KR100224582B1 - Error detecting apparatus and method of mpeg-2 audio - Google Patents

Abstract

The present invention relates to an error detection apparatus and method for MPEG-2 audio (Moving Picture Experts Group-2 Audio).

According to the error detection apparatus and method of MPEG-2 audio according to the present invention, in the MPEG-2 audio, the input unit separates the CRC check word from the basic frame and the extended frame of the input bit stream, respectively, to make 16-bit parallel data. As a result of multiplexing and outputting the error, the error detector outputs the CRC input of the CRC input unit and CRC check data of the basic data and the extended data included in the input bit string. The comparison between the comparison words detects whether an error occurs in the input bit string. In addition, in carrying out the present invention, CRC processing is performed for the basic frame and the extended frame using the CRC counter while sharing the CRC register.

Therefore, it is easy to solve a problem such as the need for a CRC checking method and a hardware CRC checking means capable of satisfying the rather complicated condition of CRC checking between the basic frame and the extended frame, and sharing the CRC register. This has the advantage of reducing design costs.

Description

Error detection apparatus and method of MPEG-2 audio. {ERROR DETECTING APPARATUS AND METHOD OF MPEG-2 AUDIO}

The present invention relates to a data error detection apparatus, and more particularly, to a Cyclic Redundancy Check (CRC) for basic and extended frames of MPEG-2 audio (Moving Picture Experts Group-2 Audio). CRC processing by sharing a cyclic redundancy check register (CRC register, hereinafter referred to as a CRC register) and using a cyclic redundancy check counter (hereinafter referred to as a CRC counter). The present invention relates to an error detection apparatus and method for MPEG-2 audio that can be easily performed.

MPEG-2 audio is an audio coding method established in November 1994 by the ISO / IEC JTC1 / SC29 / WG11 audio group. It is a multi-channel audio system that realizes high quality stereo sound through left and right stereo sound, left and right surround sound and center sound. An audio coding method corresponding to channel audio.

MPEG-2 audio was drafted in November 1993 by the Committee Draft (CD), in March 1994 by the Draft International Standard (DIS), and in November 1994 by the International Standard (IS). Standard) is completed.

MPEG-2 audio is collectively known as the official standard number ISO / IEC 13818-3, and can be used for audio-video encoding in combination with MPEG-2 video, as well as digital audio broadcasting (DAB) alone. In addition, it is possible to create a wide range of services for high-quality multi-channel audio transmission, digital storage means, etc. by an Integrated Service Digital Network (ISDN).

In other words, the multi-channel audio system based on MPEG-2 audio includes CATV (CAble Television), CDAD (Cable Digital Audio Distribution), ENG (ENG) in addition to broadcasting fields such as satellite or terrestrial television broadcasting and digital audio broadcasting. Electronic News Gathering, Electronic Cinema (EC), InterPersonal Communication (IPC), Interactive Storage Media (ISM), Network DataBase service (NDB), Digital Storage Media (DSM), Home Television Theater (HTT), Integrated Service Digital Non-broadcasting media such as networks have various application areas.

MPEG-2 audio is an upgraded version of MPEG-1 audio that absorbs most of the MPEG-1 audio's ability to provide stereo sound, while the left and right surround channels, center sound channel, and low-frequency effects Supports audio input / output formats that combine channels, etc., provides multilingual functions, and lower sampling frequencies of 16 kHz, 22.05 kHz, and 24 kHz to the sampling frequencies provided by MPEG-1 audio. Is additionally provided.

It is clear that MPGE-2 audio shares a core algorithm with MPEG-1 audio as it supports backward comkpoibility. Accordingly, the main features of MPEG-1 audio shared with MPEG-2 audio will be briefly described below.

MPEG-1 audio is based on a masking-kpotern universal subband integrated coding and multiplexing (MUSICAM) scheme based on subband coding. The MUSICAM system is a 16-bit uniform PCM using a subband coding scheme. Pulse Code Modulation data can be received and variably encoded from 64 kbps to 196 kbps per channel, using critical bands and masking effects, which are human hearing characteristics, according to performance and complexity. Three types of layers are defined and used.

The basic principle of the MUSICAM system is to increase the data compression efficiency by eliminating the components of the audio signal that are not perceived by the audio using human hearing characteristics.In the case of transform coding, the masking threshold is directly used by using coefficients. Although a masking threshold may be obtained, in the case of split band coding, a detailed frequency analysis is required to obtain a masking threshold so that an additional fast Fourier transform (FFT) is used for frequency characteristics of each subband. Find it.

Here, the masking threshold refers to a limit at which noise is masked by the original sound or lower than an audible limit, and thus cannot be heard. To obtain a masking threshold, the linear frequency domain must be converted to a nonlinear critical band. The masking threshold is obtained using the frequency spectrum changed in units of Bark.

