RU2793198C2 - Multi-mode channel coding - Google Patents

Abstract

FIELD: channel coding.

SUBSTANCE: technical result consists in increasing the data transfer rate and increasing the reliability of the transmission of the coding mode. The technical result is achieved due to the fact that the channel decoder determines the decoding mode without separately receiving information regarding the decoding mode and parameters, since the coding mode indicator is included in the code word by applying a colored sequence, which is selected in accordance with the coding mode.

EFFECT: increasing the data transfer rate and increasing the reliability of the transmission of the coding mode.

43 cl, 26 dwg, 1 tbl

Description

The present application relates to multi-mode channel coding.

In digital communications, channel coding (also referred to as error correction coding) is used to manage errors in data over unreliable or noisy communication channels, and thus channel coding has become an essential part in digital communications. The purpose of channel coding is to protect information from disturbances during transmission. Because of this, redundancy is added for error correction and error detection, i. redundancy is added to the sequence of data packets, such as frames of an audio/video encoder, that are sent over an error-prone channel to allow a certain amount of error correction in transmission at the receiver side. Error correction capability correlates with redundancy ratio, which means that improved error correction performance is usually accompanied by a higher amount of redundancy.

In the context of audio data frames, three effects must be considered:

1. The same frame of audio data can typically be encoded with a flexible number of bits when the audio quality is scaled with the bit rate.

2. Lost frames can be masked because the transmitted data exhibits a temporal pattern that is accompanied by a certain amount of quality degradation that increases with the frame error rate (FER).

3. Packet loss concealment (PLC) methods generally provide much better results than undetected bad frame decoding.

Therefore, channel coding is therefore very practical for audio data, since it can improve the sound quality by:

- bad frame detection (PLC instead of bad frame decoding),

- correction of bad frames (reduction of FER).

However, positive effects are observed only in the presence of errors, while the negative impact of a lower data rate is present at any time. In addition, the signal strength of wireless networks, such as a DECT (Advanced Digital Wireless Standard) based system, typically changes over the duration of the connection, i.e. for a phone call in which the speaker moves during a conversation or due to external temporary disturbances. Because of this, it is suboptimal to apply a fixed Forward Error Correction (FEC) scheme during the connection. Instead, it is highly desirable to have a flexible channel encoder providing multiple FEC modes ranging from low security and high data rate to high security and low data rate (assuming that the full rate, i.e. the sum of the data rate and the redundancy factor is fixed).

From an audio codec point of view, such a switchable system is not a difficult task, since modern audio codecs typically support "real-time" bit rate switching for speech and audio signals. But this causes a technical transmission problem in frame-based FEC mode signaling. For easy integration into existing systems, the FEC mode must be signaled in-band. If this is done explicitly, it also reduces the data transfer rate. In addition, the signaling of mode information must also be subject to transmission errors and must not be protected by error correction codes, since the channel decoder requires knowledge of the mode before being able to decode the encoded data. Because of this, it is necessary to protect the FEC mode separately to avoid having an Achilles' heel in the FEC scheme, which again reduces the data rate for the audio frame.

A well-known channel encoder for audio data is an EP (error protection) tool specified in the MPEG-4 Part 3 standard ("Information technology - Coding of audio-visual objects - Part 3: Audio Standard, International Organization for Standardization, Geneva, CH , 2009. It contains many protection classes ranging from error detection to FEC schemes of various strengths. It also contains flexible frame architectures and non-uniform error protection (UEP). The basic idea of UEP is to divide the frame into subframes according to sensitivities to bit errors and protect these subframes with an appropriate FEC and/or cyclic redundancy check (CRC) rate In order to apply UEP to audio frames, information is required at least a) the number of classes, b) the number of bits that each class contains, c) the CRC code to be applied for each class, which may be represented as the number of CRC bits, and d) the FEC code to be applied for each class as frame configuration parameters. As explained above, UEP requires both out-of-band basic configuration signaling and a significant number of configuration parameters that are signaled in-band. Also, in-band configuration parameters are protected separately from the data, since they are needed before the data is decoded.

Therefore, it is an object of the present invention to provide an efficient and error-tolerant channel coding concept.

This problem is solved by the objects of the invention - a channel encoder according to paragraph 1 of the formula, a channel decoder according to paragraph 14 of the formula, a method for encoding a frame according to paragraph 43 of the formula, a method for channel decoding at least one transmitted code word according to paragraph 46 of the formula, a computer program according to paragraph 52 of the formula, and data flow according to paragraph 53 of the formulas of the present application.

According to the present invention, the channel encoder comprises a coloring module for applying a coloring sequence to at least one codeword, i.e. the codeword includes coding mode information/indicator. Therefore, the transmission bits used to indicate the coding mode to the channel decoder are not required, and thereby the transmission rate is increased and the codeword can be efficiently transmitted. In addition, the coding mode information/indicator is included in the codeword by applying a coloring sequence which is selected according to the coding mode, and thereby error-resistant signaling of the mode information can be achieved.

According to the present invention, the channel decoder receives at least one transmitted codeword, i.e. a codeword including coding mode information/indicator (decoding mode). Thus, the coding mode information/indicator is distributed in the codeword by applying a coloring sequence, and therefore the coding mode information/indicator is received in the channel decoder in an error-tolerant manner. In addition, the channel decoder includes a decoding mode detector for generating a decoding mode indicator indicating a certain decoding mode to be used for redundant decoding, and the decoding mode indicator is associated with at least one coloring sequence as a specific coloring sequence used for coloring the transmitted code. words. Therefore, it is possible to detect the decoding mode by determining a specific coloring sequence, i. e. the channel decoder is able to determine the decoding mode without separately receiving specific decoding mode information. Therefore, the data transfer rate is improved.

According to embodiments of the present application, a channel encoder for encoding a frame, comprising: a multi-mode redundant encoder for redundantly encoding a frame in accordance with a certain coding mode from a set of different coding modes, wherein the coding modes differ from each other with respect to the amount of redundancy added to the frame wherein the multi-mode redundant encoder is configured to output an encoded frame including at least one codeword; and a coloring module for applying a coloring sequence to at least one codeword; wherein the coloring sequence is such that at least one bit of the codeword is changed by applying at least one of the coloring sequences, wherein a specific coloring sequence is selected in accordance with a certain coding mode.

According to embodiments of the present application, the channel coding may change every frame based on the mode selection indicator. The pointer includes the mode selection and the coloring sequence to be applied or the coloring sequence bypass pointer.

According to embodiments of the present application, the channel encoder further comprises a data splitter for splitting a frame into a plurality of data words, wherein the multi-mode redundant encoder is configured to encode each of the plurality of data words according to a specific coding mode to obtain a plurality of codewords, wherein the module The coloring sequence is configured to apply a specific coloring sequence to each codeword in a predefined subset of the plurality of codewords. Thus, the redundancy factor may differ for different data words, i.e. the redundancy factor may be different for each data word. In addition, the length of the data word included in the codeword is changed based on the calculated number of codewords as well as based on the codeword index.

According to embodiments of the present application, a channel decoder for channel decoding at least one transmitted codeword, comprising: a coloring module for applying at least one coloring sequence to at least one transmitted codeword or to at least one transmitted codeword with error correction so as to obtain at least one colored codeword, wherein the coloring sequence is such that at least one bit of the codeword is changed by applying at least one coloring sequence, and wherein the at least one coloring sequence associated with a specific decoding mode as a specific coloring sequence; a redundant decoder for redundantly decoding at least one colored codeword to obtain a decoded output codeword; and a decoding mode detector for generating a decoding mode indicator indicating a specific decoding mode to be used by the redundant decoder to obtain a decoded output codeword, wherein the decoding mode indicator is associated with at least one coloring sequence as a specific coloring sequence used for coloring. transmitted code word. Thus, different coloring sequences (bleaching) are applied to the plurality of transmitted codewords, and the colored (decolorized) words are decoded by using different decoding modes, and one of the used decoding modes is selected as the specific decoding mode based on the test result.

According to embodiments of the present application, the redundant decoder comprises a bit reduction module for reducing the number of bits of a colored codeword and an error correction module for correcting a colored word error, or the channel decoder further comprises an error correction module for correcting an error of the transmitted codeword.

Thus, in case there is an error in the transmitted codewords, the error correction process operates in part of the decoding process in the redundant decoder, or the error correction process operates before the application of coloring (discoloration) independently of the redundant decoder.

In accordance with embodiments of the present application, the coloring module is configured to use in addition to the coloring sequence at least an additional coloring sequence, or the channel decoder is configured to bypass the coloring module in an additional decoding mode without coloring, for example, the coloring sequence has only null as values; wherein the redundant decoder is configured to redundantly decode at least one additional colored codeword colored using the additional coloring sequence, obtaining an additional decoded codeword, an additional colored codeword that is obtained from the transmitted codeword using the additional coloring sequence or codeword transmission without coloring to obtain another additional decoded codeword, and wherein the redundant decoder is configured to output a reliability metric for the decoded codeword, an additional reliability metric for the additional decoded codeword, or another additional reliability metric for the other additional codeword, for example, reliability is computed for each decoded codeword using a different coloring sequence and decoding mode, wherein the decoding mode detector is configured to determine a decoding mode indicator based on the reliability metrics, and the redundant decoder is configured to receive the decoding mode indicator and output, as decoded output codeword, or decoded codeword, or additional decoded codeword, or other additional decoded codeword. Thus, in case there is an error in the transmitted codeword, a reliability index such as a risk value (reliability index) is calculated, and the decoded mode used for the decoded codeword having the smallest risk value is selected as the determined decoding mode.

According to embodiments of the present application, a decoding mode detector is configured to store a list of possible decoding modes indicating a predetermined number of possible decoding modes, wherein one decoding mode candidate may be indicated without a coloring sequence, or all decoding mode candidates are associated with a coloring sequence. , and selecting one decoding mode candidate as the determined decoding mode to be used by the redundant decoder to obtain a decoded output codeword to be used, wherein the decoding mode detector is configured to perform a first operation in the decoding mode and a second operation in the decoding mode, wherein the decoding mode detector for performing an operation in the first decoding mode is configured to evaluate the determined decoding mode, which is a possible variant of the decoding mode without a coloring sequence, i.e. before the hash is evaluated to see if the first codeword is uncolored, to calculate the syndromes of the codeword, check if the computed syndromes have a value of zero, and when the computed syndromes have a value of zero, calculate the hash value of the transmitted codeword words, compare the computed hash value and the hash value included in the transmitted codeword, if the computed hash value is equal to the included hash value, generate a decoding mode indicator to indicate a candidate coding mode without a coloring sequence as the determined decoding mode, or, if the computed hash value differs from the included hash value, excluding the decoding mode candidate without the coloring sequence from the candidate list and continuing with the operation in the second decoding mode. Thus, the decoding mode detector performs two operations, for example, the decoding mode detector comprises a first decoding mode detector for performing an operation in the first decoding mode and a second decoding mode detector for performing an operation in the second decoding mode, and in case a specific decoding mode is not selected in operation in the first decoding mode, the selection process continues with the operation in the second decoding mode. Therefore, if there are no errors in the transmitted codewords, and the mode associated with no coloring sequence is used in the encoder, it is not necessary to continue further, and thus a specific decoding mode is effectively selected.

According to embodiments of the present application, in operation in the second decoding mode, a transmitted codeword error is detected by using a syndrome, an error symbol is calculated using an error locator polynomial, and the error symbol is corrected. During the procedure, if the detected errors are not correctable, then the decoding mode associated with the coloring sequence applied to the transmitted word, which includes an uncorrectable error, is excluded from the list of possible options. In addition, if the error locator polynomial for a transmitted word having a correctable error cannot be determined, then the additional decoding mode is excluded from the list of possible options. Thus, the listed decoding mode options are eliminated step by step, and the remaining decoding mode in the list is selected as the determined decoding mode at the end. Therefore, a certain decoding mode is reliably selected in view of the risk of an error occurring.

In the preferred embodiments of the present application, the FEC mode is signaled by modifying known line codes such that the redundancy factor is effective, while the decoder is able to determine the FEC mode through partial probe decoding. This implicit transmission in null-byte signals is deterministic if no transmission errors occur, and otherwise finds the correct mode with high probability, i.e. frame loss due to signaling errors is negligible compared to frame loss due to uncorrected frames. More specifically, it refers to a scheme (FEC) that provides a plurality of N _modes for encoding data sequences into code sequences of a given length N _target . Here, for simplicity, binary sequences are assumed, but a similar scheme should also apply to the general case in which the data characters represent elements in any field, such as a finite Galois field.

