The encoding/decoding digital data of audio/video signals and devices for their implementation

 

(57) Abstract:

The invention relates to encoding and decoding digital data divided into blocks of digits, in order of importance digits. The technical result is more accurate data recovery. The encoding method includes the representation of the corresponding digital data to the same specified number of digits and coding digital data represented by the same number, using the specified encoding from the most senior digital sequences to the most Junior digital sequences. Method for decoding digital data coded in order of importance, by assessing the significance of the digits of the digital data includes an analysis of the relevance of the encoded digital data and decoding the analyzed digital data from senior figures to the younger digits using the specified method of decoding. Even if the bit streams are lost or distorted, due to the fact that the first encoded important information, deterioration of sound quality can be reduced. 14 C. and 25 C.p. f-crystals, 10 ill., table 2.

The technical field to which the invention relates

The invention relates to the encoding and decterov the simulation/decoding digital data, divided into blocks of digits, in order of importance digits.

Description of the prior art

Typically, the signal containing the information is essentially a continuous analog signal. For representation of a signal in the form of a digital signal to an analog-to-digital (A/D) conversion.

To perform A/D conversion requires two procedures: sampling procedure for converting a continuous time signal into a discrete signal and the quantization procedure to limit the number of possible amplitudes up to a limited value, that is, to convert the input amplitude x(n) in the element y(n), belonging to a finite set of possible amplitudes at time n.

Because of the quantized signal is simply encoded using pulse-code modulation (PCM), which does not require additional processing, the quantized data is very convenient to use. However, from the point of view of the data module required for storing or transmitting, this simple way of coding is not optimal, even if the input samples are statistically independent. In addition, if the input samples are statistically independent from each other, the encoding method samples the emer statistical coding or specific type of adaptive quantization. Therefore, the coding procedure becomes much more complicated compared to the simple way of storing PCM data.

The bit stream includes the quantized data and the additional information to compress the signals. However, this information forms the bit streams by encoding the quantized data in the block of samples in the data sequence without regard to significance. If the bit streams just remembered, and then restored without any errors, then these bit streams can be generated regardless of the value.

If the bit streams are transmitted via the communication network, the separate parts of the bit streams can be lost depending on the status of the network connection. Also, if during the transmission of bit streams generated error information bit streams, formed after the error is recovered as incorrect information due to a propagating error. If the number of transmitted bit streams as the correct information is reversed only a separate part of the bit streams, the bit streams recovered in the form of correct information, recover more meaningful signals compared to unrestored streams of bits, resulting in p is the set of fixed bit rate, searched optimum condition for a given transmission speed in bits, and then to perform the quantization and encoding, resulting in the formation of the bit streams in accordance with the bit rate. The known method of encoding are formed bit streams having a size suitable for a given transmission rate, without regard to the order of bit streams.

Indeed, if the resulting bit streams are transmitted over a communication network, these bit streams are divided into several segments for transmission. If the transmission channel, there is an overload, or only part of the segments and not all the segments sent from the transmitting end, are at the receiving end due to the narrow band transmission channel, the data cannot be correctly restored. Also, if you are recovering only some of the data streams, the quality is significantly degraded. In the case of digital audio data is reproduced sound unpleasant to the ear. In the case of digital video image that is restored on the screen is greatly distorted.

The invention

To resolve the above problems, the aim of the invention is to provide a method for encoding/decoding digital data is the first data recovery data close to the original data, by minimizing the deterioration of some of the data streams.

To achieve this goal, we propose a method of encoding a given quantity of digital data, containing the following steps: (a) representation of the corresponding digital data by the specified number of digits and (b) encoding the digital data represented by the same number of digits from the sequence of the most senior figures of the sequence to sequences youngest digits.

Step (a) is the representation of digital data in the form of binary data having the same number of bits, and the step (b) is the encoding represented in binary form data sequences from the most high bits (MSB-bit) to sequence the most significant bit (LSB bit).

Step encoding is done by combining the bits forming the bit sequence into blocks with a specified number of bits.

According to another aspect of the invention proposes a method of coding a specified number of digital data formed from the data of the sign and level data, containing the following steps: (a) representation of the respective digital data of the same is senior figures-level data, forming represent digital data; (C) coding the data of the character corresponding to the nonzero data from among coded the most senior digital sequences; (d) coding the most senior digital sequences from a number of unencrypted data level digital data; (e) encoding unencrypted data character from among the data of the character corresponding to non-zero data level, from among the digital sequences encoded in step (d); and (f) performing steps (d) and (e) for the respective digits of the digital data.

Step (a) is the representation of digital data in the form of binary data having the same number of bits, and the numbers in steps (b) to (f) are the bits.

The encoding steps (b) through (f) are performed by combining the bits forming the corresponding bit sequence for the data level and the data mark in the blocks with the specified number of bits.

To achieve this goal a device of the coding sequence of a specified number of digital data, comprising: a separation unit bits to represent the corresponding digital data binary data formed from the same specified number of bill divided bit, coming from the separation unit bit, and the subsequent collection and coding bits and the block sealing bits to generate bit streams in order of importance encoded data out of the unit of encoding.

If the digital data formed from the data of the sign and level data, the coding block collects and encodes the data bits having the same significance level from among divided into data bits, encodes a noncoding data character from among the data of the character corresponding to non-zero data level, and the data coding module and the data of the mark is performed sequentially from the MSB bits to lower bits.

When the coding block collects and encodes the bits according to significance, the coding is performed by combining the bits in the block with the specified number of bits.

