KR20070011100A - Methods for energy compensation for multi-channel audio coding and methods for generating encoded audio signal for the compensation - Google Patents

Abstract

A method of compensating the energy of an audio signal in multi-channel audio coding and a method of generating an encoded audio signal for the compensation are provided to compensate channel and frequency distortions generated when a multi-channel signal is reconstructed by using energy compensation information. A method of compensating the energy of a multi-channel audio signal by using energy compensation information includes steps of analyzing the energy compensation information, converting the analyzed information into a value applicable when the multi-channel audio signal is decoded, and performing energy compensation of the multi-channel audio signal when the multi-channel audio signal is decoded by using the converted value.

Description

METHODS FOR ENERGY COMPENSATION FOR MULTI-CHANNEL AUDIO CODING AND METHODS FOR GENERATING ENCODED AUDIO SIGNAL FOR THE COMPENSATION}

1 is a block diagram of a decoding apparatus for decoding an audio signal using an energy correction method of an audio signal in multi-channel audio coding according to an embodiment of the present invention.

FIG. 2 is a detailed block diagram of the multichannel decoder illustrated in FIG. 1;

3 is a block diagram of an encoding apparatus for encoding an audio signal by using an energy correction method of the audio signal in multichannel audio coding according to an embodiment of the present invention.

4 and 5 are structural diagrams of a bit stream generated by a method of generating an encoded audio signal according to an embodiment of the present invention.

Explanation of symbols on the main parts of the drawings

10: signal separation unit 20: audio decoder

30: multi-channel decoder 31: high frequency band energy compensation unit

32: full band energy compensation unit 33: multi-channel generator

50: spatial encoder 60: spatial bit stream generator

70: energy compensation determining unit 100: audio bit stream

110: downmix bit stream 120: energy correction information

130: spatial information 140: energy correction data

150: multi-channel audio signal 160: downmix audio signal

170: downmix signal information

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to multi-channel (multi-channel) audio coding, and more particularly, to an energy correction method for performing energy correction of a multi-channel signal and an audio signal generation method for the correction.

Recently, various coding techniques and methods for digital audio signals have been developed, and related products have been produced. In addition, coding methods for converting a mono or stereo audio signal into a multichannel in a decoding step using spatial information of a multichannel audio signal have been developed, and a product for this has been put into practical use.

However, the multi-channel audio signal processing technique using the above products has the advantage of reducing the amount of data in the signal processing process, but there is a problem that the signal distortion occurs over time in a specific channel or a specific frequency band.

Accordingly, the present invention has been proposed to solve the above conventional problems, and an object of the present invention is to reconstruct a multichannel signal with downmix signals and spatial cues for a multichannel audio signal coding method. The energy correction information obtained based on the difference between the original signal and the down-mixed signal or the difference between the original signal and the final output signal is obtained by the distortion of each channel and frequency generated in the process. An energy correction method is provided.

In addition, another object of the present invention is to provide a method for generating encoded audio including energy correction information for the above correction.

An energy correction method of an audio signal in multichannel audio coding according to the present invention includes the steps of: correcting energy of a multichannel audio signal using energy correction information, comprising: interpreting the energy correction information; Converting the interpreted information into an applicable value when decoding the multichannel audio signal; Performing energy correction of the multichannel audio signal when decoding the multichannel audio signal using the converted value.

In addition, in the multi-channel audio coding according to the present invention, the method for correcting the energy of an audio signal includes encoding an energy between a downmix signal downmixing the multichannel audio input signal and the multichannel audio input signal when encoding the multichannel audio input signal. Obtaining energy correction information using the difference; It may be performed by combining the obtained energy correction information and spatial information, and correcting the energy using the combined result.

Meanwhile, the encoded audio signal generating method according to the present invention downmixes a multichannel audio input signal in multichannel audio coding, extracts spatial information from the multichannel audio input signal, and encodes the downmix signal and spatial information. In the method of generating a compressed audio signal, the encoded audio signal includes energy correction information, wherein the energy correction information represents whether or not to perform energy correction, or when performing the energy correction energy of the previous frame Flag information indicating whether to use the correction information as it is or to use the new energy correction information; Correction performing band information for displaying information on a band to perform energy correction; And correction level information indicating a degree of energy correction in the correction performing band.

Hereinafter, an embodiment according to the present invention, an energy correction method of an audio signal in a multi-channel audio coding and a coded audio signal generation method for the correction will be described with reference to the drawings.