Preferred aspects of the present are the subject of the dependent claims. Preferred embodiments of the present application are described below with reference to the drawings, in which:

Fig. 1 shows a block diagram illustrating an example of a channel encoder for encoding a frame to be transmitted according to embodiments of the present application;

Fig. 2 shows a block diagram illustrating an example of another channel encoder for encoding a frame to be transmitted according to embodiments of the present application;

Fig. 3 shows a block diagram illustrating an example of an encoder including a channel encoder for encoding a frame to be transmitted according to embodiments of the present application;

Fig. 4 shows a block diagram illustrating an example of the additional channel encoder shown in FIG. 3;

Fig. 5A, 5B show a block diagram illustrating examples of a channel decoder for channel decoding at least one transmitted codeword according to embodiments of the present application;

Fig. 6 shows a flowchart for an example of a channel decoding operation implemented by the inventive channel decoder shown in FIG. 5A or FIG. 5B according to embodiments of the present application;

Fig. 7A, 7B show a block diagram illustrating changes to a channel decoder for channel decoding of at least one transmitted codeword according to the present application, as indicated by FIG. 5A and 5B;

Fig. 8 shows a block diagram illustrating further modification of a channel decoder for channel decoding of at least one transmitted codeword according to the present application, as indicated by FIG. 7A or 7B;

Fig. 9 shows a block diagram illustrating another example of a channel encoder for encoding a frame to be transmitted according to embodiments of the present application;

Fig. 10 shows a flowchart for an example of a channel encoding operation implemented by the channel encoder shown in FIG. 9 according to embodiments of the present application;

Fig. 11 shows a block diagram illustrating a change in a channel encoder for encoding a frame to be transmitted according to the embodiments of the present application shown in FIG. 9;

Fig. 12 shows a schematic illustration of an example for a frame architecture in the channel coding operation shown in FIG. 11 according to embodiments of the present application;

Fig. 13 shows a schematic illustration of an example for frame architecture versus coding mode according to embodiments of the present application;

Fig. 14 shows a block diagram illustrating a further example of a channel encoder for encoding a frame to be transmitted according to embodiments of the present application;

Fig. 15 shows a block diagram illustrating a further example of a channel decoder for channel decoding at least one transmitted codeword according to embodiments of the present application;

Fig. 16 shows a block diagram illustrating an example of a decoding mode detector of a channel decoder located in the channel decoder shown in FIG. 15 according to embodiments of the present application;

Fig. 17 shows a flowchart for an example of the decoding mode detection operations of the decoding mode detector implemented by the decoding mode detector shown in FIG. 16 and an example of the decoding operation of the channel decoder implemented by the channel decoder shown in FIG. 15 according to embodiments of the present application;

Fig. 18 shows a flowchart illustrating an operation example in the second decoding mode of the decoding mode detector implemented by the decoding mode detector shown in FIG. 16 according to embodiments of the present application;

Fig. 19 shows a flowchart for an example of a second decoding mode operation procedure implemented by the second decoding mode detector shown in FIG. 18 according to embodiments of the present application;

Fig. 20 shows a block diagram illustrating the modification of a channel decoder for channel decoding at least one transmitted codeword according to the embodiments of the present application shown in FIG. 15;

Fig. 21 shows a block diagram illustrating the change of the decoding mode detector of the channel decoder located in the channel decoder shown in FIG. 20 according to embodiments of the present application;

Fig. 22 shows a schematic illustration of an operation for an operation in the first decoding mode implemented by the mode detector shown in FIG. 20;

Fig. 23 shows a schematic illustration of an operation for operation in the second decoding mode implemented by the second decoding mode detector shown in FIG. 18; And

Fig. 24 shows an enlarged schematic illustration of an operation for the operation in the second decoding mode shown in FIG. 23.

The same or equivalent elements, or elements with the same or equivalent functionality, are referred to in the following description by the same or equivalent reference numerals.

In the following description, many details are provided to provide a more complete explanation of the embodiments of the present application. However, those skilled in the art will appreciate that embodiments of the present application may be practiced without these specific details. In other cases, known structures and devices are shown in block diagram form rather than in detail so as not to obscure the embodiments of the present application. In addition, the features of the various embodiments described below may be combined with each other unless expressly stated otherwise.

Fig. 1 illustrates an embodiment of a channel encoder 2 for encoding a frame, comprising a multi-mode redundant encoder 4 for redundantly encoding a frame according to a certain coding mode from a set of different coding modes, and a coloring module 6 for applying a coloring sequence to at least one codeword. Fig. 2 illustrates another example of a channel encoder 2 further comprising a controller 8 for providing coding criteria such as frame redundancy ratio, required data protection (determined coding mode, data length, target size N _target , number of data words constituting a frame, etc. In case the channel encoder 2 does not include the controller 8 as indicated in Fig. 1, the coding criteria are passed in association with the frame to the multi-mode redundant encoder 4. Alternatively, the channel encoder 2 also includes a hash extension to calculate the hash value of the frame to be transmitted.

In a further embodiment, for example, an audio data frame is transmitted to channel encoder 2. In case controller 8 is not included in channel encoder 2 (as shown in FIG. 1), the audio data frame is associated with an indicator of target size N _target and specific coding mode m _fec . Then, the transmitted frame is encoded in the multi-mode redundant encoder 4, and a plurality of codewords are output from the multi-mode redundant encoder 4. The output codewords are sent to the coloring module 6, and at least one codeword of the output codewords is colored by applying a specific coloring sequence sig _m that is selected according to the specific coding mode m _fec . Regarding the coloring sequence, when the value of the coloring sequence is zero, then the coloring is bypassed for the codewords. Thus, for example, encoding mode m _fec =1 is associated with a coloring sequence sig ₁ having a value of zero, i. the coloring sequence sig ₁ is the null sequence, and therefore coloring for codewords encoded by coding mode m _fec =1 is bypassed.

Fig. 3 shows an example of an encoder including a channel encoder according to an embodiment of the present application. The encoder contains an audio encoder 1, a channel encoder 2 and a switching module 3. The audio encoder 1 is configured to encode the input audio data, and the channel encoder 2 is configured to encode the channel according to the specified mode. The switching unit 3 inputs into the audio encoder 1 the payload size N _p and into the channel encoder 2 the FEC mode and the slot size N _s . As shown in FIG. 3, the encoder outputs the encoded frame.

Fig. 4 shows a further example of the channel encoder shown in FIG. 3. Channel encoder 2 contains a hash extension 10, a line encoder 4' and a coloring module 6'. The audio data encoded in the audio encoder 1 is input as input a _in , and the hash extension 10 generates the extended data a _ext by adding the hash data. The spread data a _ext is input to a line encoder 4' and the line encoder 4' generates a code word a _enc in a line code. The code word a _enc is input to the coloring module 6', and the coloring module 6' generates the colored code word a _out to be transmitted.

. Первая стадия кодирования данных в силу этого состоит из преобразования определенного слова данных

в кодовое слово в C_m.Channel encoder 2 in FIG. 4 performs implicit signaling of the mode information, i. e. transmission of the encoded frame to the decoder without explicitly specifying the coding mode. In particular, metadata other than the target size N _target need not be transmitted to the decoder (eg, as shown, for example, in FIGS. 5A and 5B). For each mode N _tmode , a linear codeword C _m of length N _target is assigned, which means

. The first stage of data encoding therefore consists of converting a particular data word

into a codeword in C _m .

в C_k. Тем не менее, поскольку коды являются линейными, пересечения

не являются пустыми. В частности, многие семейства кодов, к примеру, примитивные BCH-коды в узком смысле или коды Рида-Соломона с одинаковым базовым полем, даже должны формировать включения между различными кодами при упорядочении согласно расстоянию Хэмминга, т.е.:The implicit signaling of mode information is based on the fact that it can be efficiently checked whether a given word is

in C _k . However, since the codes are linear, the intersections

are not empty. In particular, many families of codes, such as primitive BCH codes in the narrow sense or Reed-Solomon codes with the same base field, even have to form inclusions between different codes when ordered according to the Hamming distance, i.e.:

(1)

(1)

Because of this, it may happen that the codeword C _m obtained by the data encoding word in mode m is also in C _n for n≠m, so that it is generally not possible to determine the mode based on such a test alone. This problem is resolved by applying a mode-dependent transformation:

(2)

(2)

для m≠n, что разрешает вышеуказанную неоднозначность. Поскольку преобразования не изменяют длину кодовых слов, этот способ эффективно передает в служебных сигналах режим канального кодера с нулевыми битами.after coding. These transformations are calculated in such a way that

for m≠n, which resolves the above ambiguity. Because the transforms do not change the length of the codewords, this method effectively signalizes the zero-bit channel encoder mode.

. Таким образом, при кодировании в режиме m канальный кодер выводит окрашенное кодовое слово:In the present embodiment, these transforms are specified by computing the bitwise XOR of the encoded codeword c and the signaling sequence

. Thus, when encoding in mode m, the channel encoder outputs a colored codeword:

. (3)

. (3)

и незнании представления в (3), декодер затем может эффективно проверить, верно ли следующее:When you receive

and not knowing the representation in (3), the decoder can then effectively check if the following is true:

(4)

(4)

представляет собой нулевую последовательность, декодер должен иметь возможность верифицировать то, что:Because the

is a zero sequence, the decoder must be able to verify that:

(5)

(5)

is in C _m . In addition, provided:

(6)

(6)

overhead sequences sig _k may be chosen such that:

(7)

(7)

for k≠l, enabling deterministic mode detection at the decoder if no transmission errors occur. Therefore, it is not necessary to separately transmit data to indicate a specific coding mode and parameters for a channel decoder.

As shown in FIG. 5A, the channel decoder 20 includes a coloration (bleaching unit) 22 , a redundant decoder 24 , and a mode detector 26 . The redundant decoder 24 includes a bit reduction unit 24a and an error correction unit 24b; however, it is not necessary that the bit reduction unit 24a and the error correction unit 24b be provided in the redundant decoder 24. Thus, as shown in FIG. 5B, the error correction module 24b may be connected to the coloring module 22, and the bit reduction module 24b may be located between the coloring module 22 and the mode detector 26. The mode detector 26 is connected to the redundant decoder 24 (bit reduction unit 24a) and the colorization unit 22. In the channel decoder 20, more specifically, in the decoding mode detector 26, a decoding mode indicator is determined indicating a specific decoding mode to be used by the redundant decoder 24 (bit reduction unit 24a) to obtain a decoded output codeword and a specific coloring sequence, used for coloring the transmitted code word in the coloring module 22 .

The channel decoder 20 receives the transmitted codewords transmitted from the channel encoder 2. The predetermined number of transmitted codewords is then used/checked to generate a decoding mode indicator in the decoding mode detector 26, as explained below. The decoding mode detector 26 has information regarding the decoding modes that can be used by the channel encoder 2, such as a list of possible decoding mode options.

Fig. 6 is a flowchart for an example of a channel decoding operation implemented by the inventive channel decoder shown in FIG. 5A or FIG. 5b. The given number of transmitted codewords is decolorized by applying a coloring sequence (S2), for example, a coloring sequence sig ₁ associated with decoding mode m _fec =1 in the candidate list. Then, the colored (decolorized) codewords are decoded in the bit reduction unit 24a (S4), and a predetermined number of decoded output codewords is obtained. In case no transmission errors occur (S6), a decoding mode indicator is generated in the decoding mode detector 26 based on the decoded output codewords (S10). The generated decoding mode pointer is passed to the coloring unit 22 and the redundant decoder 24 (bit reduction unit 24a), i. in case the decoding mode indicator indicates a certain decoding mode, it is set that the decoding mode is detected (S12). Then, transmission codewords decoded by using the determined decoding mode are output as decodable output codewords (S16).

As indicated in S6, if errors are detected, a reliability index (risk value) is calculated in the decoding mode detector 26 (S7). Thus, in case transmission errors occur, errors are detected and an attempt is made to correct them by calculating and using syndromes in the error correction module 24b, and the error correction result is transmitted to the decoding mode detector 26 from the error correction module 24b. In the case where the error correction unit 24b is independent, as shown in FIG. 5B, the error correction result is associated with the transmitted codeword. Next, the error detection and error correction procedure will be explained in detail.

Then, in case all possible decoding mode options in the list are checked (S8), the method proceeds to S10 as described above. If there is a remaining decoding mode candidate in the list (S8), S2-S7 are repeated until all decoding mode candidates have been checked. In the case where a specific decoding mode is not determined (S12) although all possible decoding modes are checked, a frame that consists of a transmitted codeword used/checked to determine the specific decoding mode is registered as a bad frame.

Fig. 7A and 7B show variants of channel decoder 20 as shown in FIG. 5A and 5B. As shown in FIG. 7A, the channel decoder 20 further comprises a controller 28 that is connected to the coloring module 22, the redundant decoder 24, and the mode detector 26. In addition, the channel decoder 20 in FIG. 7B corresponding to the channel decoder 20 described in FIG. 5B further comprises a controller 28 that is connected to an error correction module 24b, a coloration module 22, a bit reduction module 24a, and a mode detector 26.

In the channel decoder 20 in FIG. 7A and also in FIG. 7B, a plurality of mode tests may be processed in parallel. For example, if there are four possible modes in the candidate list, a mode test is performed for each mode candidate, i.e. for option 1: coloring sequence sig ₁ and decoding mode m _fec =1, for option 2: coloring sequence sig ₂ and decoding mode m _fec =2, for option 3: coloring sequence sig ₃ and decoding mode m _fec =3 , and for option 4: coloring sequence sig ₃ and decoding mode m _fec =4, are executed in parallel as described below (S2-S8 in FIG. 6 are executed for each option in parallel). Mode detector 26 then generates a decoding mode indicator, and controller 28 causes redundant decoder 24 (bit reducer 24a) to output decoded codewords decoded using the indicated decoding mode as output decoded codewords.

Fig. 8 shows a modification of a channel decoder including a plurality of coloring modules 221-224, a plurality of decoders 241-244 (redundant decoders), a mode selector (detector) module 25, and a switch 27. As described in FIG. 8, the transmitted code word, ie. encoded data a _in is input to a channel decoder and copied for transmission to a plurality of coloring modules 221-224. The coloring sequence sig ₁ is applied to the copied coded data a _in in coloring module 221, and decoding mode m _fec =1 is applied to colored codeword a _in (1) in decoder 241 to obtain decoded codeword a _dec (1). In this embodiment, the coloring sequence sig ₁ is the null sequence, i. the staining operation is bypassed. The coloring sequence sig ₂ is applied to the copied coded data a _in in coloring module 222, and decoding mode m _fec =2 is applied to colored codeword a _in (2) in decoder 242 to obtain decoded codeword a _dec (2). The coloring sequence sig ₃ is applied to the copied coded data a _in in coloring module 223, and the decoding mode m _fec =3 is applied to colored codeword a _in (3) in decoder 243 to obtain decoded codeword a _dec (3). The coloring sequence sig ₃ is applied to the copied coded data a _in in coloring module 224, and decoding mode m _fec =4 is applied to colored codeword a _in (4) in decoder 244 to obtain decoded codeword a _dec (4).

