null packets re-insertion, is a valid TS (the input stream synchronizer may optionally be activated). To fully exploit the potential ACM advantages, the additional control-loop delays introduced by the TS-specific equipment (Gateways, TS Muxes) should be minimized. The Merger/Slicer in figure 21 cyclically polls the input buffers, and conveys to the ACM modulator a block of users' data ready to fill (or partially fill) a Data Field. A timeout may be defined in order to avoid long delays in each merger/slicer buffer. During traffic peaks, overloading the physical channel, a simple round-robin policy may not fulfil the requirements of suitable distribution of the available throughput among users. Therefore, alternative policies to profile the round-robin priority may be adopted. It is important to note that the queuing time spent in the Merger/Slicer inputs buffers contributes to the overall ACM control-loop delay. In fact, channel variations occurring within the time interval between physical layer mode decision and signal transmission can lead to the utilization of a sub-optimum physical layer mode or to unsuccessful decoding of UPs. To the purpose of minimizing such waiting time, a time out can be introduced. Moreover, other approaches can be followed, such as appropriately limiting the dimension of the Merger/Slicer input buffers according to the system requirements. By limiting the input buffer size to a small number of data fields, the queuing time would be consequently reduced. However, an additional buffer is needed before the ACM router to absorb traffic peaks, whose size shall be designed taking into account the IP traffic flow at the GW input. Figure 22 shows the block diagram of a possible system architecture, which follows the approach described above. If we assume N possible QoS levels in the network, the input data packets are first split according to their QoS requirements. As an individual ST can support simultaneously several services with different requirements, its packets can be routed to different buffers. The ACM router, driven by the ACM Routing Manager, reads the UPs in the input FIFO queues and separates them according to the channel status indications sent by the STs via the return channel. As for the architecture described in the previous clause, the ACM routing manager policy may also take advantage of the information concerning buffer occupation in order to increase traffic aggregation and thus maximizing efficiency. ETSI 45 ETSI TR 102 376 V1.1.1 (2005-02) Figure 22: Block diagram of a possible system architecture with multiple DVB-S2 input streams (one per protection level) In the multiple input streams configuration, unlike the single input stream configuration, the scheduling functions are not completely decoupled with respect to the DVB-S2 modulator. On the contrary, part of the scheduling functionalities is associated to the merging operation, and part takes place outside the DVB-S2 modulator and can be therefore system and application dependent. The overall system performance is thus driven both by the merger policy (cyclic polling in the reference case) and by the routing/scheduling algorithms applied outside the modulator. In the system architecture case presented here the buffer organization is definitely less complex than the one described in the previous clause. However, simple FIFO queues, where UPs are aggregated without any differentiation, coupled with a round robin merging policy, can present some performance limitations when adaptive systems are considered. The impact on throughput and packet delay of this system architecture is analysed in depth in clause 6.1.4. 6.1.3 Encapsulation efficiency of ACM modes The Internet Protocol (IP) interconnects multiple networks attached to the Internet and the DVB-S2 network can be seen as another upcoming access network offering adaptive physical layer. When an IP packet enters/exits a particular access network, it can be encapsulated into a local packet or capsule, having only meaning within the local network and adding some additional overhead. The capsule of DVB-S2 is called BBFRAME (in information bits) or FECFRAME (in encoded bits). The layered architecture of the DVB-S2 interface is presented in figure 23. ETSI 46 ETSI TR 102 376 V1.1.1 (2005-02) IP MPEG/MPE 80 bits 0 1+α. In [9] a carrier spacing of BS/Rs > 1+α was therefore adopted as a general rule and it was indicated that lower values should be applied only with caution after a case-by-case study. This is further analyzed below for DVB-S2 carriers. The degradation due to adjacent channels depends on the C0/(N0+I0) failure point of the DVB-S2 mode (MODCOD) and on the relative level of the interfering adjacent carriers. These effects are illustrated in figure 30, and a recommendation for carrier spacing is given in table 14. In deriving figure 30 and table 14, it was assumed that: • TX and RX filters satisfy the normative/recommended upper masks; • all carriers have the same symbol rate and roll-off factor; • the C0/(N0+I0) failure point is 2 dB above the value listed in [2], table 13 (pessimistic assumption as safety margin); • the TX HPA is operated in linear mode with 40 dB regrowth rejection (typically corresponding to 7 dB to 11 dB HPA OBO); • the transmit frequency uncertainty is less than 0,5 % of the symbol rate. It is seen that the value BS/Rs > 1+α is about right for 16APSK 4/5 with two adjacent carriers at +8 dB, but that significantly lower carrier spacings are possible in many cases. Additional cases are summarized in table 14. Situations where the carrier under test and the adjacent carriers do not have identical symbol rate and roll-off factor should be analyzed separately. NOTE: Curves are shown for all roll-off factors and for an interfering carrier level +0 dB and +8 dB above the carrier under test. Use this figure with caution if the assumptions listed in the main text are not satisfied. Figure 30: Expected degradation caused by adjacent channel interference for QPSK 3/4 and 16APSK 4/5 modes and assuming two adjacent carriers using the same symbol rate and the same roll-off factor ETSI 55 ETSI TR 102 376 V1.1.1 (2005-02) Table 14: Typical minimum carrier spacing BS/Rs Mode of carrier under test Roll-off factor Adjacent carrier level Adjacent carrier level +4 dB each +8 dB each QPSK 3/4: Failure point α = 0,20 1,10 (see note 2) 1,11 assumed α = 0,25 1,10 (see note 2) 1,14 C0/(N0+I0) = 4 dB+2 dB α = 0,35 1,14 (see note 2) 1,22 8PSK 3/4: Failure point α = 0,20 1,11 1,17 assumed α = 0,25 1,14 1,21 C0/(N0+I0) = 8 dB+2 dB α = 0,35 1,22 1,30 16APSK 4/5 Failure point α = 0,20 1,15 1,20 assumed α = 0,25 1,19 1,25 C0/(N0+I0) = 11 dB+2 dB α = 0,35 1,27 1,35 NOTE 1: Use this table with caution if the assumptions listed in the main text are not satisfied. NOTE 2: Values below 1,10 were not allowed in the table. 7.2.5 Link budget examples for DSNG 7.2.5.1 Generic Hypothesis In order to illustrate some potential examples of the use of the system, link budget analysis have been carried out assuming the following hypothesis: • It is considered a full 36 MHz transponder loaded with 4 equals digital carriers each one in a 9 MHz bandwidth slot. • A symbol rate of 7,20 Mbaud in 9 MHz (BW/Rs = 1,25) is considered. In table 15 the useful bit rate at the DVB-S2 modulator input for some transmission modes is summarized. Table 15: Useful bit rate (Mbit/s) for a symbol rate of 7,2 Msymb/s QPSK 8PSK 16APSK 1/2 2/3 3/4 5/6 8/9 2/3 3/4 5/6 3/4 4/5 5/6 7,12 9,52 10,71 11,91 12,72 14,26 16,04 17,85 21,36 22,79 23,76 The satellite TWTA overall operating point is related to typical multicarrier per transponder operation in a non-linearized transponder: total IBO = 8 dB, total OBO = 3,4 dB. Satellite resources per carrier are, assuming that the percentage of the power consumption per carrier is the same as the percentage of bandwidth consumption per carrier: • Power resources per carrier = 1/4 of the total power taking into account the overall operating point. • Bandwidth resources per carrier = 1/4 of the total transponder bandwidth (i.e. 9 MHz). • Quality of the links: PER equal to 10-7. • Es/N0 performance as summarizes in table D.2 (clause D.3). Transmit DSNG transportable earth station characteristics: • Location: at beam edge. • Antenna diameter Φ = 0,9, 1,2 and 1,8 m. • 65 % antenna efficiency, 0,3 dB coupling losses, 0,5 dB pointing losses. • Equipped with a TWT amplifier of 250 W. • Maximum operational EIRP (for 3 dB OBO) = 67 dBW for 1,8 m, 63 dBW for 1,2 m and 60 dBW for 0,9 m). ETSI 56 ETSI TR 102 376 V1.1.1 (2005-02) Receive earth station characteristics for DSNG transmissions: • Location: at beam edge. • Antenna Diameter Φ = 2,4 m (G/T = 25 dB/K); 4,5 m (G/T = 30 dB/K) and 8,1 m (G/T = 35 dB/K). • 65 % antenna efficiency, 0,3 coupling losses, 0,5 pointing losses, 1,2 dB noise figure. Satellite characteristics: • G/T = 5,5 dB/K at beam centre (-0,5 dB/K at beam edge). • EIRP (at saturation) = 50 dBW at beam centre (42 dBW at beam edge). • IPFD = -85,5 dBW/m2 (Nominal gain - NG) at beam centre (-79, dBW/m2 at beam edge). A low gain setting of the IPDF = -82,5 dBW/m2 is also considered. Other assumptions are: • Reference satellite orbital location: 0ºE. • Uplink frequency: 14,25 GHz. • Downlink frequency:11,75 GHz. • Sea level height for the transmit and receive earth stations: 100 m. • Atmospheric absorption: 0,3 dB for uplink and 0,2 dB for the downlink. • Worst case polarization (Linear horizontal). In figure 31 an example link budget is given. ETSI 57 ETSI TR 102 376 V1.1.1 (2005-02) Figure 31: Example DSNG link budget ETSI 58 ETSI TR 102 376 V1.1.1 (2005-02) 7.2.5.2 DSNG Examples Clear Sky Margin Figures 32, 33 and 34 summarize the results in terms of clear sky margins obtained for the DSGN links using the above hypothesis for the transponder nominal gain operation. 