WO2011104144A1

WO2011104144A1 - Encoder and encoding method providing incremental redundancy

Info

Publication number: WO2011104144A1
Application number: PCT/EP2011/052231
Authority: WO
Inventors: Nabil Muhammad; Lothar Stadelmeier; Joerg Robert; Malek Kechaou
Original assignee: Sony Corporation; Sony Deutschland Gmbh
Priority date: 2010-02-26
Filing date: 2011-02-15
Publication date: 2011-09-01
Also published as: TW201143301A

Abstract

The present invention relates to an encoder for error correction code encoding input data words (D) into codewords (Z1, Z2), comprising: an encoder input (1451) for receiving input data words (D) each comprising a first number K_ldpc of information symbols, an encoding means (1452) for encoding an input data word (D) into a codeword (Z1, Z2, Z3, Z4) such that a codeword comprises a basic codeword portion (B) including a data portion (D) and a basic parity portion (Pb) of a second number N_ldpc - K_ldpc of basic parity symbols, and an auxiliary codeword portion (A) including an auxiliary parity portion (Pa) of a third number M_IR of auxiliary parity symbols, wherein said encoding means (14) is adapted i) for generating said basic codeword portion (B) from an input data word (D) according to a first code, wherein a basic parity symbol is generated by accumulating an information symbol at a parity symbol address determined according to a first address generation rule, and ii) for generating said auxiliary codeword portion (A) from an input data word (D) according to a second code, wherein an auxiliary parity symbol is generated by accumulating an information symbol m at a parity symbol address γ, wherein said parity symbol addresses γ are determined according to a second address generation rule N_ldpc - K_ldpc + {x + m mod G_a x Q_IR } mod M_IR if x > N_ldpc - K_ldpc, wherein x denotes the addresses of a parity symbol accumulator corresponding to the first information symbol of a group of size G_a and Q_IR is an auxiliary code rate dependent, predefined constant, and an encoder output (1454) for outputting said codewords (Z1, Z2).

Description

Encoder and encoding method providing incremental redundancy

FIELD OF INVENTION

[0001] The present invention relates to an encoder and a corresponding encoding method for error correction code encoding input data words into codewords. Further, the present invention relates to a transmitter and a corresponding transmission method for broadcasting data in a broadcasting system. Still further, the present invention relates to a computer program for implementing said transmission method on a computer. Finally, the present invention relates to a broadcasting system com- prising such a transmitter and one or more receivers for receiving data broadcast by said transmitter.

The present invention relates particularly to an LDPC encoder as, for instance, used in a transmitter of a Digital Video Broadcasting (DVB) system utilizing Orthogonal Frequency Division Multiplexing (OFDM). Further, the present invention can be applied in other systems, which apply the same or similar LDPC encoding schemes from DVB.

BACKGROUND OF THE INVENTION

[0002] The transmission parameters of known broadcasting systems, such as the broadcasting systems in accordance with the DVB-T2 standard (second generation digital terrestrial television broadcasting systems standard), are generally optimized for fixed reception with stationary receivers, e.g. with roof-top antennas. In future broadcasting systems, such as the upcoming DVB-NGH (DVB Next Generation Handheld; in the following also referred to as NGH) standard, a mobile receiver (which is the main focus of this upcoming standard) shall be enabled to receive data correctly also in bad reception situations, e.g. despite suffering from multipath propagation, fading effects and Doppler shifts. Such broadcasting systems are particularly characterized by the fact that there is generally no feedback channel and no signalling from receivers to transmitters.

SUMMARY OF INVENTION

[0003] It is an object of the present invention to provide an encoder and a corresponding encoding method as well as a transmitter and a corresponding transmission method for broadcasting data in a broadcasting system by which the probability of error-free reception/reconstruction of data by a mobile receiver is increased compared to transmitters and transmission methods in known broadcasting systems, even under bad reception conditions. It is a further object of the present invention to enable the use of different types of decoders and receivers, but to ensure compatibility. It is a still further object of the present invention to provide a computer program for implementing said encoding method and a broadcasting system.

[0004] According to an aspect of the present invention there is provided an encoder for error correction code encoding input data words into codewords, comprising:

an encoder input for receiving input data words each comprising a first number K_Upc of information symbols,

an encoding means for encoding an input data word into a codeword such that a codeword comprises a basic codeword portion including a data portion and a basic parity portion of a second number N_ldpc - K_ldpc of basic parity symbols, and an auxiliary codeword portion including an auxiliary parity portion of a third number M_IR of auxiliary parity symbols,

wherein said encoding means is adapted

i) for generating said basic codeword portion from an input data word according to a first code, wherein a basic parity symbol is generated by accumulating an information symbol at a parity symbol address determined according to a first address generation rule,

and

ii) for generating said auxiliary codeword portion from an input data word according to a second code, wherein an auxiliary parity symbol is generated by accumulating an information symbol m at a parity symbol address γ, wherein said parity symbol addresses y are determined according to a second address generation rule

N_ldpc - K_ldpc + {x + m mod G_a x Q_1R } mod M_1R if x > N_ldpc - K_ldpc ,

wherein x denotes the addresses of a parity symbol accumulator corresponding to the first information symbol of a group of size G_a and Qj_R is an auxiliary code rate dependent, predefined constant, and

an encoder output for outputting said codewords. [0005] According to a further aspect of the present invention a transmitter for broadcasting data in a broadcasting system is provided, comprising:

a data input for receiving at least one transmitter input data stream segmented into input data words,

an encoder as defined above for error correction code encoding the input data words into codewords,

a data mapper for mapping the codewords onto frames of a transmitter output data stream, and

a transmitter unit for transmitting said transmitter output data stream.

[0006] According to further aspects of the present invention there are provided a corresponding encoding method, transmission method and a computer program comprising program means for causing a computer to carry out the steps of said encoding method, when said computer program is carried out on a computer.

[0007] Preferred embodiments of the invention are defined in the dependent claims. It shall be understood that the claimed apparatus, the claimed methods and the claimed computer program have similar and/ or identical preferred embodiments as the claimed encoder and as defined in the dependent claims.

[0008] The present invention is based on the idea to enable a mobile receiver, for instance an in-car receiver or a handheld receiver (e.g. in a mobile phone or a PDA) in a broadcasting system to decode broadcast data even under severe transmission channel conditions by improving the provided error correction measures. In particular, it is proposed to provide a sufficient amount of redundancy by the encoder to increase the robustness of the code. Said additional redundancy is provided by the transmitter such that a receiver can, but does not mandatorily, use it if reception or reconstruction (decoding) of received broadcast data is erroneous or could only be made with insufficient quality. The broadcasting operator also has the ability to choose from a variety of different coding and modulation schemes, thereby trading off throughput versus robustness. [0009] To ensure that a receiver (e.g. an existing legacy receiver), in particular its decoder, can correctly decode received data without any additional redundancy provided according to the present invention, a first code is applied by the error correction code encoder (generally applying forward error correction) to generate a basic codeword portion for the data words of the data to be broadcast. This way of encoding can be the known standard way of encoding input data words into codewords, for instance the forward error correction (FEC) encoding as applied in DVB- T2, DVB-S2, DVB-C2, or the upcoming DVB-NGH transmitters (e.g. LDPC encoding), i.e. the basic codeword portion may correspond to a ("normal" error correction code) codeword according to the DVB-T2 standard (in the following also referred to as T2). In addition, however, it is proposed according to the present invention to provide incremental redundancy for the decoder of the receiver by generating an auxiliary codeword portion from the input data words according to a second code. The "total" code, i.e. the code according to which the "total" codeword (comprising the basic codeword portion and the auxiliary codeword portion) is generated, thus has a lower code rate than the first code. Hence, said "total" code, in particular said auxiliary codeword portion, provides a higher robustness and enables (better) decoding than the first code even under bad reception conditions.

[0010] Hence, under normal reception conditions the decoder does generally not (need to) use the auxiliary codeword portion at all, but only uses the basic codeword portion to decode the received data. In situations where the decoder realizes that the decoding of received data is erroneous or with insufficient quality it uses part of or the complete auxiliary codeword portion to better decode the received data. Thus, the basic codeword portion can be used as it is by a receiver / decoder for decoding, and the auxiliary codeword portion needs only to be used if really needed for decoding.

[0011] Furthermore, the auxiliary codeword portion represents an additional measure for improving the decoding abilities, particularly for mobile receivers in case of bad reception conditions. Both the basic codeword portion as well the auxiliary codeword portion are mapped into the transmitter output data stream by an appropriate data mapper of the transmitter, said transmitter output data stream generally being segmented into frames. For instance, a framing structure as applied according to the DVB-T2 system using T2 frames and FEF (Future Extension Frames) frames can be used for transporting the two codeword portions in an appropriate way.

[0012] Preferably, the encoder and the decoder apply a systematic code for generating said codewords such that said basic codeword portion comprises a data portion, in particular the input data word, and a basic parity portion, and said auxiliary codeword portions comprises an auxiliary parity portion. For instance, the basic codeword portion may be a combination of information symbols (e.g. information bits or information bytes) of an input data word and generated basic parity symbols (e.g. basic parity bits or bytes), said combination representing a basic codeword of a first code, which can be decoded by the decoder. In this example, the auxiliary codeword portion may comprise auxiliary parity symbols (e.g. auxiliary parity bits or bytes), said auxiliary codeword portion representing an auxiliary codeword of a second code, which can be used, to improve the likelihood of decoding said first codeword.

[0013] For generating said (basic and auxiliary) parity symbols of said basic and auxiliary codeword portion parity symbol accumulators are used, as generally known and applied, e.g. according to DVB-T2 (ETSI EN 302 755 VI.1.1 (2009 - 09) "Digital Video Broadcasting (DVB); Framing structure Channel Coding and Modulation for a Second Generation Digital Terrestrial Television Broadcasting System (DVB- T2)"), DVB-C2 (DVB BlueBook A138 "Digital Video Broadcasting (DVB); Frame structure channel coding and modulation for a second generation digital transmission system for cable systems (DVB-C2) or DVB-S2 (ETSI EN 302 307 VI.2.1 (2009 - 08) "Digital Video Broadcasting (DVB); Second generation framing structure, channel coding and modulation systems for Broadcasting, Interactive Services, News Gathering and other broadband satellite applications (DVB-S2)"). For such a parity symbol accumulation an address generation rule is used to determine the addresses of the parity symbols, at which an information symbol is accumulated ("accumulating" being in general also referred to as "encoding"). Generally, however, only a single address generation rule is applied, which is applied for determining all parity addresses of the basic codeword, which shall not exclude that more address generation rules are applied for generating the basic code. According to the present invention, however, different address generation rules are used for generating the basic parity symbols and the auxiliary parity symbols to obtain the desired features of the first and second codes. Particularly the second address generation rule to obtain each auxiliary parity symbol by accumulating an information symbol m at a parity symbol address y, wherein said parity symbol addresses y are determined according to the second address generation rule, is given as

N_ldpc - K_ldpc + {x + m mod G_a x Q_1R } mod M_1R if x > N_ldpc - K_ldpc , wherein x denotes the addresses of a parity symbol accumulator corresponding to the first information symbol of a group of size G_a and Qj_R is an auxiliary code rate dependent, predefined constant, is designed such that the desired backward compatibility of the codes is obtained, as explained later. Thus, the total code also comprises the first code. Furthermore, the second address generation rule follows the same blockwise (and quasi-cyclic) encoding principle (in groups of length G_a) as is done for all LDPC codes of of the DVB family of standards. In particular, both groups can have the same length.

[0014] The encoder and the encoding method of the present invention can be generally applied to extend a basic code into an extended code (having a lower code rate than said basic code), but where different types of decoders can decode at least the basic code, i.e. where backward compatibility is ensured for legacy decoders which cannot decode the extended code, but only the basic code.

[0015] According to a preferred embodiment said encoding means is adapted for generating a basic parity symbol by accumulating an information symbol m at a parity symbol address y, wherein said parity symbol addresses y are determined according to the first address generation rule

{x + m mod G_b x Q_ldpc } mod(N^_c - K_ldpc ) if x < N_ldpc - K_ldpc wherein x denotes the addresses of a parity symbol accumulator corresponding to the first information symbol of a group of size G_b and Qj_dpc is a basic code rate dependent, predefined constant. Hence, according to this embodiment the same address generation rule is applied that is defined in the DVB-T2 (or DVB-C2 or DVB-S2) standard, which enables receivers according to the DVB-T2 standard to decode the first code. Another advantage of this backward compatibility is that it simplifies decoding of several data input streams, which are either encoded by the basic encoding rule or by the extended encoding rule, which outputs both basic and auxiliary codeword portions in the following way: if successful decoding of the basic codeword portions is possible, the decoder can use the same (basic) decoding operation. Only, if decoding of the basic code fails, the decoder has to change the decoding operation according to the extended code (and can ignore those input data streams, which are encoded only by the basic encoder only).

[0016] According to a further embodiment said group sizes used in said first and second address generation rules are identical, i.e. G_a = G_h. Preferably a group size of 360 is selected (according to the LDPC code generation of DVB). Thus, a blockwise (or groupwise) encoding is performed by subsequently taking groups of G = G_a = G_b (preferably = 360) information symbols (preferably, information bits) and encoding them into parity symbols. In this way the quasi-cyclic structure for both the basic parity portion and the auxiliary parity portion is maintained, which enables block- wise and, thus, simpler decoding based on said groups having a fixed group size.