The process of obtaining the masking threshold value is called psychoacoustic modeling process. If the amount of quantization noise is smaller than the masking threshold value, a bit is applied to each subband so as to be optimal at a predetermined bit rate under the assumption that the noise is not recognized by the ear. The subband signals are encoded using Advanced Pulse Code Modulation (APCM).

Meanwhile, as described above, MPEG-1 audio defines three types of layers shown in Table 1 according to performance and complexity, and thus allows freedom of selecting and using layers appropriately for each application. do.

As can be seen from Table 1, the higher the layer, the higher the quality and the higher the compression ratio, but the larger the hardware scale.

First, the characteristics of each layer will be described based on the coding scheme used in the layer I of MPEG-1 audio.

In order to effectively use hearing psychology such as critical bands, it is effective to first divide the signal into frequency components. To do this, 32-band subband encoding is performed, which is divided into 32 bands. In Layer I of MPEG-1 Audio, the entire band is divided into 32 equally spaced frequency widths, and each signal is divided into It adopts 32 subband coding to sample with.

Hierarchical Characteristics of MPEG-1 Audio hierarchy Bit rate Coding method Hardware scale Layer I 128 Kbps (monaural) 256 Kbps (stereo) 32 subband scale factor bit allocation384 sample / frame small Layer II 96 Kbps (monaural), 192 Kbps (stereo) 32 subband scale factor bit allocation, 3 sets 1152 sample / frame medium Layer III 64 Kbps (monaural) 128 Kbps (stereo) 32 subband + MDCT scale factor bit allocation 1152 sample / frame Huffman coding versus

In normal filter When extracting the frequency band, if it is not an ideal filter, since aliasing may occur at the subsampling point, 32-band repetitive noise may be removed by using a poli-phase filter bank which reduces aliasing. By canceling each other, the degradation caused by the filter is minimized, and 384 samples, which are multiples of 32 and the highest coding efficiency in the range of temporal masking effects, are set to 1 frame, that is, 12 per subband. Subband encoding is performed by allocating samples to convert them into a total of 32 bands.

In the layer I of MPEG-1 audio, a scale factor and a bit allocation method are used for encoding.

Accordingly, the 12 sample data in one band are divided into waveforms and magnifications, and the waveforms are normalized to have a maximum amplitude of 1.0 and the magnifications are encoded as a scale factor. Since this coding method limits the occurrence of quantization noise, the effect of auditory psychology is difficult to detect the noise.

In addition, the bit allocation can be adjusted independently for each frame and each subband. In combination with the scale factor, the quantization precision can be specified up to the masking level limit in consideration of the critical band so that the masking effect can be used more effectively. do.

In addition, the layer-1 of MPEG-1 audio adopts a stereo encoding method for encoding efficiency by combining two channels together.

This is a coding method that increases the encoding efficiency in stereo encoding by transmitting the waveform in monaural at high frequencies and stereo transmitting only amplitude information indicating a loudness at high frequencies by using the hearing phase detection ability at high frequencies.

On the other hand, in the layer II of MPEG-1 audio, the following two methods are additionally adopted in the layer I described above.

That is, one efficiently represents bit allocation information using bit allocation tables prepared according to the respective bit rates, and the other one encodes a waveform by adopting a mode in which three samples of the same band are collected and encoded. It is to increase the efficiency.

Layer III of MPEG-1 audio is characterized by adopting the MS side (Middle Side stereo) method, Modified Discrete Cosine Transform (MDCT), Huffman coding, and the like.

In the MS stereo system, instead of encoding two channels of stereo signals, that is, the left signal Left and the right signal R, respectively, the sum signal L + R and the difference signal LR) is encoded. Since the sum signal is treated more importantly in the stereo signal, this method is efficient in increasing the coding efficiency at a low bit rate by allocating a large number of bits to the sum signal (L + R). In other words, the effect is good for a signal having a small component of the difference signal L-R.

The transform discrete cosine transform divides the frequency band more precisely than 32 bands, thereby making it possible to use the hearing psychology more efficiently by dividing the frequency band more precisely, and the Huffman coding is obtained by the transform discrete cosine transform. Entropy coding the discrete cosine transform coefficients removes the statistical redundancy inherent in the audio data.

Layer I, Layer II, and Layer III of MPEG-1 audio having the above characteristics commonly have an AAU (Audio Access Unit) as a basic unit of audio decoding.

Here, AAU refers to one frame of the MPEG-1 audio bit string, which is the smallest unit that can decode the audio signal alone, and is a sample of layer I, for example, 384 x bit rate [sample], which is always a constant sample even in other conflicts. It consists of a number of data.

1 shows the configuration of an AAU of MPEG-1 audio.