0,9m terminal (maximum EIRP= 60 dBW) (NG) 9,00 8,00 Clear Sky margin (dB) 7,00 6,00 Rx:2,4 m 5,00 Rx: 4,5 m 4,00 3,00 Rx: 8,1 m 2,00 1,00 0,00 QPSK 1/4 QPSK 1/3 QPSK 2/5 QPSK 1/2 QPSK 3/5 QPSK 2/3 QPSK 3/4 QPSK 4/5 QPSK 5/6 QPSK 8/9 8-PSK 2/3 8-PSK 3/4 8-PSK 5/6 16-APSK 3/4 16-APSK 4/5 16-APSK 5/6 Modulation Scheme Figure 32 1,2m terminal (maximum EIRP= 63 dBW) (NG) 9,00 8,00 Clear Sky margin (dB) 7,00 6,00 Rx:2,4 m 5,00 Rx: 4,5 m 4,00 3,00 Rx: 8,1 m 2,00 1,00 0,00 QPSK 1/4 QPSK 1/3 QPSK 2/5 QPSK 1/2 QPSK 3/5 QPSK 2/3 QPSK 3/4 QPSK 4/5 QPSK 5/6 QPSK 8/9 8-PSK 2/3 8-PSK 3/4 8-PSK 5/6 16-APSK 3/4 16-APSK 4/5 16-APSK 5/6 Modulation Scheme Figure 33 1,8m terminal (maximum EIRP= 67 dBW) (NG) 9,00 8,00 Clear Sky margin (dB) 7,00 6,00 Rx:2,4 m 5,00 Rx: 4,5 m 4,00 3,00 Rx: 8,1 m 2,00 1,00 0,00 QPSK 1/4 QPSK 1/3 QPSK 2/5 QPSK 1/2 QPSK 3/5 QPSK 2/3 QPSK 3/4 QPSK 4/5 QPSK 5/6 QPSK 8/9 8-PSK 2/3 8-PSK 3/4 8-PSK 5/6 16-APSK 3/4 16-APSK 4/5 16-APSK 5/6 Modulation Scheme Figure 34 ETSI 59 ETSI TR 102 376 V1.1.1 (2005-02) Figures 35, 36 and 37 provide the results for the computations in transponder low gain setting. 0,9m terminal (maximum EIRP= 60 dBW) (LG) 8,00 7,00 Clear Sky margin (dB) 6,00 5,00 Rx:2,4 m 4,00 Rx: 4,5 m 3,00 Rx: 8,1 m 2,00 1,00 0,00 QPSK 1/4 QPSK 1/3 QPSK 2/5 QPSK 1/2 QPSK 3/5 QPSK 2/3 QPSK 3/4 QPSK 4/5 QPSK 5/6 QPSK 8/9 8-PSK 2/3 8-PSK 3/4 8-PSK 5/6 16-APSK 3/4 16-APSK 4/5 16-APSK 5/6 Modulation Scheme Figure 35 1,2m terminal (maximum EIRP= 63 dBW) (LG) 9,00 8,00 Clear Sky margin (dB) 7,00 6,00 Rx:2,4 m 5,00 Rx: 4,5 m 4,00 3,00 Rx: 8,1 m 2,00 1,00 0,00 QPSK 1/4 QPSK 1/3 QPSK 2/5 QPSK 1/2 QPSK 3/5 QPSK 2/3 QPSK 3/4 QPSK 4/5 QPSK 5/6 QPSK 8/9 8-PSK 2/3 8-PSK 3/4 8-PSK 5/6 16-APSK 3/4 16-APSK 4/5 16-APSK 5/6 Modulation Scheme Figure 36 Clear Sky margins for DVB-S2-DSNG links (beam centre-to-Beam centre) using 1,8m terminal (maximum EIRP= 67 dBW) (LG) 10,00 Clear Sky margin (dB) 8,00 Rx:2,4 m 6,00 Rx: 4,5 m 4,00 Rx: 8,1 m 2,00 0,00 QPSK 1/4 QPSK 1/3 QPSK 2/5 QPSK 1/2 QPSK 3/5 QPSK 2/3 QPSK 3/4 QPSK 4/5 QPSK 5/6 QPSK 8/9 8-PSK 2/3 8-PSK 3/4 8-PSK 5/6 16-APSK 3/4 16-APSK 4/5 16-APSK 5/6 Modulation Scheme Figure 37 ETSI 60 ETSI TR 102 376 V1.1.1 (2005-02) Link availability In figures 38 and 39 it is summarized the availability (% of the averaged year) in terms of the required margin (dB) for Ku band (14/12 GHz), following the ITU model (P.618-7, P.837-3, P.839-3), for a number of cities. ITU rain characteristics for the examples are summarized in table 16. Table 16: ITU rain characteristics City Country I0,001 mm/h ho (m) Sea level height (P.837-3) P.839-3 (m) Miami USA 95,8 4 209 0 La Paz Bolivia 84,7 4 829 2 761 Sao Paulo Brasil 68,1 4 181 708 Rome Italy 41,0 2 685 73 Paris France 25,4 2 226 92 Madrid Spain 18,7 2 647 786 8 MIAMI LA P AZ 7 S AO P AULO ROME 6 P ARIS MADRID 5 4 3 2 1 0 99 99.1 99.2 99.3 99.4 99.5 99.6 99.7 99.8 99.9 100 Figure 38: Attenuation (dB) versus availability (% a.y.) @ 14 GHz 6 MIAMI LA P AZ S AO P AULO 5 ROME P ARIS MADRID 4 3 2 1 0 99 99.1 99.2 99.3 99.4 99.5 99.6 99.7 99.8 99.9 100 Figure 39: Attenuation (dB) versus availability (% a.y.) @ 12 GHz ETSI 61 ETSI TR 102 376 V1.1.1 (2005-02) Overall Availability Considerations For a given service an overall availability is usually requested. Considering a 99,5 % availability for a link, for example from Rome (up-link) to Paris (down-link ), the global availability can be computed as follows. It can be assumed that the rain does not occur simultaneously in both places (up and down-link). The total unavailability (0,5 %) can be equally divided: 0,25 % in uplink and 0,25 % for the downlink. The margin required to assure this availability (99,75 %) could be estimated following the above graphics: about 2 dB for Rome in the up-link (14 GHz) figure 38, and 0,6 dB for Paris in the down-link (12 GHz) (figure 39). Following this example, to assure an overall availability of 99,5 %, a clear sky margin of about 2,6 dB should be required. In the DSNG examples, for 0,9 m DSNG transmitting antenna and 2,4 m receiving antenna in NG (see figure 32 for the 99,5 % ), a maximum useful rate of 11,96 MHz (QPSK 5/6) can be transmitted (2,6 dB clear sky margin operation). Taking advantage of the new DVB-S2 features, under clear sky conditions the margin can be used to increase the useful data rate up to 16,4 Mb/s with 8PSK 3/4. 7.2.6 DSNG transmitting station identification The proliferation of SNG stations and, in particular, the fact that operators sometimes do not strictly adhere to standard operating regulations, has created the problem of how to identify the interference origin. This happens also for fixed earth stations, that, having to be repositioned when working with several satellites, are very often not correctly aligned. Solutions to this problem are nowadays mostly based on operational rules (by recommending mandatory procedures when operating satellite stations). A technical device which gives information about the transmitting station identification is necessary to facilitate application of the operational rules. The DVB-S2 specification allows to use the physical level scrambling sequence as signature sequence for identification of the interfering signal. The number of possible codes offered by this process seems to be largely sufficient. The operational requirements on: • A registration procedure must be established. The elements to be registered are (i) a country code and (ii) identification of the owner of the station. A similar procedure already exists in the DVB structure. • The identification code has to be set up by the supplier of the equipment and access to the identification code has to be protected. This will guarantee the system against tampering to disable or modify the identification maliciously. • The identification process for a previously unknown carrier should be as rapid as possible. • Transmission on the satellite could be of very short duration; "a few minutes" can be taken as a typical minimum value. In addition to this, it would be useful if information about the geographical location of the station could be included in the stream. Inclusion of this information must be fully automatic in order to avoid error and falsification. The knowledge of geographical location information is essential, in any case, to permit correct setting up of the station antenna. 7.2.7 DSNG Services using ACM In point-to-point ACM links, where a single TS is sent to a unique receiving station (e.g. DSNG), the TS packets protection must follow the C/N+I variations on the satellite channel in the receiving location. When propagation conditions change (see figure 40, yellow arrow), the PL frames Fi switch from protection mode Mj to protection mode Mk to guarantee the service continuity. Constant Transport Stream bit-rate and end-to-end delay, as required by MPEG, may be guaranteed by using DVB-S2 stream adaptation tools which are described in detail in clause 4.5.2.1. ETSI 62 ETSI TR 102 376 V1.1.1 (2005-02) Mj Mk F1 F2 F3 F4 F5 Figure 40: PLFRAMEs changing protection during a rain fading The DVB-S2 system may operate as follows (see figure 41): 1) the bit-rate control unit keeps the video encoder bit-rate at the maximum level compatible with the actual C/N+I channel conditions. In parallel, it may set the DVB-S2 modulator transmission mode via the "ACM Command" input port. 2) The variable bit-rate (VBR) source encoder outputs a constant bit-rate transport stream, where rate variations of the useful bit-rate are compensated by the insertion of MPEG null-packets. 