Also proposes a method of decoding digital data coded in order of importance, by assessing the significance of the digits of the digital data containing the following steps: analysis of the significance of the encoded digital data and decoding the analyzed digital data using a predetermined encoding method from senior figures to the younger digits.

These figures are bits. the x bits to lower bits in the block vectors and recovery is divided into data bits from the data decoded in a block of vectors.

In addition, to achieve the above objective, it is suggested a method of decoding digital data coded in order of importance, by assessing the significance of the digits of the digital data containing the following steps: (a) analysis of the significance of the encoded digital data; (b) decoding the data level of the analyzed digital data in the form specified method of decoding from senior figures to the younger digits and (C) decoding the data of the sign of the analyzed digital data for combining the decoded data of the mark with the decoded data.

These figures are bits. Step (b) includes the steps of: decoding lossless analyzed digital data from bits to lower bits in the block vectors and recovery is divided into data bits from the data decoded in the block vectors.

Also a device decoding digital data coded in order of importance, by assessing the significance of bits of digital data that contains the unit of analysis of the bit stream for analyzing the significance of bits of the bit stream of encoded digital data; a block decoder for decoding the analyzed digital data from senior figures to mladih data the output from the block decode and recover digital data corresponding samples.

The block decoding block contains the decoded level data for decoding the data level of the analyzed digital data using the specified method of decoding from senior figures to the low numbers and the block data is decoded symbol to decode the sign of the analyzed digital data and combining the decoded data of the mark with the decoded data of the module.

The block decoding decodes lossless analyzed digital data in a block of vectors from bits to lower bits and restores divided into bits of data.

According to another aspect of the invention proposes a device for encoding the audio signal containing the display unit time/frequency to convert the input audio signal time domain into a signal of a frequency domain; block quantization to the quantization signal frequency domain for each frequency band; a separation unit bit to separate chantavanich data block bits; block coding for collecting and coding the MSB bit of the number is divided into data bits output from the separation unit bit, and the subsequent collection and to the significance of the bits of the coded data and additional information for the coded data.

The block generating bit streams to generate streams of bits sequentially from low frequency to high frequency.

Also features a method of coding an audio signal, containing the following steps: dividing the audio data quantized for each given frequency band, in units of bits; collection and coding MSB bit of the number is divided into data bits and the subsequent collection and coding bits sequentially from bits and generating bit streams in order of significance bits of the coded data and additional information for the coded data.

The bit streams are generated sequentially from low frequency to high frequency and high bits to lower bits.

In addition, a device decoding bit streams of encoded audio data containing the block of data flow analysis for analyzing the significance of bits forming the data streams; a block decoder for decoding the additional information having at least quantization bits and the step size of quantization and the quantized data from bits to lower bits according to significance analyzed by the analysis block of the bit stream; block of the inverse quantization to restore the decoded step size is obrazovaniya signals, past the inverse quantization signals in the time domain.

According to another aspect of the invention proposes a method of decoding audio data, containing the following steps: analyzing the significance of bits forming the bit streams and decoding the additional information having at least quantization bits and the step size of quantization and the quantized data from bits to lower bits; restore decoded step size of quantization and quantized data into signals having the original levels; and converting the signals passed inverse quantization signals in the time domain.

As a variant, it is proposed a device for encoding video data containing block DCT (discrete cosine transform) for the DCT transform of the input video signal; a block of quantization to quantize the data of the past DCT; block seal bit to separate the additional information and the quantized values into blocks of bits in accordance with the importance of the quantized data and generating bit streams.

Also according to another aspect of the invention proposes a method of coding a video signal, containing the following steps: quantization of the input video signal prettii with the importance of the quantized data and the coding is divided into data bits in order of importance and generating bit streams.

Also according to the invention proposes a device for decoding a video signal for decoding bit streams of encoded video data containing a block of data flow analysis for analyzing the significance of bits forming the bit streams; a block decoder for decoding the additional information having at least quantization bits and the step size of quantization and the quantized data from bits to lower bits according to significance, the analyzing unit to analyze the flows of bits; block of the inverse quantization to restore the decoded step size of quantization and quantized data into signals having the original levels; and block IDCT (inverse discrete cosine transform) to perform the inverse DCT signals passed inverse quantization.

The coding block performs encoding sequentially from low frequency to high frequency.

Method of decoding a video signal for decoding bit streams of encoded video data includes the following steps: analyzing the significance of bits forming the bit streams and decoding the additional information having at least quantization bits and the step size of quantization and the quantized data from starsi the crystals, with the original levels; and performing inverse DCT signals passed inverse quantization.

Brief description of drawings

The above-mentioned objectives and advantages of the invention will become more apparent from the subsequent detailed description of the preferred variant of its embodiment with reference to the drawings, in which:

Fig. 1 is a block diagram of the digital encoding according to the invention;

Fig.2 is a conceptual diagram showing a known procedure digital encoding;

Fig.3 is a conceptual diagram showing the procedure of the digital encoding according to the invention;

Fig.4 is a block diagram of a decoding device according to the invention;

Fig.5 is a block diagram of the device audio encoding according to the invention;

Fig. 6 is a detailed block diagram of block seal bit shown in Fig. 5;

Fig.7 is a block diagram of a decoding device of the sound according to the invention;

Fig. 8 is a block diagram of the encoder of the video signal according to the invention;

Fig.9 shows the procedure for processing the video signal in a device for encoding video and

Fig.10 is a block diagram of a device for decoding a video signal according to the invention.

Description prepac described preferred embodiments of the invention.

