KR20160078321A - Apparatus for encoding/decoding multichannel signal and method thereof - Google Patents

Abstract

Encoding/decoding devices of multichannel signals and methods thereof are disclosed. The encoding device of a multichannel signal and the method thereof encode phase information of a multichannel signal through quantization and lossless encoding. The decoding device of a multichannel signal and the method thereof decode phase information through inverse quantization and lossless decoding. Therefore, the encoding/decoding devices efficiently encode/decode phase parameters to reduce the number of bits used in expressions of the phase parameters.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to an apparatus and a method for encoding / decoding multi-channel signals,

Embodiments of the present invention relate to an apparatus and method for encoding / decoding multi-channel signals, and more particularly, to a multi-channel signal encoding apparatus / method for encoding phase information through quantization and lossless encoding And a device / method for decoding a multi-channel signal that decodes phase information through inverse quantization and lossless decoding.

A method used to encode a stereo signal is parametric stereo (PS) technology. The parametric stereo technique generates a mono signal by downmixing an input stereo signal, extracts a stereo parameter indicating side information on the stereo signal, codes the generated mono signal and the extracted stereo parameter, And encodes the signal.

In this case, the stereo parameters used include Inter-channel Intensity Difference (IID) or Channel Level Differences (CLD) indicating intensity differences according to energy levels of at least two channel signals included in the stereo signal, at least two channels (Inter-channel Coherence or Inter-channel Correlation) indicating the correlation between two channel signals according to the similarity of the waveform of the signal, an inter-channel coherence or inter-channel correlation And OPD (Overall Phase Difference) indicating how the phase difference between at least two channel signals included in the stereo signal is distributed between the two channels based on the mono signal.

Embodiments of the present invention can provide an apparatus and method for encoding / decoding multi-channel signals for efficiently encoding / decoding phase parameters.

An apparatus for encoding a multi-channel signal according to an embodiment of the present invention includes a parameter extraction unit that extracts a plurality of parameters indicating a characteristic relationship between a plurality of channels constituting a multi-channel signal, a parameter quantization unit that quantizes the plurality of parameters, A parameter encoding unit for encoding a plurality of quantized parameters, a mono signal encoding unit for encoding a mono signal obtained by downmixing the multi-channel signal, and a multi-channel signal encoding unit for encoding the multi-channel signal using the plurality of encoded parameters and the encoded mono signal. And the parameter encoding unit encodes the quantized phase parameters using a Huffman coding scheme. The parameter encoding unit may encode the quantized phase parameters using a Huffman coding scheme.

Also, an apparatus for decoding a multi-channel signal according to an embodiment of the present invention includes a mono signal decoding unit for decoding a mono signal, which is a downmix signal of the multi-channel signal, from a bitstream of a multi- A parameter decoding unit for restoring a plurality of parameters indicating a characteristic relationship between a plurality of channels constituting a multi-channel signal; a parameter decoding unit for decoding a parameter (OPD) related to a phase difference between the restored monaural signal and the multi- : A parameter estimator for estimating an overall phase difference, an inverse quantization unit for inversely quantizing the reconstructed parameter and the estimated OPD, and a parameter inverse quantization unit for inversely quantizing the monaural signal using the inversely quantized parameter and the inversely quantized OPD. Upmixing unit.

According to the apparatus for encoding / decoding a multi-channel signal according to an embodiment of the present invention, the number of bits used for expressing a phase parameter can be reduced by efficiently encoding / decoding a phase parameter.

1 is a block diagram showing a detailed configuration of an apparatus for encoding a multi-channel signal according to an embodiment of the present invention.
FIG. 2 is a diagram for explaining a concept of encoding a parameter by applying 2D Huffman coding to parameter pairs constituted by parameters having no wrapping property. FIG.
3 is a diagram for explaining a concept of coding a parameter by applying 2D Huffman coding to parameter pairs constituted by phase parameters having a wrapping property.
4 is a block diagram illustrating a detailed configuration of an apparatus for decoding a multi-channel signal according to an embodiment of the present invention.
FIG. 5 is a flowchart for explaining the operation of a parameter dequantizer for dequantizing a parameter based on USAC (Unified Speech and Audio Coding).
FIGS. 6 and 7 are flowcharts for explaining the operation of a parameter de-quantization unit for dequantizing a parameter based on an MPS (MPEG Surround) lossless coding scheme.
8 is a flowchart for explaining an operation of a parameter de-quantization unit for dequantizing a parameter based on a PS (Parametric Stereo) lossless coding scheme.

