dvb s2 standard with normal and short frames
play

DVB-S2 standard with normal and short frames Guy Lesthivent Alban - PDF document

DVB-S2 standard with normal and short frames Guy Lesthivent Alban Duverdier 1 Scope This document is aimed at summarizing the main features and performance of the channel coding scheme selected by the DVBS2 working group regarding short


  1. DVB-S2 standard with normal and short frames Guy Lesthiévent – Alban Duverdier 1 Scope This document is aimed at summarizing the main features and performance of the channel coding scheme selected by the DVBS2 working group regarding short frames. It is based on the draft version of the future standard registered as DVBS2-74r9 (DRAFT Digital Video Broadcasting: Second generation framing structure, channel coding and modulation systems for broadcasting, interactive services, news gathering and other broadband satellite applications). This work was performed with the help of the consulting office NOVACOM (Paris 15 e ) in the context of a CNES R&D study. 2 System overview The system overview is presented by Figure 1. K BCH user frame EXOR ⊕ scrambling sequence r-user frame Systematic BCH N encoder BCH r-user frame BCH FEC Systematic LDPC = N 64800 or 16200 bits LDPC encoder η = r-user frame BCH FEC LDPC FEC 1 BPSK S=720 η = Mapping 2 QPSK S=360 η η = N / 8PSK 3 S=240 LDPC η = S slots of 90 symbols 4 16APSK S=180 η = 5 32APSK S=144 channel frame Header π ( 2 BPSKmodulated) 36 symbols 1 slot ( π / 2BPSK) 16 slots (selected modulation) (unmodulated) &Pilot insertion PLHEADER Pilot block Slot 1 Slot... Slot 16 Slot S (if requiered) IQ Randomization 90(S+1)+36 int(S/16) symbols PLHEADER Randomized sequence (2 PN generators on I and Q) Figure 1: System overview � CNES 20/10/2003 Page 1 sur 5

  2. 3 Code features The redundancy introduced by the BCH code followed by the LDPC code is presented by Figure 2. = K N LDPC BCH − − N K N K K LDPC LDPC BCH BCH BCH redudancy redudancy Scrambled user frame BCH LDPC N = 64800 or 16200 LDPC Figure 2: DVB-S2 information block after channel coding t The code parameters are given by Figure 3, where is the bit error correction capability of corr the BCH code. Short frame Normal frame LDPC K N = K t K N = K t BCH BCH LDPC corr BCH BCH LDPC corr Rate ½ 7032 7200 12 32208 32400 12 3/5 9552 9720 12 38688 38880 12 2/3 10632 10800 12 43040 43200 10 ¾ 11712 11880 12 48408 48600 12 4/5 12432 12600 12 51648 51840 12 5/6 13152 13320 12 53840 54000 10 8/9 14232 14400 12 57472 57600 8 Figure 3: DVB-S2 code parameters The trellis of the selected LDPC code is described by Figure 4. R = N − K LDPC LDPC p 0 p 1 p 2 p R-1 p i : LDPC parity bit nodes d°=2 d°=1 c 0 c 1 c 2 c R-1 c i : Check nodes … … … d°= ρ ρ ρ -1 ρ d°= ρ ρ ρ ρ u 0 u 1 u K LDPC -1 u i : User / BCH parity bit nodes … … … … d°=D>3 d°=3 d°=3 Figure 4: DVB-S2 LDPC trellis The trellis is check regular: all the check nodes but one have the same degree ρ that depends on the LDPC code rate. On the contrary, the trellis is not bit node regular: all LDPC parity bits but one have degree 2, BCH parity bits and a fraction of the user bits have degree 3, the user bits of the remaining fraction have a degree D greater than 3. The fraction of user bits of degree D and this degree depend on the code rate. The trellis is easily described through congruence relationships. As a consequence, the encoding process is very simple. Besides, on the decoding side, the Belief Propagation Algorithm can be efficiently implemented (high parallelism and reduced connectivity). � CNES 20/10/2003 Page 2 sur 5