1 is a block diagram of a decoding apparatus for decoding an audio signal by using an energy correction method of an audio signal in multi-channel audio coding according to an embodiment of the present invention. An audio decoder 20, a multichannel decoder 30, and an energy level compensator 40 are included.

As shown in FIG. 2, when the audio bit stream 100 including energy correction information is applied to the signal separator 10, the signal separator 10 may use a downmix bit stream ( 110, the energy correction information 120, and the spatial information 130 are separately extracted and output. An example of the downmix bit stream 110 is a mono or stereo bit stream.

The downmix bit stream 110 is applied to the audio decoder 20 and output as the downmix audio signal 160. The audio decoder 20 includes AAC (Advanced Audio Coding) or MP3. In addition, the audio decoder 20 may output downmix signal information 170 to generate the multi-channel audio signal 150 separately from the downmix audio signal 160. The downmix signal information 170 refers to information generated when the audio decoder 20 decodes. For example, the information may be referred to as information generated during a transformation process such as Quadrature Mirror Filter (QMF), Modified Discrete Cosine Transform (MDCT), and Pulse Code Modulation (PCM).

Meanwhile, the multichannel decoder 30 combines the downmix signal information 170 applied by the audio decoder 20 and the spatial information 130 applied by the signal separator 10 to combine the multichannel audio signal 150. Output In this case, the energy correction information 120 extracted by the signal separation unit 10 is applied to the energy level compensation unit 40, and then an appropriate value 140 (energy correction data) that can be applied to the multi-channel decoder 30. Are converted to the multi-channel audio decoder 30. Then, the multichannel audio decoder 30 performs correction to output the multichannel audio signal 150. The energy level compensator 40 interprets the energy correction information 120 and converts the energy correction information 120 into a value 140 applicable to the multi-channel decoder 30. In this case, there are various methods for interpreting the energy correction information. However, for example:

The energy correction information 120 transmitted to the energy level compensator 40 may be interpreted as it is without processing, or the energy correction information 120 may be applied on the time axis, or the energy correction information 120 may be applied. Interpolation techniques can be applied on the time base. In addition, the smoothing technique may be applied to the energy correction information 120 between bands, or the interpolation technique may be applied to the energy correction information 120 between bands.

Smoothing is the use of lowpass filters, etc., to smoothly distribute the values that exist. However, the value that is distributed is also affected, and the value may change slightly.

In the case of interpolation, smoothly bridging the existing values is similar to smoothing, but the distributed values do not change. Smoothing and interpolation are both types of broad smoothing techniques. It merely illustrates that a specific application in the method of interpreting the energy correction information 120 may be performed using a lowpass filter or interpolation.

In addition, the application of the smoothing technique of the energy correction information 120 on the time axis or the interpolation technique of the energy correction information 120 on the time axis may be solved as follows.

At any given time, the band has several bands, from low to high, and smooths in time as the value of a particular band changes over time.

In addition, applying the smoothing technique between the energy correction information 120 between the bands or applying the interpolation technique between the energy correction information 120 between the bands smoothes the correction values from the low band to the high band at a specific time. will be.

Since the bands used for MPEG Surround have different resolutions and different uses for each band, the energy correction information 120 is interpreted and converted into values applicable to the multi-channel decoder 30 to be applied for each domain. Can be. For example, it is applied in the subband domain, applied in the hybrid band domain, or applied in the quadrature mirror filter (QMF) band domain. The energy correction information 120 may be applied to the value of the spatial information 130.

FIG. 2 is a detailed block diagram of the multi-channel decoder shown in FIG. 1. The multi-channel decoder includes a high frequency band energy compensator 31, a full band energy compensator 32, and a multichannel generator 33. do.

As shown in the drawing, in the multi-channel audio coding according to an embodiment of the present invention, a method of correcting the energy of an audio signal is applied to the high frequency band energy compensator 31 by applying the energy correction information data 140 to correct only the high frequency band. There is a method for outputting the multi-channel audio signal 150 from the multi-channel generating unit 33 by applying the energy correction information data 140 to the full band energy compensation unit 32 to correct the entire band. The multichannel generator 33 may also output the multichannel audio signal 150. In this case, the reason for mentioning only the high frequency band is because a lot of distortion occurs in the high frequency band.

3 is a block diagram of an encoding apparatus for encoding an audio signal using an energy correction method of an audio signal in multichannel audio coding, according to an embodiment of the present invention. 70 and a bit stream generator 60.

