Method and arrangement for encoding or decoding a sequence of digital data

Method and device for encoding and decoding digital data sequence

The invention relates to a method and device for encoding a digital data sequence and a method and device for decoding a digital data sequence.

A real-time transport protocol (RTP) is disclosed in [1], which controls the encoding, transmission, and decoding of real-time data, such as audio and video data. According to [1], the RTP header contains a 16-bit sequence number, which increases with each RTP packet. This allows the receiver to recognize packet loss during transmission and arrange the packets in the correct order. Due to data security reasons, a random number is used as the starting value in the first data packet. According to [1], the RTP header contains options, which characterize the boundaries of the data block through the setting of flag bits.

In [2], a so-called interleaving block is disclosed, in which data is read into a matrix row by row and read from the matrix column by column in an encoding device. Before using the data contained in the interleaving block, the entire interleaving block is recombined in the decoding device.

A kind of redundant information (also called "redundancy") is disclosed in [3], and transmission errors can be compensated according to the redundant information. Therefore, the data to be transmitted (content data) is additionally equipped with redundant information before being transmitted through the disturbed transmission channel, and the redundant information is generated from the data to be transmitted by means of a known method. Then, the content data is sent to the receiver through the transmission channel together with the redundant information. The receiver can now compensate for transmission errors in such a way that the content data is reconstructed based on the redundant information. For this, a known error correction method is used (see [4]).

In [5], a data uneven error prevention (UXP) method is described, in which the error prevention in a data block is changed in the following way, that is, different amounts of redundant information are allocated to the data. According to [5], the number of data packets in a data block is variable and is specified in a separate data field belonging to the UXP header of each data packet.

Data packet loss occurs during data transmission through an error channel. This is particularly unfavorable when the data packets that respectively contain the flags indicating the boundaries of the data block are also lost. In this case, the data packet needs to be stored in the middle before decoding the data packet until the position of the data packet in a single data block or the boundary of different data blocks can be reconstructed.

Therefore, the task on which the present invention is based is to make it simple to reconstruct the data block boundaries.

This task is solved according to the features of the independent claims. Improvements to the invention are given by the dependent claims.

In order to solve this task, a method for encoding a sequence of digital data is presented. A part of the digital data sequence is equivalent to a data block and contains multiple data packets. At least two data packets of each data block respectively contain a characteristic mark, and the position of the data packet in the data block to which it belongs is determined according to the characteristic mark. The data is coded in consideration of the signature.

In addition, this task is solved by a method for decoding a sequence of digital data. A part of the digital data sequence is equivalent to a data block and contains multiple data packets. At least two data packets of each data block respectively contain a characteristic mark, and the position of the data packet in the data block to which it belongs is determined according to the characteristic mark. The data is decoded in consideration of the signature.

An advantage of the present invention is that at the beginning of data transmission and/or when the data packet is lost, the position of the received data packet in the data block to which it belongs can be directly determined according to the characteristic mark. As a result, real-time applications, such as video telephony or every other multimedia application, can also be implemented, because the cost of intermediate storage of data is significantly reduced. Therefore, it is also possible to significantly reduce the size of the intermediate storage for data in the decoding device or even completely eliminate the intermediate storage.

An improvement is that the digital data sequence contains progressively coded data (progressive data), such as progressively coded images or a sequence of image data streams, where the progressive data may also be image data. Progressive data is hierarchical in terms of its level of detail, that is, an image with a rough resolution is transmitted at the beginning, so although the image can be displayed, the details cannot be identified. The refinement of the image is gradually transmitted, so the resolution of the image becomes better and better as the transmission duration increases.

In an additional improvement, the data block contains redundant information. In this way, the error correction method can correct data errors that occur during transmission and reconstruct the data.

Another improvement is that the start and end of the data block are determined according to the characteristic mark of the data packet.

In an additional improvement, the data blocks respectively contain information about the width of the data blocks.