First will be described the General concept of the encoder. The input digital signal is generated in the form of bit streams through the encoding device shown in Fig.1. First, for the formation of bit streams are encoded high-order bits. In other words, the priority encoding components of the data to be encoded is determined by the relative importance of the components. Components with a higher priority have precedence over the components with a lower priority. Since the first encoded important information, if the number used to date a bit more or equal to the tolerance on the generation of bits, the encoding at this point, stops, and generating bit streams is completed. If the generation of the bit streams ends in the middle of the encoding process, the data in the process of their recovery in the decoder partially lost, so that the original digital data is distorted. However, since the first encoded more important information, even if the flow bit in the middle of the encoding process is not formed, the entire encoding process can be supported in the same manner as in the known method.

In Fig. 2 shows a known method of coding. According to what should be used only part of the head streams of bits from among all the threads a bit, the information is less important than that which is included in the unused rear threads a bit, largely contained in the head threads.

For this reason, in the invention of the digital data is divided into blocks of bits, as shown in Fig.3. Usually the significance of the 1st MSB (the high bit) is much higher than the significance of 1 LSB (the least significant bit). Thus, since the high-order bits of quantization are considered more important, the coding is performed from the MSB to the LSB.

In Fig.1 shows a block diagram of the encoder of the digital data according to the invention, which includes a separation unit bit 100, the coding block 110 and block seal bit 120. Here, the digital data is formed from the binary data and presented the same number of bits. Digital data that is represented by the same number of bits, is converted so that they were presented the same number of bits. Although in this embodiment, the digital data is limited to binary data, the data can be decimal, hexadecimal, and other, non-binary data. In this case, the digital data is represented by blocks of digits (not bits).

The separation unit bits 100 parts posledovatelya digital data is such, to sign all the data was positive. When the data is taken from the absolute value represented by binary data, as shown in Fig.3, the values corresponding to the respective bits are separated selectively in accordance with the position of corresponding bits, and the data corresponding to the respective bits are going to generate new sequences. For example, if the input digital data are -31, 12, -9, 7, 17, -23, ..., the absolute value of each data is such to receive digital data in the form of 31, 12, 9, 7, 17, 23, ... their subsequent representation in binary 11111, 01100, 01001, 00111, 10111 10001 and respectively.

From among the values represented by binary data is the information for the relevant bits for subsequent consistent data collection, resulting in generating new sequence. First, data are collected for the respective MSB-bits, that is, 1,0,0... for 31,12,9... respectively. Therefore, divided into bits of data corresponding to the respective MSB bits are 1,1,1,0,0,0,.... After this can be obtained sequences corresponding to the respective bits. And finally, posledovatelnostyakh on bits of binary data, the output of the separation unit bit 100, and encodes them. Following this, the coding block 110 collects bits from bits and encodes them. Bits having the highest importance, preferred are the corresponding MSB-bits of the corresponding digital data represented by binary data, and the bits having the lowest value, preferably represent the corresponding LSB bits. The coding is performed using the algorithm of lossless encoding, suitable for storing or transmitting data.

Usually for more efficient data compression data to the MSB bits are collected consistently and few data are combined, forming the vector. MSB bits are formed by the vectors, and these vectors are encoded using the encoding is lossless. Method of lossless encoding can be an arithmetic coding or Huffman coding. Then gather the following most significant bits for subsequent encoding, which represents the coding method is divided into data bits.

If the digital data formed from the data of the sign and level data, the coding block 110 collects every level data for the MSB bit of the data taken from the separation unit bi is x level data. This procedure is performed from the top down to the LSB only Nemirovskii data character to be coded.

In the invention, as are the absolute values for the respective sample values, information for the value of the sign (positive or negative) must be encoded first or later. In this case, the encoding of the sign of the first leads to the fact that the first coding is performed without any information. Because the values quantieme from the MSB to the first high order bit equal to 1 are considered to be zero, the value of the mark have no meaning. In other words, if the quantized value represented by 5 bits 00011 and uses only 3 bits, the quantized value is restored as 00000. Therefore, even if this value is the sign bit, the information is useless. However, if you are using 4 bits 5, the quantized value is 00010. Thus, the value of the sign will have much more sense as a value of 1, which comes first in the leftmost two bits means that the quantized value is decoded to a value other than zero.

When the representation of the corresponding frequency components from the respective MSB-bits if the first Priceline whether the value of the sign or negative.

For example, when encoding MSB first encoded 1010, and then determines whether to encode the sign bit. At the same time, since the first encoded non-zero value in the first and third frequency components, bits of character for these two components are encoded sequentially, and then to encode 0000.

Block seal bit 120 generates bit streams having a desired size, the encoding order of the data coded by the coding block 110, depending on the importance. First, in the form of bit streams generated coded data is divided into data bits to the MSB bit, and then in the manner described above and encoded information for the sign to be added to the streams of bits, which are generated bit streams in General.

The bit streams generated by the encoding procedure, is restored to the original digital data by decoding procedure, as shown in Fig.2. So as to obtain the bit streams in the invention of the first encoded data with greater significance, the decoder interprets the bit streams in order of importance, i.e. in the order of generation, to perform decoding.

In Fig. 4 shows a block diagram of the device 400, the block decoding lossless 410 and block associations bit 420.

The unit of analysis flows bits 400 analyzes the coded data of the number of input bit streams depending on the importance. The bit streams are generated from the data encoded in the encoding device, depending on the value of the data. Thus, in the device decoding the bit streams are analyzed sequentially from the head of the bit stream depending on importance. First of bit streams analyzed the coded data is divided into data bits to the MSB bit, and then in the manner described above and analyzed information for the sign. Analyzed the coded data is transferred to the block decode lossless 410 and block associations bit 420 is passed information about the token for later use in restoring the original signal.