As shown therein, the multichannel audio input signal 200 is applied to the spatial encoder 50, and the applied multichannel audio input signal 200 is downmixed and output as the downmix audio signal 210. One example of the downmix audio signal 210 is a mono audio signal or a stereo audio signal.

In addition, the spatial encoder 50 receives the multi-channel audio input signal 200 to extract the spatial information 220. The energy compensation determination unit 70 determines whether to perform energy correction and outputs energy correction information 230.

The energy compensation determination unit 70 outputs the energy correction information 230 in various ways and applies it to the spatial bit stream generator 60. Then, the spatial bit stream generation unit 60 outputs the spatial bit stream 240 by combining the energy correction information 230 and the spatial information 220 into one bit string.

The energy correction information 230 may be information about an energy difference between the two signals by comparing the downmix audio signal 210 and the multichannel audio input signal 200.

Specifically, the corresponding signal (multi-channel audio input signal 200 and downmix audio signal 210) is divided into an appropriate band range, each band energy is divided by total band energy, and normalized, and each normalized band The energy difference between two signals can be obtained from the ratio of energy values.

There are various ways to divide the band range appropriately. For example, low frequencies can be densely divided and high frequencies can be divided relatively sparsely. There is also a method of dividing by an equivalent rectangular bandwith (EBR) scale.

In addition, the method of obtaining the energy correction information 230 is as follows.

First, for each band of the multi-channel audio input signal 200 and the downmix audio signal 210, the ratio of the energy magnitude of each band to the total energy (conventionally referred to as the 'first energy magnitude ratio') is obtained. In addition, the ratio of energy magnitudes (for convenience, referred to as a 'second energy magnitude ratio') of the multichannel audio input signal 200 and the downmix audio signal 210 in each band is obtained, and the first energy magnitude ratio and the second energy ratio are obtained. The total energy size ratio is used as the energy correction information 230.

Here, the meaning of synthesizing may include an operation of adding the first energy size ratio and the second energy size ratio, and may include forming a table according to each case. This can be specifically determined by the experimental results.

In addition, the spatial bit stream 240 has a degree of energy relative to the entire band of each band or a degree of difference between the multichannel audio input signal 200 and the downmix audio signal 210 in each band. The energy correction information 230 may or may not be included. For example, if the band's energy is an absolutely small amount, it can be ignored and no energy compensation can be given for that band. Also, if the total energy is very small, for example a very small sound may not perform the correction because there is no effect of energy correction. Of course, even if the difference between the multi-channel audio input signal 200 and the downmix audio signal 210 is negligibly small, the correction may not be performed. The reason for this selective energy correction is that the human ear does not recognize all the voices, but selectively. Therefore, when the energy difference or absolute amount of energy not recognized by the human ear is small, the step of adding energy correction information 230 may be omitted to reduce the amount of multi-channel data.

Specifically, when the relative energy magnitude of the entire band of each band exceeds a predetermined value (reference value), the energy correction information 230 is added to the spatial bit stream 240, otherwise it is not added. Can be taken.

In addition, when the difference between the multi-channel audio input signal 200 and the downmix audio signal 210 in each band exceeds a predetermined value (reference value), the energy correction information 230 is added to the spatial bit stream 240. If not, a method may be taken so that it is not added.

In addition, the addition and interruption of the correction performance information is severely changed with time by setting the value of the ratio that determines the moment of adding the energy correction information 230 and the value that determines the moment when the addition of the energy correction information 230 is stopped to different values. You can also take steps to avoid repetition.

In addition, the spatial bit stream 240 combines the degree of energy magnitude relative to the entire band of each band and the degree of difference between the multichannel audio input signal 200 and the downmix audio signal 210 in each band. Depending on the information, the energy correction information 230 may or may not be included.

For example, one of the factors corresponding to the degree of energy magnitude relative to the entire band of each band and the factor corresponding to the degree of difference between the multichannel audio input signal 200 and the downmix audio signal 210 in each band. If only the factor exceeds the reference ratio to be corrected, and the other factor does not exceed the reference ratio, another condition factor may be defined, and accordingly, correction method information may be added.

For example, if one factor exceeds the criterion and the other does not exceed the criterion, the value of the two plus the value is defined as another factor, and the energy when the value of the other factor exceeds a certain criterion. For example, the correction information 230 may be added to the spatial bit stream 240.

Hereinafter, a method of generating an encoded audio signal according to another embodiment of the present invention will be described.

4 and 5 are structural diagrams of a bit stream generated by a method of generating an encoded audio signal according to an embodiment of the present invention.