As shown in Fig. 1, an AAU is composed of a 32-bit header, audio data and ancillary data, and a 16-bit CRC is selectively inserted between the header and the audio data. Can be.

The head is a 12-bit syncword of all 1s in a sync pattern, an 1-bit identification always having 1, a 2-bit layer bit specifying a layer, 1-bit protection bit to indicate the presence or absence of error check (or error check if 0), 2-bit bit rate index bit to indicate bit rate, and sampling frequency 2 bits of sampling frequency bits, 1 bit of padding bits indicating the presence or absence of padding (padding frame if 1), and 1 bit of private use bits. A 2-bit mode bit, a 2-bit mode extension bit, a 1-bit copyrighter indicating copyright (if 1 is copyrighted), and an original Or 1 bit original / copy bit indicating whether or not to copy ), And a total of 32 bits in the order of two bits of emphasis bits showing emphasis characteristics.

In the above, the 2-bit hierarchical bit designates a layer by the combination of each bit as shown in Table 2, the 2-bit sampling frequency bit indicates the sampling frequency by the combination of 2 bits as shown in Table 3, As shown in Table 4, the 4-bit bit rate index bit designates the bit rate in each layer by a combination of four bits.

Structure of Layer Bits in MPEG-1 Audio AAU Header Layer bits Designated hierarchy 1 1 Tier I 1 0 Tier II 0 1 Tier III 0 0 Reserved

Configuration of Sampling Frequency Bits in MPEG-1 Audio AAU Headers Sampling frequency bits Specified sampling frequency 1 1 Reserved 1 0 32 kHz 0 1 48 kHz 0 0 44.1 kHz

Structure of Bitrate Index Bits in MPEG-1 Audio AAU Headers Bitrate_index bit Bitrate Tier I Tier II Tier III 0 free format free format free format One 32 kbps 32 kbps 32 kbps 10 64 kbps 48 kbps 40 kbps 11 96 kbps 56 kbps 48 kbps 100 128 kbps 64 kbps 56 kbps 101 160 kbps 80 kbps 64 kbps 110 192 kbps 96 kbps 80 kbps 111 224 kbps 112 kbps 96 kbps 1000 256 kbps 128 kbps 112 kbps 1001 288 kbps 160 kbps 128 kbps 1010 320 kbps 192 kbps 160 kbps 1011 352 kbps 224 kbps 192 kbps 1100 384 kbps 256 kbps 224 kbps 1101 416 kbps 320 kbps 256 kbps 1110 448 kbps 384 kbps 320 kbps 1111 forbidden forbidden forbidden

In addition, the audio data area includes a bit allocation area, a scale factor area, and a sample area, and each data amount may vary according to a value of a header or a bit allocation area.

Auxiliary data area can be defined and used by the user, and when the audio data does not fill a fixed frame length in forming a single frame, the auxiliary data area is filled with additional data having no special meaning so that the frame has a fixed length. Play a role, but not a regulation.

For further explanation of MPEG-1 audio, see ISO / IEC JTC / SC29 / WG11 Coding of Moving Pictures and Associated Audio for Digital Media at up to about 1.5Mbps, part3, Audio (ISO / IEC). 11172-3).

MPEG-2 audio accommodates most of the header configuration and tradeoffs of MPEG-1 audio described above. Accordingly, the AAU of MPEG-2 audio is similar to MPEG-1 audio, and the AAU is composed of a 32-bit header, audio data, and ancillary data. It can optionally be inserted between the header and the audio data.

The audio data area is divided into an MPEG-1 audio area and a multi-channel audio area. The multi-channel area includes a 32-bit multichannel header, a 16-bit multichannel CRC, a multichannel composite status information area, a multichannel data area, and a multiword It consists of a multilingual data area.

MPEG-2 audio can maintain compatibility with MPEG-1 audio as it is recorded in the auxiliary data area in MPEG-1 audio. The data amount of multi-channel audio can vary beyond the basic frame as it is variable. Accordingly, the first 1 bit of the multi-channel header has an extension bit indicating the presence or absence of an extended frame to indicate whether the remaining data is to be loaded in the extended frame when the multi-channel data exceeds the size of the basic frame. As mentioned above, MPEG-2 audio maintains compatibility with MPEG-1 audio, but when the extended frame is added, compatibility with MPEG-1 audio is not provided.

The header of MPEG-2 audio is composed of 32 bits identically to the header of MPEG-1 audio, and the bits allocated to each area are also the same, except for the ID bits, bit rate index bits, and sampling frequency bits. It has the same meaning and role. Here, the ID bit serves as an identifier for each algorithm. That is, as described above, the ID bit is always set to 1 in the header of MPEG-1. If the ID bit is 1, the MPEG-1 audio is used. If the ID bit is 0, the MPEG-2 audio is stored. This represents an extension to the sampling frequency.