The block decoding lossless 410 decodes divided into bits signals from among the signals output from the block analysis of streams of bits 400, from bits to lower bits. Divided into bits of data can be restored from the encoded data by a reverse procedure of the algorithm, adapted to the encoding device to encode divided into sledovatelno, and few data are combined to form vectors. Then these vectors are encoded using the encoding is lossless. Thus, these vectors are decoded from coded data, and is divided into data bits for the respective samples are recovered from the decoded vectors.

Block associations bit 420 restores the bit information of the respective samples sequentially from MSB to LSB of the split on the bits of the binary data that is restored at block decoding lossless 410, and outputs the original digital data. Values corresponding to positions of the respective bits of the digital data of the respective samples, supplemented with data corresponding to the respective decoded bits, resulting in values that are the absolute values of the original data.

Information about the sign for the respective samples obtained from a block of data flow analysis 400, if the sign is negative, the absolute value is multiplied by -1 to make the value negative.

For example, assuming that the sequences to be decoded MSB bit is a 1,0,0,0,1,1,..., and if the submission is 0000(2), 10000(2), 10000(2), ... of recovered divided into data bits. Then refers to a sequence of the following bits. If the sequence of the following bits is a 1,1,1,0,0,0, . .., the restored digital data represent 11000(2), 01000(2), 01000(2), 10000(2), 10000(2), ... . Thus, recovery is performed continuously until divided into data bits to the LSB bits, and information about the sign, resulting recover the original input data.

In Fig.5 shows a block diagram of the device audio encoding according to the invention, which includes the imaging unit time/frequency 500, psychoacoustic unit 510, block quantization 520 and block seal bit 530.

The display unit time/frequency 500 converts the audio signal from the time domain into a signal of a frequency domain. Psychoacoustic unit 510 connects the audio signal converted in the signal frequency domain, using the signal components of the frequency bands and calculates the masking threshold for each frequency band. Block quantization 520 quantum signal to the frequency domain in each frequency band, so that the 510. Block seal bit 530 encodes the additional information and the quantized signal components of the frequency domain in each frequency band for generating bit streams.

The main feature of the invention is the block seal bit 530 for generating bit streams by encoding the quantized data in the device, audio encoding. The method of encoding digital data according to the invention can be used in the device, audio encoding, shown in Fig.5, for encoding the quantized data.

Before quantization of the audio signal using the encoder of the first sound psychoacoustic unit 510 using a psychoacoustic model generates the types of groups (long group, primary group, short band, stop-band, and so on ) frames of the input data of the current processing, the values of SMR (signal-to-threshold masking) of the respective bands quantization, information about the area in the case of small groups or delayed in time PCM data to be consistent synchronization psychoacoustic model and display the time/frequency to transmit them to the display unit time/frequency 500. When calculating the psychoacoustic model used Model 2 to ISO/IEC 11 is Zuya modified DCT (MDCT), depending on the group type, the output of the psychoacoustic unit 510. Here the group size is 2048 for long/start/stop group and 256 for short group, and MDCT is performed 8 times. This procedure is the same as that used in the well-known MPEG-2 NBC.

Block quantization 520 combines the frequency components of the data, the converted data in the frequency domain, using strips of quantization are shown in Table 1, and quantum them, increasing the step size, so that the value of SNR (signal/noise) band quantization has become less than the value of the SMR, the output of the psychoacoustic unit 510. Quantization by quantization scale, and the base size of the quantization step is 214. The quantization is performed so that the NMR value was less than 0 dB. The resulting output represents the information of the quantized data and the size of the quantization step of each processed strip. For encoding the quantized signals in each band coding is searched for a quantized signal having the largest absolute value, and then calculates the largest bit quantization for coding.

In Fig.6 shows a detailed block diagram of block seal bit 530, which is used spout 600, the coding block 610 and block the generation of the bit stream 620.

The separation unit bit 600 divides the quantized data output from the block quantization 520, into blocks of bits. The coding block 610 collects MSB bits from among divided into data bits output from the divided into data bits and encodes them. Then the bits are going in order from bits for subsequent encoding. Block the generation of the bit stream 620 generates streams of bits for the encoded data output from the coding block 610, and additional information on the encoded data in the order of significance of the bits.

Now will be described the operation of encoding additional information and quantized data occurring in the separation unit bit 600 and the coding block 610. For more information to start threads a bit to the bit streams are added to the synchronization signal bit streams. Then encoded size of all streams of bits. Next should be coded group type. The subsequent encoding process may differ depending on types of the group. For coding input signals of the frame in accordance with the characteristics of the signal is long, the group may be converted or can be converted to 8 shorter groups. Because of razmerov in each band coding gain maximum value bit quantization, and the coding is performed from the maximum value of the bit quantization encoding method is divided into data bits, the proposed invention. Then encoded information about the step size of the quantization band quantization output during quantization. To encode bits of quantization or information about the size of the quantization step get the minimum and maximum values of bits of quantization or size of the quantization step, to then determine the difference between these two values, whereby get the number of required bits. In practice, before encoding additional information first using arithmetic coding encodes the minimum value and the value needed to represent the bits, and then they are stored in the bit streams. If the encoding is actually executed later, it is the difference between the minimum value and additional information. Then coded these quantized signals.