3) The Null Packets (NP) are deleted in the Mode Adapter, so that the actual bit-rate on the channel corresponds to the source bit-rate [15]. The deleted NPs are signalled in the DNP byte. 4) The receiver re-inserts Null Packets exactly in the original position, and the Transport Stream clock is regenerated using the Input Stream Clock Reference (see clause 4.5.2.1). With reference to figure 41, during a deep fading the bit rate control unit may impose a rate reduction first on the source encoder, and only after the command has been executed (e.g. after 100 ms to 500 ms), to the DVB-S2 modulator (via ACM Command). A drawback of this configuration is that the video encoder and MUX delays (D5 and D6 in figure 41) are included in the control loop, with the risk of service outage under deep fading conditions. To overcome this additional delay the ACM Command can be instantly delivered also to the modulator, but to avoid packet losses large buffers have to be inserted in the DVB-S2 modulator and demodulator (see clause 6.1.1). TS Satellite RTS=50 Mbit/s channel TS DVB-S2 Modulator D1=260 ms DVB-S2 Demodulator CBR RSource= Framing decoding ÷ 10÷50 Mbit/s NP Deletion ACM Framing NP TS Video TS & & ACM Re- DE VBR MUX Buffer Mod signalling Dem insertion MUX Source & FIFO D6= Buffer D5= 1-2 ms 100-500 ms SNIR measure RTS D7=10 ms D7=1ms D8=10ms recovery ACM Command D2=1 ms Short loop Bit-rate D4=200-300 Return channel D3=100 ms control modem unit Figure 41: Single TS - uniform protection for long periods: transmission and receiving schemes ETSI 63 ETSI TR 102 376 V1.1.1 (2005-02) Annex A: Low Density Parity Check Codes LDPC codes [11],[22],[23] are linear block codes with sparse parity check matrices H ( N − K ) × N , where each block of K information bits is encoded to a codeword of size N. As an example, an LDPC code of codeword size N = 8 and rate 1/2 can be specified by the following parity check matrix. n1 n2 n3 n4 n5 n6 n7 n8 1  0 0 1 1 0 0 1  m1 0  1 1 0 1 0 1 0  m2 H=   1  0 1 0 0 1 0 1  m3   0  1 0 1 0 1 1 0  m4 The same code can be equivalently represented by the bipartite graph in figure A.1 which connects each check equation (check node) to its participating bits (bit nodes). n1 n2 n3 m1 n4 m2 n5 m3 n6 m4 n7 bit nodes n8 check nodes Figure A.1: Bipartite graph of an LDPC code The purpose of the decoder is to determine the transmitted values of the bits. Bit nodes and check nodes communicate with each other to accomplish that. The decoding starts by assigning the received channel value of every bit to all the outgoing edges from the corresponding bit node to its adjacent check nodes. Upon receiving that, the check nodes make use of the parity check equations to update the bit node information and send it back. Each bit node then performs a soft majority vote among the information reaching from its adjacent check nodes. At this point, if the hard decisions on the bits satisfy all of the parity check equations, it means a valid codeword has been found and the process stops. Otherwise bit nodes go on sending the result of their soft majority votes to the check nodes. The decoding algorithm is briefly introduced in [2], clause G.2, and a detailed explanation can be found in [12]. ETSI 64 ETSI TR 102 376 V1.1.1 (2005-02) A.1 Structure of Parity Check Matrices of Standardized LDPC Codes LDPC codes can be specified through their parity check matrices. However, in general, generator matrices are needed for encoding. Of course, for any linear code once the parity check matrix is known, a generator matrix can be derived using for instance Gaussian elimination method. But, even though parity check matrices of LDPC codes are sparse, the resulting generator matrix would no longer be sparse leading to storage and encoding complexity problems, since the standardized LDPC codes are tens of thousands of bits long. Therefore, to facilitate the description of the codes and for easy encoding, certain structure has been imposed on the parity check matrices H of the DVB-S2 codes. Even though there are methods that partially solve the problem [24], restricting a sub-matrix of the parity check matrix to be lower triangular eliminates the need to derive a generator matrix and leads to linear encoding complexity. More specifically, the parity check matrix of the DVB-S2 codes was restricted to the form H ( N − K ) × N = [ A( N − K ) × K B( N − K ) ×( N − K ) ] where B is staircase lower triangular as in figure A.2. 1 1 1 1 1 0 B= 1 . . . 0 1 1 1 Figure A.2: Submatrix of Parity Check Matrix Then any information block i = (i0 , i1 ,..., ik −1 ) is encoded to a codeword c = (i0 , i1 ,..., ik −1 , p0 , p1 ,... pn −k −1 ) using HcT = 0, and recursively solving for parity bits. a 00 i0 + a 01i1 + ... + a 0,k −1i k −1 + p 0 = 0 ⇒ Solve p0 a10 i0 + a11i1 + ... + a1,k −1i k −1 + p0 + p1 = 0 ⇒ Solve p1 : : : a N − K −1,0 i0 + a N − K −1,1i1 + ... + a N − K −1,k −1ik −1 + p N − K −2 + p N − K −1 = 0 ⇒ Solve p N − K −1 The matrix A is sparse; as a result encoding has linear complexity with respect to the block length. Simulations show that the above lower-triangular restriction on the parity check matrix lead to negligible (within 0,1 dB) performance loss with respect to a general parity check matrix for the cases of relevant interest. Furthermore to reduce the storage requirement of the matrix description by a factor of M, the following restriction on the A submatrix of parity check matrix design has been applied. For a group of M bit nodes, if the check nodes connected to the first bit node of degree, say d v , are numbered as a1 , a2 ,..., a d v then the check nodes connected to i th bit node (i ≤ M ) are numbered as, {a1 + (i − 1) q} mod( N − K ), {a2 + (i − 1) q} mod( N − K ),....., {a d v + (i − 1) q} mod( N − K ) where N − K = total number N −K of check nodes and q = . M For the following groups of M bit nodes, the check nodes connected to the first bit node of the group are in general randomly chosen so that the resulting LDPC code is cycle-4 free and occurrence of cycle-6 is minimized. From the above description, it is clear that adjacent check nodes of only one bit node need to be specified in a group of M. In DVB-S2, M is equal to 360. ETSI 65 ETSI TR 102 376 V1.1.1 (2005-02) A.2 Description of Standardized LDPC Codes In DVB-S2, a wide range of bandwidth efficiency from 0,5 bits/symbol up to 4,5 bits/symbol is covered by defining ten different code rates 1/4, 1/3, 2/5,1/2, 3/5, 2/3, 3/4, 4/5, 5/6, 8/9 and 9/10 with four different modulation schemes QPSK, 8PSK, 16APSK and 32APSK. Rate 1/4, 1/3, 1/2 and 3/5 codes are also used in the low priority branch of hierarchical 8PSK of backward compatible mode. These codes are optimized for broadcast modes. For each code rate, a parity check matrix is specified by listing adjacent check nodes for the first bit node in a group of M = 360. The coded block length is N = 64 800 bits for all rates for broadcast mode. To improve the performance, irregular LDPC codes are used where degrees of bit nodes are varying [25]. The list of bit node degrees and the total number of nodes with those degrees are shown in table A.1 for all the code rates of this length. Table A.1: Number of Bit Nodes of Various Degrees for N = 64 800 codes Code Rate 13 12 11 8 4 3 2 1 1/4 5 400 10 800 48 599 1 1/3 7 200 14 400 43 199 1 1/2 12 960 19 440 32 399 1 3/5 12 960 25 920 25 919 1 2/3 4 320 38 880 21 599 1 3/4 5 400 43 200 16 199 1 4/5 6 480 45 360 12 959 1 5/6 5 400 48 600 10 799 1 8/9 7 200 50 400 7 199 1 9/10 6 480 51 840 6 479 1 For non-broadcast applications, a set of codes with N = 16 800 has also been generated with the same value of M, i.e. M = 360. Table A.2 shows the list of bit node degrees and the total number of node with those degrees. The q value for the 64 800 bit codes and the 16 200 bit codes are listed in tables A.3 and A.4, respectively. Constellation labellings for all the modulations specified are shown in figure 1. Table A.2: Number of Bit Nodes of Various Degrees for N = 16 200 codes Code Rate 13 12 11 8 4 3 2 1 1/5 360 2 880 12 959 1 1/3 1 800 3 600 10 799 1 2/5 2 160 4 320 9 719 1 4/9 1 800 5 400 7 999 1 3/5 3 240 6 480 6 479 1 2/3 1 080 9 720 5 399 1 11/15 360 11 520 4 319 1 7/9 12 600 3 599 1 37/45 360 12 960 2 879 1 8/9 1 800 12 600 1 799 1 Table A.3: q Values for Codes with N = 64 800 bits Code Rate q 1/4 135 1/3 120 2/5 108 1/2 90 3/5 72 2/3 60 3/4 45 4/5 36 5/6 30 8/9 20 9/10 18 ETSI 66 ETSI TR 102 376 V1.1.1 (2005-02) Table A.4: q Values for Codes with N = 16 200 bits Code Rate q 1/4 36 1/3 30 2/5 27 1/2 25 3/5 18 2/3 15 3/4 12 4/5 10 5/6 8 8/9 5 A.3 Performance Results Even though the above design restricts the parity check matrix to be structured, the performance is still very good due to the careful choice of check node/bit node connections. Performance of various code rates with different constellations on AWGN channel is depicted in figure A.3 for N = 64 800, for QPSK, 8PSK, 16APSK and 32APSK modulation. Each LDPC frame is divided to form multiple MPEG packets, 188 bytes each. Since the error rate requirements of DVB-S2 are rather stringent (10-7 packet error rate), an outer BCH code with the same block length as LDPC frame and an error correction capability of up to 12 bits is employed, as defined in [2]. Performance of the N = 16 200 bit codes for QPSK modulation is shown in figure A.4. Typically, these codes are about 0,25 dB to 0,3 dB worse than the N = 64 800 bit codes. ETSI 67 ETSI TR 102 376 V1.1.1 (2005-02) QPSK 8-PSK 1.E-01 3/5 1.E-01 8/9 3/5 3/4 2/3 9/10 9/10 1/2 5/6 8/9 1.E-02 1.E-02 3/4 2/3 5/6 Packet Error Rate Packet Error Rate 1.E-03 4/5 1.E-03 1.E-04 1.E-04 1.E-05 1.E-05 1.E-06 1.E-06 1.E-07 1.E-07 0 1 2 3 4 5 6 7 5 6 7 8 9 10 11 12 Es/No (dB) Es/No (dB) (a) (b) 16-APSK 32-APSK 1.E-01 3/4 8/9 1.E-01 5/6 4/5 8/9 4/5 3/4 1.E-02 1.E-02 2/3 9/10 5/6 Packet Error Rate Packet Error Rate 1.E-03 1.E-03 1.E-04 1.E-04 1.E-05 1.E-05 1.E-06 1.E-06 1.E-07 1.E-07 8 9 10 11 12 13 14 12 13 14 15 16 Es/No (dB) Es/No (dB) (c) (d) Figure A.3: Performance of LDPC+BCH Codes over AWGN Channel, N = 64 800 bits, (a) QPSK, (b) 8PSK, (c) 16APSK, (d) 32APSK ETSI 68 ETSI TR 102 376 V1.1.1 (2005-02) QPSK 1.E+00 4/5 1.E-01 1/2 1/3 8/9 5/6 1.E-02 3/5 Packet Error Rate 2/5 2/3 1.E-03 3/4 1.E-04 1.E-05 1.E-06 1.E-07 -2.0 -1.