Hereinafter, embodiments according to the present invention will be described in detail with reference to the accompanying drawings. However, the present invention is not limited to or limited by the embodiments. Like reference symbols in the drawings denote like elements.

1 is a block diagram showing a detailed configuration of an apparatus for encoding a multi-channel signal according to an embodiment of the present invention.

The apparatus 100 for encoding a multi-channel signal according to an exemplary embodiment of the present invention includes a parameter extracting unit 110, a parameter quantizing unit 120, a parameter encoding unit 130, a downmixing unit 140, A bitstream generation unit 150, and a bitstream generation unit 160. Hereinafter, the function of each component will be described in detail.

Here, the multi-channel signal means a signal of a plurality of channels. In the present specification, each of a plurality of channels included in the multi-channel signal is referred to as a channel signal.

Hereinafter, for convenience of description, it is assumed that a multi-channel signal input to the multi-channel signal encoding apparatus 100 is a stereo signal including a left channel signal and a right channel signal. However, the apparatus 100 for encoding a multi-channel signal according to an exemplary embodiment of the present invention can be used not only for a stereo signal but also for encoding a multi-channel signal. .

The parameter extracting unit 110 extracts a plurality of parameters indicating the characteristic relationship between the left channel signal and the right channel signal constituting the stereo signal. The plurality of parameters may include the above-mentioned CLD, ICC, IPD, OPD, and the like. Here, IPD and OPD are examples of phase parameters related to the phase information between the left channel signal and the right channel signal.

The parameter quantization unit 120 quantizes the extracted plurality of parameters, and the parameter encoding unit 130 encodes the plurality of quantized parameters.

In this case, since OPD can be estimated from other parameters, according to an embodiment of the present invention, the parameter encoding unit 120 encodes only CLD, ICC, and IPD among the extracted parameters, OPD does not encode . That is, the multi-channel signal encoding apparatus 100 can reduce the number of bits of the bitstream to be transmitted by not encoding and transmitting the OPD. A more detailed description of the OPD estimation will be made with reference to the description of the multi-channel signal decoding apparatus 400 of FIG.

According to an embodiment of the present invention, the parameter quantization unit 120 may quantize CLD or ICC. In this case, CLD can be quantized to 8 levels or 16 levels, and ICC can be quantized to 8 levels.

Further, according to an embodiment of the present invention, the parameter quantization unit 120 can quantize phase parameters such as IPD or OPD in addition to CLD and ICC. In this case, the phase parameter may be quantized to 8 levels or 16 levels.

When the parameter quantization unit 120 quantizes CLD or ICC, the parameter encoding unit 130 can encode the quantized CLD or the differential value of the quantized ICC between the current frame and the previous frame. That is, the parameter encoding unit can encode CLD or ICC using a differential coding technique.

For example, when CLD or ICC is quantized to 8 levels in the parameter quantization unit 120, CLD or ICC has a 15-step difference range from -7 to 7, and the parameter encoding unit 130 It is possible to encode CLD or ICC having a differential range of 15 steps. In another example, when the CLD is quantized to 16 levels in the parameter quantization unit 120, CLD or ICC has a differential range of 31 steps from -15 to 15, and the parameter coding unit 130 calculates the differential range CLD < / RTI >

When the parameter quantization unit 120 quantizes the phase parameter, the parameter encoding unit 130 can encode the change amount (difference) of the quantized phase parameter between the current frame and the previous frame.

For example, when the parameter quantization unit 120 quantizes the phase parameter to 8 levels, the parameter encoding unit 130 can encode the phase parameter having the differential range of 15 steps. As another example, when the parameter quantization unit 120 quantizes the phase parameter to 16 levels, the parameter encoding unit 13 can encode the phase parameter having the difference range of 31 steps

Since the phase has a wrapping property with respect to 360 degrees (for example, 370 is equal to 10 degrees), if the phase parameter is quantized using the wrapping property, the bit allocated to the representation of the phase information It is possible to reduce the number.

According to an embodiment of the present invention, the parameter encoding unit 130 may apply a modulo operation to the difference of the phase parameter having the wrapping property so that the difference of the phase parameter has the enhancement positive value. Accordingly, the parameter encoding unit 130 does not transmit the sign bit assigned to the phase parameter, so that it is possible to reduce the number of bits used for expressing the phase parameter. In addition, when the modular operation is applied, discontinuity occurs between the phase parameters due to the wrapping property, thereby solving the problem that the entropy of the phase parameter distribution increases.

