JP2006003580A - Device and method for coding audio signal - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce a bit rate by appropriately conducting grouping of frequency data along a time direction or a frequency direction in audio signals of a plurality of channels and efficiently conducting coding. <P>SOLUTION: Analysis filter banks 100 and 101 receive left and right channel input audio signals and convert the signals into time series of frequency data. Strength computing sections 110 and 111 compute strength of the frequency data. A degree of similarity determining section 130 determines the degree of similarity of the plurality of the channels. When the degree of similarity is determined to be high by the degree of similarity determining section 130, a grouping section 122 commonly conducts grouping of the frequency data along the time direction or the frequency direction for the plurality of the channels based on the strength average of the frequency data for the plurality of the channels. Coding sections 150 and 151 conduct coding of the frequency data based on the grouping. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention converts an input audio signal of a plurality of channels into a time series of frequency data, and performs common time grouping of the frequency data or frequency direction grouping to efficiently encode the audio signal. The present invention relates to an encoding device and an audio signal encoding method.

  In recent years, the input audio signals of a plurality of channels are converted into a time series of frequency data, and the grouping of the frequency data or its envelope data in the time direction or the frequency direction is performed in common for a plurality of channels. There has been proposed a method for reducing the information about and improving the coding efficiency.

  As such a proposal, for example, AAC (Advanced Audio Coding) of MPEG (Moving Picture Experts Group) described in Non-Patent Document 1 and Non-Patent Document 2 is known. In AAC, an input signal is converted into frequency data (MDCT coefficient) by MDCT (Modified Discrete Cosine Transform). There are two types of MDCT block length, a long-time block and a short-time block, and an appropriate block length is selected according to the nature of the input signal. In the case of a transient input signal, a short time block is selected.

  In the AAC standard, one short-time block consists of 128 frequency data, and 8 blocks are encoded as one frame. When encoding short-time blocks, a plurality of short-time blocks that are temporally continuous are grouped, and quantization and encoding processes are performed in units of groups. In the AAC standard, one frame is divided into a minimum of 1 group and a maximum of 8 groups. If the quantization step size of the frequency data in band units is a scale factor, this scale factor is made common to the same group. Thereby, the number of bits required for the scale factor can be reduced, and the encoding efficiency can be improved.

  Also, when the input signal is a stereo signal and the common window flag is on, grouping is performed in common on the left and right channels, thereby reducing the information related to grouping to one channel and encoding efficiency. Can be further improved.

  Another example of the proposal is SBR (Spectral Band Replication) described in Patent Document 1, Non-Patent Document 3, and Non-Patent Document 4. In SBR encoding, a low frequency band signal obtained by deleting a high frequency band from an input audio signal is encoded, and the deleted high frequency band envelope data is encoded. In SBR decoding, frequency data in a low frequency band is reproduced by decoding encoded data in a low frequency band. Then, the reproduced frequency data of the low frequency band is copied to the high frequency band, and the frequency data of the high frequency band is restored by adjusting the gain of the envelope of the high frequency band based on the decoded envelope data. To do. Thus, the input audio signal can be reproduced by synthesizing the frequency data of the two frequency bands.

  In SBR, encoding efficiency is improved by performing grouping in the time direction and the frequency direction for encoding the deleted envelope data in the high frequency band (see, for example, Patent Document 1 and Non-Patent Document 4). In addition, when the input signal is a stereo signal and the coupling mode flag is on, grouping is performed in common on the left and right channels, thereby reducing grouping information and further improving coding efficiency. can do.

  The above-described prior art documents do not describe a specific method or apparatus for efficiently encoding frequency data or envelope data in a group by using a plurality of channels.

  In order to clarify the problem of the present invention, a two-channel audio signal encoding apparatus which is recognized as a prior art but not recognized as a publicly known document will be described as a conventional example. FIG. 3 is a block diagram showing the configuration of a conventional 2-channel audio signal encoding apparatus. The audio signal encoding apparatus includes analysis filter banks 100 and 101, intensity calculation units 110 and 111, grouping units 120 and 121, encoding units 150 and 151, a multiplexing unit 160, and a grouping match determination unit 190. Composed. The operation of the audio signal encoding apparatus having such a configuration will be described.

  The time axis audio signal in the left channel (Lch) input to the analysis filter bank 100 is converted into a time series of frequency data. Similarly, the time axis audio signal in the right channel (Rch) input to the analysis filter bank 101 is converted into a time series of frequency data.