In another refinement, the signature and the information about the width of the data block are transmitted in a data field alternately, in particular according to a predeterminable repeating pattern. This is advantageous because only one data field is needed to transmit the signature and information about the width of the data block, and despite the additional function, the amount of data that needs to be transmitted does not increase.

An additional improvement is that the number of data packets with a signature is predetermined so that every nth data packet obtains the signature.

Another improvement is that the number of data packets with a signature is predetermined such that the data field of every nth data packet contains the signature and a part of the remaining data packets contains the data block width in the data field.

In another improvement, the predeterminable number of data packets with signature is every second data packet.

In an additional improvement, the data block is an interleaved block. Here, in the coding device, data is read into the interleaving block row by row, for example, and read out of the interleaving block row by row, for example, and then transmitted. If a data packet, that is, a column of the interleaving block is lost during data transmission, the data loss is scattered in each row of the interleaving block. If the data line contains redundant information, an error correction method can be used to correct errors within a specific number, where the number of redundant information directly affects the number of errors that can be corrected.

An additional improvement is that the sequence of the data blocks can be identified, especially based on time stamps or based on consecutive numbers. The time stamp is a digital mark which, for example, gives the data, here the time of transmission of the data block.

An additional improvement is that the signature used to determine the position of the data packet in the data block is a sequence number. The sequence number is, for example, a consecutive number of data packets. For data security reasons, you can choose a random number or the number "0" or "1" as the starting value.

In an additional improvement, the real-time transport protocol (RTP) is used. RTP provides services to transmit real-time data, such as multimedia data. Belonging to this service are the time stamp and sequence number assigned to the data packet.

According to this improved solution, the sequence number of the RTP is used to specify the signature for determining the position of the data packet in the data block.

An additional improvement is the use of uneven error-proofing methods, such as UXP. Therefore, the progressive data in the data block are respectively equipped with different amounts of redundant information, so that the progressive data of the unit to be displayed, such as the image, should be considered in particular, that is, the progressive data of different steps should be considered when the redundant information is allocated. Advantageously, a lot of redundant information can be set at the beginning of the progressive data, and less and less redundant information can be set as the details increase. The number of data packets in a data block, that is, the so-called data block width, can vary for each data block.

In order to use the same data field for the feature mark for determining the position of the data packet in the data block and for the width of the data block, preferably the size of the feature mark can match the size of the data field. If the data field for the width of the data block has, for example, a size of 8 bits and the signature is determined from the 16-bit long sequence number of the RTP header, an 8-bit signature can be generated from the 16-bit signature by deleting the upper 8 bits mark.

In order to solve this task, a device for encoding a sequence of digital data is also provided. A processor unit is provided in the device, which is arranged such that a part of the digital data sequence is a data block and contains a plurality of data packets. At least two data packets of each data block respectively include a characteristic mark, and the position of the data packet in the data block can be determined according to the characteristic mark. The data can be coded in consideration of the signature.

In order to solve this task, a device for decoding the digital data sequence is also provided. A processor unit is provided in the device, which is arranged such that a part of the digital data sequence is a data block and contains a plurality of data packets. At least two data packets of each data block respectively contain a characteristic mark, and the position of the data packet in the data block can be determined according to the characteristic mark. The data can be decoded in consideration of the signature.

The device is particularly suitable for implementing the method of the present invention or one of the above-mentioned improvements.

The present invention or each of the above-mentioned improved solutions can also be implemented by a computer program product having a storage medium on which a computer program that can run on a computer and execute the present invention or the improved solution is stored.

The embodiments of the present invention will be described and illustrated below based on the drawings.

Wherein: FIG. 1 shows a schematic diagram illustrating the principle of interleaving blocks, FIG. 2 shows a schematic diagram illustrating a method for encoding digital data, and FIG. 3 shows a schematic diagram illustrating a method for decoding digital data, Fig. 4 shows the structure of the transmission packet, Fig. 5 shows the location determination of the data packet in the data block, Fig. 6 shows the structure of the transmission system, and Fig. 7 shows the processor unit.