As shown in Table 5, the 4-bit bit rate index bit designates the bit rate in each layer by a combination of four bits.

Structure of Bitrate Index Bits in MPEG-2 Audio AAU Headers Bitrate_index bit Bitrate Tier I Tier II, Tier III 0 free format free format One 32 kbps 8 kbps 10 64 kbps 16 kbps 11 96 kbps 24 kbps 100 128 kbps 32 kbps 101 160 kbps 40 kbps 110 192 kbps 48 kbps 111 224 kbps 56 kbps 1000 256 kbps 64 kbps 1001 288 kbps 80 kbps 1010 320 kbps 96 kbps 1011 352 kbps 112 kbps 1100 384 kbps 128 kbps 1101 416 kbps 144 kbps 1110 448 kbps 160 kbps 1111 forbidden forbidden

The two-bit sampling frequency bits represent the sampling frequency by a combination of two bits as shown in Table 6.

Configuration of Sampling Frequency Bits in MPEG-2 Audio AAU Headers Sampling frequency bits Specified sampling frequency 1 1 Reserved 1 0 16 kHz 0 1 24 kHz 0 0 22.05 kHz

That is, the syntax, semantics, and coding schemes of ISO / IEC 11172-3 (MPEG-1 Audio) are excluded except for the newly defined sampling frequency field, bit rate index field, and bit allocation table in MPEG-2 audio. coding techniques) continue to be used. Here, the sampling frequency field, the bit rate index field, and the bit allocation table newly defined in ISO / IEC 13818-3 (MPEG-2 audio) have an identification bit set to 1 in the ISO / IEC 11172-3 header. It is obvious that it is valid when it is zero.

Hereinafter, the present invention is related to CRC processing of MPEG-2 audio, and the CRC of MPEG-2 audio will be described in detail with reference to the above-described description of MPEG-2 audio AAU.

For a detailed description of MPEG-2 audio except CRC, refer to ISO / IEC JTC / SC29 / WG11 Generic Coding of Moving Pictures and Associated Audio, part3, Audio (ISO / IEC 13818-3).

In MPEG-2 audio, the CRC processing is performed in the same format as in MPEG-1 audio, except that each processing is performed independently over the base frame and the extended frame.

If the protection bit of the header of the basic frame is set to 0, a 16-bit CRC check word is inserted into the bit string immediately after the header. The error detection technique used for MPEG-2 audio includes the generator polynomial (G (X)).

G (X) = X ^ 16 + X ^ 15 + X ^ 2 + 1

CRC-16, which is well known as Equation 1, is used.

Among the data of the basic frame, bits included in the CRC check are shown in Table 7.

Protected fields included in CRC check in basic frame hierarchy (layer) Bits Included in CRC-Check (protected field) Tier I Header bit allocation information from 16 bits to 31 bits Tier II Header bit allocation information from 16 bits to 31 bits Tier III Headers from 16 to 31 bits Side information:-Single channel mode: 0 to 135 bits of audio data-Other modes: 0 to 255 bits of audio data

CRC processing is performed using the CRC buffer shown in FIG. 2, and an initial state of the shift register is 1111 1111 1111 1111, and bits included in a CRC check are applied as inputs, and the CRC buffer is MPEG-. 2 To generate a word to be compared with the CRC check word included in the bit string of the audio, the generated polynomial is generated as b_15 ..... b_0 as an output.

The 16-bit multi-channel CRC in the multi-channel region is not only used for error detection of the multi-channel region but also indicates whether the multi-channel information or the multilingual information is useful.

The data included in the multi-channel CRC check includes the first bit of the multi-channel header up to the last bit of scale factor selection information. Accordingly, the basic frame performs a CRC check using a CRC check word and a multi-channel CRC check word.

On the other hand, the extended frame is composed of a header for the extended frame, an extended frame CRC region and extended data. As described above, the multi-channel header includes bits indicating the presence or absence of the extended frame to indicate the presence or absence of the extended frame.

The CRC check on the extended frame is to perform a CRC check on the initial 128 bits including the head area and the extended data area. When the length of the extended frame is not 128 bits, the CRC check is performed on the mode bits included within 128 bits. Perform a CRC check on the

In the CRC check for the basic frame and the extended frame, if the CRC check word and the output of the CRC buffer (b_15 ..... b_0) do not coincide, an error occurs in a protected field of the bit string. Error concealment techniques, such as muting the sound of an errored frame or repeating the previous frame, are used to avoid audible loud unpleasant distortion.

However, for the CRC check of MPEG-2 audio, if the bit string of the MPEG-2 audio includes an extended frame, the CRC check needs to be performed on both the basic frame and the extended frame. There is a problem that a CRC checking method and a hardware CRC checking means are required.