  The intensity calculators 110 and 111 calculate the intensity of the frequency data in the encoding processing unit before grouping each channel. Here, the encoding processing unit is a unit for performing frequency data quantization and encoding processing, and a common quantization step size is set for all frequency data belonging to the same encoding processing unit. Use.

  FIG. 4A shows an encoding processing unit per frame before grouping. In FIG. 4, one frame before grouping is composed of a total of 64 coding processing units, 8 in the time direction and 8 in the frequency direction. The intensity calculators 110 and 111 calculate and output the intensity of the frequency data of the encoding process unit in each channel by the sum of squares of all the frequency data belonging to the encoding process unit.

  The grouping units 120 and 121 perform grouping in the time direction and the frequency direction based on the strength of the frequency data of the encoding processing unit in each channel, and information on grouping (hereinafter referred to as grouping information). Is output. First, grouping in the time direction is performed. The grouping in the time direction is performed based on the amount of change in intensity in the frequency data in the time direction. The strength of the frequency data of the encoding processing units belonging to the same time is added in the frequency direction to calculate the total strength of the encoding processing units belonging to the time.

  Next, the ratio of the total intensity of the encoding processing units adjacent in the time direction is calculated and set as the amount of change in the time direction. However, when the value of the ratio is smaller than 1, the amount of change is obtained by taking the reciprocal of the value of the ratio so as to be 1 or more. When the amount of change is less than or equal to a predetermined threshold value, the encoding processing units adjacent in the time direction are merged into one group to form a new encoding processing unit. By repeating such processing in the time direction, grouping in the time direction is performed.

  Next, frequency grouping is performed. There are two types of grouping in the frequency direction: high frequency resolution and low frequency resolution. High frequency resolution is a unit of encoding processing before grouping in the frequency direction. The low frequency resolution is a new coding processing unit obtained by merging several coding processing units in the frequency direction. The grouping in the frequency direction is performed on the encoding processing unit after the grouping in the time direction, and is performed based on the amount of change in the intensity of the frequency data in the frequency direction. And the ratio of the intensity | strength of the encoding process unit adjacent to a frequency direction is computed as a variation | change_quantity. However, when the value of the ratio is smaller than 1, the reciprocal of the value of the ratio is used so as to be 1 or more.

  Next, within the coding processing unit with low frequency resolution, when the amount of change in the intensity of the frequency data in the coding processing unit with high frequency resolution is greater than a predetermined threshold, the frequency is set to high frequency resolution. Use low frequency resolution. The grouping units 120 and 121 output grouping information in the time direction and the frequency direction in each channel calculated as described above.

  FIG. 4B is a diagram illustrating an example of an encoding processing unit of one frame after grouping in the time direction and the frequency direction. The encoding units 150 and 151 in FIG. 3 constitute the encoding processing unit after grouping based on the grouping information from the grouping units 120 and 121, respectively, and the frequency data of each channel Quantization and encoding are performed in units of encoding processing, and encoded frequency data is output.

  The grouping match determination unit 190 determines whether the left channel grouping information output from the grouping unit 120 completely matches the right channel grouping information output from the grouping unit 121. I do. If they match, the grouping match determination unit 190 turns on and outputs a common group flag indicating that common grouping is performed on the left and right channels.

The multiplexing unit 160 multiplexes the encoded frequency data from the encoding units 150 and 151, the grouping information from the grouping units 120 and 121, and the common group flag from the grouping match determination unit 190. This is output as encoded data. When the common group flag is on, the left and right channel grouping information is the same, so for example, only the grouping information from the grouping unit 120 is multiplexed as encoded data.
Special table 2003-529787 gazette Bosi, 9 others, “ISO / IEC MPEG-2 Advanced Audio Coding”, J.I. Audio Eng. Soc. 45, No. 10, October 1997, pages 789-814. ISO / IEC 13818-7, “Information Technology, Generic Coding of Video and Accompanying Acoustic Signals, Part 7 Advanced technology—Generic coding of moving pictures and associations and associations. Part 7 Advanced Audio Coding (AAC)), 1997 Martin, 3 others, “Spectral Band Replication, a novel approach in audio coding”, 112th AES Conference (May 2002), Paper No. No. 5553 ISO / IEC MPEG 14496-3: 2001 / FDAM1 (FDAM1), “Information Technology, Coding of Audio-Visual Objects, Part 3 Audio, Modification 1: Information technology − Coding of audio-visual objects, Part 3 Audio, Amendment 1: Bandwidth Extension), Document N5570, March 2003.