Although not illustrated in FIG. 8, the decoder includes an error correction module, and the decoding of statistical information such as the number of corrected bits or symbols and the number of corrected symbols in code subwords, if the code is constructed from smaller codes, is transmitted from each decoder 241-244 to the mode selection module 25, as indicated by the dotted line. Next, the mode selector 25 generates a decoding mode indicator indicating a specific decoding mode to be used by the decoder to obtain a decoded output codeword. The mode selection unit 25 selects the decoding mode as the determined decoding mode based on the decoding of the historical information, i.e. if the mode decodes without error correction, this mode is selected as the determined decoding mode. In this case, it should be noted that an incorrect selection cannot occur if errors do not occur. If errors occur, a risk value (reliability index) is calculated for probability-based decoding modes to generate a codeword with a similar number of errors by random guessing. The mode detector then determines for the decoding mode with the smallest risk value. In addition, the mode detector may also require the risk value of the selected mode to be less than a predetermined threshold. Therefore, if there is no decoding mode having a risk value less than a predetermined threshold, a bad frame indicator is generated instead of a decoding mode indicator. The bad frame indicator is passed to the encoder along with the size/length N _data of the data.

The decoder determines/calculates/calculates a risk value (reliability index) of a decoding mode based on the number of corrected symbols during a colored codeword decoding operation. If a predetermined number of codewords are colored, for example, 6 codewords are colored, then the risk value (reliability index) for the decoding mode m _fec =2 is calculated based on the number of corrected symbols during the decoding operation of the 6 colored codewords. In the same way, a risk value (an additional reliability metric) for a decoding mode m _fec =3 is calculated based on the number of corrected symbols during a decoding operation of 6 colored codewords, or a risk value (another additional reliability metric) for a decoding mode m _fec =1 is calculated based on the number of corrected symbols during the decoding operation of 6 codewords without coloring.

When a certain decoding mode is selected, then the decoding mode indicator is transmitted from the mode selection unit 25 to the switch 27. The switch 27 switches the connection with the decoder to obtain a decoded output codeword based on the decoding mode indicator. Thus, for example, if the decoding mode m _fec =3 is indicated as the determined decoding mode, then the switch 27 is connected to the decoder 243 to obtain a decoded output codeword.

In case transmission errors occur, decoder 20 receives:

(8)

(8)

with an erroneous sequence ε. This can lead to ambiguities in the sense that:

(9)

(9)

обозначает наборы всех последовательностей, которые могут корректироваться посредством схемы FEC (например, в модуле 24b коррекции ошибок, показанном на фиг. 7a и 7b), чтобы формировать кодовое слово в C_k. Если k≠m представляет собой такой режим, то стратегия декодирования состоит в том, чтобы рассматривать член ошибок:for different k, where

denotes the sets of all sequences that can be corrected by the FEC scheme (eg, in error correction module 24b shown in FIGS. 7a and 7b) to form a codeword in C _k . If k≠m is such a mode, then the decoding strategy is to consider the error term:

(10)

(10)

as a random guess and evaluate the regimen from the list of possible regimens according to the risk value, i.e. indicator of reliability. This risk value is derived from decoding statistics, such as the number of corrected symbols, and reflects the probability that the decoder's input is a random guess in C ₀ (ie, a sequence of tossing the correct coin).

Thus, the risk of selecting the wrong mode is limited by the risk of erroneously decoding a codeword that is corrupted beyond the error correction characteristics of the base code. In cases where this risk is considered too great, the hash value may be added as an option to the data prior to encoding and taken into account in the mode detection procedure. While this, like explicit signaling, reduces the data rate, it equally increases the risk of mode selection and the risk of incorrect decoding. Therefore, the proposed FEC scheme is very suitable for an application in which undetected corrupted frames typically result in greater degradations than detected and concealed corrupted frames.

Further details regarding the channel encoder 2 and the channel decoder 20 according to the present application are explained below.

Channel encoder

The channel encoder of the invention operates on bytes and uses GF(16) Reed-Solomon codes in such a way as to construct a family of line codes. It takes as input a target size in bytes, denoted N _s , also called "interval size", a mode number m _fec between 1 and 4, and an input sequence a(k) of data bytes, k=0, 1, ..., N _p -1, which are interpreted as integers between 0 and 255. The input size N _p is extracted from the parameters N _s and m _fec as follows. Below, the target size is assumed to be at least 40 bytes.

Fig. 9 shows a block diagram illustrating another example of a channel encoder 2 for encoding a frame to be transmitted according to embodiments of the present application. Channel encoder 2 contains controller 8, preprocessor 10, data partitioning module 12, multi-mode redundant encoder 4, coloring module 6 and universal mixer 14.

Fig. 10 shows a flowchart for an example of a channel encoding operation implemented by the channel encoder shown in FIG. 9. As shown in FIG. 10, the hash value of the input, i.e. input frame data is calculated and added to the input data (S20). The frame including the input data and the added hash value is split by the data splitter 12 into a plurality of data words (S22). The number of data words is calculated based on the target frame size. A plurality of data words are encoded by the multi-mode redundant encoder 4 (S24) and the encoded data words, i. e. the code words are transmitted to the staining unit 6 for applying the staining sequence (S26). The colored code words are then interleaved in the universal mixer 14 (S28).

Fig. 11 shows a block diagram illustrating the change of the channel encoder illustrated in FIG. 9. The channel encoder of FIG. 11 additionally contains a frame configurator 11 for transmitting information to other devices. For example, the frame configurator 11 receives parameters such as target codeword size N _s and coding mode m _fec . In addition, the _hash size N of the frame hash value and the Hamming distance are calculated in the frame configurator 11 . As illustrated in FIG. 11, the input data a _in and data length information N _p and the computed _hash size N hash and the Hamming distance are transmitted to the preprocessor 10. In addition, codeword lengths Li, Hamming distances d _i and number N _cw of code words are transmitted to the data partitioning module 12 , a Reed-Solomon (RS) encoder 4, a coloring module 6, and a multiplexer 14 (interleaving module). From the preprocessor 10, for example, data a ₂ having a size of N ₂ =N _p +N _hash is transferred to the data splitter 12 . The data splitter 12 splits the frame into a plurality of codewords D ₀ , D ₁ -D _Ncw-1 , which are transmitted to the encoder 4 RS. RS encoder 4 encodes data words to obtain code words C ₀ , C ₁ -C _Ncw-1 , which are transmitted to coloring module 6 . The coloring unit 6 applies a coloring sequence to the codewords to obtain colored codewords, and the colored codewords are interleaved in the multiplexer and output as a _out . The detailed procedure is explained below.

Data preprocessing

In the preprocessing step, a hash value of N _hash bytes is computed for the input, e.g. for a CRC (cyclic redundancy check) hash, where N _hash = N _hash (N _s , m _fec ) is a number depending only on the size of N _s interval and mode m _fec FEC. In a preferred embodiment, the number is given by:

(11)

(eleven)

The hash value is used to validate the data, since the error detection of Reed-Solomon codes is not very rigorous.

, и пусть rem(n, m) обозначает остаток в делении в столбик n на m. Хэш и данные конкатенируются и разбиваются на последовательность чисел от 0 до 15 (далее называемую «номерами в единицах по 4»), например, согласно:Let hash bytes be denoted as h(k),

, and let rem(n, m) denote the remainder after n divided by m. The hash and data are concatenated and split into a sequence of numbers from 0 to 15 (hereinafter referred to as "numbers in units of 4"), for example, according to:

(12)

(12)

And:

(13)

(13)

, и:For

, And:

(14)

(14)

And:

(15)

(15)

For

The input data extended by the computed hash, i.e. a frame including a hash value is split into a plurality of data words. The number of data words is calculated, for example, based on the target size for the frame and the codeword index.

Fig. 12 shows a schematic illustration of an example for a frame architecture in a channel encoding operation performed by the channel encoder shown in FIG. 11. Input data, i.e. the input frame including the hash value is split into the computed number of data words (also the number of codewords), eg 6 data words D ₁ -D ₅ .

Reed-Solomon coding

(Source reference "Error Correction Coding: Mathematical Methods and Algorithms", Todd K. Moon, 2005).

Line codes are constructed from a set of GF(16) Reed-Solomon codes with 1, 3, 5, and 7 Hamming distances. The number of codewords is calculated from the interval size as follows:

, (16)

, (16)

and the length of codeword i in data symbols is given by:

(17)

(17)

. Расстояния Хэмминга для различных кодовых слов являются зависимыми от режима и от кодового слова и задаются посредством:where i varies from 0 to N _cw -1. The condition N _cw ≥40 implies that

. The Hamming distances for different codewords are mode and codeword dependent and are given by:

(18)

(18)

The data array is split into N _cw data words according to the following:

(19)

(19)

и

.where the sequence of partition points is inductively given by

And

.

This limits the input length to:

(20)

(20)

which depends only on N _s and m _fec .

кодов С_i. Кодирование по Риду-Соломону зависит от генератора для базового поля и генератора для единичной группы этого базового поля (см., например, работу "Error Correction Coding: Mathematical Methods and Algorithms", Todd K. Moon, 2005 года), а также от преобразования данных в символы. Ниже по тексту, поле GF(16) предположительно должно формироваться посредством

, и генератор α единичных групп предположительно представляет собой остаточный класс x в

. Кроме того, преобразование данных в символы (преобразование номеров в единицах по 4 в элементы GF(16)) принимается следующим образом:Then, the data words Di are encoded into

codes С _i . Reed-Solomon coding depends on the generator for the base field and the generator for the identity group of this base field (see, for example, "Error Correction Coding: Mathematical Methods and Algorithms", Todd K. Moon, 2005), as well as on the transformation data to characters. Below, the field GF(16) is supposed to be formed by

, and the identity group generator α is presumably the residual class x in

. In addition, the conversion of data to characters (conversion of numbers in units of 4 to GF(16)) is taken as follows:

(21)

(21)

where bit _k (n) denotes the kth bit in the binary representation of n, given as follows:

(22)

(22)

The codeword C _i is then in this case a uniquely defined sequence corresponding to:

(23)

(23)

, а также соответствующую тому, что полином:For

, and also corresponding to the fact that the polynomial:

(24)

(24)

is divided by the generating polynomial RS:

(25)

(25)

As explained above, data words are encoded and code words are output as also indicated in FIG. 12. The length of each codeword is not necessarily the same, for example, the length of codewords C ₀ and C ₁ is 14 nibbles, and the length of codewords C2-C ₅ is 13 nibbles, and therefore the length of a frame including codewords C ₀ -C ₅ is 40 bytes. It should be noted that the redundancy factor may differ for different data words because the length of the data words is not exactly the same.

Fig. 13 is a schematic illustration of an example for a frame architecture depending on the coding mode. Fig. 13 shows the relationship between the data to be transmitted and the redundant layout when the target size is 40. For example, when the codewords are encoded using m _fec =1, no errors are expected, when the codewords are encoded using m _fec =2, one error in symbol per codeword may occur when codewords are encoded using m _fec =3, two symbol errors per codeword may occur, and when codewords are encoded using m _fec =4, three symbol errors per calculation per code word may occur. The channel decoder 20 is informed of the data to be transmitted and the redundant arrangement since decoding requires mode knowledge in order to spread the encoded data.

In the next step, mode information signaling is performed by coloring the code words according to:

(26)

(26)

for k=0, 1, ..., L _i , where sig _k is one of the following sequences:

In this embodiment, the number of code words (data words) is 6 if N _s =40. If the number of codewords is large, it is not necessary to color all the codewords in order to reliably signal the FEC mode. In this case, the coloring may be limited to a predefined subset of codewords, eg, the first 6 codewords, as indicated by i<6 above. However, the number of codewords may vary depending on the parameters, i.e. target size, codeword index, codeword length, etc. In addition, the numbers of the coloring sequence are not limited to the above examples and may be other numbers.

The function bitxor(n, m) is defined for natural numbers and denotes the result of a bitwise XOR for binary representations of n and m, i.e.:

(27)

(27)

является декодируемым с d символов в C_j(15), когда c_i извлекается случайно равномерно из C_i(15).Table 1. Probability that

is decodable with d symbols in C _j (15) when c _i is extracted randomly uniformly from C _i (15).

d: 0 1 2 3 not decodable

0 0 - - 1

0 0

-

0 0

0 0

-

0 0

0 0 0 0 1

Where:

(28)

(28)

The sequences are chosen in such a way as to maximize the separation of the different codes with respect to the Hamming distance. When denoted by:

,

и

, окрашивание кодовых слов соответствует векторному сложению

в GF(16)^Li. Для детерминированного обнаружения режима в отсутствие ошибок при передаче, условие

для k≠l является достаточным, поскольку оно обеспечивает возможность уникального определения передаваемого в служебных сигналах режима. Последовательности служебных сигналов соответствуют следующим еще более строгим условиям, которые рассчитываются с возможностью ограничения риска неоднозначностей режима, когда возникают ошибки при передаче.corresponding vectors in GF(16) ^Li and the specification

,

And

, coloring of code words corresponds to vector addition

in GF(16) ^Li . For deterministic mode detection in the absence of transmission errors, the condition

for k≠l is sufficient since it allows unique determination of the signaling mode. The overhead sequences comply with the following even more stringent conditions, which are calculated to limit the risk of mode ambiguities when transmission errors occur.

, слово

не является декодируемым в Cl(Li), что означает то, что его расстояние Хэмминга до Cl(Li) превышает характеристики коррекции ошибок кода.- For k<l and

, word

is not decodable in Cl(Li), which means that its Hamming distance to Cl(Li) exceeds the error correction performance of the code.

, слово

может быть декодируемым в Cl, но требует по меньшей мере коррекции двух символов.- For k<l and

, word

may be decodable in Cl, but requires at least two symbols correction.

в C_l(15) с предписанным числом коррекций символа, когда c_k извлекается случайно равномерно из C_k(15), приводятся в таблице 1. Они обеспечивают верхнюю границу для соответствующих вероятностей для меньших значений Li.Decoding Capability Probabilities

in C _l (15) with a prescribed number of symbol corrections when c _k is extracted randomly uniformly from C _k (15) are given in Table 1. They provide an upper bound on the corresponding probabilities for smaller values of Li.

Colored codewords are shown in FIG. 12. To codewords encoded by using coding mode m _fec =1, a coloring sequence sig ₁ is applied, i. e. staining is bypassed because the staining sequence is the null sequence. The coloring sequence sig 2 is applied to codewords encoded by using coding mode m _fec =2 , the coloring sequence sig ₃ is applied to codewords encoded by using coding mode m _fec = ₃ , and to codewords encoded by using mode m encoding _fec =4, coloring sequence sig ₃ is applied. It should be noted that not all codewords need to be colored, i. a coloring sequence can only be applied to a given subset of codewords in a frame.