[0017] Preferably, said encoding means is adapted for blockwise generating said basic parity symbols and said auxiliary parity symbols by use of a group of subsequent information symbols, wherein each information symbol i of said group of subsequent information symbols is accumulated at a set of different parity symbol addresses γ, wherein the set of parity symbol addresses, at which the first information symbol of said group is accumulated, is taken from a predetermined address table and wherein the symbol addresses, at which the subsequent information symbols of said group are accumulated, are determined from said set of parity symbol addresses according to said first or said second address generation rule, respectively, and wherein a separate set of parity symbol addresses is taken from said address table for generating each new block of basic parity symbols and auxiliary parity symbols. The use of such predetermined address tables enables that the parity symbol addresses contained therein can be optimized in advance for each desired combination of values of the data rate, group sizes G_a, G_hl code rate dependent predefined constants Q_ddpa Q _R, third number of auxiliary parity symbols M_IR and length of the basic codeword portion N_ldpC; so that the obtained code is as strong as possible and allows correcting of as many errors as possible. Further, the addresses provided in said address tables (whose number are valid for groups of G_a and G_b of information symbols) provide the advantage that these address generation rules (and also the address tables) can be formulated and stored efficiently and with a reasonable amount of storage space in the transmitter and the receivers. Otherwise, the address generation rules and the address tables would be much larger and could have an extremely large size, if for each information symbol the parity address had to be provided in a table explicitly.

[0018] In a particular advantageous application the following parameter values are used: M_IR = N_ldpc = 16200, Qj_R = 45, G = G_a = G_h = 360, and Qi_R = 45. Further, for the code rate identifiers the values 1/2, 3/5, 2/3, 3/4, 4/5, 5/6 and 8/9 and for the parameter Qj_dpc the values (in the same sequence, i.e. corresponding to the respective code rate) 25, 18, 15, 12, 10, 8 and 5 are selected. As is known from the DVB-T2 and DVB-S2 standards the code rate identifiers are not always identical to the true code rate (e.g., the identifier 4/5 of the 16k code from DVB-T2 or DVB-S2 has a true code rate, which is 7/9). All basic codewords are encoded according to the DVB-S2, DVB- T2, DVB-S2 or upcoming DVB-NGH standard, except the rate 3/5 code, which is according to the DVB-T2 standard. Optimized address tables including the parity symbol addresses for these different code rates and these parameter values are defined in further dependent claims. Hence, according to such embodiments the encoding means is adapted for subsequently taking a new row of the following address table as a new set of different parity symbol addresses y for accumulating a new group of subsequent information symbols.

[0019] While the present invention can be used in a general way to extend an existing code, in a preferred application said basic codeword portion is provided for regular decoding and said auxiliary codeword portion is provided as incremental redundancy if regular decoding of the codeword by use of the basic codeword portion is erroneous.

[0020] According to a preferred embodiment of the transmitter the data mapper is adapted for mapping the basic codeword portion of a codeword onto a different portion of the transmitter output data stream, in particular onto a different frame, than the auxiliary codeword portion of the same codeword. This provides the advantage that the auxiliary codeword portion might not be affected by disturbances of the channel that affect the basic codeword portion, such as time selective fadings or noise bursts. Generally, the amplitude and phase of different reception paths also depend on the position of the receiver. Further, in case of a moving receiver, especially the phase of the signals of different reception paths changes, which causes a time-selective channel. The changes in the time direction can also have a very regular structure, the change rate of which over the time axis being proportional to the relative velocity of the receiver to the transmitter and the transmission frequency of a signal. Also other disturbances, such as impulsive noise, can have a regular structure, e.g. caused by the line cycle frequency of the power grid or by bursts from other data transmission systems, e.g. a GSM communications system. Mapping the basic codeword portion on the one hand and the corresponding auxiliary codeword portion on the other hand onto different portions of the transmitter output data stream may in such situations avoid that all data related to a particular codeword are affected by such regular disturbances and may thus ensure correct decoding of the codeword at the receiver. Further, as will be explained below in more detail, a receiver can fall into sleeping mode during the transmission of the auxiliary data portion, if said auxiliary data portion is not required for decoding.

[0021] According to another preferred embodiment of the transmitter said data mapper is adapted for mapping the basic codeword portion of a codeword onto a frame of a first type of the transmitter output data stream, in particular onto a T2 frame of a transmitter output data stream in accordance with a DVB broadcasting system, and for mapping the auxiliary codeword portion of the respective codeword onto a frame of a second type, arranged between frames of the first type, of the transmitter output data stream, in particular onto an FEF frame of a transmitter output data stream in accordance with a DVB broadcasting system. This provides the advantage that stationary receivers only access the data transmitted in the frames of the first type, for instance that receivers in accordance with the DVB-T2 standard only access the data transmitted in the T2 frames. Mobile receivers generally also access the data transmitted in the frames of the first type which are also, according to the present invention, sufficient for correctly decoding the received codewords and for reproducing the encoded data words. However, if the decoder realizes that the decoding is erroneous or that the decoded data have insufficient quality, it may then also access the data, i.e. the auxiliary codeword portion, transmitted in the frames of a second type and use these additional data as incremental redundancy, i.e. use the basic codeword portion and (part or all of) the auxiliary codeword portion for decoding, which finally provides an increased probability that the data are decoded correctly since the total code of a codeword, i.e. a combination of the basic codeword portion and the auxiliary codeword portion, has lower code rate than the first code according to which the first codeword is encoded.

[0022] This embodiment provides the additional advantage that the existing framing structure as defined in the DVB-T2 standard may be used, for instance in such a way that stationary receivers (in accordance with the DVB-T2 standard) only access the data transmitted in the T2 frames and that mobile receivers access the data transmitted in the T2 frames and, if needed, in addition the auxiliary codeword portions transmitted in the FEF frames. Stationary receivers may, of course, also make use of the auxiliary codeword portions transmitted in the FEF frames if needed, but will generally ignore these data.

[0023] In a preferred embodiment the encoder comprises a first encoding unit for encoding, according to said first code, the input data word into a basic codeword including said basic codeword portion and a second encoding unit for encoding, according to said second code, the input data word into an auxiliary codeword including said auxiliary codeword portion. Further, the data mapper is adapted for mapping the basic codeword onto a different portion of the transmitter output data stream, in particular onto a different frame, than the auxiliary codeword. Hence, according to this embodiment, an already existing encoder, i.e. the first encoding unit, can be used without any changes, and simply a second encoder, i.e. the second encoding unit, is added which is also provided with the input data words from which the auxiliary codewords are generated according to the second code, which itself may also have a lower code rate than the first code applied by the first encoding unit, but which can also have the same or a higher code rate.

[0024] While it is generally possible to map both the basic codewords and the auxiliary codewords completely onto the transmitter output data stream, it is in certain embodiments (in particular if the auxiliary codewords do not only contain auxiliary parities but also parts of or the complete input data word and/or basic parities of the basic codeword) sufficient for achieving the desired object that only the auxiliary parity portion is mapped onto the transmitter output data stream in addition to the basic codeword. The receiver then exploits, in case of need, the auxiliary parity portion as redundancy for decoding a received basic codeword that could not be decoded correctly.

[0025] The second encoding unit may further be adapted for encoding, according to said second code, the input data word into an auxiliary codeword including said basic codeword portion and said auxiliary codeword portion. Hence, the basic codeword portion is part of both the basic codeword and the auxiliary code- word, but is not further used in the auxiliary codeword, from which basically the auxiliary parity portion is embedded into the transmitter output data stream. Such an embodiment has the advantage that the first encoding unit can be the (conventional) encoder of a known receiver, e.g. a receiver in accordance with the DVB-T2 standard, which can generally be used without changes, and to which the second encoding unit is added according to the invention.

[0026] Alternatively, the second encoding unit may further be adapted for encoding, according to said second code, the input data word into an auxiliary codeword including only said auxiliary codeword portion. Such an encoding unit can be easily added to an existing encoder and performs only the minimum required steps for implementing the present invention on the transmitter side.

[0027] According to an alternative embodiment of the transmitter said encoder comprises a single encoding unit for encoding the input data word into a codeword including said basic codeword portion and said auxiliary codeword portion, and wherein said data mapper is adapted for mapping the basic codeword portion of said codeword onto a different portion of the transmitter output data stream, in particular onto a different frame, than the auxiliary codeword portion of said codeword. This embodiment requires less processing capacity since only a single encoding process is carried out for generating the codewords, which are thereafter split up into the basic codeword portion, on the one hand, and the auxiliary codeword portion on the other hand.

[0028] According to a further embodiment said encoder is adapted for encoding the input data words into codewords, a codeword comprising a basic codeword portion and an auxiliary codeword portion, said auxiliary codeword portion including at least two auxiliary codeword sub-portions, said basic codeword portion being provided for regular decoding and said at least two auxiliary codeword sub- portions being provided as incremental redundancies if regular decoding of the codeword by use of the basic codeword portion and less redundancies is erroneous. This embodiment provides the advantage that a receiver may decide how many additional incremental redundancies it requires for correct decoding if the regular decoding, i.e. the decoding by use of only the basic codeword portion, is erroneous. This is enabled by encoding the input data words such that two or more auxiliary codeword sub-portions (e.g. two or more groups of auxiliary parity symbols) are generated which can stepwise be used as such incremental redundancies, i.e. the auxiliary codeword sub-portions are generated such that not all sub-portions are required completely to perform a correct decoding of a received codeword, but one or more sub-portions thereof are also sufficient for a correct decoding. If more sub- portions are used, however, the code rate decreases and the probability of correct decoding increases.

[0029] Thus, each receiver may decide by itself (which decision may also change from time to time) how many additional incremental redundancies, i.e. how many of those at least two auxiliary codeword sub-portions shall be used for increasing the quality of decoding, if needed. Hence, if only a small sub-portion is additionally required, the other auxiliary codeword sub-portions may be ignored and may not even be received or at least demapped at all so that during time when these other auxiliary codeword sub-portions are transmitted the receiver can be switched into sleeping mode saving battery power and processing time.

[0030] This embodiment may be developed further such that the data mapper is adapted for mapping the at least two auxiliary codeword sub-portions of a codeword onto said transmitter output data stream such that the auxiliary codeword sub-portion used as first incremental redundancy is received by a receiver after reception of the respective basic codeword portion, but before further auxiliary codeword sub-portions. Such an embodiment ensures that a receiver can be switched into sleeping mode after having received sufficient auxiliary codeword sub-portions for enabling a correct decoding if, as proposed, the "most helpful" auxiliary codeword sub-portions (e.g. the largest auxiliary codeword sub-portions) are transmitted first. In another embodiment, the auxiliary codeword sub-portions may be mapped on the transmitter output data stream such that at first a small auxiliary codeword sub-portion is provided, and that thereafter auxiliary codeword sub-portions are provided with increasing size.

[0031] The sequence of the auxiliary codeword sub-portions can preferably be derived by the receiver from the sequence of the related basic codeword portions transmitted before so that no additional signalling is required for signalling the sequence of the auxiliary codeword sub-portions, e.g. the sequence of the auxiliary codeword sub-portions is identical to the then already known sequence of the related basic codeword portions.

[0032] Further, if in the frames (e.g. T2 frames) preceding a frame (e.g. FEF frame) containing auxiliary codeword sub-portions data from various transmitter input data streams (e.g. PLPs) are contained, the auxiliary codeword sub-portions may be grouped together such that, for instance, the largest or smallest sub-portions of all auxiliary codeword portions are transmitted first, and that thereafter further groups of sub-portions are transmitted.

[0033] Generally, the basic codeword portions and the auxiliary codeword portions are mapped onto the transmitter output data stream such that a basic codeword portion of a codeword is received by a receiver before the corresponding auxiliary codeword portion is received. Alternatively, however, the data mapper is adapted for mapping the basic codeword portions and the auxiliary codeword portions onto frames of a transmitter output data stream such that the auxiliary codeword portion of a codeword is received by a receiver before the corresponding basic codeword portion is received. The auxiliary codeword portions are thus buffered in the receiver in a respective buffer. If it is clear that the decoding of the corresponding codeword based on only the basic codeword portion was correct, the corresponding auxiliary codeword portion (if any) can be deleted from the buffer. Even if decoding based on the basic codeword portion is not correct, the corresponding auxiliary codeword portion is already available in the buffer. This provides the advantage that no or reduced waiting times (for waiting for the auxiliary parity portion, if the decoding was erroneous based on the basic codeword portion) occur, which is particularly important for reducing zapping times and for mobile receivers in case of sudden signal distortions. Hence, this embodiment also provides the advantage that no interruption of the service occurs (due to waiting for reception of auxiliary codeword portions) in case of (e.g. sudden) bad reception conditions of the basic codeword portions.

[0034] According to another embodiment the data input is adapted for receiving at least two transmitter input data streams segmented into input data words and the encoder is adapted for selectively encoding a transmitter input data stream only according to the first code, according to the second code or according to both codes. Hence, according to this embodiment the transmitter, for instance under control of the system operator of the broadcasting system, has the freedom to apply the idea of the present invention to provide incremental redundancies into the transmitter output data stream only for selected transmitter input data streams but must not generally apply it to all transmitter input data streams. For instance, a system operator may decide to apply the present invention for data streams that are provided for reception by both a stationary receiver and a mobile receiver, while other data streams that shall generally only be received by stationary receivers are not processed according to the present invention, i.e. no auxiliary codeword portions are generated and transmitted.

[0035] Generally, the transmitter output data stream can be transmitted by any kind of transmitter, for instance having only a single antenna or multiple antennas such as, for instance, used in MIMO (Multiple Input Multiple Output) systems. In a preferred embodiment, however, the data mapper is adapted for mapping the basic codeword portions of the codewords onto a first transmitter output data stream and for mapping the auxiliary codeword portions of the codewords onto a second transmitter output data stream, and the transmitter unit comprises a MIMO precoder for MIMO precoding said first and second output data streams, and at least two antennas including a first antenna and a second antenna for transmitting said MIMO precoded first and second transmitter output data streams. For instance, the MIMO precoder is adapted for spatial multiplexing such that a first antenna transmits said first transmitter output data stream and a second antenna transmits said second transmitter output data stream. Generally, however, any other embodiment of MIMO precoding (e.g. Alamouti precoding) can also be applied on the two transmitter output data streams before outputting them by two or more antennas of a MIMO system.