When the phase parameter is quantized to 8 levels in the parameter quantization unit 120, the phase parameter has a differential range from 0 to 7. When quantizing the phase parameter at 16 levels in the parameter quantization unit 130, And has a difference range of 0 to 15. Table 1 shows an example of quantization values of phase parameters quantized at 8 levels, and Table 2 shows an example of quantization values of phase parameters quantized at 16 levels.

As an example, when the phase parameter is the IPD quantized to 8 levels, the parameter encoding unit 130 can calculate the difference value of the phase parameter by applying the modulo-8 operation. This can be expressed as Equation (1).

'는 현재 프레임에서의 IPD 값, '

'는 이전 프레임에서의 IPD 값, '%8'은 Module-8 연산을, '

'는 Module?8 연산을 통해 도출된 IPD의 차분 값을 각각 의미한다. here, '

'Is the IPD value in the current frame,'

'Is the IPD value in the previous frame,'% 8 'is the Module-8 operation,

'Means the difference value of the IPD derived from the Module? 8 operation.

값의 일례들을 나타내고 있다. 표 3에서, 'Previous'는

과, 'Current'는

와, 'Diff.'은

과, 'Diff with 모듈러'는

와 각각 대응된다. Table 3 shows the results of Module-8 computation.

Values. ≪ / RTI > In Table 3, 'Previous'

And 'Current'

And 'Diff.'

And 'Diff with modular'

Respectively.

Generally, when 'Diff.' Is '0', it shows the lowest number of bits. As 'Diff.' Is closer to '0' or '7', the number of bits decreases. 0 is the difference from the previous frame, 7 is the difference from the previous frame is -1, 1 is the highest frequency because the difference from the previous frame is 1 You can allocate fewer bits.

The parameter encoding unit 130 may hatch the difference of the quantized parameter through lossless encoding. For example, the parameter encoding unit 130 may encode quantized parameters using Huffman coding.

Huffman coding is a type of entropy coding used for lossless coding and is a coding algorithm that uses codes of different lengths depending on the occurrence frequency of a symbol corresponding to each data.

Huffman coding schemes include 1D Huffman coding and 2D Huffman coding. 1D Huffman coding refers to a method of coding each data as one symbol, and 2D Huffman coding groups a plurality of data into a plurality of data groups including two data, and a plurality of data groups A pair of data consisting of a plurality of data) is regarded as one symbol and encoded.

In other words, when 1D Huffman coding is applied to parameters, the parameter coding unit 130 directly codes each parameter according to the Huffman coding scheme.

For example, when CLD or ICC is to be encoded, the parameter encoding unit 130 performs Huffman coding by considering a plurality of CLD values or a plurality of ICC values as one symbol.

In another example, when it is desired to encode a phase parameter, the parameter encoding unit 130 regards a plurality of phase parameter values as one symbol, and performs Huffman coding.

In addition, when parameters are encoded by applying 2D Huffman coding, the parameter encoding unit 130 encodes parameter pairs constituted by two parameters according to the Huffman coding scheme.

When coding a parameter by applying 2D Huffman coding, the correlation between the two parameters constituting the parameter pair is reduced. When a parameter is encoded by applying 2D Huffman coding, the bit rate is reduced because the joint probability of two symbols between two parameters constituting the parameter pair is considered.

Hereinafter, the operation of the parameter encoding unit 130 for encoding a parameter according to 2D Huffman coding will be described in detail with reference to FIG. 2 and FIG.

FIGS. 2 and 3 are diagrams for explaining the concept of 2D Huffman coding for coding parameter pairs.

2 is a diagram for explaining a concept of coding parameters by applying 2D Huffman coding to parameter pairs constituted by parameters having no wrapping property such as CLD or ICC.

Here, the parameter pair is represented by (X, Y). X and Y correspond to one parameter, respectively. For example, when the parameter type is CLD, X and Y correspond to one CLD value, respectively.

The parameter encoding unit 130 performs 2D Huffman coding according to the following three steps.