Codeword Multiplexing

The colored codewords are interleaved by the multiplexer 14. Thus, a bit from the colored codeword is placed in another codeword in an additional bit of at least one other codeword to obtain a frame.

Thus, the lengths Li of the codewords are chosen such that the elements of the colored codewords C _i can be completely interleaved, with the multiplexed codeword given:

(29)

(29)

, которое преобразуется в выходную последовательность:For

, which is converted to the output sequence:

(30)

(thirty)

for k=0, 1, ..., N _s -1. Interleaving increases the strength of protection if bit errors are not evenly distributed over the frame, such as when bursts of errors occur. It should be noted that the coloring of codewords can also be described at this stage by calculating the bitwise XOR of the final output and the corresponding sequences extracted from sig _m .

Fig. 14 shows a block diagram illustrating a further example of a channel encoder for encoding a frame to be transmitted. The channel encoder 2 shown in FIG. 14 contains a controller 8, a channel encoding core 2' and an audio encoder 16, i. e. this channel encoder is used to encode the audio frame to be transmitted. In the case of encoding a video frame to be transmitted, channel encoder 2 contains a video encoder instead of an audio encoder 16.

Channel decoder

Fig. 15 shows a further example of a channel decoder 20. Channel decoder 20 includes a demultiplexer 60, a decoding mode detector 26, an error correction module 62, a coloring module 22, a redundant decoder 24, a data combining module 64, and a post-processor 68. An error correction module 62 is coupled to a mode detector 26. decoding in Fig. 15. However, error correction module 62 may be included in redundant decoder 24, for example, as shown in FIG. 5A, or be located between the decode mode detector 26 and the staining unit 22, or the staining unit 22 and the redundant decoder 24. However, in addition to this, in FIG. 15 shows that channel decoder 20 may further comprise a controller and/or a memory/storage device for storing a list of possible decoding mode options.

Fig. 20 shows the change of the channel decoder 20 shown in FIG. 15. FIG. 20 illustrates channel decoder 20 in accordance with its operation similar to FIG. 11. The channel decoder 20 further includes a frame configurator 61 that transmits the number of code words, N _cw , and the length of the code words, L _i , to the demultiplexer 60, the mode detector 26, which includes an error correction module, a bleaching module 22 ( coloring module), an RS decoder 24 and a data combining module 64 . In addition, the frame configurator 61 transmits the target frame size, i. e. slot size N _s to post processor 68. Demultiplexer 60 extracts the interleaved codewords. The mode detector 26 detects the determined decoding mode and generates a decoding mode indicator. The decoding mode indicator includes at least information to indicate a specific decoding mode, and may include additional information such as the number of corrected symbols. The decoding mode indicator is passed from mode detector 26 to decolorization module 22, RS decoder 24, and postprocessor 68. Decolorization module 22 decolorizes according to the determined decoding mode by XORing a given number of first encoded words, e.g., the first 6 codewords with a coloring sequence associated with a certain decoding mode. The RS decoder 24 extracts only parts of the codeword data, i.e. error correction is performed earlier for colored codewords. The amount of redundancy is indicated by the decoding mode indicator. The data fusion module 64 concatenates the input, i. e. combines data words to produce output data. The post processor 68 validates the output using the hash value if the determined decoding mode is not mode 1. Detailed operation is explained below.

The channel decoder 20 receives as input a byte sequence a _in and a slot size N _s in bytes. On the other hand, bytes are interpreted as numbers between 0 and 255.

Demultiplexing codewords

The demultiplexer 60 extracts the interleaved codewords in the decoder 20, i. e. the frame configurator 61 calculates the number of codewords, N _cw , and the lengths of codewords, L _i , from the input size N _s as specified in the Reed-Solomon Coding section, and extracts the code words C _i according to the arrangements described in the section " Code word multiplexing.

Mode detection

Direct mode detection according to the channel decoder 20 shown in FIG. 5A and the above associated description can be done by going from C _i to:

(31)

(31)

обозначает число символов, которые скорректированы в кодовом слове C_{i, m} во время декодирования RS. Для рассматриваемых кодов RS(Li, Li-2t) Рида-Соломона (расстояние Хэмминга составляет 2t+1, и в силу этого число корректируемых символов составляет t), вероятность выбора случайного слова w из n символов в GF(16)ⁿ, которое может корректироваться в кодовое слово RS(n, n-2t) посредством модификации τ≤t символов, задается посредством:using trial decoding for all possible modes and, if successful, validating the decoded data by recomputing the hashes of the transmitted codewords (ie, the decoded frame) as explained above (data pre-processing). If this procedure is successfully performed for several modes, then the risk value can be attached to the mode classes as follows: let

denotes the number of symbols that are corrected in codeword C _{i, m} during RS decoding. For the considered RS(Li, Li-2t) Reed-Solomon codes (the Hamming distance is 2t+1, and therefore the number of corrected symbols is t), the probability of choosing a random word w out of n symbols in GF(16) ⁿ , which can corrected into the RS(n, n-2t) codeword by modifying τ≤t symbols, is given by:

(32)

(32)

Therefore, the risk value for mode m can be taken as:

, (33)

, (33)

and m _fec must be chosen such that ρ _mfec is minimal.

The proposed decision on the choice of mode is made along a slightly different path. Instead of aiming for complete decoding for all possible modes, the mode detector takes a parallel approach, narrowing down the list of possible modes in steps and reaching a final decision after processing the first 6 codewords. This approach has the advantage of less computational complexity on average.

Fig. 16 shows a block diagram illustrating an example of the decode mode detector 26 of the channel decoder located in the channel decoder 20 shown in FIG. 15. The decode mode detector 26 comprises a first decode mode detector 30, a second decode mode detector 32, and a controller 34. The decode mode detector 26 is configured to perform a first decode mode operation in the first decode mode detector 30 and a second decode mode operation in the second detector. 32 decoding modes. The controller 34 may include a memory/data storage device for storing a list of possible decoding mode options.

Fig. 17 shows a flowchart for an example of the decoding mode detection operations of the decoding mode detector implemented by the decoding mode detector shown in FIG. 16 and an example of the decoding operation of the channel decoder implemented by the channel decoder shown in FIG. 15.

The first operation in the decoding mode is performed by checking whether the determined decoding mode is mode 1. First, the syndromes of the code word are calculated, and when the calculated syndromes disappear, i.e. calculated syndrome without value (S30), the hash value is calculated and evaluated (S31). Thus, if the decoding mode is mode 1, no error should occur, and therefore the syndrome has a value of zero. When the calculated syndrome has a value, the first operation in the decoding mode ends and proceeds to the second operation (S38) in the encoding mode. When the calculated hash value is not equal to the included hash value (received hash value) in the codeword (S34), the first decode mode operation ends and proceeds to the second decode mode operation (S38). When the hash values are the same (S34), the first decoding mode detector 30 generates a decoding mode indicator (S36), and the controller 34 performs the previous additional steps to output decoding data (S82).

Fig. 18 shows a block diagram illustrating an example of the second decoding mode detector 32 of the decoding mode detector implemented by the decoding mode detector shown in FIG. 16. The second decoding mode detector 32 comprises a syndrome calculation module 40, a syndrome coloration module 42, a syndrome check module 44, an error locator polynomial calculation module 46, a risk value calculation module 48, an error position calculation module 50, and an error symbol calculation module 52. In operation in the second decoding mode, the syndrome staining module 42 applies a staining sequence to decolorize the calculated syndrome, since the staining execution for codewords or syndromes is essentially the same.

Fig. 19 shows a flowchart for an example of a second decoding mode operation procedure implemented by the second decoding mode detector 32 shown in FIG. 18. As shown in FIG. 19, the syndrome is calculated in the syndrome calculation unit 40 (S40), and the syndrome is stained in the syndrome staining unit 42 (S42). For example, the syndromes are stained with the syndromes of the coloring sequence associated with the next candidate mode in the list of possible variants, for example, the coloring sequence sig ₂ associated with mode 2. word before computing its syndromes, but it is computationally less complex. The discolored syndrome is checked, i.e. the syndrome value is checked in the syndrome checking unit 44 (S44). If all syndromes disappear for one mode, then that mode is selected directly, without checking additional modes. When the discolored syndrome is significant, an error occurs (S46), and therefore, error locator polynomials (ELP) are calculated in error locator polynomial calculation unit 46 (S48). In case ELP does not exist (S50), i.e. ELP cannot be calculated, the decoding mode is excluded from the candidate list (S52), and if there is any mode not yet checked (S58), return to step S40 is performed. For example, there is no ELP for the discoloration syndrome applied to the sig ₂ coloring sequence, and then decoding mode 2 is excluded from the list of options. When the ELP exists (S50), the risk value is calculated in the risk value calculation unit 48 (S54). When the calculated risk value exceeds the predetermined threshold value (S56), the mode is excluded from the candidate list (S52). When the risk value is less than the predetermined threshold value (S56), it is checked whether there are modes that have not yet been tested in the candidate list (S58). If there are modes in the candidate list (S58) that have not yet been checked, then the process returns to step S40. If all modes in the candidate list are checked (S58), the error position is calculated by the error position calculation unit 50 (S60). When the error is not correctable based on the calculated error position (S62), the decoding mode is excluded from the candidate list (S52). If the error is correctable (S62), an error symbol is calculated by the error symbol calculation unit 52 of error symbols (S64), and a decoding mode indicator is generated (S66).

Then, as illustrated in FIG. 17, the error is corrected by the error correction unit 62 (S68). When error correction is not successful, a frame including a codeword having an uncorrectable error is registered as a bad frame (S70). When error correction is successfully performed, the codewords are colored based on the decoding mode indicator by the coloring unit 22 (S72). Then, the colored (bleached) codewords are decoded based on the decoding mode indicator by the redundant decoder 24 (S74), and the data words are concatenated by the data unit 64 (S76). The hash value of the concatenated data is calculated and compared with the included hash value to evaluate the hash value (S78). When the hash values match (S80), the decoded data is output. When the hash values do not match (S80), the decoded frame is registered as a bad frame (S70).

Fig. 21 shows a block diagram illustrating changes of the decoding mode detector of the channel decoder located in the channel decoder shown in FIG. 20. FIG. 21 shows a circuit diagram that illustrates the operation performed by the mode detector 26 in FIG. 20. Thus, mode detector 26 in FIG. 20 includes a first decoding mode detector performing step 1 (first operation in decoding mode), a second decoding mode detector performing step 2 (second operation in decoding mode), a mode selection unit, and an error correction unit.

First Mode Detection Operation (Stage 1)

Fig. 22 shows a flowchart of step 1. As shown in FIG. 22, the first stage mode detection is performed to check if the mode m _fec FEC is equal to 1. The following notation is used to denote the i-th code polyline in GF(16)[x] corresponding to the i-th codeword:

(34)

(34)

The choice of FEC 1 mode depends on two conditions. First, the first two syndromes must disappear, i.e.:

(35)

(35)

and second, the recomputed hash value must match the received hash value. If both conditions are met, a decoding indicator is generated to indicate that mode 1 is the determined decoding mode, and data is decoded according to the data configurations in the encoder. If at least one of these conditions is violated, mode 1 is excluded from the list of possible mode options, and mode detection proceeds to stage 2.

Thus, as shown in FIG. 22, if both conditions are met, the first decoding mode detector notices that the determined decoding mode is mode 1 (as indicated by "is_mode_1" in FIG. 21). In this case, additional procedures, i.e. stage 2 and error correction are omitted.

Second mode detection operation (stage 2)

. Операция для режима 2 выполняется посредством модуля 1 обесцвечивания, модуля 1 проверки синдромов, модуля 1 вычисления полиномов локатора ошибок, модуля 1 оценки риска, модуля 1 вычисления положений ошибки и модуля 1 вычисления ошибочных символов. Операция для режима 3 выполняется посредством модуля 2 обесцвечивания, модуля 2 проверки синдромов, модуля 2 вычисления полиномов локатора ошибок, модуля 2 оценки риска, модуля 2 вычисления положений ошибки и модуля 2 вычисления ошибочных символов. Операция для режима 4 выполняется посредством модуля 3 обесцвечивания, модуля 3 проверки синдромов, модуля 3 вычисления полиномов локатора ошибок, модуля 3 оценки риска, модуля 3 вычисления положений ошибки и модуля 3 вычисления ошибочных символов.Fig. 23 shows a diagram of step 2 performed by the second decoding mode detector. As shown in FIG. 23, when a certain decoding mode is already selected in step 1, the process in step 2 is skipped. In addition, if the determined decoding mode is mode 1, it is assumed that no errors occur, and therefore no error correction is performed. In FIG. 23, three processes are illustrated, i.e. the syndromes of 6 codewords are calculated, the decoding mode detection operation for mode 2, for mode 3 and for mode 4 is illustrated. Thus, the syndromes for a given number of codewords are calculated, for example, for six codewords C ₀ -C ₅ , six syndromes for each codeword are calculated, i.e.

. The operation for mode 2 is performed by the bleaching unit 1, the syndrome checking unit 1, the error locator polynomial calculation unit 1, the risk assessment unit 1, the error position calculation unit 1, and the error symbol calculation unit 1. The operation for mode 3 is performed by the bleaching module 2, the syndrome checking module 2, the error locator polynomial calculation module 2, the risk assessment module 2, the error position calculation module 2, and the error symbol calculation module 2. The operation for mode 4 is performed by the bleaching module 3, the syndrome checking module 3, the error locator polynomial calculation module 3, the risk assessment module 3, the error position calculation module 3, and the error symbol calculation module 3.