For encoding the quantized signal can be used coding method is divided into data bits with which the corresponding quantized signals are combined into blocks of bits, and the bits have the e coding of the quantized signals necessary information about the sign of the respective samples, the information on the sign is also encoded. If the number of bits used up to this point, becomes greater than or equal to the allowable number of bits during encoding, depending on the significance, then the encoding at this point ceases to end the thread a bit. Therefore, the overall complexity can be significantly reduced compared with the known encoding method. However, to perform a more efficient encoding regardless of the complexity of the number of bits generated after encoding by an appropriate increase in the step size of each band quantization becomes greater than the number of allowed bits. The coding efficiency can be improved through repetition quantization and coding, while the number of generated bits is less than the allowable number of bits.

Similarly, 8 short of groups obtained by dividing the long group having a size of about one-eighth of the size of the long group, are displaying the time/frequency and quantization, and then the quantized data is lossless encoding. Here, the quantization is not performed separately on each of the 8 groups. Instead, using and what tawania in these sections, shown in Table 2, for further processing as a single lane in the long group. Thus, it can be obtained information about the size of the quantization step for each band in these three sections.

First encoded information bit quantization, and then get the maximum bit quantization. Then encode divided into data bits according to the invention, as long group. If bit quantization of some bands is less than the bit encoded in the moment, then the encoding is not performed. If bit quantization of a certain band becomes equal to the bit encoded in the moment, then the encoding is performed. When encoded strip, first encoded information on the size of the step for the band quantization, and then for subsequent encoding selected values corresponding to the bits of quantization, from among the quantized frequency components.

Through this procedure, the coding is performed in order of importance, and in block generating bit streams 620 are generated bit streams. In other words, to generate bit streams are encoded in order from the MSB bit to LSB-bit information for synchronization, frame size, group type, bit quantization for each of the bands is>In Fig. 7 shows a block diagram of a device for decoding audio decoding bit streams generated by the encoding device of the sound, which includes the analysis block bit streams 700, block decoding 710, block inverse quantization 720 and the display unit frequency/time 730.

The order of decoding audiobot using device decoding the audio signal is the reverse order of the encoding procedure, device, audio encoding. The unit of analysis flows bits 700 examines the significance of the bits forming the bit streams. The block decoding 710 for decoding bits of quantization step size quantization or quantized data generation bit streams input from the device, audio encoding decodes the additional information having at least quantization bits and the step size of quantization and the quantized data from bits to lower bits according to significance analyzed by the analysis block bit streams 700. The block of the inverse quantization 720 restores the decoded step size of quantization and quantized data into signals having the original levels. The display unit frequency/time 730 converts the signals passed back quanta is mA device encoding a video signal, which includes a DCT block 800, the block quantization 810 and block seal 820 bit. Characteristic of the invention is the block seal bits 820 for generating bit streams by encoding the quantized data unit encoding the video signal. The encoding of the quantized digital signals according to the invention can be used in the device, audio encoding, shown in Fig.8.

Block DCT 800 performs DCT video spatial region having an arbitrary level, data in the frequency domain. Block quantization 810 quantum data converted into the frequency domain. Block seal bits 820 separates the additional information and information about the quantized values into blocks of bits according to significance of video data to be encoded, and encodes them sequentially from low frequency to high frequency for generating bit streams.

As shown in Fig.8, the data in the frequency domain is obtained using DCT of video spatial region having an arbitrary level. And then through the quantizer performs quantization to the extent appropriate for the location of the quantized data, as shown in Fig.9.

Data , (usually 1616), as shown in Fig.9B. Then Px group are divided into 4 subgroups (usually 88) having the same size. Then using block DCT 800 DCT is performed for converting the spatial domain into frequency domain data, thereby obtaining the coefficients of the frequency components having the strip 64, as shown in Fig.9C. The resulting coefficients denoted as Fi (i=0, 1, 2, . .., 63) and quanthouse using block quantization 810. The quantized data is indicated as P(Fi) and are arranged in a one-dimensional matrix, as shown in Fig. 9D. Usually the low-frequency components are in most groups, and high-frequency components only in a few groups. Thus, the size of the data is obtained as shown in Fig.9F.

The quantized data is grouped in a similar fashion, and then you can use the method of digital encoding according to the invention for reordered quantized data using block seal 820 bit. Then the original quantized data can effectively be encoded in accordance with the importance, thereby generating the bit streams.

In Fig.10 shows a block diagram of an apparatus of decoding a video signal for decoding bit streams, zakodirovannye 20, the block of the inverse quantization 30 and the IDCT block 40.

The unit of analysis flows bits 10 examines the significance of the bits constituting the coded bit streams. The block decoder 20 decodes the additional information having at least quantization bits and the step size of quantization and the quantized data from bits to lower bits according to significance, the analyzing unit to analyze the flows of bit 10. In other words, get the bit streams generated by the encoding device of the video signal is divided into bits of the quantized data decode according to importance using the method proposed in the invention, and quantized data of the respective frequency components combine with the decoded data. Combined quantized data 64 frequency components are rearranged in the reverse order in the encoding device, resulting in their transformation into quantized data of frequency components to the original subgroups. The block of the inverse quantization 30 restores the decoded step size quantization and the quantized data in the form of signals having the original levels. The IDCT block 40 performs IDCT signals passed inverse quantization, for vos is mportant information that reduces the deterioration of the quality of the audio, even if some threads a bit lost or distorted.

Also, a new method for lossless encoding of digital data according to the invention is compatible with the known method of lossless encoding.

In addition, since the first encoded more important information, the invention can be used for encoding signals of various types, such as audio signals or video signals.