Fig. 1 shows a schematic diagram illustrating the working principle of the interleaving block.

The progressive digital data sequence 101 with data 1 to 12 is, for example, divided into three fine levels (Verfeinerungsschritte), among which data 1 to 3 are the most important data, and data 4 to 7 are less important data. Finally, in this example Data 8 to 12 have the smallest significance in the progressive data sequence 101.

The interleaving block 102 includes 3 rows and 6 columns. In each row of the interleaving block 102, data of the fine level is stored in the interleaving block 102, and redundant information is generated for each data of the fine level and stored in the interleaving block 102 together. In Figure 1, the interleaving block contains 3 rows. In the first row, data 1 to 3 are equipped with redundant information R1, R2, R3, and in the second row, data 4 to 7 are equipped with redundant information R4, R5 and are Data 8 to 12 in the third row are equipped with redundant information R6. Therefore, the data 1 to 3 of the first row of the interleaving block 102 can be reconstructed with the greatest probability (compared to the respective data of the other two rows) because the first row is allocated the largest amount of redundant information.

In this way, a redundancy profile 110 is generated in the interleaving block 102, which is derived from the allocation of data 1 to 12 and the redundant information R1 to R6 generated in the interleaving block 102.

Then, the data 1 to 12 and the redundant information R1 to R6 are read out from the interleaving block 102 column by column, and the contents of each column are formed into data packets 103 to 108: the data packet 103 contains data 1, 4, and 8. Data packet 104 contains data 2, 5, 9, data packet 105 contains data 3, 6, and 10, data packet 106 contains redundant information R1 and data 7, 11, and data packet 107 contains redundant information R2, R4 and data 12. , And the data packet 108 contains redundant information R3, R5, R6.

Therefore, the read data sequence 109 is: {1, 4, 8}, {2, 5, 9}, {3, 6, 10}, {R1, 7, 11}, {R2, R4, 12}, {R3, R5, R6}.

A schematic diagram illustrating a method for encoding digital data is shown in FIG. 2.

Here, the progressive data sequence 210 exemplarily composed of one data block is input to the encoding unit 202, and the encoding unit includes an optimization unit 203 and a packing unit 210. First, the sequence 201 is input to the optimization unit 203 and analyzed there. This analysis provides a progressive structure of the data, according to which the size of the interleaving block 204 and the redundancy profile 205 are determined. The redundancy profile 205 belongs to the management data 206 required to analyze the interleaving block 204 at the receiver. In the optimization unit 203, redundant information 207 is generated for the management data 206 and redundant information 208, 209 is generated for the digital data 201. The more important the data, the more redundant information is set, that is, it also depends on the progressive fineness. Level (see example above).

First, the management data 206 is stored in the interleaving block 204, and the most redundant information 207 is allocated to the management data, so as to be able to correct as many transmission errors as possible, for example. Next, the interleaving block 204 is filled line by line with progressive data 201 and the associated redundant information 208 and 209.

If the interleaving block 204 is filled with data and redundant information, the content of the interleaving block 204 is read out column by column and input to the packing unit 210 as shown in FIG. 1. The packing unit 210 exemplarily shows the packing of the content of the column 211 of the interleaving block 204. Therefore, the data packet 211 corresponding to the content of the column 211 is provided with a header 217 and constitutes a data packet 218, which is referred to as a transmission packet 218 hereinafter.