  3. 4 Modulation In addition to usual BPSK , QPSK and 8PSK , two modulation schemes have been added: 16APSK and 32APSK. R 1 , R The 16APSK constellation is given Figure 5 with real numbers such that 2 + = γ = R R R 2 / R 2 2 4 12 1 and is a function of the code rate. 1 2 1 Q 1010 1000 0010 MSB 0000 R 2 LSB 0110 1100 0100 1110 R 1 I text 0111 γ 1111 1010 Code rate 1101 2/3 3.1 3/4 2.85 0011 0001 5/6 2.6 1011 8/9 2.58 1001 Figure 5: DVB-S2 16APSK constellation R 1 , R R real numbers and The 32APSK constellation is given Figure 6 with , 2 3 + + = γ = γ = R R R R / R R / R 2 2 2 4 12 16 1 where and are function of the code rate. 1 2 3 1 2 1 2 3 1 Q 01101 11101 01001 MSB 01100 11001 LSB 00101 00001 R 3 11100 01000 00000 00100 R 2 10000 10001 10100 10101 11110 11000 R 1 text I 10110 10111 10010 10011 γ 1 γ 2 01110 Code rate 11010 00010 00110 3/4 2.84 5.26 4/5 2.7 4.8 00111 11111 00011 01010 5/6 2.6 4.5 2.5 =R 2 / R 1 γ γ γ γ 1 1 8/9 4.2 1 1 =R 3 / R 1 01111 11011 γ 2 γ γ γ 2 2 2 01011 Figure 6: DVB-S2 32APSK constellation � CNES 20/10/2003 Page 3 sur 5

  4. 5 Header and pilot insertion A header of 90 symbols long has been devised to ensure fast and secure frame detection and synchronization. The header comprises a unique word (SOF) followed by 64 symbols which carry 7 control bits, Reed-Müller encoded, as described. HEADER Service information SOF Figure 7: DVB-S2 header They inform the receiver about: - the modulation type, - the LDPC code rate, - the frame type (normal or short), - the use of pilot symbols. For modes requiring pilots, pilot blocks of 36 symbols are inserted every 16 slots to help receiver synchronisation. 20/10/2003 � CNES Page 4 sur 5

  5. 6 Performances The performances of the code have been validated by DVB-S2 for normal frames in an AWGN channel (ideal carrier synchronization and no phase noise). Table 1 presents them. Ideal E s /No performance at Quasi Error Free PER=10 -7 (AWGN channel) Mode Spectral E s /No (dB) E b /No (dB) BPSK 1/2 0,495114 -2.00 1,05 BPSK 3/5 0,594976 -0.77 1,49 BPSK 2/3 0,662042 0.10 1,89 QPSK 1/2 0,988857 1.00 1,05 QPSK 3/5 1,188303 2.23 1,48 QPSK 2/3 1,322251 3.10 1,89 QPSK 3/4 1,487472 4.03 2,31 QPSK 4/5 1,587195 4.68 2,67 QPSK 5/6 1,654662 5.18 2,99 QPSK 8/9 1,766451 6.20 3,73 QPSK 9/10 1,788612 6.42 3,89 8PSK 3/5 1,779989 5.50 3,00 8PSK 2/3 1,980633 6.62 3,65 8PSK 3/4 2,228122 7.91 4,43 8PSK 5/6 2,478560 9.35 5,41 8PSK 8/9 2,646012 10.69 6,46 8PSK 9/10 2,679207 10.98 6,70 16APSK 2/3 2,637197 8.97 4,76 16APSK 3/4 2,966726 10.21 5,49 16APSK 4/5 3,165621 11.03 6,03 16APSK 5/6 3,300181 11.61 6,42 16APSK 8/9 3,523142 12.89 7,42 16APSK 9/10 3,567341 10.13 4,61 32APSK 3/4 3,703293 12.73 7,04 32APSK 4/5 3,951568 13.64 7,67 32APSK 5/6 4,119537 14.28 8,13 32APSK 8/9 4,397854 15.69 9,26 Table 1: DVB-S2 performances for normal frames (DVB-S2 validation) For short frames, CNES validation is given for the 8PSK 2/3 (belief propagation algorithm for decoding, 6 bits for I/Q and branch metric quantification, maximum of 48 iterations). 8PSK 2/3 3,4 3,5 3,6 3,7 3,8 3,9 1,E-02 1,E-03 PER - MPEG2 Normal frames 1,E-04 Short frames 1,E-05 1,E-06 Eb/N0 Figure 8: DVB-S2 performances for short frames (CNES validation) The estimated SNR degradation between normal frames (64800 bits) and short frames (16200 bits) is less than 0.3 dB at PER of 10 -6 . 20/10/2003 � CNES Page 5 sur 5

Recommend


More recommend