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 Es/No (dB) Figure A.4: Performance of LDPC+BCH Codes over AWGN Channel, N = 16 200 bits ETSI 69 ETSI TR 102 376 V1.1.1 (2005-02) Annex B: DVB-S2 Physical Layer Frame and pilot structure Frame synchronization is needed to indicate the start of each FEC block for the decoder. It also provides the necessary information for the receiver to apply the appropriate demodulator and decoder to demodulate and decode the transmitted information. Given that some overhead is necessary for frame synchronization, it is also designed such that it can be used to reduce initial frequency and phase uncertainty of the modulated signal. The frame synchronization is designed to provide reliable operation in the worst case Es/N0 with minimum overhead. It is also used to minimize the demodulator implementation loss in the presence of consumer quality low-noise-block (LNB) phase noise. In fact, phase noise is particularly detrimental to demodulator performance for higher-order modulation such as 8PSK, 16APSK, and 32APSK. To preserve the near Shannon limit performance of the DVB-S2 FEC, pilot symbols may be added to assist the demodulator to minimize probability of cycle-slips and to provide more accurate phase estimates. These pilot symbols are also designed to use a minimum overhead of the overall bandwidth, and can be turned on or off as desired. The frame synchronization structure and the pilot structure are described in this annex. The frame and carrier synchronization algorithms that make use of this framing structure are described in annex C. B.1 Structured PLS code for Frame Synchronization Figure B.1 illustrates the general structure of DVB-S2 physical layer frames. Each LDPC coded block is preceded by the Start of Frame (SOF) and the Physical Layer Signalling (PLS) code (PLSCODE). SOF is a known 26-symbol pattern. PLSCODE is a 64-bit linear binary code, which conveys 7 bits of information with a minimum distance 32, i.e. a [64, 7, 32] code. In total, SOF and PLSCODE occupy one slot (90 symbols). SOF PLSCODE LDPC coded frame Figure B.1: Frame structure of DVB-S2 Independent from the modulation scheme of the LDPC coded block that follows, these fields are modulated by π/2-BPSK modulation to reduce the envelope fluctuation in comparison with the classic BPSK scheme. /2-BPSK π rotates the signal constellation 90 degrees every symbol. The 7 bits carried by the PLSCODE inform the receivers about the modulation scheme, code rate, pilot configuration, and length of the LDPC coded data. In the broadcasting mode, the length of LDPC codes is 64 800 coded bits regardless of code rates and modulation schemes. In the Adaptive Coding and Modulation (ACM) mode, LDPC codes of length 16 200 can also be used. Once frame and phase synchronization are acquired, the probability of incorrectly decoding the PLSCODE is negligible. As long as the PLSCODE is correctly decoded, the next SOF can be located and the frame synchronization can be maintained. In the broadcast mode, since modulation and coding scheme will not change on a frame-by-frame basis, maintaining frame synchronization does not require the correct decoding of PLSCODE. Instead, frame synchronization can be maintained as long as the symbol timing loop is in lock. In both cases, the frame synchronization can be maintained after it is initially acquired. Therefore, we will focus our discussions on the initial frame synchronization. Extensive analysis showed that SOF by itself is too short to provide reliable and rapid frame synchronization. Given that PLSCODE uses a rather low rate code, it is natural to consider embedding certain structures into this code to aid the initial frame synchronization. There are many ways to construct a [64, 7, 32] linear block code. For instance, it can be constructed as an extended BCH code, the dual of an extended Hamming code, an extended maximum length code, or a first-order Reed-Muller code [39]. In the DVB-S2 standard, a construction is used that is particularly useful for rapid frame synchronization and efficient maximum likelihood (ML) decoding. ETSI 70 ETSI TR 102 376 V1.1.1 (2005-02) The construction utilizes the first-order Reed-Muller code of parameters [32, 6, 16]. A generator matrix for [32, 6, 16] Reed-Muller code is shown as follows: 01010101010101010101010101010101 00110011001100110011001100110011 00001111000011110000111100001111 00000000111111110000000011111111 00000000000000001111111111111111 11111111111111111111111111111111 The generator matrix can be constructed recursively by the well-known |u|u+ v| construction. This notation indicates how to use two codes of length n to construct a code of length 2n, i.e. u and v are drawn from each of the component codes respectively. In the case of a first order Reed-Muller code, v uses the trivial linear code 0 and 1, the all-zero and all-one vectors of length n, as codewords, and u belongs to a first-order Reed-Muller code of length n. The formulation of a PLSCODE codeword is shown in figure B.2. For 7 information bits, we encode the first six bits by the [32,6,16] first-order Reed-Muller code to obtain a binary vector Y. The vector Y is further duplicated into two identical vectors. Figure B.2: Construction of PLSCODE Every bit at the lower branch is binary summed with the seventh information bit. The upper and the lower branches are multiplexed bit by bit to form a vector of length 64. In other words, let Y = (y0,y1,·· ,y31) be a codeword of the first-order Reed-Muller code [32,6,16]. Then two code words of the [64,7,32] code can be generated as ( y0 , y0 , y1 , y1 ,..., y31 , y31 ) and ( y0 , y0 , y1 , y1 ,..., y31 , y31 ) respectively, where y represents the binary complement of y. Instead of bit by bit multiplexing the upper and lower vector, if the two vectors were cascaded together, it would have resulted in the |u|u+ v| construction of the first-order Reed-Muller code of parameters [64, 7, 32]. This shows that the PLSCODE constructed in such a way is actually an interleaved first-order Reed-Muller code of parameters [64, 7, 32], which is known to achieve the best minimum distance for a binary [64,7] code, therefore, the code as constructed in figure B.2 is an optimal binary[64,7] code.  0 if y 2i + 1 = y 2 i A very useful property of the code for frame synchronization is that y 2i ⊕ y2i +1 = y 2i + 1 = y 2 i   1 if for i = 0, 1,…, 31. Therefore, if we take a 64-bit codeword and form 32 pair-wise differences between the adjacent bits, we will obtain the same result for all 32 pairs. We can improve the reliability of the difference by averaging all 32 values. If the modulation is BPSK, the pair-wise difference can average to a value equal to ∆ or ∆+180°. We can easily detect the seventh information bit and ∆, since the pair wise difference represents frequency offset, and if the frequency offset is small. For /2-BPSK modulation, a constant /2 shift need to be subtracted out from each pair-wise difference, the π π result is the same. ETSI 71 ETSI TR 102 376 V1.1.1 (2005-02) As DVB-S2 requires the receivers to acquire with a initial frequency offset up to 5 MHz, i.e. as large as 25 % of the symbol rate for a typical 20 Msymbol/s data stream, the carrier phase can rotate up to 90° in one symbol interval. There are only two options in such a scenario: frequency offset resistant non-coherent differential detection or searching through multiple hypotheses. The latter is clearly less preferred since it takes much longer time to acquire, up to 2 seconds (based on earlier analysis carried out during the standard development). That is too long during initial installation, when antenna pointing often requires an indication of proper reception of the signal. The differential detection scheme described above using information embedded in the PLSCODE