As shown, in the case where signal distortion occurs in the process of processing a multi-channel audio signal, when energy correction of the signal is performed using the energy correction information, a method of inserting the energy correction information into the bit string for the correction is taken. All. In this case, the energy correction information included in the bit string includes the following information.

The energy correction information inserted in the bit string may include flag information 200, correction performance band information 210, and correction level information 220. The flag information 200 expresses whether or not to perform energy correction, or if the energy correction is performed, whether the energy correction information of the previous frame is used as it is or whether new energy correction information is used. In the correction performance band information 210, information on a band to be corrected is displayed. The correction level information 220 displays the degree of energy correction.

In this case, whether or not to perform energy correction can be made in the following manner. For example, there is a method of determining a reference energy, and performing energy correction if an energy difference occurs below the reference ratio compared to the reference energy, and performing an energy correction if it occurs above. Such a method can be performed because when the energy difference is less than the reference energy ratio, the distortion of the signal is insignificant, so even if the difference is ignored, there is no difference.

In addition, the correction band information 210 includes at least one of whether to perform correction for each band, the start and end of the band to be corrected, the start value of the band to be corrected, and the end value of the band to be corrected. One may be included.

In addition, the correction level information 220 includes an energy correction value for each band to be corrected, one value for the entire band for energy correction, one slope value for the bands for energy correction, and energy. It is preferable to include at least one of several values for correction value interpolation for the bands performing the correction.

As an example of inserting the energy correction information into the bit string, when performing the energy correction, the flag information is displayed as 0 when the correction information of the previous frame is used as it is, and the flag information is set to 1 when the new correction information is used. There is a way to display it.

In addition, by dividing the starting point of the band to be corrected into two places, the correction performing band information 210 is displayed as 0 in order to perform the correction from one starting point, and to perform the correction from another starting point, the correction is performed. There is a method of marking performance band information as 1 and inserting the information into a bit string. For example, when performing correction from band 10, correction band information 210 may be displayed as 0, and when performing correction from band 15, correction band information 210 may be displayed as 1. . The above method can be performed because one of two cases can be roughly modeled.

In addition, the method of inserting the correction information into the bit string according to an embodiment of the present invention may take a method that is divided for each channel.

For example, as shown in FIG. 5, when performing the correction channel band information and the correction level information into two, and applying it to the left channel, a corresponding bit is inserted into band information 1 (310) and energy level information 1 (320). When applied to the right channel, a method of inserting corresponding bits of band information 2 (310 ') and energy level information 2 (320') may be applied. In this case, the correction value applied to the center channel is 1/2 of the sum of the correction values applied to the left channel and the right channel. For example, when the correction value of the left channel is 3 and the correction value of the right channel is 6, it means that (3 + 6) /2=4.5 is applied as the correction value to the center channel.

In addition, in the case where the flag information 200, 300 is displayed as 1, when the correction is not performed, the correction level information 210, 310, 310 'is displayed as 0, and the energy difference in the end band is 1.5. In order to linearly attenuate dB, the correction level information (210, 310, 310 ') is indicated as 1, and linear attenuation is made so that the energy difference in the end band is 3dB. In this case, the correction level information 210, 310, 310 ′ may be displayed as 2 and inserted into the bit string. As described above, the reason for linearly attenuating the energy decay into two parts, 1.5 dB and 3 dB, can be attenuated in more detail. . Also, linear decay is because energy often has linearity in relation to frequency.

In addition, when the energy difference in the end band needs to be further divided into 1.5 dB and 3 dB, the correction level information 210, 310, and 310 ′ may be displayed as 3 and stored.

In addition, the flag information, the correction performance band information, and the correction level information may be displayed as a single value by grouping two or three together without being displayed independently and inserted into the bit string.

Although the preferred embodiment of the present invention has been described above, the present invention may use various changes, modifications, and equivalents. It is clear that the present invention can be applied in the same manner by appropriately modifying the above embodiments. Accordingly, the above description does not limit the scope of the invention as defined by the limitations of the following claims.

As described above, in the multi-channel audio coding according to the present invention, the energy correction method of the audio signal and the coded audio signal generation method for the correction of the channel-specific distortion generated during the reconstruction of the multi-channel signal There is an effect that can be compensated using the energy correction information.