[0036] A receiver may also have only a single antenna or multiple antennas, wherein generally each antenna of the receiver receives signals from all antennas of the transmitter. It may, however, also be possible that generally the antennas of the receiver are tuned to receive the signals from the first antenna only (i.e. for receiving the first transmitter output data stream) and that only in case of decoding problems or errors the antennas are additionally tuned to receive the signals from the second antenna (for receiving the second transmitter output data stream). According to a further embodiment it may be possible that the receiver comprises one antenna that is tuned to the first antenna of the transmitter and that the receiver comprises a second antenna that is tuned to receive the signals from the second antenna of the transmitter, and that the signals received at the second antenna are only evaluated in case of need.

[0037] In a further embodiment the transmitter comprises a modulator for modulating the basic codeword portion of a codeword differently than the auxiliary codeword portion of the same codeword before or after mapping the codewords onto the frames of the transmitter output data stream. In general, the auxiliary codeword portion can be transmitted in any dimension, which is orthogonal to the basic codeword portion of a codeword, such as time, frequency, space (MIMO) or spreading code. The term "orthogonal" shall be understood as generally known in the art of coding and modulation, i.e. to provide separable data streams (which might even overlap before separation, such as OFDM subcarriers and their related partial spectra). This provides the advantage that the auxiliary codeword portion is additionally protected against disturbances by which the basic codeword portion may be disturbed. Another possibility is the application of a hierarchical modulation. In this case, the basic codeword portion addresses the more reliable bits of the modulation scheme, while the auxiliary codeword portion addresses the less reliable bits.

BRIEF DESCRIPTION OF DRAWINGS

[0038] These and other aspects of the present invention will be apparent from and explained in more detail below with reference to the embodiments described hereinafter. In the following drawings

Fig. 1 shows a schematic block diagram of an embodiment of a transmitter in accordance with the present invention,

Fig. 2 shows a schematic block diagram of a first embodiment of an encoder used in the transmitter,

Fig. 3 shows the arrangement of a data portion, a basic parity portion and an auxiliary parity portion according to the present invention,

Fig. 4 shows the format of a BBFrame according to the DVB-T2 standard,

Fig. 5 shows the format of an FEC codeword in accordance with a DVB-T2 standard,

Fig. 6 shows the format of a codeword in accordance with the present invention,

Fig. 7 shows a diagram illustrating the DVB-T2 framing structure,

Fig. 8 shows a diagram illustrating the arrangement of T2 frames and FEF frames within a superframe in accordance with the DVB-T2 standard, Fig. 9 shows a schematic block diagram of a second embodiment of an encoder,

Fig. 10 shows a schematic block diagram of a third embodiment of an encoder,

Fig. 11 illustrates the arrangement of auxiliary parity portions according to the present invention,

Fig. 12 shows the mapping of data PLPs onto a T2 frame according to the

DVB-T2 standard,

Fig. 13 shows the mapping of segmented auxiliary parity portions onto an

FEF frame,

Fig. 14 shows a schematic block diagram of a fourth embodiment of an encoder,

Fig. 15 shows a schematic block diagram of a receiver,

Fig. 16 shows a schematic block diagram of a first embodiment of a decoder used in the receiver,

Fig. 17 shows a schematic block diagram of a second embodiment of a decoder,

Fig. 18 shows a schematic block diagram of a broadcasting system in accordance with the present invention,

Fig. 19 shows a diagram illustrating the generation of the bits of a codeword according to the present invention, Fig. 20 shows the structure of a transmission frame as used according to DAB,

Fig. 21 shows a block diagram of another embodiment of a receiver,

Fig. 22 shows a block diagram of another embodiment of an encoder according to the present invention,

Figs. 23 to 29 show various code rate dependent address tables used for generation of the parity bits,

Fig. 30 shows a diagram illustrating the performance of the total code proposed according to the present invention,

Fig. 31 shows a block diagram of another data transmission system, and

Fig. 32 shows a block diagram of a receiver used in such a data transmission system.

[0039] Fig. 1 shows an example block diagram of a transmitter 10 in accordance with the present invention. Such a transmitter 10 can, for instance, be a Coded OFDM (COFDM) transmitter which may be used to transmit video, images and audio signals in accordance with the DVB-T2 (or an upcoming DVB-NGH) standard and in which the invention can be used. Said data to be transmitted by the transmitter 10 are generally provided as at least one transmitter input data stream II, 12, In, which data streams are generally segmented into input data words. Said transmitter input data streams II, 12, In may be one or more (e.g. MPEG-2) Transport Stream(s) and/or one or more Generic Stream(s), and the data may be carried therein in individual Physical Layer Pipes PLPs. [0040] From the data input 12, in which some input processing may be performed on the transmitter input data streams II, 12, In, such as CRC (Cyclic Redundancy Check) encoding, BB (BaseBand) header insertion, padding insertion and BB scrambling, the input data are provided to an encoder 14 in which the input data words of the transmitter input data streams II, 12, In are encoded into codewords as will be explained in more detail below. From the encoder 14 the encoded data are then provided to a data mapper 16 for mapping the generated codewords onto frames of a transmitter output data stream O, which is then output- ted by a transmitter unit 18. Generally (but not mandatorily), a modulator 17 is provided for modulating the data before output and transmission.

[0041] Fig. 2 depicts a first embodiment 141 of an encoder 14 according to the present invention. Said embodiment of the encoder 141 comprises two branches, i.e. two encoding units 20, 30 to which a transmitter input data stream II, called Physical Layer Pipe (PLP) in the context of DVB, is fed. With the exception of the time interleavers 27, 37, the processing is preferably performed on a fixed frame level. An input frame of the transmitter input data stream II, which shall be looked at in the following as an example, is denoted as a BBFrame in the context of DVB.

[0042] The first encoding unit 20, i.e. the upper branch, in this embodiment corresponds to the Bit Interleaved Coding and Modulation (BICM) chain as described in the DVB-T2 standard. Hence, it comprises an FEC encoding block 21 for subsequent LDPC encoding, a bit interleaver 22, a demultiplexer 23 for demultiplexing bits to cells, a constellation mapper 24 for mapping cells to constellations according to Gray mapping, a unit 25 for constellation rotation and cyclic Q-delay, a cell interleaver 26 and a time interleaver 27. The function and operation of these units 21 to 27 is generally known and, for instance, described in the DVB-T2 standard, which is herein incorporated by reference, so that no further explanations are provided here. [0043] The second encoding unit 30, i.e. the lower branch, is also provided with the transmitter input data stream II in this embodiment. The FEC encoding block 31 is generally not identical to the FEC encoding block 21 of the first encoding unit 20. While said FEC encoding block 21 appends to the input data words parity bits of an LDPC codeword, said LDPC parity bits being generally referred to herein as basic parity portion of a first code, the FEC encoding block 31 generates additional redundancy to increase the robustness of the overall channel code, said overall channel code referring to the redundancy from both FEC encoding blocks 21 and FEC encoding blocks 31. In other words, the FEC encoding block 31 generates auxiliary parity bits that can, in addition to the basic parity bits, be used by a receiver to decode a received codeword as will be explained in more detail below.

[0044] The consecutive blocks 32 to 37 can generally be identical to the blocks 22 to 27 and can thus be adopted from the DVB-T2 standard, but can also be adjusted according to the specific circumstances and needs of the second encoding unit 30. The application of a time interleaver 37 is optional since applying time interleaving within just one frame of auxiliary parity bits is already covered within the cell interleaver 36. However, applying time interleaving over more than one auxiliary parity data frame allows for more time diversity.

[0045] In this embodiment, the input of the two FEC encoding blocks 21, 31 is identical, in particular an input data stream II*, which substantially corresponds to the transmitter input data stream, but wherein to the input data words (BBFrames in the context of DVB) parity bits of a BCH codeword have been added by a BCH encoder 40 (as is generally known in the art of DVB). Hence, the input data stream II has already been encoded by a BCH code, before further encoding is performed in the FEC encoders 21 and 31. It should, however, be noted that the encoder 40 is generally not an essential element of the present invention. In certain application the encoder 40 can be completely omitted, can be replaced by a different encoder or this initial encoding can be part of the encoding performed in the encoders 21 and 31. [0046] Further, it shall be noted that hereinafter it is generally referred to parity "bits" and input data "bits". The same idea is, however, also applicable using parity "bytes" and input data "bytes" or, generally, parity "symbols" and input data "symbols".

[0047] The outputs of the first and second encoding units 20, 30 are fed forward to a data mapper 16, generally including a frame builder, and, optionally, an OFDM generator. The data mapper 16 and the OFDM generator may generally operate according to the DVB-T2 standard which particularly shows embodiments of these blocks. For mapping the outputs of the first and second encoding units 20, 30, however, various embodiments exist which will also be explained below in more detail.

[0048] By way of Fig. 3 the encoding performed in the FEC encoding blocks 21 and 31 shall be explained in more detail. Fig. 3A shows a first codeword Zl which is the output of the FEC encoding block 21. Said first codeword Zl comprises a data portion D, which generally corresponds to the input data word of the FEC encoding block 21 and which, according to this embodiment, comprises k input data bits s_1; s₂, s_k, and a basic parity portion Pb comprising, in this embodiment, m parity bits p_1; p₂, p_m. This first codeword Zl generally corresponds, in the context of DVB-T2, to the LDPC codeword belonging to a code Q with code rate Rj = k/(k+m), where k is the amount of the systematic bits (input data bits) s and m is the amount of the basic parity bits p. By use of these codewords a receiver, in particular a stationary receiver and/or a mobile receiver that is not affected by too many disturbances, is able to decode the input data encoded therein.

[0049] The second codeword Z2 shown in Fig. 3B belongs to a second code C₂ of a lower code rate R₂ = k/(k+m+v) < R_x. Said second codeword Z2 comprises, in addition to the data portion D and the basic parity portion Pb, an auxiliary parity portion Pa of v auxiliary parity bits i_1; i₂, i_v. Said auxiliary parity portion Pa may be used by a receiver, for instance in case of transmissions errors, decoding errors and/or an insufficient quality of the decoding, as incremental redundancy in addition to the codeword Zl for decoding. Hence, if the first codeword Zl cannot be decoded correctly at the receiver, some or all of the auxiliary parity bits i_1; i₂, i_v may be used for decoding, in addition to the first codeword Zl, to increase the probability of correct and error-free (or at least an increased quality of) decoding. For this purpose at least the auxiliary parity portion Pa will also be mapped onto the transmitter output data stream for eventual reception and use by the receiver. Said mapping will be explained below in more detail.

[0050] Hence, the combination of the codeword Zl and the auxiliary parity portion Pa (of the second codeword Z2) can also be regarded as a "total" codeword of a "total" code having a lower code rate than the first code of the codeword Zl, i.e. the codeword Zl can be regarded as a basic codeword portion B of this "total" codeword and the auxiliary parity portion Pa can be regarded as a auxiliary parity portion A of this "total" codeword. Here in this embodiment shown in Fig. 3 this "total" codeword is identical to the codeword Z2. This does, however, not hold for all embodiments as will be shown below.

[0051] In the context of DVB-T2 the transmitter input data stream to the encoder 14 is generally segmented into frames referred to as BBFrames comprising K_bch bits as exemplarily depicted in Fig. 4. A first codeword Zl (in this context) generated therefrom by the BCH encoder 40 and the first encoding block 21, i.e. in accordance with the encoding as performed according to the DVB-T2 standard, is schematically depicted in Fig. 5. This codeword is a standard FEC codeword comprising the (systematic) input data portion, which itself consists of K_bch bits, followed by N_bch - K_bch parity bits of the BCH encoder, followed by N_ldpc - K_ldpc parity bits of the LDPC encoder. In total, this codeword comprises N_ldpc bits. Hence, the basic LDPC code has a code rate of R_c = K_ldpc/N_ldpc. Referring to the above explanations provided with reference to Figs. 3 and 4 K_ldpc corresponds to k (i.e. the portions BBFRAME and BCHFEC are regarded as the input data word D) and N_ldpc corresponds to k+m (i.e. the portion LDPCFEC is regarded as the basic parity portion Pb). [0052] The second encoding block 31 computes auxiliary parity bits for use as incremental redundancy, based on its input, which generally is the same as the input of the FEC encoding block 21. In general, these are v auxiliary parity bits which can be partitioned into q sub-portions (it should be noted that in the main aspect of the present invention only a single sub-portion is used). The k-th sub-portion is of length v^(k). Thus, it holds

[0053] If the bits from the first x sub-portions are then appended to the first basic codeword (Zl) generated by the first encoding unit 20 and generally received and evaluated by a receiver, an auxiliary codeword (Z3*) of the "total" code is generated which stems from an encoder with the overall code rate R_c* of

which is smaller than R„ which means that this overall code is more powerful.

[0054] Fig. 6 shows such an embodiment of a codeword Z2 (in the context of DVT-T2) generated by the second FEC encoding block 31 which, according to this embodiment, also comprises BCH and LDPC encoding, but additionally generates v auxiliary parity bits for use as incremental redundancy at the receiver in case of need.

[0055] The generation of additional LDPC parity bits, e.g. of a known LDPC code, and their use as incremental redundancy is generally known, e.g. from Kim J. et al.„Design of Rate-Compatible Irregular LDPC Codes for Incremental Redundancy Hybrid ARQ Systems", ISIT 2006, Seattle, USA, July 9-14, 2006. A diagram illustrating such an "extended" code and its generation is shown in Fig. 19. Therein, it is shown how each of the bits of the basic codeword portion B and of the auxiliary parity portion A are generated from other, in particular all "previous" bits in the codeword by use of modulo-2 units 45. This way of generating the codeword ensures that a decoder can decode a codeword by use of only the basic codeword portion B or by additional use of one or more of the auxiliary parity bits of the auxiliary parity portion A. Below, particular ways of generating the LDPC parity bits according to preferred embodiments of the present invention will be explained in detail.