  However, in the conventional 2-channel audio signal encoding apparatus, if the grouping of both channels does not completely match, the common grouping of the two channels cannot be performed. For this reason, there has been a problem that the frequency of occurrence of common grouping is low and the improvement effect of the coding efficiency by the common grouping is small.

  Further, since the grouping is performed independently for each channel based on the intensity of the frequency data of each channel, there is a problem that it is not suitable for performing a common grouping on two channels.

  The present invention solves such a conventional problem, and the time direction or frequency direction grouping of the frequency data of the input audio signal in a plurality of channels is made according to the strength and similarity of the frequency data in the plurality of channels. An object of the present invention is to realize an audio signal encoding apparatus that encodes a plurality of channels of audio signals, which is appropriately performed, improves encoding efficiency, and particularly improves sound quality at a low bit rate.

  In order to solve this problem, an audio signal encoding device according to the present invention inputs an audio signal of a plurality of channels and converts it into a time series of frequency data, and an analysis of the frequency data of the plurality of channels. A strength calculation unit that calculates strength, a similarity determination unit that determines the similarity of the plurality of channels based on the strength of frequency data in the plurality of channels, and the similarity determination unit determines that the similarity is high In this case, at least one grouping of the frequency data in the plurality of channels in the time direction or the frequency direction is performed in common for the plurality of channels based on an average intensity of the frequency data in the plurality of channels. A grouping unit; and frequency of the plurality of channels based on the grouping. It is characterized in that it comprises a data or coding unit to encode the envelope data.

  In order to solve this problem, an audio signal encoding apparatus according to the present invention inputs audio signals of a plurality of channels and generates encoded data for a signal obtained by deleting a predetermined frequency band from the input audio signal. And an audio signal encoding device that generates envelope data of the signal of the deleted frequency band, and multiplexes the encoded data with the encoded data for transmission or storage, wherein the audio signals of the plurality of channels are received. An analysis filter bank that inputs and converts to a time series of frequency data, an intensity calculation unit that calculates the intensity of frequency data of the deleted frequency band in the plurality of channels, and the deleted in the plurality of channels Based on the frequency data intensity of the frequency band, When it is determined by the similarity determination unit that determines similarity and the similarity determination unit determines that the similarity is high, at least one grouping in the time direction or frequency direction of the frequency data of the deleted frequency band is performed. A grouping unit that is commonly used for the plurality of channels based on an intensity average value of frequency data of the deleted frequency band, and the deleted frequencies of the plurality of channels based on the grouping. And an encoding unit that encodes the envelope data of the frequency data of the band.

  Here, the intensity of the frequency data may be calculated by a power of the absolute value of the frequency data.

  In order to solve this problem, an audio signal encoding method according to the present invention includes a frequency conversion step of inputting audio signals of a plurality of channels and converting the signals into a time series of frequency data, and the frequency data of the plurality of channels. A strength calculation step for calculating strength, a similarity determination step for determining the similarity of the plurality of channels based on the strength of the frequency data in the plurality of channels, and a determination that the similarity is high by the similarity determination step In this case, at least one grouping of the frequency data in the plurality of channels in the time direction or the frequency direction is performed in common for the plurality of channels based on an average intensity of the frequency data in the plurality of channels. Based on the grouping step and the grouping, the previous Is characterized in that comprises a coding step of coding the frequency data or envelope data of a plurality of channels, the.

  In order to solve this problem, the audio signal encoding method of the present invention inputs audio signals of a plurality of channels and generates encoded data for a signal obtained by deleting a predetermined frequency band from the input audio signal. And a stereo audio signal encoding method for generating envelope data of the deleted frequency band and multiplexing or transmitting the envelope data to the encoded data, wherein the audio signals of the plurality of channels are recorded. A frequency conversion step for inputting and converting to a time series of frequency data; an intensity calculating step for calculating an intensity of frequency data of the deleted frequency band of the plurality of channels; and the deleted frequency of the plurality of channels. Similarity of the multiple channels based on the frequency data intensity of the band When the similarity is determined to be high by the similarity determination step and the similarity determination step, at least one grouping in the time direction or frequency direction of the frequency data of the deleted frequency band is A grouping step commonly performed for the plurality of channels based on an average intensity of frequency data of the deleted frequency band, and a frequency of the deleted frequency band of the plurality of channels based on the grouping And an encoding step for encoding the envelope data of the data.