First, the parameter encoding unit 130 sets the value of X or Y so that the absolute value of X has a value larger than the absolute value of Y, that is, the relation of 'X |> Y' Change it. If the relation is satisfied, the pair of X / Y symbols has a larger value on the X axis, so that the degree of skew of the data increases, Can be reduced. Accordingly, the parameters existing in the second region are the first region, the parameters existing in the third region are the fourth region, the parameters existing in the sixth region are the fifth region, the parameters existing in the seventh region are Respectively. In this case, 1-bit additional information is assigned to each of the parameter pairs moving in the region.

In the next step, the parameter encoding unit 130 changes the X or Y value so that the sum of X and Y has a value smaller than '0', that is, a relation of '(X + Y) <0'. '(X + Y) <0', by changing the sign of X and Y, the value of the generated symbol has a positive value.

Accordingly, parameters existing in the fourth region move to the first region and parameters existing in the fifth region move to the eighth region, respectively. In this case as well, 1-bit additional information is given to each of the parameter pairs moving in the region. The pair of (X, Y) in which the region is shifted in the second step is referred to as (X ', Y').

In the next step, the parameter encoding unit 130 derives a parameter pair finally encoded through LAV coding using a parameter having a maximum absolute value (LAV) among a plurality of parameters. LAV is additional information for 2D Huffman coding, which is also transmitted via Huffman coding. LAV means a value having the largest absolute value among the symbols corresponding to the current frame to be coded, and thus the codebook of 2D Huffman coding can be determined.

LAV is additional information for 2D Huffman coding, and LAV is encoded and transmitted through Huffman coding. LAV denotes a symbol value having the largest absolute value among the symbols corresponding to the current frame to be encoded. The parameter encoding unit 130 determines the codebook of 2D Huffman coding through the LAV.

LAV coding using LAV is a coding method for reducing the size of a table used for 2D Huffman coding and is a coding method in which each parameter is composed of '1', '3', '5', and '7' It is represented by four symbols.

In case of CLD or ICC, the frequency of occurrence of LAV is '7' is the least since the frequency of occurrence of CLD or ICC of the previous frame is larger than the difference of CLD value or ICC value of the current frame. CLD and ICC can apply the same Huffman codebook. The syntax for the Huffman codebook is shown in Table 4.

Here, {0x0, 0x2, 0x6, 0x7} denotes a codebook, {1, 2, 3, 3} denotes a length of a codebook, and {1, 3, 5, 7} denotes a LAV.

The parameter pair (X ", Y") finally encoded according to LAV coding can be derived according to the algorithm shown in Table 5. [

Here, 'LAV' denotes the maximum absolute value, and '% 2' denotes the modulo-2 operation.

3 is a diagram for explaining a concept of coding parameters by applying 2D Huffman coding to parameter pairs constituted by phase parameters having wrapping properties such as IPD or OPD.

In this case, 2D Huffman coding can be performed through the following two steps, unlike the 2D Huffman coding described above with reference to FIG.

First, the parameter encoding unit 130 sets the value of X or Y so that the absolute value of X has a value larger than the absolute value of Y, that is, the relation of 'X |> Y' Change it.

As described above, when the modulo operation is applied to the difference of the quantized phase parameters, the phase parameter always exists in a positive value, so that the phase parameter exists only in the first and second areas. Therefore, in the first step, the phase parameter existing in the second region moves to the first region. In this case, 1-bit additional information is assigned to each of the parameter pairs moving in the region.

If the pair of (X, Y) in which the region is shifted in the above step is (X ', Y'), X 'and Y' can be derived according to the following equation (2).

In the next step, the parameter encoding unit 130 derives a parameter pair finally encoded through LAV coding using a parameter having a maximum absolute value (LAV) among a plurality of parameters.

The parameter pair (X ", Y") finally encoded according to LAV coding can be derived according to the algorithm shown in Table 5 described above.

When the parameter encoding unit 130 encodes the phase parameter, unlike in the case of encoding a parameter such as CLD or ICC, one region shift occurs. Therefore, when the parameter encoding unit 130 encodes the phase parameter, the number of transmitted bits is smaller than that in the case of encoding a parameter such as CLD or ICC.

IPD has a different distribution of CLD or ICC and LAV. IPD has the highest frequency when the difference is '7'. For example, the IPD may have a LAV distribution as shown in Table 6. [

The codebook length allocation according to the probability distribution, which is a characteristic of Huffman coding, must reflect the length of the actual symbols and the codebook. If the actual distribution of the symbol and the length of the Huffman codebook are not matched, then optimal performance can not be achieved. Therefore, the parameter encoding unit 130 can solve the problem of distribution mismatching of the symbol distribution by remapping the symbols and encoding the remapped symbols. In this case, the parameter encoding unit 130 can remap symbols according to Equation (3).