[0056] Next, an embodiment of the data mapper 16 shall be explained. Generally, it is sufficient if the first codeword (generally, the basic codeword portion B) generated by the first encoding unit 20 and the auxiliary parity portion (generally, the auxiliary codeword portion A) generated by the second encoding unit 30 for the same input data word are mapped in any way onto the frames of the transmitter output data stream O. In other words, using the terminology from Fig. 3, for each input data word the corresponding data portion D, the basic parity portion Pb and the auxiliary parity portion Pa are mapped onto the transmitter output data stream O according to the present invention. Preferably, however, the data mapper 16 is adapted such that the data portion D and the basic parity portion Pb are mapped onto a different portion, for instance onto a different frame, than the corresponding auxiliary parity portion Pa of the same codeword. This provides the advantage that regular channel disturbances do generally not affect both the data portion D and the basic parity portion Pb on the one hand and the auxiliary parity portion Pa on the other hand.

[0057] A particular embodiment for such a mapping structure shall be illustrated with reference to Figs. 7 and 8. Fig. 7 illustrates the framing structure as applied according to the DVB-T2 standard. In particular, according to DVB-T2, a superframe structure is applied where each superframe is subdivided into a multitude of T2 frames. After each predetermined number of consecutive T2 frames an FEF part (Future Extension Frame part) is inserted for future use. This is also schematically illustrated in the data stream structure shown in Fig. 8. When applying such a framing structure in the transmitter 10 according to the present invention, the data mapper 16 is adapted in one embodiment such that the data portion and the basic parity portion of a codeword, which can also be regarded as the basic codeword portion B and, in this embodiment, the basic codeword Zl (see Fig. 3A), are mapped onto the T2 frames and that the auxiliary parity portion Pa (generally the auxiliary codeword portion A) of the same codeword (taken from the auxiliary codeword Z2) is mapped onto the FEF part, preferably the FEF part following next to the T2 frame(s) into which the corresponding data portion D and basic parity portion Pb are mapped.

[0058] Such a mapping provides the advantage that a conventional receiver in accordance with the DVB-T2 standard simply ignores the data transmitted in the FEF parts and only evaluates the data transmitted in the T2 frames as usual. Mobile receivers, however, for instance in accordance with the upcoming DVB-NGH standard, whose decoding and reproduction capability may often be affected by disturbances, may also access the T2 frames and decode, in a first step, the codewords embedded therein. In addition, however, particularly in case of disturbances and decoding errors resulting therefrom, such mobile receivers access the FEF parts and use parts or all of the auxiliary parity data contained therein for decoding, in a second step, the codeword received in the corresponding T2 frame again, as will be explained in more detail below.

[0059] According to still another embodiment of the data mapper 16 all data required for decoding by a mobile receiver are transmitted in the FEF parts, i.e. a complete codeword comprising the data portion D, the basic parity portion Pb and the auxiliary parity portion Pa is mapped onto the FEF part. Such mobile receivers thus ignore the data contained in the T2 frames which are only accessed by stationary receivers, in particular receivers in accordance with the DVB-T2 standard.

[0060] In such a situation, however, the auxiliary parity portion Pa is preferably modulated in a different way than the data portion D and the basic parity portion Pb. Preferably, an auxiliary modulation code is applied for modulating the auxiliary parity portions, said auxiliary modulation code being orthogonal to a basic modulation code that is used for modulating the data portions and the basic parity portions, generally after the data mapping. For instance, an orthogonal time, frequency, space (MIMO) or a spreading code may be applied. Another possibility would be the application of hierarchical modulation.

[0061] Of course, there may be further embodiments of the data mapper 16. The framing structure applied by the transmitter 10 may also be completely different than the framing structure used according to the DVB-T2 standard as shown in Figs. 7 and 8. Generally, any framing structure, e.g. a newly created framing structure, may be applied as long as the receiver is able to detect or know in advance where to find the data portions and the various parity portions. Further, in embodiments of the invention BCH and LDPC encoding is not required, but other codes (e.g. other FEC codes) can be applied.

[0062] A simplified block diagram of another embodiment of an encoder 142 and a data mapper 16 is shown in Fig. 9. According to this embodiment the encoder 142 comprises a single encoding unit by which the input data words are encoded, i.e. by which both the basic parity portions Pb and the auxiliary parity portions Pa are generated. In other words, in said single encoding unit of the encoder 142 the complete codeword Z2 (see Fig. 3B) is generated. These codewords Z2 are provided to the data mapper 16 which splits off the auxiliary parity portion Pa and maps it onto a different portion of the transmitter output data stream than the data portion D and the basic parity portion Pb. For the data mapper 16 generally the same embodiments exist as have been explained above.

[0063] As shown in Fig. 1 the data input 12 may be adapted for not only receiving a single transmitter input data stream, but may generally receive a number n of transmitter input data streams, for instance a number n of physical layer pipes. The encoder 14 may, in such a case, however be adapted to select if a transmitter input data stream is encoded as usual, i.e. according to the basic code and without the generation of any auxiliary parity data, or if another code having a lower code rate shall be applied and auxiliary parity data for use as incremental redundancies by the receiver shall be generated. It may also be possible that various codes with different (e.g. decreasing) code rates may be available for application by the encoder so that even more than two possibilities exist. Which code and which code rate to apply may be prescribed, e.g. by the operator of the transmitter or the owner of the broadcast channel. But the choice of the code may also depend on the kind of data to be transmitted. For instance, audio data may be encoded with a code having a higher code rate than video data so that only for video data such auxiliary parity data are generated or vice versa. As another example, any decoding errors may be acceptable when watching news, but may not be acceptable when watching a movie for which auxiliary parity data may thus be generated and transmitted.

[0064] A still further embodiment of an encoder 143 is illustrated in Fig. 10. Said encoder 143 is adapted such that it generates two or more auxiliary parity sub- portions Pal, Pa2, Pa3 in addition to the data portion D and the basic parity portion Pb thus forming a codeword Z3. Hence, comparing the codewords Z2 and Z3, the auxiliary parity sub-portions Pal, Pa2, Pa3 of the codeword Z3 can be seen as segments of the auxiliary parity portion Pa of the codeword Z2, having in total the identical content, although generally the auxiliary parity sub-portions Pa2 and Pa3 can also be additional auxiliary parity portions in addition to the auxiliary parity portion Pal that (alone) corresponds to the auxiliary parity portion Pa.

[0065] These auxiliary parity sub-portions Pal, Pa2, Pa3 are generated such that they can be stepwise used by a decoder as incremental redundancies. In other words, generally it is possible to decode the codeword by use of only the data portion D and the basic parity portion Pb (i.e. the basic codeword portion). If such decoding fails, the first auxiliary parity sub-portion Pal (i.e. a part of the auxiliary codeword portion) may be used in addition for decoding. If this again fails (or provides insufficient quality) the second auxiliary parity sub-portion Pa2 may be added and so on. [0066] All the auxiliary parity sub-portions Pal, Pa2, Pa3 may be grouped together and mapped onto a single portion of the transmitter output data stream. However, it is also possible and advantageous to distribute the various auxiliary parity sub-portions of a single codeword Z3, preferably such that the first auxiliary parity sub-portion Pal is received before the second auxiliary parity sub-portion Pa2, which again is received before the third auxiliary parity sub-portion Pa3. This provides the advantage that a receiver which, after using the first auxiliary parity sub- portion, can decode the codeword with sufficient quality, can fall into sleeping mode for the time period during which other auxiliary parity sub-portions are transmitted that are no longer needed and/ or from other data streams that shall currently not be decoded. This provides some power saving and less calculation efforts at the receiver.

[0067] An embodiment of the arrangement of auxiliary parity sub-portions of various input frames (BBFrames, generally referred to as input data words) is shown in Fig. 11. The input frames are enumerated in this embodiment by two indices (e, f), where the index e corresponds to the PLP_ID (PLP number, also referred to here as the number of the transmitter input data stream), and where f relates to the input frame (input data word). The index e is part of a set S₂, i.e., the set of PLPs that are protected by additional incremental redundancy. Assuming that n different PLPs are transmitted, i.e. e e Sj = {1, n}, S₂ is a subset of Sj of PLPs that are protected by additional incremental redundancy according to the present invention since, as mentioned above, not all PLPs need necessarily use this idea.

[0068] Thus, the f-th input frame of the e-th PLP is denoted as I_e,f. The index f e {1, FJ, where F_e is the number of input frames of the e-th PLP, which precede the FEF, starting from the end of the previous FEF. Hence, in one embodiment, the auxiliary parity sub-portions Pal_e,f up to Pay_e,f may be mapped onto an FEF frame in the sequence as shown in Fig. 11 and may belong to codewords mapped onto preceding T2 frames. [0069] Fig. 12 illustrates more details of the arrangement of data within a T2 frame in accordance with the DVB-T2 standard. The PLPs in a T2 frame underlie a certain order. After the preambles PI, P2 and after common PLPs the PLPs of type 1 are transmitted without subslicing, then PLPs of type 2 are transmitted with subsiding. Both type 1 and type 2 PLPs have a fixed order which is signalled in the P2 preamble. This is also illustrated and explained in detail in the DVB-T2 standard, which explanation is herein incorporated by reference.

[0070] Fig. 13 shows a proposal for arranging the auxiliary parity sub- portions in an FEF frame that is made similar to the arrangement of the PLPs in the T2 frame. In this embodiment the FEF also starts with a PI preamble, i.e. an OFDM symbol, which is made to be used for (time, frequency) synchronization purposes, for channel estimation and for signalling the most important transmission parameters. The subsequent (one or more) P2 preamble(s) contain(s) more detailed information about the content of the FEF. Using the notation explained with reference to Fig. 11 and assuming that there are two PLPs of three input frames each, the arrangement of the auxiliary parity sub-portions in the FEF in this embodiment is such that the order of the partitioned auxiliary parity sub-portions is derived from the slicing/ ordering of the T2 frames, even if not all PLPs from T2 frames have auxiliary parity sub-portions to be mapped onto the FEFs. Thus, the order of the auxiliary parity sub-portions does not have to be signalled explicitly.

[0071] The auxiliary parity portions are sorted in the time domain, in particular such that the first portion Pal_e,f, for instance the more robust portion, of all PLPs having auxiliary parity portions is inserted at the beginning of the FEF, i.e. right after the preambles PI, P2. The second portion Pa2 of all PLPs with auxiliary parity portions follows afterwards etc. As mentioned above, if basic codewords are not decodable the related first portion Pal of the FEF is evaluated. If the receiver can now correctly decode the overall codeword without error it falls into sleeping mode to save power. Otherwise it will additionally include the second auxiliary parity portion Pa2 etc. [0072] It shall be noted also that the FEFs may contain signalling information, e.g. in the preamble or in a ModCod header, regarding the link of the auxiliary parity portions used in the FEFs and the PLPs mapped onto the T2 frames (which itself remain unchanged in this embodiment). Further, other data may be contained in the FEFs as well, for instance low-bitrate information for use by a mobile receiver.

[0073] Fig. 14 shows still another embodiment 144 of an encoder according to the present invention. Similar to the embodiment shown in Fig. 2 the encoder 144 comprises a first encoding unit 20 for encoding, according to said first code, the input data word into said basic codeword Zl as explained above. Further, a second encoding unit 30a is provided for encoding, according to said second code, the input data word into an auxiliary codeword portion (which can be regarded as an auxiliary codeword Z4) comprising only said auxiliary parity portion Pa. Hence, the second encoding unit 30a only generates the data ultimately necessary for improving decoding at the decoder, if required, but does not generate other codeword portions that are already generated by the first encoding unit 20.

[0074] Fig. 15 shows a schematic block diagram of a receiver 50 for use in a broadcasting system comprising a transmitter 10 as illustrated above. The receiver 50 is particularly adapted for making use of the auxiliary parity portion (generally, the auxiliary codeword portion) as incremental redundancy in case of erroneous or low quality decoding.

[0075] The receiver 50 comprises a data input 52 for receiving a receiver input data stream O' which generally corresponds to a transmitter output data stream O that has been transmitted over a broadcast channel of the broadcasting system by a transmitter and which may thus be affected by disturbances that may appear in such a broadcasting system, particularly in case of using mobile receivers which are the main application of the present invention on the receiver side. [0076] Optionally, a demodulator 53 is provided that is interrelated with the (optional) modulator 17 of the transmitter 10 for demodulating the received receiver data input stream O'. A demapper 54 demaps the (optionally demodulated) receiver data input stream O', particularly at least the data portions and the basic parity portions (i.e. the basic codeword portions) of the codewords mapped into the receiver data input stream O' as will be explained below in more detail. A decoder 56 then decodes these codewords by use of the basic codeword portions according to the same code as applied by the encoder 14 of the transmitter 10. As particularly in case of mobile receivers severe disturbances, for instance due to the high velocity of the moving receiver, might appear a check unit 58 is provided in the receiver 50 by which it is checked if decoding has been made correctly and/ or with sufficient quality and/or below a tolerable error level as will be explained below. If the decoding is made without errors or with sufficient quality the decoded data are provided to an output unit 60. The output thereof might be one or more receiver output data streams II' 12', In' which should as much as possible correspond to the transmitter input data streams II, 12, In.

[0077] If, however, the check by check unit 58 shows that a decoding is erroneous or that the decoded data have an insufficient quality and would, for instance, result in a noisy receiver output signal (e.g. insufficient picture quality of a movie), a feedback loop 62 is provided from the check unit 58 to the demapper 54 and/ or the decoder 56 in order to make use of the auxiliary parity portion (generally, the auxiliary codeword portion) (completely or in part) for improving the quality of decoding. Hence, in such a situation the demapper 54 then also demaps the auxiliary parity portion (completely or in part) from the (optionally demodulated) receiver input data stream O'. By use of this additional redundancy the decoder 56 will then again decode the received codewords, but now applies a code having a lower code rate which thus has a higher robustness against the disturbances. Hence, there is a high likelihood that the decoding quality will be better than before. In some embodiments, a feedback loop 62 is also provided from the check unit 58 to the demodulator 53, e.g. if auxiliary parity portions are required by the decoder 56 from another receiver input data stream, e.g. from data received at a different antenna in a MIMO receiver or from another channel (e.g. using another frequency).