The header 217 contains a field 219 that contains the sequence number of the transport packet 218. In addition, the header 217 also includes a field 220 in which a characteristic mark or the width of each interleaving block 204 is alternately given. According to the characteristic mark, the position of the data packet 211 in each interleaving block 204 can be determined. The header 217 also includes a field 221, in which a time stamp for each interleaving block 204 is given. Each interleaving block of an image sequence obtains another value as a time stamp, so individual interleaving blocks can be distinguished from each other. It should be noted here that the digital data sequence 201 contains a plurality of progressively coded units, wherein preferably each such unit is stored in an interleaving block. The size of the interleaving block can be changed in units of progressive coding. Preferably, a single image of the digital data sequence 201 is progressively coded.

Similar to the column 211, the remaining columns 212 to 216 of the interleaving block 204 are packed into transmission packets 222 to 226 and transmitted together with the transmission packet 218. Thereby, an encoded data sequence 228 corresponding to the data block 227 is generated. The data block 227 is also referred to as a transport block.

The method for decoding digital data is shown in FIG. 3.

The digital data sequence 301 includes a data block 302, which includes a plurality of transmission packets 303 to 308, wherein each transmission packet has a header and a data packet.

Therefore, the transmission packet 303 contains the header 309 and the data packet 315, the transmission packet 304 contains the header 310 and the data packet 316, the transmission packet 305 contains the header 311 and the data packet 317, the transmission packet 306 contains the header 312 and the data packet 318, and the transmission packet 307 contains The header 313 and the data packet 319, and the transmission packet 308 include the header 314 and the data packet 320. In particular, the digital data sequence received through the interfered transmission channel is input to a decoding unit, which includes an unpacking unit 322 and an analysis unit 325.

First, the transmission packets 303 to 308 are input to the unpacking unit 322 and the data packets 315 to 320 are taken out. With regard to the unpacking unit 322, it is exemplarily shown that the data packet 315 is taken out from the transmission packet 303. Therefore, the header 309 of the transport packet 303 is analyzed first, and the column position of the data packet 315 in the interleaving block 324 is determined according to the header 309, especially the characteristic mark 323 contained in the header. The problem of location determination will be explained in detail in Figure 5 below.

The transmission packet 303 will be stored in the unpacking unit 322 until the position of the data packet 315 in the interleaving block 324 can be determined.

If the position of the data packet 315 in the interleaving block 324 can be determined, the data packet 315 is stored as a column in the interleaving block 324 of the analysis unit 325. The interleaving block 324 is filled column by column with data packets 316 to 320 accordingly.

Next, the content of the interleaving block 324 is analyzed line by line, for example, image information is read out. The data 326 contains management information from which the redundancy profile 327 of the interleaving block 324 can be reconstructed. The redundant profile is used to determine the boundary between the content information of the management class (see data 326) or pure image data (see data 331).

If the data packets 315 to 320 of the interleaving block 324 have been lost due to transmission errors, for example, these errors (below a certain frequency, the upper limit of the frequency is determined by the amount of redundant information transmitted) can be analyzed The unit 325 uses the redundant information 328, 329, and 330 to correct by error correction methods, where for example (only) the loss of the data packet 316 determines that a column of the interleaving block 324 must be reconstructed, which is arranged row by row in the above-mentioned redundancy. In the case of information, it is possible, especially based on data progression, to ensure that important data in the data packet 316 can be reconstructed, and unimportant data can be discarded if necessary, without jeopardizing the function of the method. The read-out digital data sequence 332 can be further processed in an image decoder, especially a decoding that works according to image compression standards, such as MPEG1, MPEG2, MPEG4, H.261, H.263, H.26L.

The structure of the transport packet is shown in FIG. 4. The transmission packet 401 of the transmission block 402 includes a header 403 and a data field 404, and the data field includes a data packet 405. The header 403 contains: a sequence number field 406, which gives the sequence number 407 of the transmission packet 401; an identifier field 408, which gives a feature mark 409 for determining the position of the data packet 405 in the data block 413 and the transmission block 402 Width 410; and a time stamp field 411, in which the value 412 of the time stamp of the transport block 402 is given.

Figure 5 shows that the position of the data packet in the data block is determined according to the signature.