  Here, the intensity of the frequency data may be calculated by a power of the absolute value of the frequency data.

  According to the audio signal encoding device and the audio signal encoding method of the present invention, when the similarity of input audio signals in a plurality of channels is high, the grouping in the time direction or frequency direction of the frequency data of the audio signal By performing in common for a plurality of channels based on the average of the intensity of frequency data in the channel, the frequency data or its envelope data can be efficiently encoded. In particular, the sound quality at a low bit rate can be improved.

The best mode for carrying out the present invention will be described below with reference to the drawings.
(Embodiment 1)
FIG. 1 is a configuration diagram of a 2-channel audio signal encoding apparatus according to Embodiment 1 of the present invention. The audio signal encoding apparatus includes analysis filter banks 100 and 101, intensity calculation units 110 and 111, grouping units 120, 121, and 122, a similarity determination unit 130, a switching unit 140, encoding units 150 and 151, and multiplexing. The unit 160 is configured to be included.

  The analysis filter bank 100 inputs a left channel audio signal and converts it into a time series of frequency data. The analysis filter bank 101 inputs a right channel audio signal and converts it into a time series of frequency data. The intensity calculation unit 110 calculates the intensity of the frequency data in the left channel. The intensity calculation unit 111 calculates the intensity of the frequency data in the right channel. The grouping unit 120 performs grouping in the time direction or frequency direction of frequency data in the left channel. The grouping unit 121 performs grouping in the time direction or frequency direction of frequency data in the right channel. The similarity determination unit 130 determines the similarity between both channels based on the intensity of the frequency data in the left and right channels.

  When the similarity determination unit 130 determines that the similarity is high, the grouping unit 122 performs at least one grouping in the time direction or frequency direction of the frequency data in the left and right channels, and the frequency data in both channels. This is performed in common for both channels based on the intensity average. The switching unit 140 switches grouping information based on the common group flag from the similarity determination unit 130. The encoding unit 150 inputs the signal of the analysis filter bank 100 and encodes the frequency data of the left channel or its envelope data based on the grouping by the switching unit 140. The encoding unit 151 inputs the signal of the analysis filter bank 101 and encodes the right channel frequency data or its envelope data based on the grouping by the switching unit 140. The multiplexing unit 160 multiplexes the encoded frequency data from the encoding units 150 and 151, the grouping information from the switching unit 140, and the common group flag from the similarity determination unit 130, and encodes the encoded data. Is output as

  The operation of the two-channel audio signal encoding apparatus configured as described above will be described below. The input audio signal on the time axis of the left channel is converted into a time series of frequency data in the analysis filter bank 100. Similarly, the input right channel audio signal is converted into a time series of frequency data in the analysis filter bank 101.

  Intensity calculators 110 and 111 calculate and output the intensity of frequency data in a predetermined encoding processing unit in each channel. In the following description, an encoding processing unit means a unit for performing frequency data quantization and encoding processing, and a common quantization step size for frequency data belonging to the same encoding processing unit. Is used. The intensity calculators 110 and 111 calculate the intensity of the frequency data of the encoding process unit in each channel by the sum of powers of all the frequency data belonging to the encoding process unit, here the sum of squares.

  The grouping units 120 and 121 perform grouping in the time direction and the frequency direction based on the strength of the frequency data of the encoding processing unit in each channel, and output information on grouping (hereinafter referred to as grouping information). To do. The grouping units 120 and 121 first perform grouping in the time direction. The grouping in the time direction is performed based on the amount of change in the intensity of the frequency data in the time direction. The grouping units 120 and 121 calculate the total strength of the coding processing units belonging to the same time by adding the strengths of the frequency data of the coding processing units belonging to the same time in the frequency direction. Next, the ratio of the total intensity of the encoding processing units adjacent in the time direction is calculated and set as the amount of change in the time direction. However, when the value of the ratio is smaller than 1, the amount of change is obtained by taking the reciprocal of the value of the ratio so as to be 1 or more. When the change amount is equal to or smaller than the predetermined threshold, the grouping units 120 and 121 merge the coding processing units adjacent in the time direction into one group as shown in FIG. To unitize processing. By repeating such processing in the time direction, grouping in the time direction is performed.

