Method for selective decoding of a digital broadcast receiver fill code word

Abstract

The application discloses a method for selectively decoding the fill code word of a digital broadcast receiver. The fill data in a digital broadcast multiplex frame is encoded into a low-density parity-check code word together with the payload. The receiver cannot obtain the fill boundary from the frame header field before decoding is completed. The method analyzes the decoded multiplex layer frame header of the Nth logical frame to obtain the payload end position and convert it into the high water line of the valid code word. According to this, the code word after the high water line in the N+1th logical frame is predicted to be a fill code word. The low-density parity-check code decoding, constellation demapping, deinterleaving and noise estimation are omitted for the fill code word to fill the information bit with a predetermined constant. When the continuous analysis of the multiplex layer frame header fails or the transmission mode or the encoding code rate changes, the method reverts to full decoding and re-determines the high water line. The method reduces the decoding operation amount under weak signal conditions, enables the receiver to meet the real-time decoding requirements on a low-order single-core processor and does not lose the service data.

Description

Technical Field

[0001] This invention relates to the field of digital audio broadcasting reception technology, and in particular to a method for selective decoding of filler codewords in a digital broadcasting receiver. Background Technology

[0002] The FM band digital audio broadcast receiver demodulates, equalizes, demaps, and deinterleaves the orthogonal frequency division multiplexing (OFDM) symbols at the physical layer, obtaining a series of low-density parity check (LDPC) codewords. Then, forward error correction decoding is performed on each codeword to recover the multiplexing layer data. The physical layer LPC codeword length is 9216, with a code rate of 1 / 4, 1 / 3, 1 / 2, or 3 / 4. The decoded byte stream is organized into multiplexed frames. Each multiplexed frame consists of a frame header, several service multiplexed subframes (i.e., payload), and subsequent padding data, which is a constant set of all "1"s.

[0003] When running the aforementioned receiver in real-time on a low-end single-core ARM processor, low-density parity code decoding consumes most of the computational resources. In weak signal, high-net-value code rate operating modes, the number of decoding iterations increases, making it difficult for a single-core processor to meet real-time decoding requirements. In these modes, a logical frame contains over forty codewords, of which only about a quarter carry valid service content, with the remainder carrying padding data. If the decoding of padding codewords could be eliminated, the processor's computational burden would be significantly reduced.

[0004] However, the padding data and payload are encoded together using forward error correction and encapsulated within the same low-density parity codeword. Before decoding, the receiver cannot determine the starting codeword of the padding data from any header field of the current logical frame. Only by decoding all codewords and then parsing the multiplexing layer frame header can the end position of the payload be determined. Therefore, existing receivers can only perform complete decoding on all codewords, including the padding codeword.

[0005] Several existing technologies exist to reduce the computational load on receiver decoding. One approach calculates a reliability metric based on the signal-to-interference-plus-noise ratio (SINR) of the code block, disabling the forward error correction decoder when the reliability falls below a threshold. This approach relies on channel quality, independent of whether the codeword content is padding, and demodulation and noise estimation processing are still required to calculate the SINR. Another approach allows the receiver's physical layer to decide whether to continue receiving the entire frame based on transmitter characteristics, a frame-by-frame decision that does not involve decoding of codewords within the frame. A further approach uses signaling in the frame header to indicate the padding length, allowing the receiver to skip the padding portion by reading the signaling. This approach requires the padding boundary to be known before decoding, unlike the case where the padding is encoded into the forward error correction codeword. None of these approaches address the problem of skipping padding codeword decoding when the padding boundary is unknown before decoding. Summary of the Invention

[0006] The purpose of this invention is to provide a selective decoding method for padding codewords in a digital broadcast receiver, so as to solve the problems that the padding boundary is unknown before decoding, the receiver can only perform decoding on all codewords, and low-level single-core processors are difficult to decode in real time.

[0007] To achieve the above objectives, the technical solution adopted by this invention is: a selective decoding method for padding codewords in a digital broadcast receiver, applied to a digital broadcast receiver. The digital broadcast receiver performs low-density parity code decoding on a codeword-by-code basis on physical layer signal frames. The multiplexed frame of the physical layer signal frame includes payload and padding data following the payload. The padding data and the payload are encoded together into low-density parity codewords using the same forward error correction method, so that the boundary between the padding and the payload cannot be directly read from the header field of the current logical frame before the low-density parity code decoding of the codewords is completed. The method includes: for the Nth logical frame, parsing its decoded multiplexed layer frame header to obtain the payload end position. The high watermark is then converted into a valid codeword. Based on the high watermark determined in the Nth logical frame, the codewords located after the high watermark in the N+1th logical frame are predicted to be padding codewords. For codewords predicted to be padding, their low-density parity code decoding, constellation demapping, deinterleaving, and noise estimation are omitted, and their information bits are filled with a conventional constant. When the multiplexing layer frame header fails to be parsed for K consecutive logical frames, or when a change in transmission mode or coding rate is detected, the process reverts to decoding all codewords. After the revert, when the multiplexing layer frame header of any logical frame is successfully parsed, the high watermark is re-determined based on that logical frame, and selective decoding of padding codewords is resumed.