The following starts from the determination of the position of the data packet in the transmission block according to the glossary in the above-mentioned figure. The transmission block contains multiple transmission packets, each of which has a header and a data packet (see the description of FIG. 2). The data block is generated from the data packets of each block. Therefore, the transport block contains the transport packet (see Figure 4, 401) including the respective header (see Figure 4, 403). The information in this header is important for the location determination.

The data sequence 501 includes transmission blocks 502, 503, 504, and 505. The transmission block 502 includes transmission packets 506 to 513, the transmission block 503 includes transmission packets 514 to 519, the transmission block 504 includes transmission packets 520 to 525, and the transmission block 505 includes Transfer packets 526 to 529. The structure of each transmission packet is shown in Figure 4. Therefore, for the glossary of a single field, refer to the description of FIG. 4.

The transmission packets 506 to 513 belonging to the transmission block 502 are respectively marked with a time stamp value "A" in the time stamp field 411, and the transmission packets 514 to 519 belonging to the transmission block 503 are respectively marked with a time stamp value "B" in the time stamp field 411, The transmission packets 520 to 525 belonging to the transmission block 504 are respectively marked with a time stamp value "C" in the time stamp field 411, and the transmission packets 526 to 529 belonging to the transmission block 504 are respectively marked with a time stamp value "D" in the time stamp field 411 . The transmission packets 506 to 529 contain a continuous sequence number 401 in the sequence number field 406 of the transmission packet 506, for example, the sequence number starts from "10" for the transmission packet 506 and ends with "33" for the transmission packet 529.

Transport packets with even-numbered serial numbers 506, 508, 510, 512, 514, 516, 518, 520, 522, 524, 526, and 528 contain in their respective identifier fields to determine the transport block to which the transport packet belongs The characteristic mark of the position in the transmission block, where the characteristic mark is composed of the sequence number of the first transmission packet in the transmission block, that is, transmission packets 506, 508, 510, 512 contain the value "10" in the identifier field 408, and the transmission Packets 514, 516, and 518 contain the value "18" in the identifier field 408, transmission packets 520, 522, and 524 contain the value "24" in the identifier field 408, and transmission packets 526 and 528 contain the value "24" in the identifier field 408. The value "30". The transmission packets with odd serial numbers 507, 509, 511, 513, 515, 517, 519, 521, 523, 525, 527, 529 contain their respective transmission block width 410 in the identifier field, that is, transmission packets 507, 509, 511 and 513 contain the value "8" for the width of the transmission block 502 in the identifier field 408, the transmission packets 515, 517, and 519 contain the value "6" for the width of the transmission block 503 in the identifier field, and the transmission packet 521 , 523, and 525 contain the value "6 for the width of the transport block 504 in the identifier field, and the transport packets 527 and 529 contain the value "4 for the width of the transport block 505 in the identifier field 408.

It is assumed in FIG. 5 that the transmission packets 507, 508, and 512 of the transmission block 502, transmission packets 514 to 519, that is, the entire transmission block 503, the transmission packets 521 and 525 of the transmission block 504, and the transmission packet 526 of the transmission block 505 are already in the data Lost during transmission. In Figure 5, the missing data packets are marked by "X" in their respective sequence number fields.

Since the last transmission packet 519 of the transmission block 503 has not been received, and since the transmission blocks 502 to 505 have different transmission block widths, the conversion of the time stamp value from "B" to "C" cannot determine the value of the transmission block 504. Initially, because not only the transmission packet 520 but also one of the transmission packets 514 to 519 may be the first transmission packet in the transmission block 504. Because the boundary between the transmission block 503 and the transmission block 504 cannot be clearly determined, the received transmission packets 520, 522, 523, and 524 cannot be clearly allocated in the transmission block 504 without the above signature. position. If the last transmission packet 525 of the transmission block 504 and the first transmission packet 526 of the transmission packet 505 have also been lost, they cannot be converted from the time stamp value from "C" to "D" without a signature. The end of the transport block 504 is clearly determined, so more than one transport block needs to be stored intermediately before the data can be decoded.