[0078] Thereafter, again a check can be made by the check unit 58 if the decoding has now been made error-free or with sufficient quality, and, if not, a still further part of the auxiliary parity portion can be used in another iteration of de- mapping and decoding. If, on the other hand, the complete auxiliary parity portion of a codeword has already been completely used for decoding the check can also be omitted and the decoded data can be outputted directly.

[0079] Similarly as for the encoder 14 of the transmitter 10 there exist various embodiments of the decoder 56 of the receiver 50. A first embodiment 561 of the decoder 56 is schematically depicted in Fig. 16. According to this embodiment the decoder 561 comprises a first decoding unit 70 and a second decoding unit 80, similarly as the embodiment of the encoder 141 depicted in Fig. 2. Further, an additional decoder 90 (e.g. a BCH decoder) is provided for BCH decoding the output of the decoder 561, if the corresponding transmitter used in the broadcasting system applies a step of BCH encoding. The first decoding unit 70 generally corresponds to the decoding unit as used in receivers in accordance with the DVB-T2 standard. It comprises a time deinterleaver 71, a cell deinterleaver 72, a cyclic delay remover 73, a constellation demapper 74, a bit deinterleaver 75 and a first LDPC decoding block 76. Such a decoder is, for instance, explained in more detail in the DVB document A133, February 2009 "Implementation Guidelines for a Second Generation Digital Terrestrial Television Broadcasting System (DVB-T2)" which is herein incorporated by reference. This first decoding unit 70 thus is provided with the data portions D' and the basic parity portions Pb' (generally the basic codeword portion B') as demapped by the demapper 54 from a received receiver input data stream O' and decodes, as usual, these codewords.

[0080] In addition, in this embodiment of the decoder 561 a second decoding unit 80 is provided which basically comprises the same elements, in particular a time deinterleaver 81, a cell deinterleaver 82, a cyclic delay remover 83, a constellation demapper 84, a bit deinterleaver 85 and a second decoding block 86, whose function is identical to the function of the respective elements of the first decoding unit 70. However, the parameters of those blocks may differ, if different parameters are applied in the encoder, e.g. in the second encoding unit 30 (see Fig. 2). The second decoding unit 80 is, however, provided, in addition to the data portions D' and the basic parity portions Pb', with the additional auxiliary parity portion Pa' (generally the auxiliary codeword portion A') (completely or in part) for decoding the codewords with high reliability using said auxiliary parity portion Pa' as redundancy information. Hence, the second decoding unit 80 only becomes active in case of need, i.e. if "instructed" by the check unit 58 through feedback loop 62. Alternatively, the second decoding unit 80 is provided with only the auxiliary codeword portion A'.

[0081] An alternative embodiment 562 of a decoder is depicted in Fig. 17. According to this embodiment the demapper 54 accesses the T2 frames of a received receiver data input stream O' to demap the data portions D' and the basic parity portions Pb' therefrom and accesses FEF frames to demap the auxiliary parity portions Pa' therefrom, i.e. the demapper 54 is adapted to interrelate with the mapper 16 as illustrated in Fig. 9. The decoder 562, however, only comprises a single decoding unit which is both able to decode the codewords according to the first code (with a higher code rate) based on only the data portion D' and the basic parity portion Pb' and according to the second code (with a lower code rate) using as redundancy information (part of or all of) the auxiliary parity portion Pa', if needed.

[0082] A standard LDPC decoder as, for instance, provided in a DVB-T2 , DVB-S2, combined DVB-T2/S2, DVB-C2 or DVB-NGH receiver accepts at its input a (channel disturbed) codeword, as well as signalling information about the code rate and the length of the codewords (either 16200 or 64800 bits). Based on the signalling information, it applies an appropriate decoding algorithm (typically so called iterative message passing) based on that particular code realization and outputs an estimate of the data portion. [0083] The same applies for an extended LDPC decoder included in the decoder 56, in particular the extended LDPC decoder 76 and 86, which receive in addition auxiliary codeword portion, in particular auxiliary parity bits. The number of auxiliary bits is generally signalled in addition to the applied code rate and the length of the basic codeword portion to the decoder. Given these parameters, the decoder 56 applies an appropriate decoding algorithm based on this extended (or "total") code.

[0084] As mentioned above, in a preferred embodiment the demapper and the decoder are preferably adapted to stepwise add more parts ("sub-portion") of an auxiliary parity portion to improve the decoding. Preferably, once a sufficient decoding quality has been reached the demapper and the decoder are adapted for going into a sleeping mode while other parts of the (segmented; see for instance Fig. 13) auxiliary sub-portions are provided in a received receiver input data stream, i.e. those additional segments of an auxiliary parity sub-portion are preferably not demapped and used for decoding. This particularly saves power in the receiver which is especially advantageous in case of mobile receivers using a battery as power source.

[0085] A generic LDPC decoder has as input a received codeword (with or without additional parity bits) as well as signalling information about the code rate and the codeword length. The latter is in addition implicitly updated by the feedback loop 62, which signals, if (and how many) auxiliary parity bits are appended. In contrast to such a generic decoder, the LDPC decoder in the decoder 56 according to the present invention outputs its estimates on each code bit of the (received) codeword C, i.e. an estimate C" on the codeword C. Preferably, the bitwise estimates are expressed in terms of log likelihood ratios (LLRs), whose magnitude reflects the reliability of the estimate.

[0086] If the check unit 58 decides that the estimate C" is probably the transmitted codeword C, it outputs the (hard decided) estimates of the data portion D' and sets a flag S to 1, which corresponds to a decoding success. Otherwise, S = 0, which is signalled within the feedback loop 62 to initiate the suffixing of auxiliary parity bits (if still available). The indicator E from the check unit 58 is optional and gives an estimate of how many additional auxiliary parity sub-portions are still needed. In case of E > 1, the LDPC decoder 56 does not even have to try to decode the next larger codeword, but has to wait for E additional auxiliary parity sub- portions to restart decoding.

[0087] The criteria for decoding success (S=l) are:

a) Within a maximum number of allowed decoder processing steps (typically a maximum number of iterations is imposed), a valid codeword C" (after hard decision) is found.

b) The estimate of the data portion D' (can be derived from the estimate C", or is even included in C", in case of a systematic code (as in DVB-T2)) can be decoded by a BCH decoder. Note that the BCH decoder has also some error detection capabilities.

c) After BCH decoding, the stream II' should correspond to a BBFrame, whose header (BBHeader) is protected by a CRC. If this check is successful, the likelihood that the whole BBFrame is correct, is increased.

d) Preferably, the reliabilities of all LLRs (log-likelihood ratios) are checked by check unit 58. This could be done by averaging the magnitudes of all LLRs belonging to the codeword. If this average is larger than a certain threshold (which depends on the code and has to be defined), a decoding success is very likely.

[0088] If S=0, the last criterion (d) can also offer an estimate of how unreliable the codeword (after decoding) is. Assuming that the following auxiliary code portions have a similar quality than the previous codeword, an estimate E can be made about how many additional portions are needed for successful decoding.

[0089] It should be noted that there a two ways of combining the previous codeword (which the decoder wasn't able to decode correctly) with the auxiliary code word portions: 1) store the previous codeword, which entered the decoder, and append the auxiliary codeword portion at its end or

2) store the final estimate C" of the LDPC decoder 56 (e.g. after the maximum number of iterations has passed) and append the auxiliary codeword portion at its end.

[0090] In addition to the embodiments explained above the encoder of the transmitter may also be adapted such that the auxiliary parity portion (generally, the auxiliary codeword portion) may not (only) comprise "real" parity information, but that it may also comprise a repetition of (part of or all) of the information of the "basic" codeword, i.e. (some or all) bits of the data portion D and/or the basic parity portion Pb (i.e. of the basic codeword portion). Hence, in a very simple embodiment, the auxiliary parity portion Pa simply comprises a copy of the data portion D and/or the basic parity portion Pb. This will also improve decoding if the basic codeword is disturbed but the auxiliary parity portion is not (or less) disturbed. Further, even if both the basic codeword portion and the auxiliary codeword portion are disturbed, by use of both portions for decoding the result of the decoding may be improved, e.g. by applying the principle of soft combining, e.g. by improving the soft values obtained in a first decoding step using only the basic codeword portion in a second decoding step using in addition the auxiliary codeword portion.

[0091] An embodiment of a broadcasting system in accordance with the present invention is schematically depicted in Fig. 18. The broadcasting system generally comprises a transmitter (Tx) 10 and one or more receivers (Rx) 50a, 50b, 50c. While it is generally sufficient that the transmitter 10 has a single antenna for transmitting the transmitter output data stream O, here in this embodiment the transmitter 10 is provided with two antennas 19a, 19b.

[0092] In a first mode both antennas can be used for simultaneously transmitting the identical transmitter output data stream O (or a modified stream thereof, e.g. in accordance with the Alamouti scheme of the DVB-T2 standard), for instance to increase coverage.

[0093] In another mode, that is specifically depicted in Fig. 18, the data mapper 161 provided in this embodiment of the transmitter 10 is adapted for generating two transmitter output data stream Ol and 02, wherein the data portions D and the basic parity portions Pb of codewords (i.e. the basic codeword portions) are mapped onto the first transmitter output data stream Ol and wherein the auxiliary parity portions Pa of codewords (i.e. the auxiliary codeword portions) are mapped onto the second transmitter output data stream 02. In this embodiment the first antenna 19a can then be provided with the first transmitter output data stream Ol for transmission and the second antenna 19b can be provided with the second transmitter output data stream 02 for transmission. For instance, during the transmission of T2 frames only the first transmitter output data stream Ol is transmitted, while during the transmission of FEFs both transmitter output data streams Ol and 02 are transmitted.

[0094] In still another mode, the first transmitter output data stream Ol may be transmitted by a horizontally polarized antenna, whereas the second transmitter output data stream 02 my be transmitted by a vertically polarized antenna, or vice versa.

[0095] Optionally, a MIMO precoder 162 is provided to which said first and second transmitter output data streams Ol, 02 are provided from the data mapper 161 for precoding them according to any MIMO precoding scheme. For instance, the first and second transmitter output data streams Ol, 02 can be spatially multiplexed onto the precoded transmitter output data streams Ol*, 02* which are then transmitted by the antennas 19a, 19b, or Alamouti precoding can be applied to the first and second transmitter output data streams Ol, 02. The precoded transmitter output data streams Ol*, 02* may then both contain a mix of data from the first and second transmitter output data streams Ol, 02. [0096] In this embodiment illustrated in Fig. 18 the transmitter 10 is provided with two antennas 19a, 19b. It should be noted, however, that the transmitter, in particular a MIMO transmitter, comprises more than two antennas to which said precoded transmitter output data streams Ol*, 02* are provided for transmission.

[0097] A first receiver 50a having a single antenna 61 may be adapted for receiving only the first transmitter output data stream Ol (as first receiver input data stream Ol'), but not the second transmitter output data stream 02. Such a receiver 50a might be an existing, e.g. legacy or stationary, receiver that is not adapted for using any auxiliary parity portions at all. For instance, if the transmitter 10, particularly the second transmitter output data stream 02, is directed to reception by mobile receivers in accordance with a new standard, e.g. the DVB-NGH standard, the receiver 50a could be a stationary receiver in accordance with the DVB-T2 standard.

[0098] Another embodiment of the receiver 50b comprises two antennas 61a, 61b. In this embodiment the first antenna 61a is adapted for reception of the first transmitter output data stream Ol (as first receiver input data stream Ol'), and the second antenna 61b is adapted for reception of the second transmitter output data stream 02 (as second receiver input data stream 02'). For instance, if the two antennas 19a, 19b of the transmitter 10 make use of different transmission channels, e.g. transmission frequencies, the two antennas 61a, 61b of the receiver 50b can be adapted for reception on the same respective transmission channel.

[0099] A third embodiment of a receiver 50c again has a single antenna 61, but is adapted for reception of the signals from both antennas 19a, 19b. The receiver 50c comprises means for internally splitting up or decomposing the two received input data streams Ol', 02' accordingly.

[00100] The embodiment of the receiver 50b, having two separate antennas 61a, 61b for receiving the different transmitter output data streams Ol, 02 separately, provides the advantage that the second antenna 61b and the subsequent processing means within the receiver 50b need only be activated if any auxiliary parity portions are required as incremental redundancies for improvement of the decoding. This holds also for an embodiment of a transmitter, where the second transmitter output data stream 02 does not only carry the auxiliary parity portions, but also the data portions and the basic parity portions of the codewords. In the latter case the transmission can be even made more stable. For instance, if the transmission channel between the transmitter antenna 19a and the receiver antenna 61a is disturbed, it can be switched to the other transmission channel between transmitter antenna 19b and the receiver antenna 61b. The advantage of a more stable transmission is also achieved with the receiver 50c which, in the latter case, can switch between reception of the first or second transmitter output data streams Ol, 02 or which continuously receives both transmitter output data streams Ol, 02. Further, such an embodiment generally also provides an increased spectral density.

[00101] In the above, particularly with respect to the transmitter 10, various embodiments have been illustrated, particularly how the data portions, the parity portions and the auxiliary parity portions are mapped onto the transmitter output data stream. Further, various examples have been given regarding the framing structure of the transmitter output data stream. It shall be understood that the data demapper 54 of the receiver 50 is, of course, adapted appropriately for demapping the required data from the receiver data input stream, i.e. the demapper 54 is aware of the particular framing structure and/or the locations at which the respective data are placed in the receiver data input stream. Known measures for signalling this information to the receiver from the transmitter and/or for prescribing this information, for instance in a standard, and for enabling the transmitters and receivers accordingly are generally applied to ensure this.

[00102] The framing structure applied according to the present invention may generally be adapted to be in consistence with the framing structure according to an existing standard, e.g. the DVB-T2 standard, so that existing receivers in accordance with this standard may also receive and process such data streams, even if they do not make use of the auxiliary parity information contained therein as incremental redundancy. However, the framing structure may be freely selected and newly created according to specific needs of the broadcasting system.