Claims (19)

In the method for correcting the energy of the multi-channel audio signal using the energy correction information, (a) interpreting the energy correction information; (b) converting the interpreted information into an applicable value when decoding the multichannel audio signal; and (c) performing energy correction of the multichannel audio signal when decoding the multichannel audio signal using the converted value. The method of claim 1, In step (a), the energy correction information is processed as it is, the smoothing technique is applied to the energy correction information on the time axis, the energy correction information is applied to the time axis, and the energy correction information is applied between the bands. And applying at least one of a smoothing technique and applying the energy correction information between bands to the interpolation technique. The method of claim 1, And (b) converting the interpreted information into a value applicable to at least one of a subband domain, a hybrid band domain, and a QMF band domain. The method of claim 1, In the step (b), the energy correction method of the audio signal in the multi-channel audio coding, characterized in that for converting the interpreted information to a value applicable to the spatial information. The method of claim 1, The energy correction method is an energy correction method of the audio signal in the multi-channel audio coding, characterized in that to perform the energy correction by applying the energy correction information only to the high frequency band. The method of claim 1, The energy correction method is the energy correction method of the audio signal in the multi-channel audio coding, characterized in that for performing the energy correction by applying the energy correction information to the entire band. (a) obtaining energy correction information using an energy difference between the downmix signal downmixing the multichannel audio input signal and the multichannel audio input signal when encoding the multichannel audio input signal; and (b) combining spatial information with the obtained energy correction information and correcting the energy using the combined result. The method of claim 7, wherein In the step (a), the downmix signal and the multi-channel audio input signal are divided into bands, normalized by dividing the energy of the divided band by the total band energy, and then energy correction information is expressed as a ratio of energy values of the normalized band. The energy correction method of the audio signal in the multi-channel audio coding, characterized in that obtaining. The method according to claim 7 or 8, In step (a), the ratio of the energy magnitude of the divided band to the total energy with respect to the band of the divided multichannel audio input signal and the downmix signal, and the multichannel audio signal and the downmix signal within the divided band The energy correction method of the audio signal in the multi-channel audio coding, characterized in that the ratio of the energy magnitude is calculated, and the sum of the ratios of the two energy magnitudes is calculated as energy correction information. The method of claim 9, In step (b), the step of determining whether to combine the energy correction information with the spatial information is preceded, and the energy is corrected only when the combination of the energy correction information and the spatial information is permitted. Energy correction method of audio signal in multichannel audio coding. The method of claim 10, The permission of combining the energy correction information and the spatial information is determined according to the relative energy magnitude of the entire band of the divided band. The method of claim 10, The permission of combining the energy correction information and the spatial information is determined according to an energy difference between the multi-channel audio input signal of the divided band and the downmix signal. The method of claim 10, The permission of the combination of the energy correction information and the spatial information is determined according to a result of the sum of the relative energy magnitudes of the entire bands of the divided bands and the energy difference between the multi-channel audio input signals of the divided bands and the downmix signal. The energy correction method of the audio signal in the multi-channel audio coding, characterized in that. The method of claim 11, The energy correction method of the audio signal in the multi-channel audio coding, characterized in that the combination of the energy correction information and the spatial information is permitted only when the relative magnitude of energy of the entire band of the divided band exceeds a predetermined value. The method of claim 12, The energy of the audio signal in the multi-channel audio coding is allowed to combine the energy correction information and the spatial information only when the energy difference between the multi-channel audio input signal of the divided band and the downmix signal exceeds a predetermined value. Correction method. The method of claim 13, The combination of the energy correction information and the spatial information is allowed only when the sum of the relative energy magnitudes of the divided bands and the energy difference between the multi-channel audio input signal of the divided band and the downmix signal exceeds a predetermined value. The energy correction method of the audio signal in the multi-channel audio coding, characterized in that. A method for downmixing a multichannel audio input signal in multichannel audio coding, extracting spatial information from the multichannel audio input signal, and generating an audio signal encoded with the downmix signal and spatial information, The encoded audio signal includes energy correction information, wherein the energy correction information indicates whether to perform energy correction or, when performing energy correction, whether to use the energy correction information of the previous frame as it is or new energy correction. Flag information indicating whether to use the information; Correction performing band information for displaying information on a band to perform energy correction; And And a correction level information indicating a degree of energy correction in the correction performing band. The method of claim 17, The information displayed in the correction performing band information includes at least one of presence or absence of energy correction for the band, start and end of the band to be corrected, start value of the band to be corrected, and end value of the band to be corrected. Coded audio signal generation method, characterized in that. The method of claim 17, The information displayed in the correction level information includes an energy correction value for the band to be corrected, one value for the entire band for energy correction, one slope value for the bands for energy correction, and energy correction. And at least one of a predetermined number of values for interpolation of correction values for the performing bands.

Images