. Проверка синдрома 1 содержит проверку, верно ли

для k=1, 2, 2*n. Результат теста отправляется в модуль выбора режима в качестве syndrome_check_1. Полином

локатора ошибок (ELP) вычисляется в модуле n вычисления ELP, т.е. в модуле 1 вычисления ELP для режима 2. Режим 2 (режим n+1) исключается из списка возможных вариантов (помещается в черный список в списке возможных вариантов), если

не существует для одного

Модуль n оценки риска, т.е. модуль 1 оценки риска для режима 2, вычисляет значение риска из степеней

согласно

. Вычисленное значение риска отправляется в модуль выбора режима и модуль 1 вычисления положений ошибки в качестве risk_value_1. Модуль n вычисления положений ошибки, т.е. модуль 1 вычисления положений ошибки для режима 2, вычисляет положения ошибки в режиме n+1, т.е. режиме 2 посредством разложения на множители

. Если разложение на множители завершается неудачно, или если положение ошибки выходит за границы, или если значение риска выше заданного порогового значения, режим 2 исключается из списка возможных вариантов(помещается в черный список). Модуль n вычисления ошибочных символов, т.е. модуль 1 вычисления ошибочных символов для режима 2, вычисляет символы ошибок в режиме 2 (n+1) из

, и положения err_pos_{i, n+1} ошибки.Fig. 24 shows an enlarged part of FIG. 23 indicating an operation for mode 2. In an operation for mode n+1, that is, for mode 2, the bleaching modulus n, i. e. decolorization module 1 decolorizes the computed syndrome according to mode 2:

. Syndrome check 1 contains a check to see if the

for k=1, 2, 2*n. The result of the test is sent to the mode selection module as syndrome_check_1. Polynomial

error locator (ELP) is computed in module n of the ELP computation, i.e. in ELP calculation module 1 for mode 2. Mode 2 (mode n+1) is excluded from the list of possible options (placed in the black list in the list of possible options) if

does not exist for one

Risk assessment module n, i.e. risk assessment module 1 for mode 2, calculates the risk value from the degrees

according to

. The calculated risk value is sent to the mode selection module and the error position calculation module 1 as risk_value_1. Module n of calculating error positions, i.e. the error position calculation module 1 for mode 2 calculates the error positions in mode n+1, i.e. mode 2 by factoring

. If the factorization fails, or if the error position is out of bounds, or if the risk value is above a predetermined threshold, mode 2 is excluded from the list of options (blacklisted). Module n for calculating error symbols, i.e. error symbol calculation module 1 for mode 2, calculates error symbols in mode 2 (n+1) from

, and err_pos _{i, n+1} error positions.

As shown in FIG. 23, the same operation as that performed for mode 2 is also performed for mode 3 and mode 4. Then, data indicating the position of the error, err_pos _i,2 , and the error symbol data pointer err_symb _i,2 , are sent from the error calculation unit 1. characters in the switch. In addition, err_pos _i,3 and err_symb _i,3 are sent from the error symbol calculation module 2 to the switch, and err_pos _i,4 and err_symb _i,4 are sent from the error symbol calculation module 2 to the switch. The mode selection module selects a specific decoding mode according to the following steps:

1. If it is indicated that "is_mode_1", then mode 1, if the procedure fails, then check "is_mode_n", n=2, 3, 4; exit,

2. If syndrome_check=true for n, then m _fec =n+1 (no error) is chosen,

3. If all modes are blacklisted, no mode is selected, i.e. is_mode_n<0, and

4. Of the modes that remain on the list of options (not blacklisted), choose m _fec for which risk_val_(m _fec -1) is minimal.

The switch then switches between inputs according to m _fec (the output is irrelevant if no mode is selected). In the error correction module, if is_mode_1, i.e. m _fec =1, then output=input because no error occurs. Otherwise, the error correction module corrects the symbols indicated by err_pos _{i, mfec} by XORing with err_symb _{i, mfec} . The detailed processes are explained below.

At stage 2, the list of possible options for modes is further reduced in several stages. The procedure for selecting a specific decoding mode ends as soon as a valid mode is found, or if a valid mode does not remain in the list of possible options. In the second case, decoding is stopped and the frame is marked as a bad frame.

In step 2, only codewords C ₀ -C ₅ are considered for mode detection. In this embodiment, the number of codewords is six; however, this number may differ, for example, depending on the circumstances of the transmission channel, transmission environment, required security strength, and/or codec performance.

First, the syndromes:

(36)

(36)

, m=2, 3, 4 и

, т.е. для первых по меньшей мере шести передаваемых кодовых слов. Следует заметить, что:colored code words are computed for

, m=2, 3, 4 and

, i.e. for the first at least six transmitted codewords. It should be noted that:

, (37)

, (37)

where sig _k (a ^k )) can be tabulated. Thus, the syndromes of colored codewords C _i,m can be computed efficiently by coloring the syndromes of codewords C _i . This consideration of all modes does not increase the complexity of the worst-case calculation of syndromes.

для

, то режим m выбирается. Следует отметить, что это всегда имеет место, если ошибки при передаче не возникают, и что посредством выбора sig_k, такое m является обязательно уникальным.If m exists, so

For

, then the mode m is chosen. It should be noted that this is always the case if no transmission errors occur, and that by choosing sig _k , such m is necessarily unique.

Otherwise, transmission errors occur and mode detection proceeds to compute error locator polynomials for the remaining modes. At this stage, polynomials are provided:

(38)

(38)

, и где коэффициенты являются уникальным решением для:Where

, and where the coefficients are the unique solution for:

(39)

(39)

. Такой полином не обязательно существует, если кодовое слово i не является корректируемым в режиме m, и если встречается кодовое слово, для которого не может обнаруживаться такой

, режим m исключается из списка возможных вариантов режимов.For

. Such a polynomial does not necessarily exist if codeword i is not modifiable in mode m, and if a codeword is encountered for which no such

, mode m is excluded from the list of possible modes.

обнаруживается, значение риска вычисляется для всех m оставшихся режимов согласно:Otherwise, i.e. such

is detected, the risk value is calculated for all m remaining modes according to:

(40)

(40)

, поскольку L_i изменяется от 13 до 15 для рассматриваемых размеров интервалов. Режимы, для которых ρ_m превышает данное пороговое значение, исключаются из последующей обработки. Следует отметить, что пороговое значение, например, составляет 6×10^-8; тем не менее, это значение может изменяться в зависимости от требуемого качества передачи и других требуемых критериев.The length 14 of the code word is hereby used as an estimate for the real risk

, since L _i varies from 13 to 15 for the considered interval sizes. Modes for which ρ _m exceeds this threshold value are excluded from further processing. It should be noted that the threshold value is, for example, 6×10 ^-8 ; however, this value may vary depending on the desired transmission quality and other required criteria.

For the remaining modes (usually only one mode remains at this stage), the error locator polynomials are factorized in GF(16)[x]. Factoring is said to succeed if the polynomial is decomposed into various linear coefficients of the following form:

(41)

(41)

является корректируемым в допустимое кодовое слово

посредством модификации

для

. Все режимы, для которых по меньшей мере один полином локатора ошибок не разбивается согласно формуле (41), исключаются из списка возможных вариантов режимов. Если несколько режимов остаются в списке возможных вариантов, определенный режим m_fec декодирования выбирается таким образом, что риск ρ_mfec при декодировании является минимальным.If this is the case, then

is correctable into a valid code word

through modification

For

. All modes for which at least one error locator polynomial is not partitioned according to formula (41) are excluded from the list of possible modes. If several modes remain in the list of possible options, a particular decoding mode m _fec is chosen such that the decoding risk ρ _mfec is minimal.

. Если коррекция завершается неудачно, кадр помечается в качестве плохого кадра, и декодирование завершается. В противном случае, данные декодируются после выполнения обесцвечивания согласно выбранному режиму в избыточном декодере 24 согласно компоновкам данных кодера в режиме m_fec FEC.If the mode is successfully detected, codeword error correction is performed.

. If the correction fails, the frame is marked as a bad frame and decoding ends. Otherwise, the data is decoded after performing bleaching according to the selected mode in the redundant decoder 24 according to the encoder data arrangements in the m _fec FEC mode.

If m _fec >1, the decoded data is validated by recalculating the hash value and compared with the hash value included in the transmitted codewords in the post processor 68. If the validation succeeds, the decoded data is output according to the data layout in the encoder. Otherwise, the bad frame is signaled to channel encoder 2.

When a decoding mode is selected, a decoding mode indicator indicating the determined decoding mode is generated and sent to the decolorization unit 22, the RS decoder 24, and the post processor 68, as shown in FIG. 20. In addition, as shown in FIG. 20, the decolorizer 22 decolorizes according to the m _fec mode by XORing the first 6 codewords with sig _mfec(k) . The RS decoder 24 extracts only parts of the codeword data, i.e. error correction is performed earlier for colored codewords. The decoding mode indicator is also sent to the RS decoder 24 because it is necessary to know the specific decoding mode to identify the amount of redundancy. A decoding process for decoding according to a certain decoding mode in the manner disclosed in "Error Correction Coding: Mathematical Methods and Algorithms", Todd K. Moon, 2005. The data fusing module 64 concatenates the input to obtain an output, and the post processor 68 validates the data, if m _fec >1, and deletes the hash data.

According to embodiments of the present application, a channel encoder indicates an encoding mode by applying a coloring sequence to a frame codeword. Therefore, it is not necessary to separately transmit data to indicate a certain coding mode and required parameters, and thereby the data transmission rate is improved. In addition, the coding mode information/indicator is included in the codeword by applying a coloring sequence which is selected according to the coding mode, and thereby error-resistant signaling of the mode information can be achieved. In addition, the coding mode information/indicator is distributed in the codeword by applying a coloring sequence, and therefore the coding mode information/indicator is received in the channel decoder in an error-tolerant manner. In addition, the channel decoder is able to determine the decoding mode without separately receiving specific information regarding the decoding mode and parameters so as to determine the decoding mode. Therefore, the transmission ratio of the channel is effectively improved.

According to embodiments of the present application, a channel decoder performs test decoding to analyze whether an error occurs to detect a decoding mode. Therefore, in the case where a transmission error does not occur, a reliable decoding mode is determined by a simple calculation.

According to embodiments of the present application, if a transmission error occurs, the channel decoder performs error correction for a given number of codewords as a test, and calculates an error risk value (reliability index). Therefore, without receiving specific information and parameters from the channel encoder, it is possible to determine the appropriate decoding mode by checking the predetermined number of codewords and considering the reliability index.

According to embodiments of the present application, the decoding mode indicator comprises a decoding mode detector for detecting a decoding mode by checking a predetermined number of codewords to output a decoding mode candidate in a decoding mode candidate list. The candidates in the candidate list are excluded or blacklisted based on the occurring errors, and the determined decoding mode is finally selected from the remaining decoding mode candidates in the candidate list, taking into account the reliability score (risk value). In such a case, the decoding mode indicator includes a risk value of the selected decoding mode, and if the error risk exceeds a predetermined threshold, the frame is registered as a bad frame. Therefore, it is possible to select a reliable and proper decoding mode by checking only a predetermined number of codewords.

Further embodiments are described below.

Forward error correction at the application layer

1. Channel encoder

1.1. Functions and definitions

1.2. General parameters of the channel encoder

1.2.1. FEC mode

The FEC mode m is a number from 1 to 4, where m=1 provides only basic error protection, and m=2, 3, 4 provides enhanced error correction performance. In a channel encoder, the FEC mode is denoted by m _fec , and in a channel decoder it is denoted by n _fec .

1.2.2. Interval size

The slot size N _s indicates the size of the channel-encoded frame in octets (bytes). N _s can take on all integer values from 40 to 300 covering nominal bit rates from 32 to 240 kbps at a frame rate of 100 Hz.

1.2.3. CMR

The coding mode request CMR is a two-bit character represented by the numbers 0-3.

1.2.4. Combined channel coding flag

The jcc_flag of the combined channel coding takes the values 0 and 1 and indicates that the input data contains data from multiple audio channels.

1.3. Extracted channel encoder parameters

1.3.1. Number of code words

The N _cw parameter specifies the number of codewords that are used to encode the data frame. It is extracted from the interval size by:

1.3.2. Codeword lengths

Parameter L _i is set for i=0...N _cw -1 and indicates the length of the i-th code word in semi-octets. It is extracted from the size N _s of the interval as follows:

1.3.3. Hamming distances

Parameter d _i , _m indicates the Hamming distance of codeword i in FEC mode m. It is set as follows:

and for m>1, as follows:

1.3.4. Number of partial masking codewords

The parameter N _pcww indicates the number of partial concealment codewords and is derived from the interval size N _s and the FEC mode m as follows:

1.3.5. Partial mask block size

The parameter N _pc specifies the size of the partial concealment block in semi-octets and is derived from the interval size N _s and the FEC mode m as follows:

1.3.6. CRC hash sizes

The numbers N _CRC1 and N _CRC2 , which correspond to the sizes of the CRC hash values, are extracted from the interval size and mode m FEC by:

1.3.7. Working data size

The payload data size N _p indicates the data size in a channel-coded frame of size N _s encoded in m FEC mode in octets, which is given by:

1.4. Algorithmic description of the channel encoder

1.4.1. Input Output

из октетов и флаг jcc+flag объединенного канального кодирования и возвращает последовательность октетов

. Октеты интерпретируются в качестве чисел от 0 до 255 согласно указанному порядку байтов.The channel encoder takes as input the slot size N _s , the FEC mode m _fec , the CMR request for the coding mode, the data sequence, for example,

of octets and the jcc+flag flag of the combined channel coding and returns a sequence of octets

. Octets are interpreted as numbers from 0 to 255 according to the specified byte order.

1.4.2. Data preprocessing

с обратным упорядочением, где a_n(2k) справедливо для верхней половины,

и a_n(2k+1) справедливо для нижней половины. В формулах, это представляет собой:The data sequence is first split into sequence

with reverse ordering, where a _n (2k) is valid for the upper half,

and a _n (2k+1) is true for the bottom half. In formulas, this represents:

And:

Then, the CRC hash values are computed for the bit extensions of the sequences:

And:

Note that N _pc may be zero, in which case an ₂ is the empty sequence. The bit extension of a sequence of (0...n-1) semi-octets is given by the sequence b (0...4N-1), where:

where k changes from 0 to N-1, and j changes from 0 to 3, and bit _j is a function that returns the j-th bit according to the specified byte order.