[00103] Generally, the present invention can be applied in all broadcasting systems, by which data are transmitted over a channel. For instance, the invention can be applied in a DAB system which shall be explained with reference to Fig. 20.

[00104] Fig. 20 shows the structure of a transmission frame as described in the DAB standard (ETS 300 401 "Radio broadcasting systems; Digital Audio Broadcasting (DAB) to mobile, portable and fixed receivers", May 1997, RE/JFT-OODAB-4). The DAB transmission system combines three channels, in particular a Synchronization Channel which is used internally with the transmission system for basic demodulator functions (e.g. transmission frame synchronization), a Fast Information Channel (FIC) which is used for rapid access of information by a receiver, which is a non-time-interleaved data channel and which can be sub-divided into Fast Information Blocks (FIBs), and a Main Service Channel (MSC) which is used to carry audio and data service components and which is a time-interleaved data channel divided into a number of sub-channels, which are individually convolutionally coded.

[00105] The MSC can also be seen as being made up of Common Interleaved Frames (CIFs), comprising Capacity Units (CU) as smallest addressable units. Each sub-channel of the MSC occupies an integral number of consecutive CUs and is individually convolutionally encoded. More details regarding the structure of the transmission frame and its content can be found in the above cited DAB standards, which explanations are herein incorporated by reference.

[00106] According to the present invention one of the sub-channels, e.g. SubCh a could comprise the basic codeword version, whereas one or more of the subsequent sub-channels, e.g. SubCh b, comprises the auxiliary codeword portion. A receiver can now process SubCh a and, if necessary, SubCh b for improving the decoding. As illustrated in an embodiment shown above the auxiliary codeword portion can be further segmented into sub-portions all carried in the same subchannel or carried in various sub-channels. This again has the advantage that the receiver can fall into sleeping mode after successful decoding until the next basic codeword portion is transmitted.

[00107] If and for which sub-channels auxiliary parity bits are provided, could be signaled in the Fast Information Channel (FIC). Since this channel is, however, fixed and predefined, the signalling should preferably be done in another sub-channel which comprises the basic codeword portion, for instance in a newly defined header. Hence, receivers modified in accordance with the present invention can make use of this additional information. In addition, the FIC could signal, which sub-channel is adapted for reception by all DAB receivers (legacy receivers and receivers according to the present invention) and which sub-channels are decoded for reception by receivers according to the present invention (only).

[00108] The error correction code applied in DAB is a convolution code. Different code rates are generally achieved according to DAB by puncturing of a mother code. This mother code generally has a code rate of 1/4, and by puncturing of certain parity bits higher code rates are obtained. These punctured parity bits could be used as auxiliary parity bits for providing incremental redundancy in accordance with the present invention. Alternatively, a completely new mother code is also applicable, from which all the DAB code rates can be obtained by puncturing and where the punctured bits are used as auxiliary parity bits in accordance with the present invention.

[00109] Fig. 21 shows a block diagram of another embodiment of a receiver 50d. Generally, the basic codeword portions B and the auxiliary codeword portions A are mapped by the transmitter onto the transmitter output data stream O such that a basic codeword portion of a codeword is received by a receiver before the corresponding auxiliary codeword portion is received. Alternatively, however, the data mapper can also be adapted for mapping the basic codeword portions B and the auxiliary codeword portions A onto frames of a transmitter output data stream such that the auxiliary codeword portion of a codeword is received by a receiver before the corresponding basic codeword portion is received. For the embodiment of the receiver 50d shown in Fig. 21 it shall be assumed that the transmitter is adapted in this way.

[00110] In such a receiver 50d the data demapper 54d is thus adapted to de- map the (first received) auxiliary codeword portions A' from the receiver input data stream O' and forward them to a buffer 64. Thereafter (whenever received) the corresponding basic codeword portions B' are demapped and forwarded to the decoder 56d for decoding them. If the check in the check unit 58d shows that additional redundancies shall be used for improved decoding the buffer 64 is informed via the feedback loop 62 to provide the buffered auxiliary codeword portion A' to the decoder 56d and the decoder is informed to then decode the codeword (now by additional use of the (complete or partial) auxiliary codeword portion again. If it is clear that the decoding of the corresponding codeword is correct the buffered auxiliary codeword portion (if any) is deleted from the buffer.

[00111] This embodiment provides the advantage that no waiting times (for waiting for the auxiliary parity portion, if the decoding was erroneous based on the basic codeword portion) occur, which is particularly important for reducing zapping times or for mobile receivers. Hence, this embodiment also provides the advantage that no interruption of the service occurs (due to waiting for reception of auxiliary codeword portions) in case of (e.g. sudden) bad reception conditions of the basic codeword portions.

[00112] Now, with reference to Figs. 5, 6, 19 and 22 preferred embodiments of an encoder according to the present invention shall be explained.

[00113] As explained above, by the application of incremental redundancy (IR) certain existing LDPC codes of FECFRAME length N_ldpc (e.g. = 16 200) are extended, such that the new codeword consists of the original codeword (basic FEC) and M_IR (also called v above) additional (auxiliary) parity bits to be used as IR. The new codeword length is thus N_ldpc<1 = N_ldpc + M_IR. The LDPC encoding with IR can be thought of as one encoder of code rate Rj = K_ldpc/N_ldpc<1, where the output is split into a basic FEC ("basic codeword portion Pb") and an IR part ("auxiliary codeword portion Pa"). Both parts can be treated as two associated PLPs.

[00114] The relationship between original codeword and extended codeword can be seen in Figs. 5 and 6. Fig. 5 depicts conventional FEC encoding according to, e.g., DVB-T2, where the input to the LDPC encoder of code rate R₀ = K_ldpc/ N_ldpc are the K_ldpc output bits of the BCH encoder, while its output is a systematic codeword of length N_ldpc. The last N_ldpc - K_ldpc bits of this codeword are the LDPC parity bits. If a PLP applies IR, then an extended LDPC encoder of code rate Rj = K_ldpc/ N_ldpc<1 < R₀ is used, which has the same input as for conventional encoding, but outputs N_ldpc<1 = Ni_dpc + Mm bits, i.e., that amount of LDPC parity bits is increased to N_ldpc - K_ldpc+ M_IR. However, the first N_ldpc - K_ldpc parity bits thereof are identical to the parity bits of the original LDPC encoder of rate R₀. The codeword is split into two parts: the first N_ldpc bits are the basic FEC part (i.e. the basic parity portion), while the remaining M_IR bits are the IR part (i.e. the auxiliary parity portion) to be used as IR at the receiver, if necessary.

[00115] Thus, it is ensured that decoding of the received codeword is possible (for good channel conditions) with a rate R₀ decoder, which considers just the basic FEC part, while the extended codeword - consisting of both basic FEC and IR part, allows for decoding with a rate Rj decoder.

[00116] The separation of one FECFRAME, which uses IR, into a basic FEC and an IR part is shown in Fig. 22 showing only the essential elements of an embodiment 145 of an encoder 14 according to this aspect of the present invention. This encoder 145 comprises an encoder input 1451 for receiving input data words D each comprising a first number K_Upc of information symbols ϊ₀ , ϊ_λ ,...,ϊ_{κ λ} (above also called s_1; s₂, s_k, e.g. in Fig. 3). [00117] Further, an encoding block 1452 is provided for encoding an input data word ) into a codeword Z2 such that a codeword comprises a basic codeword portion B including a data portion D and a basic parity portion Pb of a second number N_ldpc - K_ldpc of basic parity symbols ρ₀ , ρ_λ ,---,ρ_{Ν κ λ} , and an auxiliary codeword portion A including an auxiliary parity portion Pa of a third number M_IR of auxiliary parity symbols (Ρ^-_Κ^ , -, Ρ_Ν^-_Κ^-_Ι ) - Said encoding unit 1451 is adapted i) for generating said basic codeword portion B from an input data word D according to a first code, wherein a basic parity symbol is generated by accumulating an information symbol at a parity symbol address determined according to a first address generation rule, and ii) for generating said auxiliary codeword portion A from an input data word D according to a second code, wherein an auxiliary parity symbol is generated by accumulating an information symbol at a parity symbol address determined according to a second address generation rule. For these generations address tables are preferably used that are stored in an address table storage 1453.

[00118] Further, the encoder 145 comprises an encoder output 1454 for out- putting said codewords, which is implemented here as a series-to-parallel converter for separating the basic codeword portion B and the auxiliary codeword portion A for subsequent independent processing. Of course, the output unit may be a simple serial output for outputting the complete codeword Z2 as it is.

[00119] The encoded bits of both streams are further processed by independent instances of bit interleavers, demultiplexer parts, and so on, as depicted in Fig. 2 for a particular PLP. Thus, it is possible to increase the robustness of the basic FEC part by applying, e.g., low order QAM constellations for this part. It shall be noted here that according to this embodiment of the present invention preferably only a single encoding unit (e.g. like the encoding unit 20 shown in Fig. 2) is provided. For instance the encoding block 1452 can implement both the FEC encoding blocks 21 and 31, whereafter a single common path with subsequent processing elements or two separate paths of subsequent processing elements (as shown in Fig. 2) follow. [00120] A scheduler will allocate the bursts of the basic FEC part at earlier time instances than the bursts of the corresponding IR part. If the receiver wants to decode a PLP that uses IR, it has to demodulate at least the basic FEC part, which corresponds to the output of the original encoder of rate R₀ = K_ldpc/ N_ldpc > R_x. If decoding fails, due to inappropriate channel conditions (SNR falls below threshold of the original code), it can, in addition, demodulate the IR part, which - together with the basic FEC part, builds the codeword of length N_ldpc<1. However, the decoder then switches to a new parity check matrix according to the code of (smaller) code rate R_1; where successful decoding is more likely, as the decoding threshold of the extended code is much smaller than that of the original code.

[00121] It should be noted that the main advantage of IR (compared to applying a low code rate Rj without IR in the first place) is that the IR part can be ignored by the receiver unless it is needed. If, e.g., Rj = 1/2*R₀, a conventional approach would apply the rate Rj code, which limits the throughput of the data by a factor of 2. So, twice as many bursts have to be detected by a receiver for a given data rate, compared to the case, when, due to favourable channel conditions, transmission would be possible with a rate R₀ code. Applying IR, however, allows the receiver to fall into sleep-mode, whenever IR bursts are transmitted, if decoding of the basic FEC part is possible and/or was successful.

[00122] The amount of IR bits shall be as large as the FECFRAME length itself, thus: M_IR = N_ldpc, which results in halving of the original code rate, Rj = 1/2*R₀. In a practical implementation, e.g. for use in mobile receivers according to the upcoming DVB-NGH standard, IR is used for short codes (N_ldpc = 16 200) for the following code identifiers: R₀ e{l/2, 3/5, 2/3, 3/4, 4/5, 5/6, 8/9}.

[00123] For each LDPC code of length N_ldpc = 16 200, in the following denoted as original code of rate R₀, an extended code of rate Rj = 1/2*R₀ is derived in the following. The extended LDPC encoder treats the output of the outer BCH encoding, 1= (i₀, z^' _l .., i_K i ) , as an information block of size K_ldpc = N_{BCH I} and systematically encodes it onto a codeword of size N_{ldpc l} =N_ldpc + M_IR , where:

The following encoding procedure ensures that the first bits of the extended codeword X_t, for i e {0,..., N_ldpc - K_ldpc -1} are the same as if the original LDPC code would have been used.

[00124] The task of the encoder is to determine N_{ldpc l} - K_ldpc parity bits

(_Po, p„...,p_{Nidpc i}__Kidpc_,) for every block of K_ldpc information bits, (i₀, i^) . In other word, the connections shown in Fig. 19 are determined by the encoder. The procedure is as follows:

. Initialize p₀ = _Pl = p₂ = ... = p_Nldpc ^_Kldp^ = 0

• Accumulate the first information bit, z₀ , at parity bit addresses specified in the first row in Figs. 23 to 29. For example, for rate 4/5 (see Figure 27), (all additions are in GF(2)):

Ps = p₅ ® i₀ Pl092 ^~ Αθ92 ® i_Q P Ί3423 ^~ 'l3423 Θ

P&96 = Ps96 ^φ *0 Θ

AS65 = As65 © ^Z0 P

rm ^~ Prm ® i₀ ^~ Θ

^3976 ^— ^3976 ® Ό P99&9 ^~ P99&9 ® i₀ A 8042 ^~ A 8042 Θ

Psm = Al77 ® 0690 ^— A 0690 ® i₀ A 8646 ^~ A 8646 Θ

ΡδδΊΟ = A₅70 ® A 1608 ^~ A i608 A 8943 ^~ A 8943 Θ

Ρό450 = Ρβ450 ^φ Pl 2337 ^~~ ^" 2337 Θι₀ It should be noted that the first three parity addresses (all smaller than N_ldpc - K_ldpc = 3240 ) are the same as defined in DVB-S2 for the 16k code of rate 4/5, while the remaining addresses are according to the extended LDPC code and are written in bold numbers in the address tables depicted in Figs. 23 to 29.

• For the next 359 information bits, i_m, m = 1, 2, 359 accumulate i_m at parity bit addresses

{x + _OT mod360 x ¾J mod(N^_c - **J if * < ¾_c (which is the first address generation rule) or

N_ldpc - K_ldpc + {x + m mod 360 x Q_IR } mod M_IR if x≥ N_ldpc - K_ldpc

(which is the second address generation rule), where x denotes the address of the parity bit accumulator corresponding to the first bit i₀ , Q_ldpc is a code rate dependent constant specified for the original LDPC codes in the DVB-T2 and DVB-S2 standards and Qm = M_IR/360 = 45.

[00125] It should be noted that the size (also called G_a) of the group of information bits that are dealt with blockwise, can also be different from 360 and can also be different for the first and second address generation rules, but K_ldpc should always be a multiple of G. The code rate dependent parameter Qj_dpc is preferably selected from the following table

Code Rate Identifier Q_jdpc

1/2 25

3/5 18

2/3 15

3/4 12

4/5 10 5/6 8

8/9 5.