is therefore preferred. In fact, as long as the difference, not necessarily a constant, is known a priori, receivers can take advantage of it. For this reason, we are able to further scramble the codeword of PLSCODE to improve the autocorrelation property. The specific sequence used for the scrambling is as follows: 0111000110011101100000111100100101010011010000100010110111111010. The scrambling sequence is essentially just an extended m-sequence. In annex C, the algorithms for rapid frame synchronization by utilizing SOF and PLSCODE are described. B.2 Pilot Structure The design goal of DVB-S2 carrier recovery scheme is to deliver channel outputs reliably to the LDPC decoder at very low SNR with small synchronization overhead. The design objectives include: • Negligible LDPC decoding performance loss due to carrier synchronization impairment (less than 0,1 dB to 0,3 dB for most modes); • Capability of working at extremely low SNR, as low as Es/N0 = -2,0 dB; • Capability of acquiring large carrier frequency offset (up to 5 MHz) with a 30 KHz/s ramp; • Robustness to LNB phase noise characteristics specified by DVB-S, which works at higher SNR and allows more implementation margin; • Rapid initial acquisition; • Simple implementation. The DVB-S2 phase noise specification is rather challenging due to the desire to reuse the millions of LNBs and antennas that have already been deployed for DVB-S reception. The major challenge for carrier recovery is to handle severe phase noise and large frequency offset at low SNR. Unlike DVB-S, that supports only QPSK modulation, DVB-S2 supports several modulation schemes, such as QPSK, 8PSK, 16APSK, and 32APSK. In order to expedite carrier recovery, the standard allows two operating modes for each modulation type: pilot-less and piloted, where pilot symbols are inserted to aid carrier synchronization. The system operators have the option to choose either operating mode. The receivers are informed of the pilot configuration from the PLSCODE residing in the PLHEADER. Thorough investigations show that even with the challenging phase noise impairment, only a handful of modes, namely 8PSK rate 2/3, 16APSK rate 2/3 and 3/4, and 32APSK rate 3/4, need pilot assistance for carrier recovery. Clearly, for most high rate codes, their operating SNRs are sufficiently high such that traditional decision-directed second order phase locked loop (PLL) should suffice. Therefore, when implementation complexity is concerned, a generic carrier recovery strategy that is suitable for all the modulation schemes with/without pilots is desired. Based on these observations, the standard adopted the following aggregated pilot structure. Each LDPC coded frame is preceded by a one-slot (90-symbol) PLHEADER containing the SOF and PLSCODE. Afterward, 36 pilot symbols follow every 16-slot coded data symbols. If the pilot symbols coincide with the PLHEADER of the following frame, they will not be inserted. In [14] it is shown that this pilot structure is a good balance between the synchronization overhead and performance of 8PSK rate 2/3 modulation. Figure B.3 shows an example of the pilot structure, for 8PSK modulation. ETSI 72 ETSI TR 102 376 V1.1.1 (2005-02) PLHEADER code seg 0 UW1 code seg 1 UW2 code seg 13 UW14 code seg 14 1 slot 16 slots 36 syms Figure B.3: Pilot structure for 8PSK modulation, where UWs (unique words) refer to pilot symbols ETSI 73 ETSI TR 102 376 V1.1.1 (2005-02) Annex C: Modem algorithms design and performance over typical satellite channels The DVB-S2 standard has been designed having in mind the peculiarities of the satellite channel, in particular the on-board satellite linear and non-linear distortions, the link fading impairments and the carrier phase noise dominated by the user terminal RF front-end. The adoption of high order modulation formats (up to 32-ary QAM) makes the potential channel impairments much more important than those encountered by the classical QPSK modulation format adopted by the former DVB-S standard [1]. In the following of the annex a design of the main modulation and demodulation units of a modem compliant with the new DVB-S2 standard is proposed, to minimize the end-to-end link losses for the reference satellite channel and receiver characteristics defined in [2]. The downlink satellite channel impairments (signal fading due to rain, scintillations, atmospheric gas absorption, etc.) have been modelled as a constant signal attenuation, taken into account within the signal Es/N0 at the demodulator input. The other impairments that are modelled within the downlink channel mostly pertain to the terminal receiver, like clock and carrier frequency errors as well as carrier phase noise. Clock frequency errors are due to long term instabilities of the terminal oscillator which provides the terminal demodulator sampling clock, while Doppler effects due to the GEO satellite movements are usually negligible. The precision of the oscillator depends on its quality (i.e. its cost) but usually it can be considered to be limited to a maximum of 10 p.p.m.. Carrier frequency errors can be attributed to several factors, but the main contributors are the terminal LNB oscillator instabilities and Doppler effects. Within DVB-S2 a maximum carrier frequency error of 5 MHz has been specified for consumer-type of terminal receivers. However, the residual differential frequency errors when tuning to a different downlink carrier can be considered much smaller, i.e. in the order of max 100 kHz. The main contributor to the carrier phase noise is the terminal LNB RF oscillator, especially in low cost equipments. The terminal tuner contribution can also be not negligible. The worst case PSD (Power Spectral Density) of the combined phase noise contribution of terminal satellite receivers' tuners and LNB's is specified in [2]. The architecture of a demodulator compliant with the DVB-S2 standard comprises of an RF/IF part whose architecture depends on the RF frequency band used by the application (Ku vs. Ka) as well as to whether low cost user equipments or professional equipments are used. Also, a number of choices can be made on the down-conversion strategy. For example, in low cost DTH terminals usually the RF received signal is first down-converted to a intermediate frequency by a LNB (Low Noise Block) that is placed in the proximity of the antenna feeder, then the signal is further down- shifted in frequency by a tuner, which usually, in low rate equipments, directly converts the signal to complex- baseband. A couple of well matched A/D converters sample the signal and feed the digital demodulator section of the receiver. The block diagram of the digital demodulator is shown in figure C.1. It is assumed that the signal from the RF/IF front end is down-converted to baseband and so that the two I-Q components are made available to the digital demodulator input, sampled at a rate high enough to avoid signal aliasing. FRAME frame sync SYNCH FR OM THE FRONT END TO THE MATCHED DECODER INTERPO FILTER + X X LAT OR DW N- BUFFER X X X SAMPL INTEGRATOR + e − jθ ( k ˆ ) s LOOK-UP TABLE LOOK-UP SYMBOL frame frame frame TABLE CLK DEMUX DEMUX sync DEMUX RECOVEY: sync sync GARDNER MODIFIED PILOT PILOT L&R PILOT SYMBOLS SYMBOLS SYMBOLS FREQUENCY ESTIMATOR 1 st ORDER FINE PA-LI LOOK- DAGC LOOP PED PHASE UP LOOP FILTER ESTIMATOR TABLE ν (k )T ˆ 2 nd C OARSE ORDER FED NCO LOOP DELAY&MUL FILTER frame sync (freeze on data symbols) Figure C.1: Block diagram of the DVB-S2 digital demodulator ETSI 74 ETSI TR 102 376 V1.1.1 (2005-02) A number of synchronization sub-systems are present in the demodulator in order to coherently demodulate the received signal. They will be described in details in the next clauses but here a general overview is given. The first correction the signal receives is about the carrier frequency. The coarse frequency correction unit provides a first correction with an all-digital second-order frequency loop. The target is for a lock-in range of up to 5 MHz, which corresponds to half the minimum hypothesized symbol rate of 10 Mbaud. Then, the next block deals with clock recovery whose timing adjustment is carried out through a digital interpolator. Typically this latter can be implemented as a parabolic or cubic interpolator (4 taps) in order to limit the distortions to the minimum. Matched filtering follows, at two samples per symbol. The number of taps of this filter depends on the roll-off factor. In the simulations an impulse response that spans 20 symbol periods was assumed with no particular optimization. As we will see, this leads to a performance degradation of less than 0,1 dB. Then, two distinct feed-forward units perform fine frequency and phase compensation. Digital automatic gain control (DAGC) is then performed in order to adjust the level of the incoming symbols to the reference constellation and further carrier phase adjustment is carried out by a DPLL. The de-rotated symbols are finally passed to the decoder where the pilot symbols are first discarded and the channel intrinsics are computed on the data symbols. All the synchronization units are fully pilot-aided, i.e. do not make any use of the data symbols, except for the clock recovery, the frame synchronizer and the last DPLL. This latter works on the data symbols only and its presence is required to lower the phase error RMS for high-order constellations (16 and 32APSK). The frame synchronizer uses a unique word correlator that works on symbol-by-symbol basis and feed with its final frame alignment the two demultiplexer of the demodulator. During initial acquisition the synchronization tasks are performed with the following order: 1) Clock recovery is activated first; in fact the Gardner algorithm can work with relatively large frequency errors. 