Accordingly, the present invention has been made to solve such a problem, and as the demand for audio providing more realistic stereo sound performance than the conventional stereo system increases, the basic frame of MPEG-2 audio and In the CRC check for extended frames, an error detection apparatus for MPEG-2 audio that can easily perform CRC processing on the basic frame and extended frame of MPEG-2 audio by sharing a CRC register and using a CRC counter. And to a method.

1 is an exemplary diagram showing an AAU configuration of MPEG-1 audio;

2 is a block diagram showing a CRC buffer of a typical MPEG-2 audio;

3 is a block diagram showing an error detection apparatus for MPEG-2 audio according to the present invention;

4 is a timing diagram showing waveforms for each part of FIG. 3;

〈Description of the symbols for the main parts of the drawings〉

10: CRC execution unit 100: input unit

110: basic register section 120: extension register section

130: multiplexer 200: error detector

210: CRC register section 220: CRC buffer section

230: CRC comparison unit 300: control unit

The apparatus and method for error detection of MPEG-2 audio according to the present invention for achieving the above object is a basic CRC check word and an extended CRC from the basic frame and the extended frame of the input bit stream in the input unit in MPEG-2 audio. The error detection unit receives the output of the CRC input unit 100 and the input bit string after outputting the 16-bit parallel data by separating the test word and outputting the multiplexed data, thereby checking CRC of the basic frame and the extended frame. After generating the basic CRC comparison word and the extended CRC comparison word by using the generation polynomial for the data, it is compared with the CRC check words of the input bit string and detects whether an error of the input data occurs.

An error detection apparatus for MPEG-2 audio according to the present invention will be described with reference to FIG.

As shown in FIG. 3, the error detection apparatus of MPEG-2 audio receives 16 bits of an input bit string and separates the basic CRC check word and the extended CRC check word from the base frame and the extended frame in the MPEG-2 audio. A basic data generated by using a polynomial generated for each CRC test data by receiving the input unit 100 and multiplexing the output data after making the parallel data of the input unit 100 and outputting the input bit string A CRC performer (10) comprising an error detector (200) for detecting whether an error occurs in the input data by comparing a CRC comparison word and an extended CRC comparison word with a basic CRC check word and an extended CRC check word, respectively; A frame discrimination signal P2EN for distinguishing the extended frame, a basic frame enable signal ASTAT indicating the basic frame section, and the extended frame The CRC counter signal CRC_CNT is generated by generating an extended frame enable signal BSTAT indicating an interval, a CRC buffer enable signal CRC_BUF_EN for enabling CRC buffering, and a CRC counter signal CRC_CNT for counting bits of a CRC check interval of an extended frame. It is composed of a control unit 300 for controlling the execution unit 10.

Here, the input unit 100 receives the input bit string and the basic register unit 110 which separates the CRC check word from the basic frame of the input bit string into 16 bits of parallel data, and expands the input bit string. An extension register unit 120 that separates an extended CRC check word from a frame into 16-bit parallel data, and a multiplexer 130 that multiplexes outputs of the base register 110 and the extension register 120. It is composed of

In addition, the error detection unit 220 receives the output of the multiplexing unit 130 and receives the CRC register unit 210 and the input bit string received by using the generated polynomial using the basic CRC check data and the A CRC buffer 220 for generating a basic CRC comparison word and an extended CRC comparison word for the extended CRC check data, the basic CRC check word and the extended CRC check word, the basic CRC comparison word, and the extended CRC It is composed of a CRC comparison unit 230 for comparing the comparison words to detect the presence of an error.

Hereinafter, the operation of the error detection apparatus for MPEG-2 audio according to the present invention configured as described above will be described with reference to FIG.

As described above, the frame structure of MPEG-2 audio is composed of a basic frame and an extended frame, and each of the 16-bit CRC check words needs to be performed twice.

First, as shown in the timing diagram shown in FIG. 4, the CRC check for the basic frame includes a frame discrimination signal P2EN that distinguishes the basic frame and the extended frame from the time t_1. And a basic frame enable signal ASTAT indicating a basic frame period is high, and at the same time, a CRC buffer enable signal for enabling the CRC buffer unit 220 at a time t_2. The CRC check is performed in the section where (CRC_BUF_EN) is high.

That is, the basic register unit 110 receives the input bitstream from the time t_1 and separates the basic CRC check word from the basic frame into 16 bits of parallel data to form the multiplexer 130. To the CRC register 210, and the CRC buffer unit 220 is enabled by the CRC buffer enable signal CRC_BUF_EN starting at time t_2 and using the CRC-16 generation polynomial to perform the basic CRC check data. When the basic CRC comparison word is generated, the CRC comparison unit 230 detects an error by comparing the output of the CRC register 210 with the output of the CRC buffer 220.