Here, 'sym' denotes a symbol and '% 4' denotes a modulo-4 operation.

The relationship between the symbol after symbol remapping and the Huffman codebook can be expressed as shown in Table 7.

Tables 8 through 10 illustrate the syntax for 1D Huffman coding and 2D Huffman coding for coding phase parameters, according to one embodiment of the present invention.

Alternatively, there is a method of defining a table that maps LavIdx as follows. LavIdx denotes a symbol output through actual Huffman decoding, and lavTabIPD denotes an actual LAV of the symbol. In this case, the remapping of the symbols can be performed based on the table in Table 11. [

LavIdx lavTabIPD [ LavIdx ] 0 7 One One 2 3 3 5

Referring again to FIG. 1, an apparatus 100 for encoding a multi-channel signal according to an embodiment of the present invention will be described.

The downmixing unit 140 downmixes a stereo signal and outputs a mono signal.

Down-mixing is to generate a mono signal of one channel from a stereo signal of two or more channels, and it is possible to reduce the number of bits of the bit stream generated in the encoding process through downmixing. At this time, the mono signal may be a signal representative of the stereo signal. In other words, in the multi-channel signal encoding apparatus 100, only the mono signal can be coded and transmitted without encoding each of the left channel signal and the right channel signal included in the stereo signal.

For example, the magnitude of the mono signal can be obtained as an average value of the magnitudes of the left channel signal and the right channel signal, and the phase of the mono signal can be obtained as an average value of the phases of the left channel signal and the right channel signal.

The mono signal encoding unit 150 encodes the mono signal output from the downmixing unit 140.

As an example, when the stereo signal is a voice signal, the mono signal encoding unit 120 may encode the mono signal in a CELP (Code Excited Linear Prediction) method.

As another example, when the stereo signal is a music signal, the mono signal encoding unit 120 may encode a mono signal using a method similar to the conventional MPEG-2/4 AAC or mp3.

The bitstream generator 160 generates an encoded bitstream for a stereo signal using a plurality of encoded parameters and an encoded monaural signal.

In the above description, the multi-channel signal encoding apparatus 100 for encoding the phase information of the multi-channel signal through the quantization and lossless encoding and generating the bit stream including the encoded phase information has been described. Hereinafter, an apparatus for decoding a multi-channel signal according to an embodiment of the present invention will be described in detail with reference to FIG.

4 is a block diagram illustrating a detailed configuration of an apparatus for decoding a multi-channel signal according to an embodiment of the present invention.

The apparatus 400 for decoding a multi-channel signal according to an exemplary embodiment of the present invention includes a mono signal decoding unit 410, a parameter decoding unit 420, a parameter dequantization unit 430, a parameter estimating unit 440, And a mixing unit 450. Hereinafter, the function of each component will be described in detail.

Hereinafter, for the sake of convenience, it is assumed that the bit stream input to the multi-channel signal decoding apparatus 300 is a coded bit stream of a stereo signal.

The mono signal decoding unit 410 restores the mono signal, which is a downmix signal of the multi-channel signal, from the encoded bit stream of the stereo signal. Specifically, when the mono signal is coded in the time domain, the mono signal decoding unit 410 decodes the coded mono signal in the time domain, and when the mono signal is coded in the frequency domain, It can be decoded.

The parameter decoding unit 420 restores a plurality of parameters indicating a characteristic relationship between a plurality of channels constituting the multi-channel signal from the encoded bit stream of the stereo signal. At this time, a plurality of parameters may include CLD, ICC, and IPD, but may not include OPD.

The parameter estimator 430 estimates the OPD using a plurality of restored parameters when the OPD is not included in the plurality of parameters.

Hereinafter, the operation of the parameter estimator 430 for estimating OPD will be described in detail. Hereinafter, the formulas described below are only examples of the present invention, and it will be apparent to those skilled in the art that the formulas described below can be modified.

First, the parameter estimator 430 calculates the first intermediate variable c using the CLD according to Equation (4).

Here, b represents an index of a frequency band. As shown in Equation (4), the first intermediate variable c can be obtained by expressing the number obtained by dividing the IID value in the specific frequency band by 20 in an exponential form of 10. At this time, the second intermediate variable c ₁ and the third intermediate variable c ₂ can be obtained by using the first intermediate variable c as shown in the following equations (5) and (6).