  Next, the grouping units 120 and 121 perform grouping in the frequency direction. In the present embodiment, there are two types of grouping in the frequency direction: high frequency resolution and low frequency resolution. High frequency resolution is a unit of encoding processing before grouping in the frequency direction. The low frequency resolution is obtained by merging several high frequency resolution encoding processing units in the frequency direction to form a new encoding processing unit. The grouping in the frequency direction is performed on the encoding processing unit after the grouping in the time direction, and is performed based on the amount of change in the intensity of the frequency data in the encoding processing unit in the frequency direction. The ratio of the intensity of coding processing units adjacent in the frequency direction is calculated as a change amount. However, when the ratio value is smaller than 1, a value obtained by calculating the reciprocal of the ratio value so as to be 1 or more is calculated as the amount of change.

  Next, within the coding processing unit with low frequency resolution, if the amount of change in the intensity of the frequency data in the coding processing unit with high frequency resolution is greater than a predetermined threshold value, set to high frequency resolution, otherwise Use low frequency resolution.

  The grouping units 120 and 121 output grouping information in the time direction and frequency direction of each channel calculated as described above. The grouping unit 122 first calculates an average intensity value of frequency data in the left and right channels. Next, the grouping unit 122 performs grouping in the same manner as the grouping units 120 and 121 based on the average intensity values of the frequency data in the left and right channels, and outputs common grouping information for the two channels. As described above, since the grouping unit 122 performs common grouping based on the average intensity value of the frequency data in the left and right channels, even when there is a difference in the intensity of the frequency data in the left and right channels, this is reflected. Common groupings.

  The similarity determination unit 130 receives the intensity of the frequency data in the left and right channels from the intensity calculation units 110 and 111, and calculates the normalized correlation coefficient of the intensity of the frequency data in the left and right channels over the entire encoding processing frame. . If the value of the normalized correlation coefficient is greater than a predetermined threshold, it is determined that the similarity between the frequency data in the two channels is high. Then, the similarity determination unit 130 turns on and outputs a common group flag indicating that grouping is performed in common on the left and right channels. This common group flag is called a common window flag in the AAC standard described in Non-Patent Document 2.

  The switching unit 140 switches the grouping information based on the common group flag from the similarity determination unit 130. When the common group flag is on, common grouping information is selected for the left and right channels from the grouping unit 122 and output to the encoding units 150 and 151 and the multiplexing unit 160. When the common group flag is OFF, the left channel grouping information from the grouping unit 120 is output to the encoding unit 150 and the multiplexing unit 160, and the right channel grouping information from the grouping unit 121 is output. Are output to the encoding unit 151 and the multiplexing unit 160.

  The encoding units 150 and 151 configure an encoding processing unit based on the grouping information from the switching unit 140, perform quantization and encoding in units of the encoding processing on the frequency data of each channel, The encoded frequency data is output to multiplexing section 160. The multiplexing unit 160 multiplexes the encoded frequency data from the encoding units 150 and 151, the grouping information from the switching unit 140, and the group common flag from the similarity determination unit 130, and encodes the encoded data. Output as. When the common group flag is on, the left and right channel grouping information is the same, so only the grouping information of one channel is multiplexed as encoded data.

  As described above, in the audio signal encoding device according to Embodiment 1, the similarity determination unit 130 that determines the similarity of audio signals in two channels, and the grouping unit 122 that performs common grouping on the two channels, Is provided. Accordingly, when it is determined that the similarity of the audio signals is high in the two channels, the grouping of the frequency data in the time direction and the frequency direction is set to 2 based on the average intensity value of the frequency data in the two channels. Common to two channels. For this reason, the audio signals of the two channels can be efficiently encoded, and the sound quality can be improved particularly at a low bit rate. In addition, the frequency of occurrence of common grouping is higher than in the conventional example, and the coding efficiency can be improved.

  When the frequency data is noisy in encoding sections 150 and 151, a process called PNS (Perceptual Noise Substation) is used to encode envelope data representing the average power of frequency data instead of frequency data. May be used. When PNS processing is used, frequency data can be reproduced by generating random noise data at the time of decoding and adjusting the average power based on the envelope data.