2) Once the clock recovery has converged, frame synchronization can be carried out using techniques that are relatively insensitive to the maximum specified carrier frequency error. 3) Coarse and fine carrier frequency is then carried out. 4) Finally, phase recovery is performed. The requirement of 100 ms acquisition time for channel zapping can be met down to QPSK 1/2. For lower code rates, a slightly longer convergence time is required. In the following of the annex the key algorithms for a DVB-S2 modem are described, from signal pre-distortion techniques to combat the channel non-linear effects, to the clock and carrier synchronization algorithms and finally the digital AGC algorithm. The following notation is adopted: z(k) are the samples at the signal matched filter output (SMF), zP(k)'s are the z(k) samples which refer to pilot symbol location within the DVB-S2 physical layer frame, c(k)'s are the complex information data symbols belonging to a QPSK, 8PSK, 16APSK or 32APSK constellation, cP(k)'s are the pilot symbols and L0 is the pilot symbols periodicity in number of symbols. DVB-S2 allows for inserting pilot fields of Lp = 36 symbols every 16 slots, i.e. L0 = 1 476 symbols in both the broadcast and unicast profile. C.1 Modulator with Pre-Distortion The selected Amplitude Phase Shift Keying (APSK) modulation for DVB-S2 is characterized by the fact that constellation points lie on concentric circles. This circular symmetry is particularly suited to non-linear amplification as it is easier to maintain the APSK circular constellation shape over a non-linear HPA than for classical squared Quadrature Amplitude Modulation (QAM) constellations. In [27] it has been shown that by proper APSK constellation parameters optimization it is possible to achieve performance that is very close or even slightly superior to QAM over AWGN channels. PSK modulation formats (QPSK and 8PSK) result to be a particular case of APSK corresponding to the case when the points are all lying on a single circumference. For 16APSK two rings of points are present with 4 points on the inner ring and 12 points on the outer ring. In the case of 32APSK three rings are present with 4 points on the inner ring, 12 points on the middle ring and 16 points on the outer ring. Over non-linear channels the APSK constellation, when observed at the demodulator symbol matched filter decimated output, is affected by two main kinds of impairment: 1) The constellation centroids (see note 1) warping due to the AM/AM and AM/PM HPA non-linear characteristic. ETSI 75 ETSI TR 102 376 V1.1.1 (2005-02) NOTE 1: By centroid we consider the compilation of received constellation cluster centre of mass conditioned to each constellation point. 2) The clustering effect due to the inter-symbol interference (ISI) experienced at the demodulator matched filter output. The warping phenomenon is responsible for the reduction of the distance among APSK rings (AM/AM compression) as well as a differential phase rotation among them (AM/PM differential phase). The ISI causing the constellation clustering is related to the demodulator SRRC filter mismatch on the received signal due to the combination of the signal band limiting introducing memory in the channel, the IMUX filter linear distortion, the HPA memoryless non-linearity, the OMUX linear filter distortions, resulting in a non-linear channel with memory. Clearly the warping effect has no impact on single ring constellations such as QPSK and 8PSK (see note 2) but is causing important degradations for 16APSK and 32APSK. This is because the decoder log-likelihood ratios are typically computed using the "standard" APSK constellation and do not take into account the centroid locations distortion caused by the HPA. In [28] two kinds of pre-distortion techniques have been considered for APSK: "static" and "dynamic". The static pre-distortion technique simply consists in modifying the APSK constellation points to minimize the demodulator SMF samples centroids distance from the "wanted" reference constellation. The static pre- compensation is preferably performed off-line in the absence of AWGN once we know the satellite channel characteristic according to the following steps [13]: 1) Generation of a S blocks of W symbols over which the SMF centroids are computed; 2) Computation of the error signal at the end of each block; 3) Pre-distorted constellation point update. NOTE 2: The phase and amplitude PSK distortion will be compensated for by the demodulator AGC and phase recovery sub-systems. The static pre-distortion is able to correct for the constellation warping effects but it is not able to compensate for the clustering phenomenon. The dynamic pre-distortion algorithm takes into account the memory of the channel conditioning the pre-distorted modulator constellation not only to the current symbol transmitted but also to the (L-1)/2 preceding and (L-1)/2 following symbols (L symbols in total). This calls for an increased look-up table size of ML points. By exploiting the APSK constellation symmetry the amount of memory size can be reduced to 3ML/16 for M = 16 and 32. The clustering effect reduction is obtained at the expenses of an increased outer constellation points amplitude, with two main drawbacks: a) the augmentation of the HPA Output Back-Off (OBO) which negatively affects the overall system efficiency; b) the possible impact on the HPA (TWTA) safe operation due to the higher peak-to-average ratio making the instantaneous signal power occasionally well beyond the saturation point. Based on the previous considerations an improved dynamic pre-distortion approach has been devised, as illustrated in [13]. The quantity to minimize is in fact not the r.m.s. of the centroids conditioned to a certain data pattern but rather the total link degradation DTOT given by: NL AWGN Es E DTOT ( s )[dB] = ( s )[dB] − s ( s )[dB] + OBO ( s )[dB]         (1) N0   req  N0  req NL AWGN  Es   Es  being   and   the average symbol energy over noise density required to achieve the target Frame  N0  req  N0  req Error Rate (FER) in the non-linear and linear channel respectively and OBO the HPA output back-off. This approach ensures the best trade-off between ISI minimization and the OBO penalty due to the increased peak-to-average ratio caused by dynamic pre-compensation. ETSI 76 ETSI TR 102 376 V1.1.1 (2005-02) C.2 Clock Recovery Symbol clock recovery can be performed first using the well-known Gardner's algorithm [31]. This algorithm is non- data aided and thus can be run without any frame synchronization in place. The performance of this algorithm is quite insensitive to the modulation format at least over the range of Es/N0 of interest as well as to a carrier frequency error up to 0,1-0,2 times the symbol rate. That means that for 25 Mbaud symbol rates, the timing recovery can work with a carrier frequency error up to 5 MHz. For lower symbol rates (down to 10 Mbaud) a possible technique is to use two timing recovery algorithms working in parallel. The first with a signal pre-rotated in frequency by +1/4 of the symbol rate (Rs) ad the second with -1/4 of Rs. Then, after pre-determined time left for the transient, the algorithm that has converged can be selected. The jitter RMS of the algorithm depends on the signal roll-off factor. In particular, it worsens as the roll-off decreases. For the DVB-S2 range of roll-offs and Es/N0, a normalized (to the symbol rate) loop bandwidth of 5 10-5 seems to be required for negligible impact to the receiver performance. With this loop bandwidth · and using a second order loop, a clock frequency error of 10 p.p.m. can be tracked with no residual error bias. Also, the overall acquisition transient of the timing recovery unit would be around 105 symbols, which corresponds to 5 ms at Rs = 25 Mbaud. C.3 Physical Layer Frame Synchronization After the receiver symbol timing has reached steady state the next step to be performed is frame synchronization