That is, the third intermediate variable c ₂ can be obtained by multiplying the value of the second intermediate variable c ₁ by c.

Next, the parameter estimator 430 obtains the first right channel signal and the first left channel signal using the restored mono signal and the second intermediate variable and the third intermediate variable obtained from Equations (5) and (6) . The first right channel signal and the first left channel signal can be expressed by the following equations (7) and (8).

는 제2 중간 변수 c₁과 복원된 모노 신호 M의 곱으로 나타낼 수 있다.Where n is the time slot index and k is the parameter band index. The first right channel signal

Can be expressed as the product of the second intermediate variable c ₁ and the restored mono signal M.

는 제2 중간 변수 c₂와 복원된 모노 신호 M의 곱으로 나타낼 수 있다.The first left channel signal

Can be expressed as the product of the second intermediate variable c ₂ and the restored mono signal M.

라고 하면, 제1 모노 신호

는 제1 우채널 신호

및 제2 좌채널 신호

를 이용하여 다음 수학식 9와 같이 나타낼 수 있다. At this time,

, The first mono signal

Lt; RTI ID = 0.0 &gt;

And the second left channel signal

The following equation (9) can be used.

Further, using Equations (6) to (9), the fourth intermediate variable p according to the time slot and the parameter band can be obtained as shown in Equation (10).

라 할 때, OPD는 다음 수학식 11과 같이 구할 수 있다.Here, the fourth intermediate variable p is a value obtained by dividing the sum of the sizes of the first left channel signal, the first right channel signal, and the first mono signal by two. At this time, the value of OPD

, The OPD can be obtained by the following equation (11).

라 할 때,

은 다음 수학식 12와 같이 구할 수 있다.Also, the value corresponding to the difference between OPD and IPD

In other words,

Can be obtained by the following equation (12).

은 복호화된 모노 신호와 업믹싱될 좌채널 신호 사이의 위상 차이고, 수학식 12에서 구한 값인

는 복호화된 모노 신호와 업믹싱될 우채널 신호 사이의 위상 차를 나타낸다. The value of OPD obtained from the equation (11)

Is the phase difference between the decoded mono signal and the left channel signal to be upmixed,

Represents the phase difference between the decoded mono signal and the right channel signal to be upmixed.

Equation (11) and Equation (12) can be equivalently implemented as Equation (13).

with

Here, l and m denote each subband index and parameter set.

In this manner, the parameter estimator 430 generates the first left channel signal and the first right channel signal for the left channel signal and the right channel signal from the restored mono signal using the IID indicating the difference between channels of the stereo signal And generates a first mono signal from the first left channel signal and the first right channel signal using the IPD representing the phase difference between channels of the stereo signal, and outputs the generated first left channel signal, the first right channel signal, The value of the OPD indicating the phase difference between the restored mono signal and the stereo signal using the first mono signal can be estimated.

The parameter de-quantization unit 440 dequantizes the parameters estimated from the decoded parameters and the decoded parameters.

Hereinafter, the operation of the parameter de-quantization unit 440 for dequantizing the parameters will be described in detail with reference to FIGS. 5 to 8. FIG.

5 is a flowchart illustrating an operation of a parameter de-quantization unit 440 for dequantizing parameters based on USAC (Unified Speech and Audio Coding).

First, in step 501, it is determined whether the type of the parameter is a phase parameter.

If it is determined at step 501 that the parameter type is not a phase parameter, at step 502, the second parameter, which is a parameter other than the phase parameter, is inverse-quantized based on an MPS (MPEG Surround) lossless coding method.

If it is determined in step 501 that the parameter type is a phase parameter, in step 503, the encoding mode of the phase parameter is determined. More specifically, in step 503, it is determined whether the coding mode of the phase parameter is 1D Huffman coding or 2D Huffman coding.

If it is determined in step 503 that the coding mode of the phase parameter is 1D Huffman coding, then in step 504, dT or dF is determined. dT means calculating the difference based on the quantized phase information of the same band in the previous frame and dF means calculating the difference based on the quantized phase information of the previous band of the current frame

Thereafter, in step 505, a Huffman codebook is determined according to the dT or dF determined in step 504, and a 1D Huffman decoding is performed on the phase parameter according to the determined Hoffman codebook.

Step 506 modulates the previous phase parameter and the current phase parameter according to dT or dF to recover the quantized phase parameter.