In FIG. 4, the time t_2 to the time t_3 is a section for performing a CRC check on a stereo signal corresponding to MPEG-1 audio, and the time t_4 to a time t_5 is a section for performing a CRC check on a multi-channel region.

On the other hand, the CRC check on the extended frame indicates that the frame division signal P2EN indicates that the extended frame is active starting at time t_6, and the extended frame enable signal BSTAT indicating the extended frame section is high ( high), and at the same time, a CRC check is performed in a section in which the CRC buffer enable signal CRC_BUF_EN that enables the CRC buffer unit 220 is high.

The CRC check on the extended frame is to perform a CRC check on the initial 128 bits including the extended frame head area and the extended data area. If the length of the extended frame is not 128 bits, the mode is included within 128 bits. When the length of the extended frame exceeds 128 bits, the controller 300 counts the initial 128 bits of the extended frame using the CRC counter signal CRC_CNT at time t_7. (A)) and provide the CRC execution unit 10, the CRC execution unit 10 performs the CRC check for the initial 128 bits using this.

On the other hand, when the length of the extended frame does not exceed 128 bits, the controller 300 counts all data included within the initial 128 bits of the extended frame using the CRC counter signal CRC_CNT at time t_7 (CRC_CNT (B)). ) And provide it to the CRC execution unit 10, the CRC execution unit 10 performs a CRC check on the mode bits included within 128 bits.

That is, the expansion register unit 120 receives a serial bitstream from time t_6 and separates the CRC check word from the expansion frame into 16 bits of parallel data to form the multiplexer 130. The CRC buffer unit 220 is provided to the CRC register 210 through the CRC buffer enable signal CRC_BUF_EN at a time t_7 and uses the CRC-16 generation polynomial. When a value to be compared with the extended frame CRC check word is generated, the CRC comparison unit 230 detects an error by comparing the value of the CRC register unit 210 with the value of the CRC buffer unit 220.

An error detection method of MPEG-2 audio according to the present invention will be described below.

The error detection method of MPEG-2 audio according to the present invention detects a basic CRC check word from an input bit string, and then uses the generated polynomial for basic CRC check data, which is data included in the CRC check, from the data in the basic frame. A basic CRC check step of performing a CRC check on a basic frame by comparing the basic CRC check word and the basic CRC comparison word after generating a basic CRC comparison word, and when the input bit string includes an extended frame, After detecting an extended CRC check word from the input bit string, counting an initial predetermined bit of an extended frame using a CRC counter to perform the extended CRC using the generation polynomial for data less than or equal to the initial predetermined bit. A ratio between the extended CRC check word and the extended CRC compare word after generating a compare word It consists of an extended CRC check step of performing a CRC check on the extended frame.

Here, in the basic CRC checking step, the frame discrimination signal P2EN that distinguishes the basic frame and the extended frame represents the basic frame, and the input is performed when the basic frame enable signal ASTAT indicating the basic frame period is activated. A basic CRC check word detection step of detecting a basic CRC check word from a bit string; and after generating the basic CRC check word, a CRC buffer enable signal (CRC_BUF_EN) that enables CRC buffering is activated, thereby generating the polynomial A basic CRC comparison word generation step of generating a basic CRC comparison word for a basic frame with respect to the basic CRC check data by using an error, and comparing the basic CRC check word with the basic CRC comparison word to obtain an error for the basic frame Basic CRC comparison detection step for detecting whether or not occurrence.

On the other hand, in the extended CRC checking step, if there is an extended frame determination step of detecting whether the extended frame is included in the input bit string, and if the extended frame is present as a result of the determination of the extended frame presence determination step, the frame is classified. An extended frame counting data included in an initial 128 bits or less of an extended frame in a bit unit while the signal P2EN indicates the extended frame and the extended frame enable signal BSTAT indicating the extended frame section is activated. As a result of the counting step and the counting step of the frame counting step, if the bit count is 128 bits or more, the CRC buffer enable signal (CRC_BUF_EN) is activated to perform CRC checking only up to 128 bits. Activate (CRC_BUF_EN) to target all data belonging to the extended frame Consists of expansion compared detection step of detecting whether an error occurs on the extended frame by comparing the comparison word and the extended CRC generation step of generating the extended CRC comparison word, the extended CRC check word with the comparison word CRC expansion.

Hereinafter, the procedure of the error detection method of MPEG-2 audio according to the present invention configured as described above will be described in detail.

First, when an input bit string of MPEG-2 audio is applied, the basic CRC check step detects a basic CRC check word from the input bit string and then selects basic CRC check data, which is data included in the CRC check, from the data in the basic frame. After generating a basic CRC comparison word using a generated polynomial, the CRC check is performed on the basic frame through comparison between the basic CRC check word and the basic CRC comparison word.