1. The encoding sequence of a specified number of digital data, wherein (a) represents the corresponding digital data having the same specified number of digits, (b) share corresponding to the digital data represented by the same number of digits, then collect the separated digital data to generate sequences composed of numbers with the same location in the corresponding digital data represented by the same number of digits, and (c) encode the sequence from the sequences of the most senior figures to sequences youngest digits.

2. The method of encoding digital data on p. 1, otlichayusta bit, and the operation (C) is the encoding represented in binary form data sequences from the most bits to sequence the most significant bit.

3. The method of encoding digital data on p. 2, characterized in that the encoding operation performed by combining the bits forming the bit sequence into blocks with a specified number of bits.

4. The method of encoding digital data on p. 2 or 3, characterized in that the method is a method of lossless encoding, by which all digital data encode lossless digital data.

5. The method of encoding digital data on p. 4, wherein the lossless encoding is Huffman coding.

6. The method of encoding digital data on p. 4, wherein the lossless encoding is arithmetic coding.

7. The encoding sequence of a specified number of digital data formed from the data of the sign and level data, wherein (a) represents the corresponding digital data having the same specified number of digits, (b) share corresponding to the digital data represented the same share of the, composed of numbers with the same location in the corresponding digital data represented by the same number of digits, (c) encode the sequence of the most senior figures, formed of the most senior figures-level data, forming represent digital data, (d) encode the data of the character corresponding to that contains the number that is in sequence encoded by the most senior figures and is non-zero, (e) encode the sequence of the most senior figures of the number of unencrypted data level digital data, (f) encode noncoding data character from among the data of the character corresponding to the data contains the digit that is in numeric sequences encoded in the operation (e), and is nonzero, (g) perform operations (e) and (f) on the respective digits of the digital data.

8. The method of encoding digital data on p. 7, wherein operation (a) provides for the representation of digital data in the form of binary data having the same number of bits, and the numbers in operations (b) to (g) are the bits.

9. The method of encoding digital data on p. 8, characterized in that the encoding operation with (C) through (g) act means the blocks with the specified number of bits.

10. The method of encoding digital data on p. 8 or 9, characterized in that the method is a method of lossless encoding, by which all digital data encode lossless digital data.

11. The method of encoding digital data on p. 10, wherein the lossless encoding is Huffman coding.

12. The method of encoding digital data on p. 10, wherein the lossless encoding is arithmetic coding.

13. The method of decoding digital data coded in order of importance, by assessing the significance of the digits of digital data, characterized in that (a) decode the digital data using the specified encoding from the most senior figures to the younger digits and (b) recover the digital data of the respective samples obtained in the operation (a), by combining the information digits of the respective samples according to the locations of the respective digits in the source data.

14. The method of decoding digital data on p. 13, characterized in that the figures are bits.

15. The method of decoding digital data on p. 13, wherein operation (a) is the decoding of lossless signal is divided into data bits from the decoded data by combining information for numbers of the respective samples according to the locations of the respective digits in the source data.

16. The method of decoding digital data on p. 14, wherein the method of decoding operation of the decoding is the decoding Huffman.

17. The method of decoding digital data on p. 14, wherein the method of decoding in the decoding operation is the arithmetic decoding.

18. The method of decoding digital data coded in order of importance, by assessing the significance of the digits of digital data, characterized in that (a) analyze the encoded digital data to determine whether the encoded digital data level data or data of the mark, (b) decode the data level of the analyzed digital data in the form specified method of decoding from the most senior figures to the younger digits, and (c) decode the data of the sign of the analyzed digital data for combining the decoded data of the mark with the decoded data.

19. The method of decoding digital data on p. 18, characterized in that the figures are bits.

20. The method of decoding digital data on p. 19, wherein in operation (b) decode lossless analyzed digital data from the highest bit to the lowest bit of the information digits of the respective samples according to the locations of the respective digits in the source data.

21. Device for encoding a sequence of a specified number of digital data, characterized in that it contains a separation unit bits to represent the corresponding digital data binary data formed from the same specified number of bits, and dividing these data into blocks of bits, the coding block for collecting and coding the oldest bit of the number is divided into data bits coming from the separation unit bit, and the subsequent collection and coding most bits and the block sealing bits to generate bit streams in order of importance encoded data out of the unit of encoding.

22. Device for encoding digital data on p. 21, wherein when the digital data formed from the data of the sign and level data, the coding block collects and encodes the data bits having the same significance level, from among divided into data bits, encodes a noncoding data character from among the data of the character corresponding to non-zero data level, and the encoding level data and data of the mark is performed sequentially from the highest bit to the lowest bit.

23. Device for encoding digital data on p. 21 or 22,neetsa by combining the bits in the block with the specified number of bits.

24. The device for decoding digital data coded in order of importance by assessing the significance of bits of digital data, characterized in that it contains the unit of analysis of the bit stream for analyzing the significance of bits of the bit stream of encoded digital data, the block decoder for decoding the analyzed digital data from the most senior figures to the younger numbers, and block associations bit for combining information bits corresponding samples from the decoded data output from the block decode and recover digital data corresponding samples.

25. The device for decoding digital data on p. 24, wherein the block decoding block contains the decoded level data for decoding the data level of the analyzed digital data using the specified method of decoding from the most senior figures to the youngest numbers and block decoding data of the mark for decoding data of the sign of the analyzed digital data and combining the decoded data of the mark with the decoded data.

26. The device for decoding digital signals on p. 24 or 25, wherein the block decoding decodes lossless analyzed the data bits.