The first CRC hash sequence computed for the extension a _n1 has length 8N _CRC1 -2 and the binary generator polynomials are given by:

And:

The second CRC hash sequence computed for an _2' has length 8N _CRC2 , and the binary generator polynomial is given by:

A CRC hash sequence of length k for a sequence b of binary data (0...n-1), since the given binary generator polynomial p(x) of degree k is given as usual such that it is the binary sequence b (0.. .k-1), so the binary polynomial is:

is divisible by p(x).

задается в качестве хэш-последовательности длины 8N_CRC1-2, вычисленной для конкатенированной последовательности:Let b _n1 denote the bit extension of a _n1 , and let b _n2 denote the bit extension of a _n2 . Then the sequence

is given as a hash sequence of length 8N _CRC1 -2 computed for the concatenated sequence:

задается в качестве второй хэш-последовательности длины 8N_CRC2, вычисленной для b_n2. Следует отметить, что r_n2 представляет собой пустую последовательность, если N_CRC2=0.Besides,

is given as the second hash sequence of length 8N _CRC2 computed for b _n2 . It should be noted that r _n2 is an empty sequence if N _CRC2 =0.

данных затем задается посредством:First preprocessing sequence

data is then given by:

And:

и k+2, т.е.:The final preprocessed data sequence is specified by swapping the CMR bits at positions 8N _CRC1 -2 and 8N _CRC1 -1 with bits at positions

and k+2, i.e.:

And:

for n other than 8N _CRC1 -2, 8N _CRC1 -1, k and k+2. Swapping these bits ensures that the CMR bits are stored in EC mode-independent bit positions located in the middle of the channel-coded bitstream.

полуоктетов посредством инвертирования битового расширения, т.е.:The bit sequence b _pp is converted to the sequence

semi-octets by inverting the bit extension, i.e.:

It should be noted that it is not necessary to actually perform the bit extensions described in this section, since the CRC hash can be computed efficiently for blocks of data.

1.4.3. Reed-Solomon coding

For Reed-Solomon (RS) encoding, the pre-processed data sequence a _pp from section 1.4.2 is split into N _cw D _i sequences, also called "data words", according to:

,

,

.where i varies from 0 to N _cw -1, and where the points Si of the partition are inductively given by S ₀ =0 and

.

, причем остаточный класс x в

выбирается в качестве генератора единичных групп, обозначаемого, как обычно, посредством a.RS codes are constructed according to

, and the residual class x in

is chosen as the identity group generator, denoted as usual by a.

Semi-octets are converted to GF(16) elements using data-to-character conversion:

, так что

The transformation is one-to-one and the inverse transformation is denoted as

, So

Under this notation, the generating Reed-Solomon polynomials for 3, 5, and 7 Hamming distances are given by:

And:

для слов D_i данных вычисляются. Они представляют собой (уникально определенные) последовательности полуоктетов, так что полином:For i ranging from 0 to N _cw -1, the redundant RS sequences

for data words D _i are calculated. They are (uniquely defined) sequences of semi-octets, so the polynomial is:

; i-ое кодовое слово C_i затем задается таким образом, что оно представляет собой последовательность Li полуоктетов, заданную посредством:divided by

; The i-th codeword C _i is then set such that it is a sequence Li of semi-octets given by:

And:

Note that if d _{i, mfec} =1, the redundant RS sequence is empty and C _i is simply equal to D _i

1.4.4. Mode information signaling

The m _fec FEC mode is transmitted implicitly, and is instead signaled implicitly by coloring the first 6 codewords with mode dependent coloring sequences, i.e.:

where bitxor(a, b) denotes the bitwise XOR operation for two semi-octets. Signal sequences sig _m are defined as follows:

It should be noted that codeword coloring leaves the CMR bits at bit positions 30 and 32 of C ₀ unchanged.

1.4.5. Codeword Multiplexing

The CC _i sequences are multiplexed into a sequence of octets first by semi-octet interleaving according to:

,

,

where i changes from 0 to N _cw -1 and k changes from 0 to Li-1, and then by pairing successive half-octets according to:

where k varies from 0 to N _s -1.

1.5. Algorithmic description of the channel decoder

1.5.1. Input Output

октетов и флаг jcc_flag объединенного канального кодирования и возвращает размер N_p рабочих данных, последовательность декодированных октетов

, указатель bfi плохого кадра, принимающий значения 0, 1 и 2, оценку CMR XNI, принимающую значения из 0-11, число error_report, принимающее значения от-1 до 480 (указывающее число скорректированных битов, если bfi=0), и указатель fbcwbp положения бита для частичной маскировки.The channel decoder takes as input the slot size N _s and the sequence

octets and the jcc_flag of the joint channel coding and returns the size N _p of payload data, the sequence of decoded octets

, a bad frame pointer bfi taking the values 0, 1 and 2, an XNI CMR score taking values from 0-11, an error_report number taking values from -1 to 480 (indicating the number of corrected bits if bfi=0), and a pointer fbcwbp bit positions for partial masking.

указываются только в том случае, если bfi≠1, и значение указателя fbcwbp положения бита указывается только в том случае, если bfi=2.The value of N _p and

are indicated only if bfi≠1, and the value of the bit position indicator fbcwbp is indicated only if bfi=2.

1.5.2. Demultiplexing codewords

From the interval size N _s , the extracted parameters N _cw and L _i are calculated according to sections 1.3.1 and 1.3.2. The input sequence z _cc (0...N _s -1) is then split into a sequence z _ij (0...2N _s -1) of semi-octets according to:

And:

for k=0...N _s -1, and codewords XX _i are extracted according to the data arrangements of section 1.4.5, i.e.:

where i changes from 0 to N _cw -1 and k changes from 0 to Li-1.

1.5.3. Mode detection

Mode detection aims to recover the m _fec FEC mode by parsing codewords XX _i , where i ranges from 0 to 5. The detected mode is denoted n _fec and takes values from 0 to 4, where 0 indicates that the mode is not detected. Once a mode is detected, all retrieved codec parameters such as payload data size, number of partial masking codewords, etc. are set according to that mode. The mode is selected from a list of possible mode options, initially containing FEC modes 1-4, which is then narrowed down step by step.

1.5.3.1. Stage 1

In step 1, an attempt is made to determine if the frame is encoded in FEC mode 1. To do this, the first two syndromes of codeword 0 are calculated, with the k-th syndrome of codeword XX _i set as the symbol GF(16) given by:

Mode 1 is selected if the following two conditions are met:

и

And

Data extracted according to framing m _fec =1 pass first CRC as specified in section 1.5.7 with

If these conditions are met, error_report and bfi are set to 0 and the channel decoder outputs data z _dat (0...N _p -1) generated in section 1.5.7. Otherwise, mode detection proceeds to stage 2 and mode 1 is removed from the list of options.

1.5.3.2. Stage 2

вычисляются для i=0...5 и k=1...6.In step 2, an attempt is made to determine if the frame is encoded in FEC modes 2, 3, or 4. To do this, the syndromes

calculated for i=0...5 and k=1...6.

условия:If for one

conditions:

выбирается, и канальный кодер переходит к разделу 1.5.6. Следует отметить, что такой m обязательно является уникальным, так что режимы могут проверяться в любом порядке.are performed for i=0...5 and k=1...d _{i, m} -1, i.e. all syndromes colored according to mode m disappear, then

is selected and the channel encoder proceeds to section 1.5.6. Note that such an m is necessarily unique, so the modes can be tested in any order.

локатора ошибок для i=0...5 и m=2...4. Это проводится согласно разделу 1.5.5.1.1 с

и:If such m cannot be found, then mode detection computes the polynomials

error locator for i=0...5 and m=2...4. This is done according to section 1.5.5.1.1 with

And:

colored syndromes according to mode m, for k=1...2t.

по меньшей мере для одного i от 0 до 5 исключаются из дальнейшего рассмотрения.All modes m for which

for at least one i from 0 to 5 are excluded from further consideration.

локатора ошибок и вычисляется в качестве пары "мантисса-экспонента":For the remaining modes, the risk value is calculated. The risk value rsk(m) for mode m is based on the powers of the polynomials

error locator and is calculated as a mantissa-exponent pair:

"мантисса-экспонента" указываются в таблице 2, и где умножение двух пар "мантисса-экспонента" задается следующим образом: С учетом двух пар

и

, где

, произведение

задается в качестве пары

, заданной посредством:where are the couples

"mantissa-exponent" are specified in Table 2, and where the multiplication of two pairs of "mantissa-exponent" is given as follows: Considering two pairs

And

, Where

, work

given as a pair

, given by:

"мантисса-экспонента" соответствует числу

.Such a couple

"mantissa-exponent" corresponds to a number

.

Tab. 2

\

\

0 1 2 3 1 (only for i=0) (16384, -8) (26880, -1) N/A N/A 2 (16384, -8) (26880, -1) N/A N/A 3 (16384, -16) (26880, -9) (20475, -2) N/A 4 (16384, -24) (26880, -17) (20475, -10) (19195, -4)

All modes m for which the corresponding risk value rsk(m) is above the interval-dependent threshold value risk_thresh are removed from the list of possible mode options. The risk threshold is given as:

, либо

и

является справедливым.The remaining modes with a risk value less than or equal to risk_thresh are listed as m _j , j=1...n, so for each j=1...n-1 either

, or

And

is fair.

для i=0...5. Если вычисление положения ошибки завершается удачно для всех кодовых слов, то n_fec=m_j выбирается, и канальный декодер переходит к разделу 1.5.5. В противном случае, если положения ошибки не могут определяться для одного индекса, такая же процедура выполняется для режима m_j+1 при J<n. В противном случае, n_fec задается равным 0, что указывает то, что режим не обнаружен.Starting from mode m ₁ , the error positions nm _{j, i, k} in codewords XX _i are determined according to Section 1.5.5.2 with

for i=0...5. If the error position calculation succeeds for all codewords, then n _fec =m _j is chosen and the channel decoder proceeds to section 1.5.5. Otherwise, if the error positions cannot be determined for one index, the same procedure is performed for mode m _j +1 with J<n. Otherwise, n _fec is set to 0, which indicates that the mode is not detected.

задается равным -1, обнаружение CMR выполняется согласно разделу 1.5.4 с

до того, как канальный декодер выходит с bfi=1.If the mode is not detected, i.e.

is set to -1, CMR detection is performed according to section 1.5.4 with

before the channel decoder exits with bfi=1.

1.5.4. CMR estimate when frame is not decodable

локатора ошибок для всех режимов

где M является данным набором возможных вариантов режимов.In case the frame is not decodable, the CMR is estimated by analyzing the first codeword XX ₀ and the corresponding polynomials

error locator for all modes

where M is the given set of possible modes.

First, all modes are removed from M for which either:

локатора ошибок не является достоверным, либо:polynomial

error locator is not valid, either:

значений риска, как указано в таблице 2, превышает -8.exhibitor

risk values, as indicated in table 2, exceeds -8.

The set of remaining modes is denoted as M ₁ .

If M1 is empty, the CMR XNI score is given by:

here, the term 8 indicates that this value fails the validation check.

значений риска является минимальной (следует отметить, что такой режим всегда существует, поскольку

и

не могут оба иметь значение -8 согласно расчетам последовательностей служебных сигналов). Затем коррекция ошибок выполняется для XX₀ согласно разделу 1.5.5 с n_fec=m, и оценка CMR извлекается из скорректированного кодового слова

как либо:If M1 is not non-empty, then let m denote the element of M ₁ for which the exponent

risk values is minimal (it should be noted that such a regime always exists, since

And

cannot both be -8 according to signaling sequence calculations). Error correction is then performed for XX ₀ according to section 1.5.5 with n _fec =m and the CMR estimate is extracted from the corrected codeword

either way:

, где слагаемое 4 указывает, что CMR проходит проверку достоверности со средним высоким доверием, либо:If

, where the term 4 indicates that the CMR passes validation with medium high confidence, or:

, что указывает, что значение CMR проходит проверку достоверности с высоким доверием.If

, indicating that the CMR value passes a high-confidence validation check.

1.5.5. Error Correction

Full error correction is performed only when the mode is successfully detected. In this case, the error positions for n _{nfec, i,k} for the first 6 codewords have already been calculated in section 1.5.3.2. Error correction may also be performed on only the first codeword for CMR recovery. In this case, the following steps are only performed for i=0.

ошибок согласно разделу 1.5.5.3 при:Codewords XX _i with i≤5 are corrected by calculating symbols

errors according to section 1.5.5.3 when:

, заданным в разделе 1.5.3.2, и:

specified in section 1.5.3.2, and:

The adjusted codewords are then given as follows:

является обратным преобразованием данных в символы, указываемым в разделе 1.4.3.Where

is the inverse data-to-character conversion specified in section 1.4.3.

For the remaining codewords with index i>5, error correction is performed by performing the usual steps:

Syndromes are calculated according to:

.for k=1...2t with

.

задается равным XX_i.If all syndromes are zero, the codeword is assumed to be error-free, and therefore the corrected codeword

is set to XX _i .

локатора ошибок вычисляется согласно разделу 1.5.5.1.1.Otherwise, the polynomial

error locator is calculated according to section 1.5.5.1.1.

, положения ошибок v_k, k=0...d-1 с

вычисляются согласно разделу 1.5.5.2.If successful (i.e.

, error positions v _k , k=0...d-1 s

calculated according to section 1.5.5.2.

ошибок вычисляются согласно разделу 1.5.5.3, и коррекция ошибок выполняется согласно:If successful, symbols

errors are calculated according to section 1.5.5.3 and error correction is performed according to:

If error correction fails for index i<N _cw -N _pcww , i.e. one of steps 3, 4, or 5 fails, then the bad frame pointer bfi is set to 1, error_report is calculated as follows, and channel decoding is terminated.

значений риска, как указано в таблице, превышает -16, значение T(i) задается равным 0, указывая то, что данные в кодовом слове XXi не являются надежными без дополнительной проверки достоверности. Если коррекция ошибок завершается неудачно, скорректированное кодовое слово

, тем не менее, задается как XX_i, но первый указатель bfi₀ плохого кадра задается равным 2.For indices i≥N _cw -N _pcww , the sequence T(N _cw -N _pcww ...N _cw -1) is given as follows. If error correction fails for index i<N _cw -N _pcww , or if the exponent

risk values as indicated in the table is greater than -16, the value of T(i) is set to 0, indicating that the data in codeword XXi is not reliable without further validation. If error correction fails, the corrected codeword

, however, is set to XX _i , but the first bad frame pointer bfi ₀ is set to 2.