[0008] Furthermore, the conversion of the high-water mark into a valid codeword includes: converting the end position of the payload into the number of bits, dividing it by the information bit length of the current codeword and rounding up to obtain the high-water mark codeword sequence number.

[0009] Furthermore, the length of the low-density check codeword is divisible by the number of bits carried by each constellation symbol, aligning the codeword boundary with the constellation symbol boundary; constellation demapping is not performed on the constellation symbols corresponding to codewords located after the high watermark.

[0010] Furthermore, the agreed constant is the same as the fixed bit pattern of the padding data.

[0011] Furthermore, K is 2.

[0012] Furthermore, the method is applied to FM band digital audio broadcasting, and the multiplexed frames, payload, and padding comply with the relevant national standards.

[0013] Furthermore, the high watermark is updated per logical frame: whenever the multiplexing layer frame header of a logical frame is successfully parsed, the high watermark is re-determined based on the logical frame, and the re-determined high watermark is used for the prediction of the padding codeword in the next logical frame.

[0014] The present invention also provides a digital broadcast receiver, including a multiplex layer frame header parsing module, an effective codeword high-water mark calculation module, a padding codeword prediction module, a selective decoding control module, and a self-correction backoff module, for implementing the above-mentioned selective decoding method for padding codewords in the digital broadcast receiver.

[0015] The present invention also provides a digital broadcast receiving device, including a processor and a memory, wherein the memory stores a computer program, and the processor executes the computer program to implement the above-described digital broadcast receiver filler codeword selective decoding method.

[0016] The present invention also provides a computer-readable storage medium having instructions stored thereon, which, when executed by a processor, implement the steps of the above-described method for selective decoding of filler codewords in a digital broadcast receiver.

[0017] Compared with the prior art, the present invention has the following advantages: 1. Skipping the padding codeword is achieved under the condition that the padding boundary is unknown before decoding, which reduces the amount of computation for low-density parity code decoding under weak signal conditions, enabling the receiver to decode in real time on a low-level single-core ARM processor; 2. Low-density parity code decoding, constellation demapping, deinterleaving and noise estimation are omitted for padding codewords, saving more computation than the scheme of simply turning off the forward error correction decoder; 3. A fallback mechanism is set up to restore full decoding and re-determine the high watermark when the multiplexing configuration changes, avoiding misjudging valid codewords as padding codewords and preventing loss of service data. Attached Figure Description

[0018] Figure 1 A flowchart illustrating the selective decoding method for padding codewords provided in an embodiment of the present invention;

[0019] Figure 2 This is a schematic diagram of the structure of a digital broadcast receiver provided in an embodiment of the present invention;

[0020] Figure 3 This is a schematic diagram of the multiplexed frame structure and the effective codeword high watermark in an embodiment of the present invention.

[0021] In the diagram: 101. Front-end demodulation module; 102. Noise estimation module; 103. Constellation demapping module; 104. Deinterleaving module; 105. Low-density parity code decoding module; 106. Multiplexing layer frame header parsing module; 107. Effective codeword high-water mark calculation module; 108. Filler codeword prediction module; 109. Selective decoding control module; 110. Self-correction fallback module. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0023] like Figure 2 As shown, the receiver in this embodiment includes a front-end demodulation module 101, a noise estimation module 102, a constellation demapping module 103, a deinterleaving module 104, a low-density parity check (LDPC) decoding module 105, a multiplexing layer frame header parsing module 106, an effective codeword high-water mark calculation module 107, a padding codeword prediction module 108, a selective decoding control module 109, and a self-correction backoff module 110. The front-end demodulation module 101 performs Fast Fourier Transform and channel equalization on the received orthogonal frequency division multiplexing (OFDM) symbols; the noise estimation module 102 estimates the noise variance of each subcarrier to generate soft bits; the constellation demapping module 103 maps the equalized symbols to log-likelihood ratios; the deinterleaving module 104 rearranges the log-likelihood ratios into a codeword sequence; the low-density parity check (LDPC) decoding module 105 iteratively decodes codewords with a length of 9216; and the multiplexing layer frame header parsing module 106 parses the service multiplexing subframe headers from the decoded byte stream.