27. Device for encoding the audio signal, characterized in that it contains the unit conversion time/frequency to convert the input audio signal time domain into a signal of a frequency domain, block quantization to the quantization signal frequency domain for each frequency band, the separation unit bits for dividing the quantized data into blocks of bits, the coding block for collecting and coding the oldest bit of the number is divided into data bits output from the separation unit bit, and the subsequent collection and coding bits sequentially from the oldest bit, and block the generation of the bit stream to generate bit streams in order of significance bits of the coded data and additional information for the coded data.

28. Device audio encoding on p. 27, characterized in that the power generating bit streams to generate streams of bits sequentially from low frequency to high frequency.

29. A method of coding an audio signal, wherein the divided audio data quantized for each given frequency band, in units of bits, collect and encode the most significant bits of the numbers divided into data bits and carry out the subsequent collection and coding bit sequence is more information for the coded data.

30. A method of coding an audio signal by p. 29, characterized in that the bit streams generated sequentially from low frequency to high frequency and from the bits to the youngest bits.

31. The device is decoding the audio signal for decoding bit streams of encoded audio data, characterized in that it contains the unit of analysis of the bit stream for analyzing the significance of bits forming the bit streams, the block decoding for decoding the additional information having at least quantization bits and the step size of quantization and the quantized data from the highest bit to the lowest bit in accordance with the importance analyzed by the analysis block of the bit stream, the block of the inverse quantization to restore the decoded step size quantization and the quantized data received from the block decoding, the signals having the original levels, and the conversion unit frequency/time to convert the signals passed inverse quantization obtained from the block of the inverse quantization signals in the time domain.

32. The method of decoding an audio signal for decoding bit streams of encoded audio data, wherein the decode additional information, the t to the youngest bits, have a feedback decoded quantization step size quantization and quantized data into signals having the original levels, and convert the signals that have passed inverse quantization signals in the time domain.

33. Device for encoding a video signal containing the block discrete cosine transform for performing discrete cosine transform of the input video signal, the block of quantization to the quantization of the data that passed the discrete cosine transform, block seal bit to separate the additional information and the quantized values into blocks of bits in accordance with the importance of the quantized data and generating bit streams.

34. A method of coding a video signal, characterized in that quantuum input video signal that has undergone the discrete cosine transform, share additional information and information about the quantized values into blocks of bits in accordance with the importance of the quantized data and code is divided into data bits in order of importance and generate streams of bits.

35. A method of coding video according to p. 34, characterized in that the significance correlated with the position of the bit, the significance of quantian meniaetsa down to the youngest bits.

36. A method of coding video according to p. 34, wherein the encoding is performed sequentially from low frequency to high frequency.

37. The device decoding a video signal for decoding bit streams of encoded video data, characterized in that it contains the unit of analysis of the bit stream for analyzing the significance of bits forming the bit streams, the block decoding for decoding the additional information having at least quantization bits and the step size of quantization and the quantized data from the highest bit to the lowest bit in accordance with the importance analyzed by the analysis block of the bit stream, the block of the inverse quantization to restore the decoded step size of quantization and quantized data into signals having the original levels, and the block of the inverse discrete cosine transform to perform inverse discrete cosine transform of the signals passed inverse quantization.

38. A device for decoding video by p. 37, characterized in that the coding block performs encoding sequentially from low frequency to high frequency.

39. Method of decoding a video signal for decoding dirout for more information having at least quantization bits and the step size of quantization and the quantized data from the highest bit to the lowest bit, carry back the decoded quantization step size quantization and quantized data into signals having the original levels, and perform inverse discrete cosine transformation of the signals passed inverse quantization.

Priority points:

02.04.1997 on PP.1-12, 21-23 and 27-30;

19.11.1997 on PP.13-20, 24-26 and 31-39.

 

Same patents:

The invention relates to television, in particular to the processing of the image data, and in particular to a method and apparatus for loop-filtering the image data

The invention relates to methods and devices for adaptive channel coding for communication systems

The invention relates to methods and devices for adaptive channel coding for communication systems

The invention relates to a data transmission method and apparatus for encoding and decoding signal containing side encoding at least two parallel coding block, and on the side of decoding at least two parallel block decoding

The invention relates to a device and method of channel coding in the system previously known bits at the input stage of the channel coding device

The invention relates to communication systems, and in particular to a device and method for adaptive speed control data transmission channel in accordance with the operating conditions of the channel in the communication system mdcr

The invention relates to a coding error correction used when sending text messages on channels of low quality, and, more particularly, to a method of parallel cascaded convolutional encoding and corresponding decoding device

The invention relates to a coding error correction used when sending text messages on channels of low quality, and, more particularly, to a method of parallel cascaded convolutional encoding and corresponding decoding device

The invention relates to turopereiting /facing the interleaver in communication system

FIELD: Witterby algorithm applications.

SUBSTANCE: system has first memory element for storing metrics of basic states, multiplexer, capable of selection between first and second operating routes on basis of even and odd time step, adding/comparing/selecting mechanism, which calculates metrics of end states for each state metric. Second memory element, connected to adding/comparing/selecting mechanism and multiplexer is used for temporary storage of end states metrics. Multiplexer selects first operating route during even time steps and provides basic states metrics, extracted from first memory element, to said mechanism to form end state metrics. During odd cycles multiplexer picks second operating route for access to second memory element and use of previously calculated end state metrics as metrics of intermediate source states.

EFFECT: higher efficiency.

2 cl, 9 dwg

FIELD: communications engineering.

SUBSTANCE: proposed device and method for mobile code-division multiple access communication system including device for transferring channel of backward-link transmission speed indicator afford generation of optimal code words ensuring optimal coding for all types of coding procedures from optimal type (24.1) up to optimal coding procedure 24.7 and supporting all optimal-coding devices.