  For the purpose of reducing the amount of operation processing as an audio signal encoding device, the grouping unit 122 operates only when the common group flag from the similarity determination unit 130 is on, and the grouping units 120 and 121 are common. It may be configured to operate only when the group flag is off.

(Embodiment 2)
Next, an audio signal encoding apparatus according to Embodiment 2 of the present invention will be described. FIG. 2 is a configuration diagram of a two-channel audio signal encoding apparatus according to the second embodiment. Here, a configuration when applied to SBR is shown.

  This audio signal encoding apparatus includes analysis filter banks 100 and 101, intensity calculation units 112 and 113, grouping units 123, 124, and 125, a similarity determination unit 131, a switching unit 141, encoding units 152 and 153, and multiplexing. Unit 161, down samplers 170 and 171, and low frequency encoding units 180 and 181. The operation of the two-channel audio signal encoding apparatus configured as described above will be described.

  The down sampler 170 deletes a predetermined high frequency band from the input left channel audio signal, and generates a low frequency band signal obtained by down-sampling the sampling frequency by half. The down sampler 171 deletes a predetermined high frequency band from the input right channel audio signal, and generates a low frequency band signal obtained by down-sampling the sampling frequency by half. The low-frequency encoding units 180 and 181 receive low-frequency band audio signals from the channel down-samplers 170 and 171, respectively, and output high-efficiency encoded data to the multiplexing unit 161.

  The analysis filter banks 100 and 101, the intensity calculation units 112 and 113, the grouping units 123, 124, and 125, the similarity determination unit 131, the switching unit 141, and the encoding units 152 and 153 in FIG. It is a block for encoding an envelope of a deleted high frequency band. These blocks are the same as those in FIG. 1, and a description thereof will be omitted. In addition, the intensity calculation units 112 and 113, the grouping units 123, 124, and 125, the similarity determination unit 131, and the switching unit 141 in FIG. 2 are the high frequency band from which the target frequency band is deleted. 1 are the same as the intensity calculation units 110 and 111, the grouping units 120, 121, and 122, the similarity determination unit 130, and the switching unit 140 in FIG. Note that the common group flag output from the similarity determination unit 131 is referred to as a coupling mode flag in the SBR standard described in Non-Patent Document 4.

  The encoding units 152 and 153 configure an encoding processing unit based on the grouping information from the switching unit 141, and calculate the average power of the frequency data in the high frequency band encoding processing unit deleted in each channel. Calculate the envelope data to represent. Next, the encoding units 152 and 153 perform quantization and encoding of the calculated envelope data, and output the encoded envelope data to the multiplexing unit 161.

  The multiplexing unit 161 includes the low frequency band encoded data from the low frequency encoding units 180 and 181, the high frequency band encoded envelope data from the encoding units 152 and 153, and the switching unit 141. The grouping information of the envelope data and the common group flag from the similarity determination unit 131 are multiplexed and output as encoded data. When the common group flag is on, the left and right channel grouping information is the same, so only the grouping information of one channel is multiplexed as encoded data.

  As described above, in the audio signal encoding device according to the second embodiment, the similarity of audio signals in two channels is obtained with respect to the high frequency band of the audio signal that has not been encoded by the low frequency encoding units 180 and 181. A similarity determination unit 131 for determination and a grouping unit 125 that performs common grouping of frequency data on two channels are provided. As a result, when it is determined that the similarity between the audio signals in the two channels is high, the grouping of the frequency data in the time direction and the frequency direction is performed based on the average of the intensity of the frequency data in the two channels. Common to two channels. For this reason, the envelope data of the high frequency band of the audio signal in the two channels can be efficiently encoded, and the sound quality at a particularly low bit rate can be improved.

  In each of the above embodiments, the intensity of the frequency data is calculated by the square sum of the frequency data. However, the intensity of the frequency data may be calculated by an arbitrary power sum of the absolute values in the frequency data. In each of the above embodiments, the input audio signal has two channels. However, this may be three or more channels.

  In the audio signal encoding device and audio signal encoding method according to the present invention, when the similarity of frequency data of audio signals in a plurality of channels is high, at least one grouping of frequency data in the time direction or frequency direction is divided into a plurality of groups. The audio signal can be efficiently encoded by performing the same for all channels. For this reason, since the sound quality at a particularly low bit rate can be improved, the present invention can be applied to highly efficient transmission or storage of audio signals in fields such as broadcasting, communication, and storage. In particular, it can be suitably used for a portable terminal that services content including a video signal and an audio signal at a limited bit rate.