The relation between G_b and Qj_dpc is as follows: Qj_dpc =( N_ldpc - K_ldpc)/G_h and the relation between G_a and Qm is: Qj_R = M_IR /G_a. If M_IR= JV_ldpc = 16 200 and G_a = 360, then ^=45.

[00126] The positions x, which are larger than or equal to N_ldpc - K_ldpc are written in bold numbers in the address tables shown in Figs. 23 to 29. Each of said address tables is designed for a particular code rate, the corresponding value for Qj_dpc selected from the above table and the values of Qj_R = 45, M_IR = 16200, N_ldpc = 16200.

[00127] It should be noted that this approach leaves the basic FEC as the original code, adds M_IR more parity checks, and maintains the quasi-cyclic LDPC structure for both basic and IR part. However, the quasi-cyclic structure is interrupted after the first N_ldpc - K_ldpcrows of the parity check matrix. But as this is a multiple of 360

(namely Q_Upc ), blockwise decoding based on 360 bit groups can still be applied.

[00128] In a similar manner, for every group of 360 new information bits, a new row from one of the above address tables (i.e. the address table of the desired data rated) are used to find the addresses of the parity bit accumulators.

[00129] After all of the information bits are exhausted, the final parity bits are obtained as follows:

• Sequentially perform the following operations starting with i = 1 .

Pi = Pi © Pi-i , i = ½ ..., ¾_c,i - K_ldpc - 1

• Final content of p_t, i = X..,N_{ldpc l} - K_ldpc - 1 is equal to the parity bit p_i .

[00130] An example of the extended LDPC code shall now be briefly explained. The bit and frame error rates (BER and FER) over SNR performance of an extended LDPC code over the (non-fading) AWGN channel is depicted in Fig. 30. The original code is the rate R₀ = 4/5 code of DVB-S2 of length N_ldpc = 16200, the extended code has code rate Rj = 2/5. The decoding threshold (or pinch-off limit) of the original 4/5 code from DVB-S2 is at about 1.7 dB. The rate 2/5 code from DVB-S2 of the same length, N_ldpc = 16200, has its limit at -3.4 dB. The extended LDPC code of the same code rate

has its threshold at -2.9 dB. The number of simulated information bits was 10⁹. While the rate 4/5 code of S2 shows some error floor behaviour after the decoding threshold, no errors could be detected for the other two codes in the simulated 10⁹ bits.

[00131] If the SNR at the receiver exceeds 1.7 dB, it can decode the original rate 4/5 code, which also corresponds to the basic FEC part of the extended LDPC code. In this case, the IR part could be neglected, which yields a reduction in processing power by a factor of 2 (as half of the bursts belonging to this PLP are from the IR part). If the SNR decreases, the receiver can demodulate the IR part and thus has a margin of additional 4.6 dB. In terms of coding gain (difference in E_b/N₀) this corresponds to 1.6 dB gain, as the code rate is halved.

[00132] It shall be noted that it is also possible to design new LDPC codes, which can be partitioned into a basic FEC and an IR part. As the FEC part of these newly designed codes does not have to match the existing (standardized) LDPC codes of the DVB family, the resulting coding gains of the IR part will be even larger than those codes proposed according to the present invention. Furthermore, it is also possible to extend LDPC codes with a different amount of additional parity bits, M_IR, or to allow for more than one IR part. The partitioning into several IR parts would enable the receiver to estimate the amount of additional parity portions, which are required for successful decoding.

[00133] The additional redundancies generated according to the present invention can also be used in another scenario to enable a mobile receiver, for instance an in-car receiver or a handheld receiver (e.g. in a mobile phone or a PDA) in a broadcast system to decode broadcast data even under severe transmission channel conditions by improving the provided error correction measures. In particular, it is proposed to provide a sufficient amount of redundancy on request of the receiver to increase the robustness of the code. Said additional redundancy is provided by the data transmission system after a feedback from the receiver is received requesting additional redundancy. Said additional redundancy is, however, not broadcasted over the broadcast system, but is transmitted over a unicast system to the requesting receiver only. This receiver can use the additional redundancy to perform another decoding of previously (through the broadcast system) received codeword. Hence, if reception or reconstruction (decoding) of received broadcast data is erroneous or could only be made with insufficient quality, the receiver can repeat decoding using, in addition to the previously received codeword, the requested additional redundancy.

[00134] Such a receiver, e.g. in accordance with the upcoming DVB-NGH standard, can, for instance, be included in a mobile phone, which is also capable of receiving data from unicasting networks, like mobile communications systems, e.g. a 3G (UMTS) or 4G (LTE) communications system, as well as WLAN (wireless local area network), if access points are in nearby range. Additional redundancy for erroneously received or decoded codewords (the term "erroneously" being not only understood as meaning completely erroneously, but also "with insufficient quality") can be retrieved from a different architecture (vertical handover) according to the present invention, e.g. via the 3G, 4G or WLAN network.

[00135] Thus, generally it is not essential for such a scenario in accordance with which particular unicast system the unicast request unit and the unicast receiving unit are implemented. Generally, any unicast system may be used, e.g. any (tele) communications system for wireless communication, and it is also possible that the receiver may be implemented to use several unicast systems for the proposed request and reception of additional redundancy, e.g. through the unicast system that is currently available in the respective situation. Further, the request and the reception of the additional redundancy can generally also be performed through different unicast systems, but preferably the same unicast system will be used. Since, generally, unicasting systems provide sufficient means for error correction and detection, it can be assumed that the transmission of the auxiliary codeword portion is error-free.

[00136] Fig. 31 shows a schematic block diagram of a data transmission system in accordance with such a scenario. The system comprises a broadcast transmitter 10 for broadcasting data, a receiver 50 for receiving data broadcast by said broadcast transmitter 10, a unicast transmitter 100 for transmitting data in a unicast system, and a data storage 150 for storing data received from the broadcast transmitter 10. As will be explained in more details below, the receiver 50 is not only able to receive data broadcast by said broadcast transmitter 10, but is - to a certain extent - also able to communicate with the unicast transmitter 100 to have a bidirectional communication, and also the unicast transmitter 100 is able to have a bidirectional communication with the data storage 150. In particular embodiments, an additional auxiliary encoder 160 is further provided for encoding data before transmission to the unicast transmitter 100. In the following the various elements of the data transmission system will be explained separately to illustrate their functionalities and interrelations as proposed according to the present invention.

[00137] In this data transmission system the broadcast transmitter 10 and the receiver 50 as well the link 200 in between are part of a broadcast system, such as a video broadcast system in accordance with any DVB standard, particularly a wireless broadcast system. The unicast transmitter 100 and the receiver 50 as well as the link 300 in between are part of a unicast system, such as a communications system in accordance with any communications system standard, particularly a wireless communications system.

[00138] The link 400 between the broadcast transmitter 10 and the data storage 150 can be part of the broadcast system so that the data to be stored in the data storage 150 are obtained by the data storage 150 from the broadcast signal. Alternatively, this link 400 can also be established by a separate transmission channel, e.g. a wired or wireless transmission link, from the broadcast transmitter 10 to the data storage 150, which can, for instance, be a server arranged next to the broadcast transmitter 10.

[00139] The link 500 between the unicast transmitter 100 and the data storage 150 as well as the link 600 between the auxiliary encoder 160 and the unicast transmitter 100 can be part of the unicast system so that the unicast transmitter 100 communicates with the data storage 150 and the auxiliary encoder 160 via the same unicast system as is used for communicating with the receiver 50. The links 500, 600 and/ or 700 between the data storage 150, the auxiliary encoder 160 and/ or the unicast transmitter 100 can be established by any transmission means. Preferably, the auxiliary encoder 160 is arranged in close proximity to the data storage 150 so that a wired transmission line is the advantageous solution. The auxiliary encoder 160 might, however, also be part of the unicast system or even part of the unicast transmitter 100.

[00140] Fig. 32 shows a schematic block diagram of a receiver 50 for use in a data transmission system as illustrated in Fig. 31. The receiver 50 is particularly adapted for requesting an auxiliary parity portion (generally, the auxiliary codeword portion) as incremental redundancy in case of erroneous or low quality decoding.

[00141] The receiver 50 comprises a data input 52 for receiving a receiver input data stream O' which generally corresponds to a transmitter output data stream O that has been transmitted over a broadcast channel of the broadcast system by a transmitter 10 and which may thus be affected by disturbances that may appear in such a broadcast system, particularly in case of using mobile receivers which are the main application of the present invention on the receiver side.

[00142] Optionally, a demodulator 53 is provided that is interrelated with an (optional) modulator of the transmitter 10 for demodulating the received receiver data input stream O'. A demapper 54 demaps the (optionally demodulated) receiver data input stream O', particularly at least the data portions and the basic parity portions (i.e. the basic codeword portions) of the codewords mapped into the receiver data input stream O' as will be explained below in more detail. A decoder 56 then decodes these codewords by use of the basic codeword portions according to the same code as applied by the encoder of the transmitter 10. As particularly in case of mobile receivers severe disturbances, for instance due to the high velocity of the moving receiver, might appear a check unit 58 is provided in the receiver 50 by which it is checked if decoding has been made correctly and/ or with sufficient quality and/or below a tolerable error level as will be explained below. If the decoding is made without errors or with sufficient quality the decoded data are provided to an output unit 60. The output thereof might be one or more receiver output data streams II' 12', In' which should as much as possible correspond to the transmitter input data streams II, 12, In. If, however, the check by check unit 58 shows that a decoding is erroneous or that the decoded data have an insufficient quality and would, for instance, result in a noisy receiver output signal (e.g. insufficient picture quality of a movie), a side loop 62 is provided from the check unit 58 to a unicast request unit 64 in order to request the auxiliary parity portion (generally, the auxiliary codeword portion) (completely or in part) for improving the quality of decoding.

[00143] Hence, in such a situation the unicast request unit 64 transmits a request R via the unicast system, i.e. via the unicast transmitter 100 of the data transmission system, to obtain the corresponding auxiliary parity portion (completely or in part) for the erroneously decoded codeword from the data storage 150. The request R includes at least the identification information of the erroneous codeword. In the meantime the decoding result C" of decoding the codeword (or the codeword C itself) is stored in a buffer (generally, a data memory unit) 66. The requested auxiliary codeword portion A' (or at least a portion thereof or another auxiliary parity data that can be used as incremental redundancy in an additional decoding step, i.e. which have been generated by use of the same code) is received by a unicast receiver 68. By use of this additional redundancy and the data stored in the buffer 66, a recombiner 69 will recombine these data so that the decoder 56 will then again decode the received codewords, but now applies a code having a lower code rate which thus has a higher robustness against the disturbances. Hence, there is a high likelihood that the decoding quality will be better than before.

[00144] Thereafter, again a check can be made by the check unit 58 if the decoding has now been made error-free or with sufficient quality, and, if not, a still further part of the auxiliary parity portion can be requested and used in another iteration of decoding. If, on the other hand, the complete auxiliary parity portion of a codeword has already been completely used for decoding the check can also be omitted and the decoded data can be outputted directly.

[00145] Preferably, the request R transmitted from the unicast request unit 64 comprises only an indication which auxiliary codeword portion is requested. However, an estimate E from the check unit 58 may also be included in the request R indicating the amount of incremental redundancy that is required for enabling better decoding of the codeword. Hence, based on said estimate E, the complete auxiliary codeword portion needs not necessarily to be retrieved and transmitted via the unicast system, but only the amount as signalled by said estimate E needs to be transmitted, thus saving bandwidth and transmission time (and, if the auxiliary parity portion must be generated on the fly in an auxiliary encoder 160, saving also some time needed for encoding).

[00146] The present invention thus provides an effective and easily imple- mentable measure for improving the reliability of decoding, particularly for mobile receivers in a broadcasting system, without any feedback from a receiver to the transmitter. If the DVB-T2 framing structure is kept unchanged and the FEFs contain the additional redundancy, i.e. the auxiliary codeword portion, there are basically two aspects that make a (mobile) receiver, or generally any receiver exploiting the present invention, more robust for mobile reception: i) The incremental redundancy as described and ii) the FEFs themselves with the embedded incremental redundancy, which can (and typically will) select transmission (e.g. OFDM) parameters that have a better behaviour in mobile channels. The most important ones are lower FFT sizes and higher pilot pattern densities (being related to the FFT and guard interval sizes). Of course the incremental redundancy data in the FEFs can additionally be protected by using lower modulation schemes, other interleaving depths etc..

[00147] The time interleaver depth chosen in the FEFs could, for example, complement the time interleaver depth of the T2 frames. If the T2 frame time interleaver fails (e.g. signal interrupted for a dedicated time (e.g. due to a tunnel etc.)), the other settings in the FEF time interleaver might be better suited and allow overall correct decoding. Different time interleaver settings of the T2 frame and the FEF frame overall improve the system performance.

[00148] The receiver in accordance with the present invention benefits therefore from the fact that in addition to the basic T2 reception the data in the FEFs (i.e. additional incremental redundancy) is more robust in mobile channels. Another main advantage of preferred embodiments of the present invention is that broadcasters do not have to transmit the data for mobile, (e.g. NGH) receivers, but only incremental redundancy is transmitted to enable a more robust reception of T2 data even with a mobile receiver. Thus, transmission bandwidth is utilized most efficiently.

[00149] The invention has been illustrated and described in detail in the drawings and foregoing description, but such illustration and description are to be considered illustrative or exemplary and not restrictive. The invention is not limited to the disclosed embodiments. Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims.

[00150] In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality. A single element or other unit may fulfill the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different depend- ent claims does not indicate that a combination of these measures cannot be used to advantage.

[00151] A computer program may be stored / distributed on a suitable medium, such as an optical storage medium or a solid-state medium supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems.

[00152] Any reference signs in the claims should not be construed as limiting the scope.