If it is determined in step 503 that the coding mode of the phase parameter is 2D Huffman coding, step 507 performs LAV decoding.

Thereafter, in step 508, a Huffman codebook is determined according to the LAV and a 2D Huffman decoding is performed on the phase parameters according to the determined Huffman codebook.

In step 509, information on the magnitude of X and Y decoded by 2D Huffman decoding is restored, and the order of X and Y is adjusted on the basis of information on the restored magnitude relation.

In step 510, the quantized phase parameters are recovered by modulatively computing the regulated X and Y.

In step 511, the quantized phase parameters reconstructed in step 506 or 508 are inversely quantized to reconstruct the phase parameters.

FIGS. 6 and 7 are flowcharts for explaining the operation of the parameter dequantizer 440 for dequantizing a parameter based on an MPS (MPEG Surround) lossless coding scheme.

First, in step 601, it is determined whether the type of the parameter is a phase parameter.

If it is determined in step 601 that the parameter type is a phase parameter, in step 603, the encoding mode of the phase parameter is determined. That is, in step 603, it is determined whether the coding mode of the phase parameter is 1D Huffman coding or 2D Huffman coding.

If it is determined in step 603 that the coding mode of the phase parameter is 1D Huffman coding, then in step 604 dT ₁ or dF ₁ is determined.

Thereafter, in step 605, a Huffman codebook is determined according to dT ₁ or dF ₁ determined in step 604, and 1D Huffman decoding is performed on the phase parameters according to the determined Hoffman codebook.

Step 606 reconstructs the quantized phase parameters from the previous phase parameters and the current phase parameters through inverse differential decoding and modular operations based on dT ₁ or dF ₁ .

If it is determined in step 603 that the coding mode of the phase parameter is 2D Huffman coding, step 607 performs LAV decoding.

Thereafter, in step 608, a Huffman codebook is determined according to the LAV and a 2D Huffman decoding is performed on the phase parameters according to the determined Huffman codebook.

In step 609, information on the magnitude of X and Y decoded by 2D Huffman decoding is restored, and the order of X and Y is adjusted based on information on the restored magnitude relation.

In step 610, the quantized phase parameters are recovered from the adjusted X and Y through inverse differential decoding and modular operations.

In step 611, the quantized phase parameters reconstructed in step 606 or step 610 are inversely quantized to reconstruct the phase parameters.

Returning to the beginning, if it is determined at step 601 that the parameter type is not a phase parameter, at step 602, the second parameter, which is a parameter other than the phase parameter, is inverse-quantized.

Referring to FIG. 7, step 602 will be described in detail as follows.

In step 60201, the encoding mode of the second parameter is determined. More specifically, in step 6021, it is determined whether the coding mode of the second parameter is 1D Huffman coding or 2D Huffman coding.

If it is determined in step 60201 that the encoding mode of the second parameter is 1D Huffman coding, then in step 60202, dT ₂ or dF ₂ is determined.

Thereafter, in step 60203, Huffman codebook is determined according to dT ₂ or dF ₂ , and 1D Huffman decoding is performed on the second parameter according to the determined Huffman codebook.

In step 60204, the sign of the difference between the current second parameter and the previous second parameter is decoded.

Step 60205 restores the second parameter quantized from the previous second parameter and the current second parameter through inverse differential decoding based on dT ₂ or dF ₂ .

If it is determined in step 60201 that the encoding mode of the second parameter is 2D Huffman coding, LAV decoding is performed in step 60206.

Thereafter, in step 60207, a Huffman codebook is determined according to the LAV, and 2D Huffman decoding is performed on the phase parameter according to the determined Huffman codebook.

In step 60208, information on the magnitude of X and Y decoded by 2D Huffman decoding is restored, and the order of X and Y is adjusted based on the information on the restored magnitude relation.

In step 60209, decoded X and Y signs are decoded.

In step 60210, the second parameter quantized from the order-adjusted X and Y is restored through inverse differential decoding.

Thereafter, in step 60211, the quantized second parameter restored in step 60205 or 60210 is dequantized to restore the second parameter.

In step 611, the quantized phase parameters reconstructed in step 606 or step 608 are inversely quantized to reconstruct the phase parameters.

8 is a flowchart for explaining the operation of the parameter de-quantization unit 440 for dequantizing a parameter based on a PS (Parametric Stereo) lossless coding scheme.

First, in step 801, it is determined whether the parameter type is IPD.