by searching for the physical layer header. This can be performed even in the presence of a strong carrier frequency error through differential detection. Reference [14] shows that the mean acquisition time is around 20 ms for 25 Mbaud transmissions at -2 dB and less than 4 ms at 0,7 dB. At 99 % probability the acquisition time for frame synchronization is around 130 ms at -2 dB and 10 ms at 0,7 dB (see note). In case of the unicast profile exploiting ACM, until the carrier frequency and phase synchronization estimators are not in steady-state condition the physical layer frame configuration field (MODCOD) [10] cannot be decoded. Hence pilot locations for successive frames shall be determined based on PLHEADER synchronization verification at the four possible locations (one per modulation format). Once carrier phase synchronization has been established, then the position of the pilot symbols is also known (by decoding the MODCOD fields of the PLHEADERs of each physical layer frame [XFECFRAME]). NOTE: Those acquisition times can be largely improved by using more complex frame detection schemes. C.3.1 An algorithm for Frame Synchronization Figure C.2 illustrates a scheme to correlate on both the SOF and PLSCODE differentially. The shift registers in the circuit can be partitioned into two sections. The first is associated with SOF, the second with PLSCODE. There are in total 57 taps associated with the 89 registers. In the first part, 25 of them are associated with the pair-wise difference of SOF. In the second part, 32 nonzero taps are associated with PLSCODE since only 32 out of the 64 differentials are known. The taps associated with the shift register for computing the correlation can be obtained as follows. First set all the registers to zero, then shift the modulated SOF and a modulated and scrambled codeword of PLSCODE into the circuit. Once the rightmost register becomes nonzero, the tap associated with a register is just the complex conjugate of the content of the corresponding register. Given that the modulated SOF and PLSCODE take only ±1, ±i, the taps only take these four possible values as well. Clearly these are trivial multiplications from the implementation point of view. ETSI 77 ETSI TR 102 376 V1.1.1 (2005-02) From Channel D D D D D D D D 25 coef. SOF 32 coef. PLSC conj summer summer max of the absolute value To peak detector Figure C.2: Differential detection of the SOF and PLSCODE When used for frame synchronization, the incoming signal arriving at the correlator is sampled at one sample per symbol. It is first differentially decoded and the resulting samples are then sequentially shifted into a shift register of length 89. The contents of the shift register are multiplied with the taps. The first 25 and the last 32 values at the output of the multipliers are separately summed together in two different branches. The outputs of the two summers are respectively added and subtracted to produce two values. The maximum of the absolute value out of the two branches is the final output of this correlation circuit. The output is then further processed by a peak search algorithm. The conventional approach is to compare the output of the correlator with a predetermined threshold. If the value is larger than the threshold, it is declared that preliminary frame synchronization has been achieved. Post verification may then follow. In [14], an algorithm is presented for rapid search and verification. C.3.2 An Alternative Frame Synchronization Algorithm When the initial frequency uncertainty is small, the acquisition of frame synchronization may use a simple finite state machine using correlators which exhibit a good probability of acquisition after a few number of physical frames. The solution described hereafter is proposed to show that an acceptable performance may be achieved for all noise levels used by the standard. This solution also has the advantage of being compatible with CCM, VCM or ACM modes, with or without pilot insertion. Tunable parameters are also introduced in this process, in order to allow fine optimization of the acquisition, assuming specific operating conditions. C.3.2.1 Acquisition procedure description The acquisition of the carrier frame structure may only rely on the detection of the SOF and PLSCODE sequences in the incoming frame. The symbol rate acquisition is assumed to be already locked, the whole process may be achieved with 1 sample/symbol. The carrier and fine frequency acquisition may not be used for frame acquisition as they often depend on the modulation which is not known before determining the PLSCODE value. Symbols are assumed to have an arbitrary phase and a small residual frequency deviation that will impact the correlator performance. The proposed procedure may be summarized as follows: 1) Unlocked state, the demodulator looks for the 26 symbols of the SOF sequence on the fly on the incoming carrier plus noise samples. ETSI 78 ETSI TR 102 376 V1.1.1 (2005-02) 2) Tentative state, the correlation of the samples with the SOF sequence exceeds a specified threshold TST (Threshold on SOF in Tentative state), the demodulator tries to decide which PLSCODE follows on the next 64 symbols. If the maximum correlation result over the 128 PLSCODE values exceeds a second threshold TPT (Threshold on PLSCODE in Tentative state), the demodulator goes in locked state, deciding when the next SOF should occur on the value of the PLSCODE. Otherwise, the demodulator remains unlocked. 3) Locked state, the demodulator tests that a SOF sequence is present at the expected position by comparing the correlation to a TSL (Threshold on SOF in Locked state) value. If the threshold is not exceeded, the demodulator goes back to unlocked state. 4) Locked state, the demodulator determines the maximum correlation result with the 128 PLSCODE pattern values and compares it to a threshold TPL (Threshold of PLSCODE in Locked mode). If the value exceeds the threshold, the demodulator remains in locked state and computes where the next SOF should occur, otherwise is goes unlocked. It is described with the state machine presented in figure C.3. Using this model, the behaviour of the acquisition depends on the combination of thresholds [TST, TPT, TSL, TPL] which determine the transition from one state to the other. TPL Locked Unlocked /Tentative Start PLSCODE (symbol clock SOF decision locked) detection PLSCODE SOF decision Tentative transition detection Fl [tk − 2 D ] →Slot efficiency loss event ˆ if Fl [tk ] < Fl [tk − 2 D ] →Slot payload loss event ˆ ETSI 98 ETSI TR 102 376 V1.1.1 (2005-02) where the Slot efficiency loss event corresponds to the event that the available link capacity has not been exploited in the current time slot. The Slot payload loss event is associated to the event that the received slot payload will be not correctly decoded (see note 2). It is apparent that, while the first event is causing a capacity reduction, the second one will affect the ST Frame Error Rate (FER) and shall be accurately controlled. In figure E.3-a and -b, a graphical explanation of the phenomena mentioned above can be found. NOTE 2: For any physical layer configuration Fl [tk ] the payload is assumed to be lost if [Ec/Nt][k] < [Ec/Nt]req( Fl [tk ] ). Figure E.3: a) Slot efficiency loss event - b) Slot payload loss event - c) Shifted threshold technique- d) Hysteresis technique The ideal physical layer configuration dependency on [Ec/Nt] is indicated by a step function. The cross represents the current true SNIR [Ec/Nt][k] and the estimator Gaussian-like [Ec/Nt] pdf is super imposed. The dot represents the estimated [Ê c/Nt][k - 2D]. In case a), as the estimated value is lower than the true one, a slot efficiency loss event will occur. On the contrary, in case b) the estimation is higher than the true value, thus a more efficient physical layer is associated to the current channel condition. In this case a slot payload loss event will occur. In the following, two possible techniques to control the two above unwanted effects are proposed. It has to be underlined that any technique aiming at the slot payload loss event reduction will increase the slot efficiency loss event. E.2.1 Shifted Threshold The first countermeasure to ST demodulator channel estimation errors is to shift to the right the physical layer { } configuration threshold to guarantee that Pr Fl[tk - 2D] > Fl[tk] < R1 . This conditioned probability can be derived as ˆ follows: ˆ Ec [k − 2D ] > S sh Ec [k ] ≤ S0 ≤ R1        Pr        Nt   Nt    where Ssh and S0 represent the new shifted threshold and the original threshold value, respectively and R1 represents the required slot error rate. Recalling the assumed short-term channel invariance and Gaussian Ec/Nt estimator error from the last expression the new value of the Shifted Threshold can be derived as follow: [k − 2D ] < S sh