[0024] It should be noted that the above modules can be implemented by hardware circuits or by a processor executing computer programs in memory; this embodiment does not limit this.

[0025] like Figure 1 As shown, this method executes the following steps according to logical frames. Step S1: After the multiplexing layer frame header parsing module 106 successfully parses the multiplexing layer frame header of the Nth logical frame, it obtains the payload end position. The payload end position is equal to the multiplexing frame start position plus the frame header length, plus the sum of the lengths of each service multiplexing subframe. Step S2: The valid codeword high-water mark calculation module 107 converts the payload end position into the number of bits, divides it by the information bit length of the current codeword, and rounds up to obtain the high-water mark H, that is, the valid content ends at the Hth codeword. In this embodiment, under the working mode of transmission mode 3, spectrum mode 10, 16QAM, and coding rate 3 / 4, a logical frame contains 44 codewords, the codewords carrying the valid service account for about the first quarter, the high-water mark H is 11, and the following 33 codewords are padding codewords.

[0026] like Figure 3 As shown, the multiplexed frame consists of a frame header, a service multiplexing subframe, and a padding area following it, with the padding area being a constant set of all "1". In step S3, the multiplexing configuration of the broadcast system (number of services, bitrate of each service) remains unchanged for a considerable period. Based on this, the padding codeword prediction module 108 predicts that the codeword following the high-water mark H in the (N+1)th logical frame will also be a padding codeword. In step S4, the selective decoding control module 109 processes the (N+1)th logical frame as follows: for codewords with a sequence number not greater than H, they are normally processed by the noise estimation module 102, constellation demapping module 103, deinterleaving module 104, and low-density parity code decoding module 105; for padding codewords with a sequence number greater than H, the above four modules are bypassed, and their information bits are directly filled with a conventional constant. This conventional constant is the same as the fixed bit pattern of the data being filled at the transmitting end, and in this embodiment, it is all "1".

[0027] In this embodiment, each constellation symbol carries 2, 4, or 6 bits, corresponding to QPSK, 16QAM, and 64QAM respectively. The codeword length 9216 is divisible by these bit numbers, thus aligning the codeword boundary with the constellation symbol boundary. Constellation symbols corresponding to codewords located after the high-water mark can be completely skipped. Deinterleaving in this embodiment is performed on an interleaving block basis: for interleaving blocks completely after the high-water mark, the entire block is skipped for constellation demapping; for interleaving blocks crossing the high-water mark, only the valid portion before the high-water mark is demapping, and the remaining portion is padded. The skipping operation does not corrupt the data of valid codewords.

[0028] In steps S5 and S6, the self-correction fallback module 110 monitors the parsing results and transmission parameters of the multiplexed layer frame header parsing module 106: when the multiplexed layer frame header fails to parse for K consecutive logical frames (K is 2 in this embodiment), or when a change in transmission mode or coding rate is detected, it determines that the previously determined high-water mark may have failed, and falls back to performing low-density check code decoding on all codewords; after the fallback, when the multiplexed layer frame header of any logical frame is successfully parsed, the high-water mark is re-determined based on that logical frame, and selective decoding of padding codewords is restored. In scenarios such as service additions / reductions, rate switching, and channel switching, this mechanism avoids misjudging valid codewords as padding codewords. In other embodiments, K can also be 3 or other positive integers: the larger the value of K, the higher the tolerance for occasional parsing failures of the multiplexed layer frame header, and the slower the recovery of full decoding after configuration changes.

[0029] To verify the effectiveness of this method, tests were conducted on a real hardware platform. The test platform used a low-end single-core ARM processor with a main frequency of approximately 1.1 GHz. The test signal was a digital audio broadcast signal in the FM band with transmission mode 3, spectrum mode 10, 16QAM, and a coding rate of 3 / 4. The measured signal-to-noise ratio of the receiver was approximately 14.5 dB, close to the decoding threshold for this mode. The average number of iterations for low-density parity code decoding was approximately 6.7. In this mode, a logical frame contains 44 codewords, with an effective codeword high-water mark of 11 and the remaining 33 codewords being padding codewords. After enabling this method, the processor time for low-density parity code decoding on each logical frame decreased from 548 milliseconds to 134 milliseconds, and the proportion of the single-core processor time decreased from approximately 76% to approximately 29%. The processor time for the entire receiving pipeline on each logical frame decreased from 723 milliseconds to 464 milliseconds. Compared to a real-time budget of 640 milliseconds, the real-time margin increased from 0.89 times to 1.38 times, changing from being unable to decode in real time to meeting the requirements for real-time decoding. Under strong signal conditions, the low-density parity code iterates approximately once on average, reducing its proportion of a single-core processor from about 35% to about 11%. The high-water mark prediction hit rate was 100% during testing. In test sequences including coding rate switching and near-threshold fading, the backoff mechanism was triggered seven times. The service bitstream decoded by this method was identical byte-by-byte to that decoded by the full decoding method, with no service data loss.