That is, in the basic CRC check word detection step, the frame discrimination signal P2EN that distinguishes the basic frame and the extended frame represents the basic frame, and the basic frame enable signal ASTAT indicating the basic frame period is activated. A basic CRC check word is detected from the input bit string, wherein the basic CRC check word is composed of a CRC check word corresponding to MPEG-1 audio and a multi-channel CRC check word corresponding to multi-channel audio.

After detecting the basic CRC check word, in the basic CRC comparison word generation step, as the CRC buffer enable signal CRC_BUF_EN that enables CRC buffering is activated, a CRC-16 generation polynomial such as Equation 1 is used. A basic CRC comparison word for a basic frame is generated for the basic CRC check data. The basic CRC comparison word includes a CRC comparison word corresponding to stereo audio (MPEG-1 audio) and a multichannel CRC corresponding to multichannel audio. It consists of a comparison word.

Subsequently, in the basic CRC comparison detection step, the basic CRC check word and the basic CRC comparison word are compared with each other to detect whether an error occurs in the basic frame data composed of the stereo audio and the multi-channel audio.

After performing the basic CRC check step, in the extended CRC check step, if an extended frame is included in the input bit string, the extended CRC check word is detected from the input bit string, and then a CRC counter is used to detect the extended frame. Counting an initial predetermined bit to generate the extended CRC comparison word using the generated polynomial for the data less than the initial predetermined bit, and then comparing the extended CRC check word with the extended CRC comparison word. Perform CRC check on extended frames.

In the step of checking the basic CRC step by step, in the extended frame determination step, it is detected whether the extended frame is included in the input bit string by decoding a multi-channel header, and as a result of the determination of the extended frame presence determination step, If there is an extended frame, in the extended frame counting step, the frame division signal P2EN indicates the extended frame, and an initial 128 of the extended frame is activated when the extended frame enable signal BSTAT indicating the extended frame section is activated. Count the data included in the bits below by bit unit.

Subsequently, if the counting result of the frame counting step is greater than or equal to 128 bits, the CRC buffer enable signal CRC_BUF_EN is activated in the extended CRC comparison word generation step to perform CRC checking only up to 128 bits, and if less than 128 bits, The CRC buffer enable signal CRC_BUF_EN is activated to generate the extended CRC comparison word for all data belonging to the extended frame.

Accordingly, in the extended comparison detecting step, the extended CRC check word and the extended CRC comparison word are compared to detect whether an error occurs in the extended frame.

As described in detail above, according to the apparatus and method for error detection of MPEG-2 audio according to the present invention, in MPEG-2 audio, the input unit separates the CRC check word from the basic frame and the extended frame of the input bit stream, respectively. After generating the 16-bit parallel data and multiplexing them, the error detector outputs the CRC input unit and generates the CRC check data of the basic data and the extended data included in the input bit string. The occurrence of an error in the input bit string is detected by comparing the CRC check word with the CRC comparison word of the input bit string. In addition, in carrying out the present invention, CRC processing for the basic frame and the extended frame can be easily performed by sharing the CRC register and using the CRC counter.

Therefore, it is easy to solve a problem such as the need for a CRC checking method and a hardware CRC checking means capable of satisfying the rather complicated condition of CRC checking between the basic frame and the extended frame, and sharing the CRC register. This has the advantage of reducing design costs.

Claims (14)