. Иначе, пусть I обозначает набор всех индексов 0<i<N_cw, для которых коррекция ошибок выполнена успешно. Значение error_report затем вычисляется следующим образом:The error_report value is set as follows. If error correction fails for index i<N _cw -N _pcww, then let i ₁ be the smallest index for which it fails, and run

. Otherwise, let I denote the set of all indices 0<i<N _cw for which error correction is successful. The error_report value is then calculated as follows:

,

,

.those. as the total number of bits corrected in XX _{i codewords,} s

.

If N _s =40, the number of bit corrections is artificially reduced to increase error detection. If all codewords are corrected successfully, the first bad frame indicator is set depending on the mode dependent error threshold emax _m given as follows:

, то первый указатель bfi₀ плохого кадра задается равным 0, и иначе он задается равным 1.If

, then the first bad frame pointer bfi ₀ is set to 0, and otherwise it is set to 1.

If N _s >40 and all codewords are corrected successfully, then the first bad frame indicator bfi ₀ is set to 0.

1.5.5.1.1. Calculation of error locator polynomials

, символов в GF(16), где t является числом от 1 до 3.The error locator polynomial is computed from the sequence

, characters in GF(16), where t is a number between 1 and 3.

для k=1...2t, полином

локатора ошибок задается равным [1].If

for k=1...2t, polynomial

error locator is set to [1].

Otherwise, the determinants of the matrices M _l are calculated for l=1...t, where:

And:

локатора ошибок задается равным [0], который является недопустимым полиномом локатора ошибок в смысле 1.5.5.2.If all determinants are [0] for l=1...t, the polynomial

Error Locator is set to [0], which is an invalid error locator polynomial in the sense of 1.5.5.2.

. Затем коэффициенты полинома локатора ошибок вычисляются следующим образом:Otherwise, τ is taken as the largest index from 1 to t such that

. The error locator polynomial coefficients are then computed as follows:

where the inverse matrices are given as follows:

, полином локатора ошибок задается равны м[0].If

, the error locator polynomial is given equal to m[0].

Otherwise, if τ=t, the error locator polynomial is set to:

and if τ<t, it is additionally checked whether:

for n=0...2 (t-τ)-1. If all these equalities are true, then the error locator polynomial is given by:

Otherwise, it is set to [0].

1.5.5.2. Calculating Error Positions

The error positions are computed from the error locator polynomial:

An error locator polynomial is said to be valid if it can be represented:

wherein in this case the error positions are given by n _k for k=0...d-1. Otherwise, the list of error positions is empty.

для n=0...Li-1. В качестве альтернативы, табулирование местоположений ошибок, индексированных посредством

, является возможным и может быть значительно быстрее.The values of n _k can be determined by checking

for n=0...Li-1. Alternatively, tabulating error locations indexed by

, is possible and can be significantly faster.

1.5.5.3. Calculation of erroneous characters

и положений

ошибок посредством решения линейной системы:Error symbols computed from syndromes

and regulations

errors by solving a linear system:

представляют собой матрицы Вандермонда:according to GF(16), where

are the Vandermonde matrices:

And:

Inverse matrices are given by:

And:

Where:

And:

1.5.6. Decolorization and RS decoding

The bleaching according to the detected n _fec FEC mode is performed by applying the appropriate signaling sequence from section 1.4.4, causing the bleached codewords:

Then redundant decoding is applied according to the n _fec mode to form data words:

,

,

which are combined into a sequence of z _pp (0...N _p -1) data, with N _p as specified in section 1.3.7 with m=n _fec , according to:

. После избыточного декодирования RS, декодер FEC переходит к постобработке данных раздела 1.5.7.for i=0...N _cw -1, where the split points Si are as specified in section 1.4.3. This results in a sequence of length

. After the redundant decoding of the RS, the FEC decoder proceeds to the post-processing of the data in Section 1.5.7.

1.5.7. Data post-processing

и битов в положениях 8N_CRC1-2 и k+2.Data post-processing is to remove and validate the hash and extract the CMR. The sequence z _pp from section 1.5.6 is expanded to the corresponding bit sequence y _pp , from which the sequence y _pp0 is extracted by inverting the bit permutation from section 1.4.2, i.e. bit swaps in positions 8N _CRC1 -2 and

and bits at positions 8N _CRC1 -2 and k+2.

и b_n2 из раздела 1.4.2, заданные посредством:The sequence y _pp0 is then split into sequences i _n1 , i _n2 , y _n1ext and y _n2 corresponding to

and b _n2 from section 1.4.2, given by:

And:

Two cyclic redundancy checks (CRCs) are performed for y _n1ext and y _n2 , performed by recomputing the hash sequences specified in section 1.4.2.

В противном случае, оценка CMR задается равной:If the computed 8N _CRC1 -2 excess bit sequence for y _n1ext specified in section 1.4.2 does not match i _n,1 , the bad frame indicator bfi is set to one and the CMR is evaluated according to section 1.5.4 with

Otherwise, the CMR score is set to:

If the first CRC is transmitted, and if bfi ₀ ≠2, the second CRC is performed to calculate the 8N _CRC2 hash sequence for y _n2 as specified in section 1.4.2. If the result does not match the sequence in i _n,2 , the bad frame indicator bfi is set to 2, indicating partial concealment data loss. If the first CRC is transmitted and bfi ₀ =2, then bfi is set to 2 without performing the second CRC.

If both CRCs are transmitted, the bad frame indicator bfi is set to 0, indicating that the decoded data is valid.

If bfi=2, the position fbcwbp of the first potentially corrupted bit in the partial concealment block is determined from the sequence T(N _cw -N _ccw ...N _cw -1) of section 1.5.5 as follows.

If index i does not exist, so that N _cw -N _ccw ≤i<N _cw , and so T(i)=1, or if T(N _cw -1)=0, then fbcwbp is set to 0. Otherwise , let i ₀ denote the largest index, so T(i)=1 for i0≤i<N _cw . Then fbcwbp is calculated as follows:

If bfi ₀ ≠1, the output zda _t is generated by inverting the pre-processing steps from section 1.4.2 by specifying:

for k=0...2N _p -N _pc -1

for k=0...N _pc -1

And:

For

Although some aspects are described in the context of equipment, it is obvious that these aspects also represent a description of the corresponding method, with the block or device corresponding to a method step or a feature of a method step. Likewise, aspects described in the context of a method step also provide a description of the associated block or element, or feature of the associated equipment. Some or all of the steps of the method may be performed by (or using) hardware such as, for example, a microprocessor, a programmable computer, or an electronic circuit. In some embodiments, one or more of the most important steps of the method may be performed by this equipment.

The data stream according to the invention may be stored on a digital storage medium or may be transmitted over a transmission medium such as a wireless transmission medium or a wired transmission medium such as the Internet.

Depending on certain implementation requirements, embodiments of the application may be implemented in hardware or in software. The implementation may be performed using a digital storage medium, such as a floppy disk, DVD, Blu-ray, CD, ROM, PROM, EPROM, EEPROM, or flash memory, having electronically readable control signals stored that interact (or are capable of interacting) with programmable computer system in such a way that the corresponding method is carried out. Therefore, the digital storage medium can be machine readable.

Some embodiments of the invention comprise a storage medium having electronically readable control signals that are capable of interacting with a programmable computer system such that one of the methods described herein is implemented.

In general, embodiments of the present application may be implemented as a computer program product with program code, wherein the program code is configured to perform one of the methods when the computer program product is executed by a computer. The program code may, for example, be stored on a computer-readable medium.

Other embodiments comprise a computer program for carrying out one of the methods described herein, stored on a computer-readable medium.

In other words, an embodiment of the method according to the invention is therefore a computer program having program code for carrying out one of the methods described herein when the computer program is running on a computer.

Therefore, a further embodiment of the methods of the invention is a storage medium (digital storage medium or computer-readable medium) containing a recorded computer program for carrying out one of the methods described herein. The storage medium, digital storage medium or recorded data medium is usually physical and/or permanent.

Therefore, a further embodiment of the method according to the invention is a data stream or sequence of signals representing a computer program for implementing one of the methods described herein. The data stream or signal sequence, for example, may be configured to be transmitted over a data connection, such as the Internet.

An additional embodiment includes processing means, such as a computer or programmable logic device, configured to perform one of the methods described herein.

An additional embodiment comprises a computer having a computer program installed to implement one of the methods described herein.

An additional embodiment according to the invention comprises equipment or a system capable of transmitting (eg, electronically or optically) a computer program for performing one of the methods described herein to a receiving device. The receiving device may, for example, be a computer, mobile device, storage device, or the like. The equipment or system, for example, may include a file server for transmitting a computer program to a receiving device.

In some embodiments, a programmable logic device (eg, a field programmable gate array) may be used to perform some or all of the functionality of the methods described herein. In some embodiments, a user-programmable gate array may interface with a microprocessor to perform one of the methods described herein. In general, the methods are preferably carried out by any hardware.

The apparatus described herein may be implemented using hardware, either using a computer, or using a combination of a hardware device and a computer.

The device described herein, or any components of the device described herein, may be implemented at least in part in hardware and/or software.

Claims (146)