[0030] Those skilled in the art will understand that this method is not limited to the working mode and parameter values ​​in the above embodiments. This method can be used for any digital broadcasting system where the padding data and payload are forward-corrected and the padding boundary is unknown before decoding.

[0031] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for selective decoding of codewords in a digital broadcast receiver, applied to a digital broadcast receiver, wherein the digital broadcast receiver performs low-density parity code decoding on a codeword-by-code basis on physical layer signal frames, characterized in that... include: The multiplexed frame of the physical layer signal frame includes payload and padding data following the payload. The padding data and the payload are encoded together into low-density parity codewords by the same forward error correction, so that the boundary between the padding and the payload cannot be directly read from the header field of this logical frame before the low-density parity code is decoded. For the Nth logical frame, the end position of the payload is obtained by parsing its decoded multiplexed frame header, and the high watermark of the valid codeword is calculated accordingly. Based on the high watermark determined by the Nth logical frame, the codeword following the high watermark in the N+1th logical frame is predicted to be the padding codeword. For the codeword predicted to be padding, its low-density parity code decoding, constellation demapping, deinterleaving, and noise estimation are omitted, and its information bits are filled with a conventional constant. When the parsing of the multiplexed frame header fails for K consecutive logical frames, or a change in transmission mode or coding rate is detected, the process falls back to decoding all codewords. After the rollback, when the multiplexed layer frame header of any logical frame is successfully parsed, the high watermark is re-determined based on the logical frame, and selective decoding of the padding codewords is resumed.

2. The method according to claim 1, characterized in that, The conversion of the high-water mark into a valid codeword includes: converting the end position of the payload into the number of bits, dividing it by the information bit length of the current codeword and rounding up to obtain the high-water mark codeword sequence number.

3. The method according to claim 1, characterized in that, The length of the low-density check codeword is divisible by the number of bits carried by each constellation symbol, aligning the codeword boundary with the constellation symbol boundary; constellation demapping is not performed on the constellation symbol corresponding to the codeword located after the high watermark.

4. The method according to claim 1, characterized in that, The agreed constant is the same as the fixed bit pattern of the padding data.

5. The method according to claim 1, characterized in that, The value of K is 2.

6. The method according to claim 1, characterized in that, The method is applied to FM band digital audio broadcasting, and the multiplexed frames, payloads, and padding comply with the relevant national standards.

7. The method according to claim 1, characterized in that, The high watermark is updated per logical frame: whenever the multiplexing layer frame header of a logical frame is successfully parsed, the high watermark is re-determined based on the logical frame, and the re-determined high watermark is used for the prediction of the padding codewords in the next logical frame.

8. A digital broadcast receiver, characterized in that, include: The multiplexed layer frame header parsing module is used to parse the decoded multiplexed layer frame header to obtain the payload end position; The effective codeword high watermark calculation module is used to convert the end position of the payload into the high watermark of the effective codeword; the filling codeword prediction module is used to predict the codeword located after the high watermark in the current logical frame as the filling codeword based on the high watermark determined in the previous logical frame; the selective decoding control module is used to omit the low-density check code decoding, constellation demapping, deinterleaving and noise estimation of the codeword predicted as filling, and fill its information bits with a conventional constant. The self-correction fallback module is used to control the selective decoding control module to fall back to performing decoding on all codewords when multiple consecutive logical frames of the multiplexed layer frame header fail to be parsed, or when a change in transmission mode or encoding rate is detected; and when the multiplexed layer frame header of any logical frame is successfully parsed after the fallback, the effective codeword high-water mark calculation module is triggered to re-determine the high-water mark and resume selective decoding of the padding codewords.

9. A digital broadcast receiving device, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program, and the processor executes the computer program to implement the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having instructions stored thereon that, when executed by a processor, implement the steps of the method as claimed in any one of claims 1 to 7.

Images