EFFECT: optimized capacity.

74 cl, 21 dwg, 44 tbl

FIELD: communications engineering; network remote measuring and control systems.

SUBSTANCE: proposed noise-immune cyclic code codec designed for data transfer without pre-phasing has on sending end code-word information section shaper incorporating shift-register memory elements, units for computing verifying parts of noise-immune code of code-word information section, and modulo two adder of code-word information section shaper; code-word synchronizing section shaper and modulo two adder of code-word synchronizing section; on receiving end it has binary filter incorporating binary-filter shift register memory elements, computing units for verifying parts of binary-filter noise-immune code, and binary-filter modulo two adder; shift register of code word information section; decoder; accumulator; error correction unit; unit for shaping synchronizing section of code word; and modulo two adder units.

EFFECT: enhanced speed of device.

1 cl, 1 dwg

FIELD: communications engineering; network remote measuring and control systems.

SUBSTANCE: proposed noise-immune cyclic code codec designed for data transfer without pre-phasing has on sending end code-word information section shaper incorporating shift-register memory elements, units for computing verifying parts of noise-immune code of code-word information section, and modulo two adder of code-word information section shaper; code-word synchronizing section shaper and modulo two adder of code-word synchronizing section; on receiving end it has binary filter incorporating binary-filter shift register memory elements, computing units for verifying parts of binary-filter noise-immune code, and binary-filter modulo two adder; shift register of code word information section; decoder; accumulator; error correction unit; unit for shaping synchronizing section of code word; and modulo two adder units.

EFFECT: enhanced speed of device.

1 cl, 1 dwg

FIELD: communication systems.

SUBSTANCE: method includes generating sets of sub-codes of quasi-additional turbo-codes with given encoding speeds, and given sub-codes are reorganized as a set of sub-codes with another encoding speed for use in next transfer of sub-code with given encoding speed.

EFFECT: higher efficiency.

9 cl, 13 dwg

FIELD: data transfer technologies.

SUBSTANCE: method includes segmentation of length N of quasi-complementary turbo-codes on preset amount of sections, determining identifiers of sub-code packets appropriate for segmented portions, setting of said packets separated for initial transfer of sub-code, calculation of number of remaining symbols in form N-Fs, where N - length of quasi-complementary turbo-codes, and Fs - position of start symbol of sub-code of quasi-complementary turbo-codes, determining position of symbol of remaining symbols in amount equal to sub-codes amount, which have to be sent and serial transfer of sub-code symbols from position of starting symbol Fs to position of last symbol Ls.

EFFECT: higher efficiency.

5 cl, 17 dwg

FIELD: communications engineering.

SUBSTANCE: method includes selecting one combination among given combinations, appropriate for several or every generated symbols of code word to transmit generated symbols of code word with length of sub-packet, determined in accordance to data transfer speed, information, appropriate for data transfer speed, is read, also based on length of sub-packet and chosen combination, from a table, wherein identification information, pointing at data transfer speed, sub-packet length and selected combination, is, is previously displayed for given information, and generated code word symbols are transmitted in accordance to read information and in accordance to selected combination.

EFFECT: possible check transmission of information by means of hybrid automatic repeat query for increasing carrying capacity during high-speed information transfer.

4 cl, 16 dwg, 6 tbl

FIELD: communications engineering; simulating digital communication channels with separate and grouping errors.

SUBSTANCE: proposed method includes evaluation of set of communication channel states S0,S1, ..., Sm - 1 and calculation of conditional error probabilities P(e/s) in each state s" i = 0, ..., m - 1 of communication channel, and error acquisition in communication channel in compliance with conditional error probability for current state of communication channel; in the process probability of error-free interval p(0i) of i bits is found, and conditional probabilities p(0i1/11), p(0i1/01) of error-free intervals of i bits are calculated with respect to them basing on probabilities p(0i) and using recurrent rules during each current time interval and preceding one on condition that for error generation use is made of two states of communication channel corresponding to combination of errors 11 or 01; random number p uniformly distributed within interval between 0 and 1 is generated; conditional probabilities p(0i1/11), p(0i1/01) are summed up starting from i = 0 resulting in sequence 0k1 that constitutes bit-by-bit stream of communication channel errors.

EFFECT: enhanced speed.

1 cl, 1 tbl

FIELD: communications engineering; data transfer, telemetering, and telecontrol systems.

SUBSTANCE: proposed codec has on sending end code-word data part shaper whose output and that of code-word synchronizing part shaper are connected to modulo two adder input; on receiving end it has binary filter whose code-word data part shaper output is connected to accumulator connected to synchronizing sequence decoder and to error connection unit whose outputs are connected to respective inverting inputs of code-word data part shaper; output of the latter functions as data output of device; output of binary-filter code-word synchronizing part is connected through switching unit to input of code-word data part shaping unit; synchronizing sequence decoder output is connected to control input of switching unit and to error correction unit input; on receiving end accumulator outputs are connected to inputs of code-word data part shift decoder whose output is connected to input of delay circuit whose output functions as second control input of switching unit and as synchronizing output of device.

EFFECT: enhanced noise immunity.

1 cl, 1 dwg

FIELD: coding in communication systems.

SUBSTANCE: proposed partial reverse bit-order interleaver (P-RBO) functions to sequentially column-by-column configure input data stream of size N in matrix that has 2m lines and (J - 1) columns, as well as R lines in J column, to interleave configured data, and to read out interleaved data from lines.

EFFECT: optimized interleaving parameters complying with interleaver size.

4 cl, 7 dwg, 3 tbl

Up!