Claims

1. Encoder for error correction code encoding input data words (D) into codewords (Zl, Z2), comprising:

an encoder input (1451) for receiving input data words (D) each comprising a first number K_Upc of information symbols,

an encoding means (1452) for encoding an input data word (D) into a codeword (Zl, Z2, Z3, Z4) such that a codeword comprises a basic codeword portion (B) including a data portion (D) and a basic parity portion (Pb) of a second number N_ldpc - K_ldpc of basic parity symbols, and an auxiliary codeword portion (A) including an auxiliary parity portion (Pa) of a third number M_IR of auxiliary parity symbols, wherein said encoding means (14) is adapted

i) for generating said basic codeword portion (B) from an input data word (D) according to a first code, wherein a basic parity symbol is generated by accumulating an information symbol at a parity symbol address determined according to a first address generation rule,

and

ii) for generating said auxiliary codeword portion (A) from an input data word (D) according to a second code, wherein an auxiliary parity symbol is generated by accumulating an information symbol m at a parity symbol address y, wherein said parity symbol addresses y are determined according to a second address generation rule

N_ldpc - K_ldpc + {x + m mod G_a x Q_1R } mod M_1R if x > N_ldpc - K_ldpc ,

an encoder output (1454) for outputting said codewords (Zl, Z2).

2. Encoder according to claim 1,

wherein said encoding means (1452) is adapted for generating a basic parity symbol by accumulating an information symbol m at a parity symbol address y, wherein said parity symbol addresses y are determined according to the first address generation rule

{x + m mod G_b x Q_ldpc } mod(N^_c - K_ldpc ) if x < N_ldpc - K_ldpc

wherein x denotes the addresses of a parity symbol accumulator corresponding to the first information symbol of a group of size G_b and Qj_dpc is a basic code rate dependent, predefined constant.

3. Encoder according to claim 1 or 2,

wherein G_a = G_b.

4. Encoder according to claim 3,

wherein G_a = G_b = 360.

5. Encoder according to any preceding claim,

wherein said encoding means (1452) is adapted for blockwise generating said basic parity symbols and said auxiliary parity symbols by use of a group of subsequent information symbols,

wherein each information symbol i of said group of subsequent information symbols is accumulated at a set of different parity symbol addresses y,

wherein the set of parity symbol addresses, at which the first information symbol of said group is accumulated, is taken from a predetermined address table and wherein the symbol addresses, at which the subsequent information symbols of said group are accumulated, are determined from said set of parity symbol addresses according to said first or said second address generation rule, respectively, and

wherein a separate set of parity symbol addresses is taken from said address table for generating each new block of basic parity symbols and auxiliary parity symbols.

6. Encoder according to claim 5,

wherein said encoding means (1452) is adapted for subsequently taking a new row of the following address table as a new set of different parity symbol addresses y for accumulating a new group of subsequent information symbols, said address table for Qj_R = 45, M_IR = 16200, N_ldpc = 16200, Qj_dpc = 25 and a code rate identifier of 1/2 being

7. Encoder according to claim 5,

wherein said encoding means (1452) is adapted for subsequently taking a new row of the following address table as a new set of different parity symbol addresses y for accumulating a new group of subsequent information symbols, said address table for Qj_R = 45, M_IR = 16200, N_ldpc = 16200, Qj_dpc = 18 and a code rate identifier of 3/5 being

71 1478 1901 2240 2649 2725 3592 3708 3965 4080 5733 6198

393 1384 1435 1878 2773 3182 3586 5465 6091 6110 6114 6327

160 1 149 1281 1526 1566 2129 2929 3095 3223 4250 4276 4612

289 1446 1602 2421 3559 3796 5590 5750 5763 6168 6271 6340 947 1227 2008 2020 2266 3365 3588 3867 4172 4250 4865 6290

3324 3704 4447 9541 11280 15341 18544

1206 2565 3089 9452 12449 16615 17605

529 4027 5891 7794 19046 19119 21536

141 1 187 3206 11442 13540 13717 14366

1990 2972 5120 14717 14754 19356 19725

752 796 5976 12613 15111 16917 19627

1 129 2377 4030 12161 15392 20670 21691

6077 6108 6231 9845 11360 15990 16564

61 1053 1781 9635 10213 14108 20157

2820 4109 5307 9245 10139 10168 13538

2088 5834 5988 9474 11519 13449 21438

3725 3945 4010 8280 13589 21139 22351

1081 2780 3389 10660 10727 13101 16117

659 2221 4822 8382 10072 16246 18001

3033 6060 6160 11286 11644 13351 14707

756 1489 2350 6953 7865 10732 19456

3350 3624 5470 14395 15852 18311 21527

357 1825 5242 10323 13913 15338 22082

585 3372 6062 10231 14400 14922 16589

561 1417 2348 12432 12887 17481 22485

971 3719 5567 7091 8079 19378 20819 21275

1005 1675 2062 8692 10722 11912 17491 18163

8. Encoder according to claim 5,

wherein said encoding means (1452) is adapted for subsequently taking a new row of the following address table as a new set of different parity symbol addresses y for accumulating a new group of subsequent information symbols, said address table for Qj_R = 45, M_IR = 16200, N_ldpc = 16200, Qj_dpc = 15 and a code rate identifier of 2/3 being

0 2084 1613 1548 1286 1460 3196 4297 2481 3369 3451 4620 2622

1 122 1516 3448 2880 1407 1847 3799 3529 373 971 4358 3108

2 259 3399 929 2650 864 3996 3833 107 5287 164 3125 2350 3 342 3529 7311 9608 13716 19002

4 4198 2147 8268 10733 15904 17839

5 1880 4836 14204 15736 16150 19640

6 3864 4910 5963 9085 10938 18686

7 243 1542 8274 12313 17889 20805

8 301 1 1436 12182 16130 16958 21462

9 2167 2512 6827 7272 14255 18304

10 4606 1003 6318 8133 13289 18978

1 1 2835 705 9140 13667 15001 16738

12 3426 2365 7909 15603 19927

13 3848 2474 13027 13777 20634

14 1360 1743 10224 10971 14907

0 163 2536 5747 10314 10395

1 2583 1 180 11198 15773 18917

2 1542 509 9291 10938 12837

3 4418 1005 5690 15533 16382

4 5212 5117 9605 15605 19763

5 2155 2922 6119 9142 16450

6 347 2696 6956 8879 9200

7 226 4296 15677 16248 20706

8 1560 487 9608 16673 20611

9 3926 1640 10581 16776 18039

10 149 2928 10531 17985 20550

1 1 2364 563 6090 12770 13369

12 635 688 10649 11835 11885

13 231 1684 11016 19381 20412

14 1 129 3894 8889 16308 17018

9. Encoder according to claim 5,

wherein said encoding means (1452) is adapted for subsequently taking a new row of the following address table as a new set of different parity symbol addresses y for accumulating a new group of subsequent information symbols, said address table for Qj_R = 45, M_IR = 16200, N_ldpc = 16200, Qj_dpc = 12 and a code rate identifier of 3/4 being

3 3198 478 4207 1481 1009 2616 1924 3437 554 683 1801

4 2681 2135 4880 8610 10402 18552 18691

5 3107 4027 7094 13641 15279 18429 18763

6 2637 3373 7959 9113 9380 12067 17043

7 3830 3449 8397 11573 11778 16710 19458

8 4129 2060 8239 16473 16672 16877 19512

9 4184 2742 5667 8630 16523 18761 20262

10 3946 1070 5058 8969 9009 13207 18960

1 1 2239 984 10122 12291 13316 20296

0 1458 3031 4596 11080 15340 15725

1 3003 1328 5428 7405 8681 18464

2 1 137 1716 10501 17975 19098 20286

3 132 3725 11130 16523 16916 16937

4 1817 638 12815 18149 18701 19052

5 1774 3447 9354 10403 11261 13858

6 3632 1257 6710 6902 7469 12837

7 542 3694 6061 13335 18664 19075

8 1015 1945 5882 9617 13423 18258

9 1948 412 4816 5319 17280 19387

10 995 2238 10836 18878 19179 19407

1 1 4141 1907 9622 12008 12998 14110

0 2480 3079 12195 12404 12898 16832

1 3021 1088 4545 6886 10669 15877

2 713 1379 5543 5731 12835 18802

3 997 3903 7495 9283 11396 12961

4 2323 3361 5879 8802 10942 13953

5 1 110 986 7231 10301 15987 17805

6 2532 142 6304 14368 15820 19299

7 1690 2405 5548 6086 14591 19907

8 1298 1881 6109 11679 13019 13940 9 615 174 6732 16552 16791 17769

10 1648 3112 11481 11822 16125 19402

1 1 1415 2808 8580 10110 16249 17972

10. Encoder according to claim 5,

wherein said encoding means (1452) is adapted for subsequently taking a new row of the following address table as a new set of different parity symbol addresses y for accumulating a new group of subsequent information symbols, said address table for Qj_R = 45, M_IR = 16200, N_ldpc = 16200, Qj_dpc = 10 and a code rate identifier of 4/5 being

11. Encoder according to claim 5,

wherein said encoding means () is adapted for subsequently taking a new row of the following address table as a new set of different parity symbol addresses y for accumulating a new group of subsequent information symbols, said address table for Qj_R = 45, M_IR = 16200, N_ldpc = 16200, Qj_dpc = 8 and a code rate identifier of 5/6 being

3 2409 499 1481 908 559 716 1270 333 2508 2264 1702 2805

4 2447 1926 4327 7050 8260 9431 16550

5 414 1224 3367 9182 12753 17113 17800

6 21 14 842 4181 7082 10029 11568 11895

7 212 573 5969 7807 8849 17784 17890

0 2383 21 12 5247 6688 11454 12266 13983

1 2286 2348 4236 6503 6949 7729 16002

2 545 819 3149 3192 6775 9444 10499

3 1264 143 3297 3334 5617 13705 16362

4 1701 2258 7013 9480 9707 14604 15367

5 964 166 5410 5801 8028 15261 18109

6 1 14 2413 3973 4019 5518 6014 17895

7 2243 81 5064 6270 7092 10342 10769

0 1245 1581 2977 9454 9498 11121 18718

1 775 169 6060 6582 6906 7876 7971

2 1696 1 104 3449 3549 4103 9526 12269

3 1914 2831 2893 6075 9588 12084 14120

4 532 1450 5719 7056 7505 11829 16647

5 91 974 3794 7948 13803 15159 17804

6 497 2228 3207 8248 9289 13188 17097

7 2326 1579 3854 9152 14036 16732 18531

0 2482 256 3557 6021 13590 17000 18066

1 1 1 17 1261 6360 6583 9662 14389 16722

2 1257 1658 7723 11424 11629 15667 18800

3 1478 1225 5639 6202 10571 13133 14537

4 251 1 980 3584 5850 6863 18281 19074

5 2320 2675 4310 7587 8461 12082 16172 6 435 1278 7441 8375 9335 10030 12047

7 228 503 4465 6278 10955 12600 13070

0 1885 2369 6758 10075 10461 13119 14176

1 57 483 4683 5617 6888 7296 12886

2 838 1050 7124 8119 9163 11583 13994

3 1231 1990 6403 7722 15200 17535 18008

4 1738 68 6924 9267 10291 15033 15248

5 2392 951 6087 6509 9032 12975 17490

6 163 645 8917 9806 11959 12979 15632

7 2644 1704 3740 10222 11078 16244 18139

12. Encoder according to claim 5,

wherein said encoding means (1452) is adapted for subsequently taking a new row of the following address table as a new set of different parity symbol addresses y for accumulating a new group of subsequent information symbols, said address table for Qj_R = 45, M_IR = 16200, N_ldpc = 16200, Qj_dpc = 5 and a code rate identifier of 8/9 being

13. Encoder according to any preceding claim,

wherein said basic codeword portion (B) is provided for regular decoding and said auxiliary codeword portion (A) is provided as incremental redundancy if regular decoding of the codeword by use of the basic codeword portion (B) is erroneous.

14. Encoding method for error correction code encoding input data words (D) into codewords (Zl, Z2), comprising the steps of:

receiving input data words (D) each comprising a first number K_ldpc of information symbols,

encoding an input data word (D) into a codeword (Zl, Z2, Z3, Z4) such that a codeword comprises a basic codeword portion (B) including a data portion (D) and a basic parity portion (Pb) of a second number N_ldpc - K_ldpc of basic parity symbols, and an auxiliary codeword portion (A) including an auxiliary parity portion (Pa) of a third number M_IR of auxiliary parity symbols,

generating said basic codeword portion (B) from an input data word (D) according to a first code, wherein a basic parity symbol is generated by accumulating an information symbol at a parity symbol address determined according to a first address generation rule,

generating said auxiliary codeword portion (A) from an input data word (D) according to a second code, wherein an auxiliary parity symbol is generated by accumulating an information symbol m at a parity symbol address y, wherein said parity symbol addresses y are determined according to a second address generation rule

N_ldpc - K_ldpc + {x + m mod G_a x Q_IR } mod M_IR if x≥ N_ldpc - K_ldpc ,

outputting said codewords (Zl, Z2).

15. Computer program comprising program code means for causing a computer to carry out the steps of encoding and mapping of the method as claimed in claim 14, when said computer program is carried out on a computer.

16. Transmitter for broadcasting data in a broadcasting system, comprising:

a data input for receiving at least one transmitter input data stream (II, 12, In) segmented into input data words (D),

an encoder (14; 141, 142, 143, 144, 145) for error correction code encoding the input data words (D) into codewords (Zl, Z2) according to any one of claims 1 to 14,

a data mapper (16) for mapping the codewords (Zl, Z2) onto frames of a transmitter output data stream (O), and

a transmitter unit (18) for transmitting said transmitter output data stream

(O).

17. Transmission method for broadcasting data in a broadcasting system, comprising the steps of:

receiving at least one transmitter input data stream (II, 12, In) segmented into input data words (D),

an encoding method for error correction code encoding the input data words (D) into codewords (Zl, Z2) according to claim 14,

mapping the codewords (Zl, Z2) onto frames of a transmitter output data stream (O), and transmitting said transmitter output data stream (O).

18. Broadcasting system comprising a transmitter as claimed in claim 16 and one or more receivers for receiving data broadcast by said transmitter.