If it is determined in step 801 that the type of the parameter is IPD, step 802 determines dT ₁ or dF ₁ .

Thereafter, in step 803, a Huffman codebook is determined according to the determined dT ₁ or dF ₁ , and Huffman decoding is performed on the IPD according to the determined Huffman codebook.

Step 804 reconstructs the IPD quantized from the current IPD and the previous IPD through inverse differential decoding and modular operations based on dT ₁ or dF ₁ .

In step 805, the quantized IPD is dequantized to restore the IPD.

Returning to the beginning, if it is determined at step 801 that the type of the parameter is not IPD, step 806 determines dT ₂ or dF ₂ .

Thereafter, in step 807, a Huffman codebook is determined according to dT ₂ or dF ₂ , and Huffman decoding is performed on the second parameter, which is a parameter other than the IPD, according to the determined Huffman codebook.

In step 808, the sign of the difference between the current second parameter and the previous second parameter is decoded.

Step 809 reconstructs the second parameter quantized from the previous second parameter and the current second parameter through inverse differential decoding based on dT ₂ or dF ₂ .

In step 810, the quantized second parameter is dequantized to restore the second parameter.

Referring again to FIG. 4, an apparatus 400 for decoding a multi-channel signal according to an embodiment of the present invention will be described in detail.

The upmixing unit 450 upmixes the mono signal using the restored parameter and the estimated parameter OPD.

Upmixing corresponds to downmixing by generating more than two channels of stereo signals from a mono signal of one channel. Hereinafter, the specific operation of the upmixing unit 450 upmixing the mono signal using CLD, ICC, IPD, and OPD will be described.

일 때, 제2 및 제3 중간 변수 c₁ 및 c₂를 이용하여 제1 위상

및 제2 위상 을 다음 수학식 14 및 15과 같이 구할 수 있다.First, the upmixing unit 450 receives the value of ICC

, The second and third intermediate variables c ₁ and c ₂ are used to determine the first phase

And the second phase Can be obtained by the following equations (14) and (15).

, 수학식 12에서 구한 값인

을 이용하여 아래의 수학식 16 및 수학식 17과 같이 업 믹싱된 좌채널 신호 및 우채널 신호를 구할 수 있다.Next, the upmixing unit 450 multiplies the first and second phases, the second and third intermediate values obtained through Equations (14) and (15) when the restored mono signal is M and the decorrelated signal is D, The variables c ₁ and c ₂ And the value of OPD obtained from the equation (11)

, The value obtained from the equation (12)

The left channel signal and the right channel signal upmixed as shown in Equation (16) and Equation (17) below can be obtained.

As described above, the present invention has been described with reference to particular embodiments, such as specific elements, and specific embodiments and drawings. However, it should be understood that the present invention is not limited to the above- And various modifications and changes may be made thereto by those skilled in the art to which the present invention pertains. Accordingly, the spirit of the present invention should not be construed as being limited to the embodiments described, and all of the equivalents or equivalents of the claims, as well as the following claims, belong to the scope of the present invention .

Claims (6)

Decoding the mono downmix signal from the bitstream;
Decoding a plurality of parameters indicating a characteristic relationship between channels from the bitstream; And
And upmixing the mono downmix signal into a stereo signal based on the decoded parameters,
Wherein the decoded plurality of parameters includes a phase parameter and the decoding of the plurality of parameters comprises obtaining an index for the LAV (La est's Absolute Value) of the phase parameter from the bitstream, Determining a Huffman codebook based on the LAV, decoding the symbols of the phase parameter based on the determined Huffman codebook, and considering the wrapping property of the phase parameter, And obtaining a quantized phase parameter from the symbol. 2. The method of claim 1, wherein the method further comprises estimating a parameter relating to a phase difference between one of the left signal and the right signal and the mono downmix signal using a plurality of decoded parameters, Mixes the mono down-mix signal based on the decoded parameters and the estimated parameters. The decoding method of claim 1, wherein decoding the plurality of parameters performs a modulo operation on a difference between adjacent quantized phase parameters. The decoding method according to claim 1, wherein the phase parameter is Inter-channel Phase Difference (IPD). 2. The method of claim 1, wherein the Huffman codebook comprises a data set (LavIdx, LavTabIPD) and the data set comprises (0,7), (1,1), (2,3) The method comprising: A computer-readable recording medium on which a program capable of executing the decoding method according to any one of claims 1 to 5 is recorded.

Images