1. Channel encoder for encoding a frame, containing: - a multi-mode redundant encoder for redundantly encoding a frame in accordance with a certain coding mode from a set of different coding modes, wherein the multi-mode redundant encoder is configured to encode a frame using each coding mode in the set, wherein the coding modes differ from each other with respect to the amount of redundancy added to the frame, wherein the multi-mode redundant encoder is configured to output an encoded frame including at least one codeword; And a coloring module for applying a coloring sequence to at least one codeword; wherein the coloring sequence is such that at least one bit of the codeword is changed by applying at least one of the coloring sequences, wherein the coloring module is configured to select a specific coloring sequence in accordance with a certain coding mode, - while the channel encoder additionally contains: - a data splitter for splitting a frame into a plurality of data words, wherein the multi-mode redundant encoder is configured to encode each of the plurality of data words according to a certain coding mode to obtain a plurality of code words, wherein the coloring module is configured to apply a specific coloring sequence to each codeword in a given number of codewords or in a given subset of a plurality of codewords. 2. The channel encoder according to claim 1, in which the multi-mode redundant encoder is configured to apply a certain coding mode in the previous frame, while the first coding mode has an associated coloring sequence, while the multi-mode redundant encoder is configured to receive a use indicator for the current frame a second coding mode having an associated additional coloring sequence, and wherein the coloring module is configured to apply the additional coloring sequence in the current frame, or bypassing the additional coloring with the coloring sequence in the current frame. 3. The channel encoder according to claim 1, further comprising: - a controller for providing coding criteria, wherein the coding criteria specifies a frame redundancy factor, wherein the multi-mode redundant encoder is configured to add redundancy to the frame in accordance with the redundancy factor specified by a certain coding mode and a variable or fixed target size of the encoded frame. 4. The channel encoder of claim 3, wherein the multi-mode encoder is configured to determine an encoding mode based on a desired data security strength, wherein the controller is configured to determine the required data protection strength based on the estimated error rate of the transmitted channel. 5. The channel encoder of claim 1, wherein the controller is configured to switch the specific coding mode used to encode the frame and generate coding mode information indicating the specific coding mode, wherein the multi-mode redundant encoder is configured to receive coding mode information and perform redundant coding to obtain at least one codeword in accordance with a certain coding mode indicated by the received coding mode information, wherein the coloring module is configured to receive a pointer to indicate a specific coloring sequence and apply said specific coloring sequence to said at least one code word. 6. The channel encoder of claim 5, wherein the coloring module is configured to receive a pointer to indicate a specific coloring sequence, either from a controller or from a multi-mode redundant encoder. 7. The channel encoder of claim 1, wherein the coloring module is configured to perform coloring on at least one codeword by computing a bitwise XOR for at least one codeword and a specific coloring sequence. 8. The channel encoder of claim 1, wherein the data splitter is configured to calculate the number of codewords based on the target size for the frame, wherein the length of the data word that is included in the codeword is changed based on the calculated number of codewords. 9. Channel encoder according to claim 1, containing: - an audio/video encoder for encoding audio/video frame data, wherein the audio/video encoder is configured to specify a set of audio/video frame data based on a specific mode. 10. Channel encoder according to claim 1, containing: - a preprocessor for computing a hash value in an audio/video frame, while: - the preprocessor is configured to concatenate the hash value and the audio/video frame. 11. The channel encoder of claim 1, wherein the data splitter is configured to split a frame, wherein at least one data word consists of at least part of a hash value and part of an audio/video frame. 12. Channel decoder for channel decoding of at least one transmitted code word, containing: - a coloring module for applying at least one coloring sequence to at least one transmitted codeword or to at least one transmitted error correction codeword so as to obtain at least one colored codeword, wherein the coloring sequence is such that at least one bit of the codeword is changed by applying at least one coloring sequence, and wherein at least one coloring sequence is associated with a certain decoding mode as a specific coloring sequence; a redundant decoder for redundantly decoding at least one colored codeword to obtain a decoded output codeword; And a decoding mode detector for generating a decoding mode indicator indicating a specific decoding mode to be used by the redundant decoder to obtain a decoded output codeword, wherein the decoding mode indicator is associated with at least one coloring sequence as the particular coloring sequence used to color the transmitted codeword, wherein the coloring module is configured to use, in addition to the coloring sequence, at least an additional coloring sequence, or the channel decoder is configured to bypass the coloring module in the additional decoding mode without coloring; wherein the redundant decoder is configured to redundantly decode at least one additional colored codeword colored using the additional coloring sequence, obtaining an additional decoded codeword, an additional colored codeword obtained from the transmitted codeword using the additional coloring sequence, or a codeword transmission words without coloring to obtain another additional decoded codeword, and wherein the redundant decoder is configured to output a reliability metric for a decoded codeword, an additional reliability metric for an additional decoded codeword, or another additional reliability metric for another additional codeword, wherein the decoding mode detector is configured to determine the decoding mode indicator based on the reliability metrics, and - at the same time, the redundant decoder is configured to receive the decoding mode indicator and output either a decoded code word, or an additional decoded code word, or another additional decoded code word as a decoded output code word, - at the same time, the coloring module is configured to perform a coloring operation with the same coloring sequence for a given number of transmitted code words to obtain a predetermined number of colored code words and perform an additional coloring operation with the same additional coloring sequence for an additional specified number of transmitted code words to obtain a predetermined number of additional colored code words, wherein the redundant decoder is configured to determine a reliability metric extracting a predetermined number of decoded codewords, an additional reliability metric extracting a predetermined number of additional decoded codewords, or another additional reliability metric extracting a predetermined number of decoded codewords. 13. The channel decoder of claim 12, wherein the redundant decoder comprises a bit reduction module for reducing the number of bits of at least one colored codeword and an error correction module for correcting an error in the colored codeword, or - the channel decoder further comprises an error correction module for correcting an error of at least one transmitted code word. 14. The channel decoder of claim 12, wherein the decoding mode detector is configured to compute a hash value of the decoded frame, compare the hash value of the decoded frame and the hash value associated with the transmitted codeword, and determine a decoding mode indicator based on the result of the comparison. hashes, wherein the staining module is configured to bypass the staining operation, and - at the same time, the redundant decoder is configured to receive the transmitted code word using the coloring sequence and perform the redundant decoding operation in accordance with the decoding mode indicator. 15. The channel decoder of claim 12, wherein the coloring module is configured to perform a plurality of coloring operations in parallel to at least obtain a colored codeword and obtain an additional colored codeword, wherein the redundant decoder is configured to perform in parallel a plurality of redundant decoding operations to at least obtain a decoded codeword, obtain an additional decoded codeword, and another additional codeword, - moreover, the channel decoder further comprises: - a controller for controlling the output of the redundant decoder, wherein the controller is configured to direct the redundant decoder to output a decoded codeword, an additional decoded codeword, or another additional decoded codeword as a decoded output codeword. 16. The channel decoder of claim 12, wherein the redundant decoder is configured to calculate a reliability metric based on the number of corrected symbols during a colored codeword decoding operation, an additional reliability metric based on the number of symbols adjusted during an additional decoding operation of the additional codeword, and a metric reliability based on the number of corrected symbols during another additional decoding operation of another additional codeword. 17. The channel decoder of claim 12, wherein the predetermined number of transmitted codewords is between 3 and 9. 18. The channel decoder of claim 15, wherein the controller is configured to: - directing the redundant decoder to select a specific decoding mode using a reliability metric, or - instructing the output interface to output the decoded codeword as the decoded output codeword. 19. The channel decoder of claim 18, wherein the controller is configured to cause the redundant decoder to select a specific decoding mode having the highest reliability score, or to cause the output interface to select, as the decoded output codeword, a decoded codeword from a group of decoded codewords further using different decoding modes, with the selected codeword having the highest associated reliability score. 20. Channel decoder for channel decoding of at least one transmitted codeword, comprising: - a coloring module for applying at least one coloring sequence to at least one transmitted codeword or to at least one transmitted error correction codeword so as to obtain at least one colored codeword, wherein the coloring sequence is such that at least one bit of the codeword is changed by applying at least one coloring sequence, and wherein at least one coloring sequence is associated with a certain decoding mode as a specific coloring sequence; a redundant decoder for redundantly decoding at least one colored codeword to obtain a decoded output codeword; And a decoding mode detector for generating a decoding mode indicator indicating a specific decoding mode to be used by the redundant decoder to obtain a decoded output codeword, wherein the decoding mode indicator is associated with at least one coloring sequence as the particular coloring sequence used to color the transmitted codeword, wherein the decoding mode detector is configured to store a list of possible decoding modes indicating a predetermined number of possible decoding modes, wherein one decoding mode candidate is indicated without a coloring sequence and other corresponding decoding mode candidates are indicated in association with a coloring sequence, and selecting one possible decoding mode option as the specific decoding mode to be used by the redundant decoder to obtain the decoded output codeword to be used, wherein the decoding mode detector is configured to perform an operation in the first decoding mode and an operation in the second decoding mode, wherein the decoding mode detector for performing an operation in the first decoding mode is configured to evaluate a certain decoding mode, which is a possible variant of the decoding mode without a coloring sequence, to calculate codeword syndromes in such a way as to check whether the calculated syndromes have a value equal to zero , - if the calculated syndromes have a value equal to zero, computing the hash value of the transmitted codeword to compare the calculated hash value and the hash value included in the transmitted codeword, and - if the computed hash value is equal to the included hash value, generating a decoding mode indicator to indicate a candidate coding mode without a coloring sequence as the determined decoding mode, or - if the computed hash value differs from the included hash value, excluding the decoding mode candidate without a coloring sequence from the candidate list and continuing with the operation in the second decoding mode. 21. The channel decoder of claim 20, wherein the decoding mode detector comprises, for performing an operation in the second decoding mode: - a syndrome calculation module for calculating a syndrome for said at least transmitted code word, - a syndrome staining module for applying a staining sequence in association with the decoding mode options in the candidate list to obtain stained syndromes, wherein the syndrome staining module is configured to apply to the syndrome the staining sequences associated with each possible decoding mode option in the list of possible options, and - a syndrome checking module to check if the colored syndromes have a value equal to zero to get the result of the syndrome checking, wherein the decoding mode detector is configured to generate a decoding mode indicator based on the result of the syndrome check. 22. The channel decoder of claim 21, wherein the decoding mode detector comprises, for performing an operation in the second decoding mode: an error locator polynomial calculation module for calculating an error locator polynomial based on the syndrome in each decoding mode option in the list of options; And - a risk value calculation module for calculating the risk value of the transmitted codeword in each possible decoding mode option in the list of possible options based on the error locator syndrome and polynomial, wherein the decoding mode detector is configured to generate a decoding mode indicator based on the risk value. 23. The channel decoder of claim 22, wherein the decoding mode detector is configured to evaluate a risk value for the transmitted codeword in each decoding mode candidate in the candidate list to compare it with a predetermined threshold value and exclude the decoding mode candidate from the candidate list. options if the corresponding decoding mode option risk value exceeds a threshold value. 24. The channel decoder of claim 21, wherein the decoding mode detector is configured to select the decoding mode candidate as the determined decoding mode, the decoding mode candidate corresponding to the risk value having the smallest value, and generating a decoding mode indicator indicating the determined mode. decoder to be used by the redundant decoder to obtain the decoded output codeword. 25. The channel decoder of claim 22, wherein the decoding mode detector comprises, for performing an operation in the second decoding mode: - an error position calculation module for calculating the error position in a possible variant of the decoding mode by factoring the error locator polynomial, - at the same time, the decoding mode detector is configured to receive the result of factorization of the error locator polynomial and generate a decoding mode indicator based on the result of factorization of the error locator polynomial. 26. The channel decoder of claim 25, wherein the decoding mode detector is configured to exclude the decoding mode candidate from the candidate list if no factorization result of the error locator polynomial in the decoding mode candidate is obtained. 27. The channel decoder of claim 25, wherein the decoding mode detector is configured to generate a decoding mode indicator for instructing the selection of a decoding mode candidate in association with a factorization result of the error locator polynomial. 28. The channel decoder of claim 25, wherein the decoding mode detector comprises, for performing an operation in the second decoding mode: an error symbol calculation module for calculating an error symbol in each decoding mode candidate based on the syndrome in the decoding mode candidate and the computed error position in the same decoding mode candidate, - at the same time, the decoding mode detector is configured to receive an erroneous symbol of the transmitted code word in the selected possible variant of the decoding mode and generate a decoding mode indicator that includes an erroneous symbol of the transmitted code word. 29. The channel decoder according to claim 28, comprising: - an error correction module for correcting the error of the transmitted code word, wherein the error correction module is configured to correct the erroneous symbol indicated by the error symbol calculation module. 30. The channel decoder of claim 28, wherein: - the decoding mode detector is configured to generate a decoding mode indicator instructing to mark the frame as a bad frame: - if all possible decoding mode options are excluded from the list of possible options, or - if the erroneous symbol in the transmitted code word is not correctable. 31. The channel decoder according to claims 12 and 20, comprising: - an input interface for receiving an encoded frame, which includes a plurality of transmitted code words, - while the set of code words is at least from two to eight. 32. The channel decoder of claim 20, wherein the decoding mode detector is configured to send a decoding mode indicator to the coloring module and the redundant decoder, wherein the coloring module is configured to select a coloring sequence indicated by the decoding mode indicator, wherein the redundant decoder is configured to select a specific decoding mode indicated by the decoding mode indicator. 33. The channel decoder of claim 32, further comprising: - a demultiplexer or deinterleaving module for demultiplexing or deinterleaving the received codeword to obtain at least one transmitted codeword. 34. The channel decoder of claim 32, further comprising: a data combining module for combining a plurality of decoded output codewords to obtain audio/video frame data. 35. The channel decoder of claim 34, further comprising: - a post-processor for determining the frame data to be output by recalculating the hash value of the received audio/video frame to be compared with the hash value included in the frame data. 36. The channel decoder of claims 12 and 20, wherein the decoding mode detector is configured to use a coloring sequence to determine a decoding mode indicator. 37. A method for encoding a frame, comprising the steps of: - perform multi-mode redundant coding of the frame in accordance with a certain coding mode from a set of different coding modes, while the coding modes differ from each other with respect to the amount of redundancy added to the frame, - display at least one code word; - apply the coloring sequence to at least one code word; wherein the coloring sequence is such that by applying at least one coloring sequence, at least one bit of the codeword is changed, wherein the coloring module is configured to select a specific coloring sequence in accordance with a certain coding mode, and - the frame is divided into a plurality of data words, while the multi-mode redundant encoder is configured to encode each of the plurality of data words according to a certain coding mode to obtain a plurality of code words, wherein a particular coloring sequence is applied to each codeword in a given number of codewords or in a given subset of a plurality of codewords. 38. The method of claim 37, comprising the steps of: - apply a certain coding mode in the previous frame, while the first coding mode has an associated coloring sequence, receiving for the current frame a use indicator of the second coding mode having an associated additional coloring sequence, and - apply an additional coloring sequence in the current frame or bypass the application of the coloring sequence in the current frame. 39. A method for channel decoding at least one transmitted codeword, comprising the steps of: - applying at least one coloring sequence to at least one transmitted codeword to obtain at least one colored codeword, wherein the coloring sequence is such that at least one bit of the codeword is changed by applying at least one coloring sequence , wherein at least one coloring sequence is associated with a certain decoding mode, - perform redundant decoding of at least one colored code word to obtain a decoded output code word, - generating a decoding mode indicator indicating a specific decoding mode to be used by the redundant decoder to obtain a decoded output codeword, wherein the decoding mode indicator is associated with at least one coloring sequence used to color the colored codeword, - at least an additional coloring sequence is used in addition to the coloring sequence, or the channel decoder is configured to bypass the coloring module in the additional decoding mode without coloring; - perform redundant decoding of at least one additional colored codeword colored using the additional coloring sequence to obtain an additional decoded codeword, an additional colored codeword, which is obtained from the transmitted codeword using the additional coloring sequence or the transmission codeword without coloring, to get another additional decoded codeword, outputting a reliability metric for a decoded codeword, an additional reliability metric for an additional decoded codeword, or another additional reliability metric for another additional codeword, determining a decoding mode indicator based on said reliability metrics, and - either a decoded code word, or an additional decoded code word, or another additional decoded code word is output as a decoded output code word based on said reliability index. 40. A method for channel decoding at least one transmitted codeword, comprising the steps of: - applying at least one coloring sequence to at least one transmitted codeword to obtain at least one colored codeword, wherein the coloring sequence is such that at least one bit of the codeword is changed by applying at least one coloring sequence , wherein at least one coloring sequence is associated with a certain decoding mode, - perform redundant decoding of at least one colored code word to obtain a decoded output code word, - generating a decoding mode indicator indicating a specific decoding mode to be used by the redundant decoder to obtain a decoded output codeword, wherein the decoding mode indicator is associated with at least one coloring sequence used to color the colored codeword, storing a candidate list indicating a predetermined number of decoding mode candidates, wherein one decoding mode candidate is indicated without a coloring sequence and other corresponding decoding mode candidates are indicated in association with a coloring sequence, and one decoding mode candidate is selected as the determined mode decoding to be used by the redundant decoder to obtain the decoded output codeword to be used, - wherein the determination process comprises an operation in the first decoding mode and an operation in the second decoding mode, wherein: - the operation in the first decoding mode comprises the steps of: - evaluate a certain decoding mode, which is a possible variant of the mode without a coloring sequence, calculate the syndromes of the code word, check whether the calculated syndromes have a value equal to zero, and if the calculated syndromes have a value equal to zero, calculate the hash value of the transmitted code word, comparing the calculated hash value and the hash value included in the transmitted code word, - while the determination process contains the steps at which: - receive the result of a comparison between the computed hash value and the included hash value, - if the calculated hash value is equal to the included hash value, a decoding mode indicator is generated to indicate as a certain decoding mode a possible variant of the coding mode without a coloring sequence, or - if the calculated hash value differs from the included hash value, exclude the decoding mode option without coloring sequence from the list of possible options and continue with the operation in the second decoding mode 41. The method of claim 40, wherein the operation in the second decoding mode comprises: - calculate the syndrome for at least one transmitted code word, - apply the coloring sequence in association with the decoding mode options in the list of possible options to obtain stained syndromes, wherein the syndrome staining module is configured to apply an association of the staining module sequence with each decoding mode candidate in the candidate list to the syndrome, and - checking if the colored syndromes have a value equal to zero to obtain the result of the syndrome check, wherein the determination process comprises the step of: generating a decoding mode indicator based on the result of the syndrome check. 42. A computer-readable medium containing instructions that, when executed on a computer, cause the computer to carry out the method of claim 37. 43. A computer-readable medium containing instructions that, when executed on a computer, cause the computer to carry out the method of claim 39 or 40.

Images