Now, using the signature in the identifier field, the serial number "24" and the signature "24" can be compared when the transmission packet 520 is received, and it can be concluded that the first transmission packet involving the transmission block 504 at this time is due to this The signature "24" is consistent with the serial number "24". Therefore, the received transmission packets 520, 522, 523, and 524 of the transmission block 504 can be directly stored in the correct position in the transmission block 504, and there is no need to intermediately store the transmission packet for a long time. If the transmission packet 520 with the serial number "24" is also lost, the characteristic mark "24" of the received transmission packet 522 with the serial number "26" can be immediately calculated in the transmission block 504. Location: 26-24=2, that is, there are two transmission packets 520 and 521 in the transmission block 504, so the received transmission packet 522 is the third transmission packet of the transmission block 504. Even in this case, there is no need to further intermediate storage of the transmission packet of the relevant transmission block.

In Figure 5, the sequence number of the first transmission packet of each transmission block is selected as the signature. Another possibility is to use the distance to the first or last transmission packet of each transmission block as the signature.

As another embodiment, an 8-bit long data field from UXP is used for the transmission block width, so that the transmission block width and the feature flag are alternately given in the transmission packet. The signature is determined from the 16-bit serial number of the RTP in the following manner, that is, the serial number of the first transmission packet of each transmission block is reduced from 16 bits to 8 bits. This is achieved, for example, by deleting the upper two digits of the hexadecimal code, for example, reducing it from 0xDC36 to 0x36.

The structure of the transmission system is shown in FIG. 6. The transmission system S includes a camera K, an encoder C, an error (interfered) transmission channel , a decoder D and a display device F. The image data generated by the camera K is encoded in the encoder C, transmitted through the interfered transmission channel , decoded by the decoder D, and displayed by the display device F. The encoder C and/or the decoder D work in particular in accordance with image compression standards, such as MPEG1, MPEG2, MPEG4, H.261, H.263, H.26L. The encoder C and/or the decoder D work especially in consideration of the method of the invention.

The processor unit PRZE is shown in FIG. 7. The processor unit PRZE includes a processor CPU, a memory MEM and an input/output interface IOS. The input/output interface is used in different ways through the interface IFC: the output is visible on the display MON through the graphic interface and/or on the printer PRT Output. Input is realized by mouse MAS or keyboard TAST. The processor unit PRZE also has a data bus BUS, which ensures the connection of the memory MEM, the processor CPU and the input/output interface IOS. In addition, additional components such as additional storage, data storage (hard disk) or scanners can be connected to the data bus BUS.

References: [1] H. Schulzrinne, S. Casner, R. Frederick, V. Jacobson: "RTP: ATransport Protocal for Real Time Applications", Chapter 5.1, RFC1889, Internet Engineering Task Force, January 1996 (2002 Internet site on March 28, 2010: http://www.freesoft.org/CIE/RFC/1889) [2] A. Falkenberg: "Kodier-/Dekodiervorrichtung zumDurchfuehren eines Block-Interleaving/Deinterleaving", patent document DE 198 44 140 C1 [3] Duden Informatik, p. 553, Dudenverlag 2001 [4] C. Schuler: "Design and Implementation of an AdaptiveError Control Protocol", Chapter 2.1, GMD Research Series; 1999, Issue 21 (Internet site on April 25, 2002: http://www.gmd.de/publications/research/1999/021/) [5] G. Baese, G. Liebl: " Generic erasure protection with inbandsignaling of protection profiles", ITU-Telecommunication Standardization Sector, H.323 Annex I, November 2000 ((Internet URL on April 8, 2002: http://standards.pictel.com/ftp/avcsite /ti11_0012/0011_Gen/APC-1992.zip)