Ec [k ] = S0 = 1 − R1 = 1 − Q S sh − S0   ˆ Ec        Pr   σ s ( S0 )        Nt   Nt         ETSI 99 ETSI TR 102 376 V1.1.1 (2005-02) from which: Ssh = S0 + σs(S0) Q-1 (R1) The graphical representation of this approach is reported in figure E.3-c. The computation has to be repeated for each possible physical layer transition as σs is dependent on the Ec/Nt corresponding to the threshold level. E.2.2 Hysteresis The shifted threshold technique allows to control the probability of slot payload loss event. However, let assume as working point a value close to one of the shifted thresholds computed according to the previous procedure. Due to the channel estimator random errors the SNIR value will be jittering around its true value that is close to Ssh. Thus the ST physical layer will often ping-pong between the two adjacent physical layer configurations causing a large amount of unwanted reverse link signalling. Nevertheless, the shifted threshold approach will keep the probability of slot payload loss below the design value R1. The problem described above can be mitigated by introducing a second threshold and operating the switch through an hysteresis cycle according to the graphical representation of figure E.3-d. The up and down thresholds indicated with Sup and Sdown are derived as follows: The down-threshold is related to the constraint to limit the slot payload loss event: ˆ Ec [k − 2 D] > Sdown Ec [k ] ≤ S0 ≤ R1        Pr        Nt  Nt     The up-threshold is related to the purpose of reducing the amount of reverse link signalling: ˆ Ec [k − 2 D ] ≤ Sdown Ec [k ] ≥ Sup ≤ R2        Pr        Nt   Nt     Adopting the same logic of the shifted threshold computation, the hysteresis thresholds can be derived: Sdown = S0 + σs Q-1 (R1) Sup = Sdown + σs Q-1 (R2) For all the cases where Sdown < Ec/Nt < Sup the decision on the physical layer configuration to use will be based on the slope of two consecutive Ec/Nt estimates. If the slope sign is positive the Sup threshold will be applied. Vice versa in case the slope is negative, the Sdown threshold applies. Two considerations are in order. First, the hysteresis reduces further the packet loss event. Second, a severe constraint on the reverse link signalling reduction further increases the efficiency loss event. E.3 Performance results In [16] the proposed channel estimation and physical layer adaptation for the downlink of a broadband multibeam satellite system has been deeply investigated. An overall system model able to accurately represent the ACM system physical layer adaptation has been devised. The simulator is capable of reproducing the useful and other beams interference signals, channel fading impairments, channel estimation and physical layer selection. Particular attention has been devoted to the modelling of the other beam interference and the fading channel. The latter has been implemented by means of a time domain simulator representative of experimental results for both fading amplitude and spectral characteristics. The DA-SNORE channel estimator has been proposed for tracking typical Ka-band channel variations with enough accuracy to drive the physical layer selector. For this unit the simple threshold design approach has been complemented by a more practical solution exploiting shifted threshold and hysteresis that allows to control the outage probability due to the estimator errors and to limit the reverse link signalling. ETSI 100 ETSI TR 102 376 V1.1.1 (2005-02) The performance of the proposed solution has been verified resorting to a realistic Ka-band multi-beam study case. Simulation results for optimized algorithm parameters indicate that for both heavy and light fading conditions we can achieve effective physical layer adaptation with acceptable losses compared to the ideal system. The technique proposed although investigated for systems at Ka-band can be easily adapted to other frequency bands such as Ku and Q/V-band. ETSI 101 ETSI TR 102 376 V1.1.1 (2005-02) Annex F: ACM receiver implementation In the following two receiver schemes are proposed to regenerate the Transport Stream clock R'TS under dynamic rate variations. F.1 Type 1 receiver With reference to figure F.1, R'TS is regenerated via a PLL which maintains constant, in the steady state, the receiver FIFO buffer filling condition (the Empty/Full E/F signal is in steady state when the buffer is halfway filled up). The DVB-S2 receiver, after receiving a useful packet (UP) and storing it in the FIFO buffer, reads the DNP field, and inserts DNP null-packets in front of it. The NCO (Numerically Controlled Oscillator) is driven by the E/F signal via a low pass filter (LPF), and generates a steady state frequency R'TS locked to RTS (clock at the transmitting side), independently of the source bit-rate. During the source rate transient the situation is different. As shown in figure F.1, a step in DTX+DRX1 (delay increase produced by a bit rate reduction) is immediately absorbed by the FIFO buffer (since the NCO frequency variation is very slow), which starts to empty up. When the PLL reacts to the rate reduction, R'TS slightly slows down till the buffer recovers the halfway filling condition, and R'TS may return to the steady-state RTS condition. After the transient, the end-to-end delay of the TS is increased by ∆D (in fact, in steady state, the reception buffer delay is constant by definition). Therefore Type 1 receiver does not guarantee the "constant delay" and "constant bit rate" conditions for Transport Streams, although, by increasing the PLL time constants, rate variations may be smoothed. This analysis has driven the design of a DVB-S2 subsystem called "input stream synchronizer", allowing the implementation of a "type 2" receiver not affected by delay and bit-rate variations. Read E/F NCO DNP LPF DE Null-packet Write TS R’TS MOD Re-insertion packets FIFO BUFFER Useful packets Read TS packets Delay=DTX+DRX1 Delay=DRX2 DTX(t)+DRX1 ∆D ∝ PLL time R’TS (t) constant DTOT(t)=DTX(t)+DRX1+DRX2 ∆D t Figure F.1: Example Type 1 Receiver: delay and bit-rate variations at the receiving buffer input and output ETSI 102 ETSI TR 102 376 V1.1.1 (2005-02) F.2 Type 2 receiver With reference to figures F.2 and F.3, R'TS is directly locked to the transmission RTS, by introducing a "time-stamp" mechanism using the symbol-rate RS as a common frequency reference (RS is recovered by the receiver and is not affected by the packet delay variations). In this way the reception FIFO buffer filling condition can change and automatically compensate the chain delay variations, provided it is sufficiently large and suitably initialized to avoid overflows/underflows. Figure F.2 shows the principle scheme of the input stream synchronizer (transmitting and receiving side): in the modulator, a counter runs at speed RS, and its content (ISCR field, Input Stream Clock Reference) is appended to each TS packet as soon as it crosses the input interface, before null-packet deletion. In the receiver, a similar counter runs at RS speed (RS is generated by the clock recovery subsystem), and its content is compared to that of the ISCR field of the packets when they are read out of the FIFO buffer to feed the demultiplexer. The R'TS clock is generated by a PLL, which is driven by the phase error between the local counter and the transmitted ISCR. Figure F.3 gives some receiver details: since the reception FIFO buffer is not self-balancing as in the Type 1 receiver, the initial receiver FIFO buffer state may be reset by the transmitter via the BUFSTAT field. Furthermore the maximum receiver buffer size assumed by the transmitter may be signalled to the receiver (BUFS field), in order to minimize end-to-end delay when required. For a Type 2 receiver the "constant delay" and "constant bit rate" conditions for Transport Streams are met, at the cost of a slightly increased complexity and a capacity loss of about 1 to 1,5 % for the transmission of ISCRs (2 or 3 bytes are appended to each TS packet). Recovered Local Time stamp Time stamp ISCR RS RS Inserted in TS Counter Counter Compare NCO packets RTS Received R’TS ISCR FEC coding TS TS Input NP ACM ACM NP DE MUX Stream Deletion modul. Dem Re-insertion Synch & Merger MUX Framing & Buffer & FIFOBuffer signalling SNIR measure Mode Adapter DVB-.S2 Modulator DVB-S2 Demodulator Satellite channel Figure F.2: ACM DVB-S2 model using time stamps to lock RTS and R'TS ETSI 103 ETSI TR 102 376 V1.1.1 (2005-02) DEM Rs Mod 222 Load first Clock Counter ISCR recovery 15 or 22 MSB NCO Compare Phase LPF Difference 15 or 22 bit Read Read ISCR DNP Null-packet Write TS R’TS Re-insertion packets FIFO +ISCR BUFFER Useful packets Read TS packets Initialise using buffer size BUFS, and buffer state BUFSTAT Figure F.3: Example Type 2 Receiver ETSI 104 ETSI TR 102 376 V1.1.1 (2005-02) History Document history V1.1.1 February 2005 Publication ETSI