In the error detection device of MPEG-2 audio, After receiving an input bit string and classifying it into a base frame and an extended frame, the base CRC check word and the extended CRC check word are separated from the base frame and the extended frame, respectively, and multiplexed according to the header information of the input bit string. An input unit; A basic CRC comparison word and an extension by receiving an output of the input unit and the input bit string and applying a predetermined error detection rule to basic CRC check data and extended CRC check data included in the basic frame and the extended frame, respectively After generating a CRC comparison word, an error detection unit for detecting whether the error occurs for the input data by comparing the basic CRC check word and the extended CRC check word and the basic CRC comparison word and the extended CRC comparison word, respectively; and, The input unit separates the basic CRC check word from the basic frame into first parallel data; an extension register unit separating the extended CRC check word from the extended frame into second parallel data; And a multiplexing unit configured to multiplex and output the output of the basic register unit and the output of the extension register unit according to header information of the input bit string. 2. The apparatus of claim 1, wherein the error detector comprises: a CRC register unit for receiving and storing an output of the multiplexer; A CRC buffer unit configured to receive the basic CRC check data and the extended CRC check data and generate a basic CRC comparison word and an extended CRC comparison word by applying a general polynomial; Comparing the basic CRC comparison word with a basic CRC check word to detect whether an error occurs in the basic frame, and comparing the extended CRC comparison word with the extended CRC check word to detect whether an error occurs in the extended frame An error detection apparatus for MPEG-2 audio, comprising a CRC comparison unit. 2. The error detection apparatus of MPEG-2 audio according to claim 1, wherein the first parallel data and the second parallel data are 16 bits. The apparatus of claim 2, wherein the generated polynomial is a generated polynomial of CRC-16. Cyclic Redundancy Check (CRC) checking of MPEG-2 audio; After detecting the basic CRC check word from the input bit stream, after generating the basic CRC comparison word using the polynomial generated for the basic CRC check data, which is data included in the CRC check in the basic frame, the basic CRC check word and the A basic CRC checking step of detecting whether an error occurs in the basic frame by comparing the basic CRC comparison words; When the extended bit is included in the input bit string, the extended CRC check word is detected from the input bit string, and counting is performed on an initial predetermined bit of the extended frame by using a CRC counter. An extended CRC check step of generating an extended CRC comparison word using the generated polynomial and then detecting whether an error has occurred for the extended frame by comparing the extended CRC check word and the extended CRC comparison word; and, The extended CRC check step, An extended frame determination step of detecting whether the extended frame is included in the input bit string; As a result of the determination of the presence or absence of the extended frame, an initial predetermined bit of the extended frame is provided when the extended frame enable signal indicates the extended frame and the extended frame enable signal indicating the extended frame section is active. An extended frame counting step of counting data included below; As a result of counting in the frame counting step, if the initial predetermined bit or more, the CRC buffer enable signal is activated to perform a CRC check up to the predetermined bit; if less than the initial predetermined bit, the CRC buffer enable signal is activated. Generating an extended CRC comparison word for all data belonging to an extended frame; And an extended comparison detection step of detecting whether an error for the extended frame occurs by comparing the extended CRC check word and the extended CRC comparison word. 6. The input CRC of claim 5, wherein the checking of the basic CRC comprises: a frame discrimination signal for distinguishing the basic frame from the extended frame indicates the basic frame and the input bit in a state in which a basic frame enable signal indicating a basic frame period is activated A basic CRC check word detection step of detecting the basic CRC check word from a column; After detecting the basic CRC check word, a basic CRC for generating the basic CRC comparison word for the basic CRC check data by using the generation polynomial as the CRC buffer enable signal for enabling CRC buffering is activated. A comparison word generation step; And a basic CRC comparison detection step of detecting whether an error has occurred in the basic frame by comparing the basic CRC check word and the basic CRC comparison word. 6. The method of claim 5, wherein the initial predetermined bit is 128 bits. 6. The method of claim 5, wherein the determining whether the extended frame is present or not comprises determining an extended frame by decoding a multi-channel header of the basic frame. 6. The method of claim 5, wherein the generated polynomial is a generated polynomial of CRC-16. An apparatus for detecting a data error of an input bit string consisting of a basic frame and an extended frame; After receiving the input bit string and classifying it into a basic frame and an extended frame, the basic error check word and the extended error check word are separated from the basic frame and the extended frame, respectively, and multiplexed according to header information of the input bit string. An input unit to perform; A basic error comparison word and an extension by applying a predetermined error detection rule to basic error check data and extended error check data included in the basic frame and the extended frame, respectively, by receiving the output of the input unit and the input bit string; After generating an error comparison word, and comparing the basic error check word and the extended error check word, the basic error comparison word and the extended error comparison word, respectively, an error detection unit for detecting whether an error occurs for the input data; and, The input unit separates the basic error check word from the basic frame into first parallel data; an extension register unit separating the extended error check word from the extended frame into second parallel data; And a multiplexing unit configured to multiplex and output the output of the basic register unit and the output of the extension register unit according to header information of the input bit string. 11. The apparatus of claim 10, wherein said error detection rule uses a generated polynomial. The method of claim 10, wherein the error detection check word comprises a basic cyclic redundancy check (CRC) check word and an extended CRC check word included in the basic frame and the extended frame of the input bit stream, respectively. The CRC check data includes basic CRC check data and extended CRC check data included in the basic frame and the extended frame, respectively. The CRC comparison word is composed of a basic CRC comparison word generated using the generated polynomial for the basic CRC check data and an extended CRC comparison word generated using the generated polynomial for the extended CRC check data. Data error detection device. 13. The apparatus of claim 12, wherein the error detector comprises: a CRC register unit for receiving and storing an output of the multiplexer; A CRC buffer unit which receives the input bit string and generates the basic CRC comparison word and the extended CRC comparison word for the basic CRC check data and the extended CRC check data by using a generation polynomial; And a CRC comparator for detecting whether an error has occurred by comparing the basic CRC check word and the extended CRC check word with the basic CRC comparison word and the extended CRC comparison word, respectively. 12. The apparatus of claim 11, wherein the generated polynomial is a generated polynomial of CRC-16.