1 is a configuration diagram of a 2-channel audio signal encoding apparatus according to Embodiment 1 of the present invention. FIG. It is a block diagram of the audio signal encoding apparatus of 2 channels in Embodiment 2 of this invention. It is a block diagram of the conventional audio signal encoding apparatus of 2 channels. (A) shows the example of the encoding process unit before grouping, (b) is a figure which shows the example of the encoding process unit after grouping.

Explanation of symbols

100, 101 Analysis filter bank 110, 111, 112, 113 Strength calculation unit 120, 121, 122, 123, 124, 125 Grouping unit 130, 131 Similarity determination unit 140, 141 switching unit 150, 151, 152, 153 Code Encoding unit 160,161 Multiplexing unit 170,171 Downsampler 180,181 Low frequency encoding unit

Claims (6)

An analysis filter bank that inputs audio signals of multiple channels and converts them into a time series of frequency data,
An intensity calculator for calculating the intensity of frequency data in the plurality of channels;
A similarity determination unit that determines the similarity of the plurality of channels based on the intensity of frequency data in the plurality of channels;
When the similarity determination unit determines that the similarity is high, at least one grouping in the time direction or frequency direction of the frequency data in the plurality of channels is based on an average intensity of the frequency data in the plurality of channels. A grouping unit that is commonly used for the plurality of channels;
An audio signal encoding apparatus, comprising: an encoding unit that encodes the frequency data of the plurality of channels or the envelope data thereof based on the grouping. Input audio signals of a plurality of channels, generate encoded data for a signal in which a predetermined frequency band is deleted from the input audio signal, and generate envelope data of the signal in the deleted frequency band, An audio signal encoding device for multiplexing or transmitting or storing envelope data in the encoded data,
An analysis filter bank that inputs the audio signals of the plurality of channels and converts them into a time series of frequency data;
An intensity calculator that calculates the intensity of the frequency data of the deleted frequency band in the plurality of channels;
A similarity determination unit that determines the similarity of the plurality of channels based on the intensity of the frequency data of the deleted frequency band in the plurality of channels;
When it is determined that the similarity is high by the similarity determination unit, at least one grouping in the time direction or the frequency direction of the frequency data of the deleted frequency band is classified into the frequency data of the deleted frequency band. Based on the average intensity value, a grouping unit commonly performed for the plurality of channels;
An audio signal encoding apparatus, comprising: an encoding unit that encodes envelope data of the frequency data of the deleted frequency band of the plurality of channels based on the grouping.   The audio signal encoding apparatus according to claim 1 or 2, wherein the intensity of the frequency data is calculated by a power of an absolute value of the frequency data. A frequency conversion step of inputting audio signals of a plurality of channels and converting the signals into a time series of frequency data;
An intensity calculating step for calculating the intensity of frequency data in the plurality of channels;
A similarity determination step of determining the similarity of the plurality of channels based on the intensity of the frequency data in the plurality of channels;
When it is determined that the similarity is high in the similarity determination step, at least one grouping in the time direction or frequency direction of the frequency data in the plurality of channels is based on an average intensity of the frequency data in the plurality of channels. A grouping step to be performed in common for the plurality of channels;
An audio signal encoding method comprising: an encoding step of encoding frequency data of the plurality of channels or envelope data thereof based on the grouping. Input audio signals of a plurality of channels, generate encoded data for a signal in which a predetermined frequency band is deleted from the input audio signal, generate envelope data of the deleted frequency band, and generate the envelope A stereo audio signal encoding method for multiplexing or transmitting or storing data in the encoded data,
A frequency conversion step of inputting the audio signals of the plurality of channels and converting the signals into a time series of frequency data;
An intensity calculating step for calculating the intensity of the frequency data of the deleted frequency band of the plurality of channels;
A similarity determination step of determining the similarity of the plurality of channels based on the intensity of the frequency data of the deleted frequency bands of the plurality of channels;
When it is determined that the similarity is high in the similarity determination step, at least one grouping in the time direction or frequency direction of the frequency data of the deleted frequency band is classified into the frequency data of the deleted frequency band. A grouping step commonly performed on the plurality of channels based on an intensity average;
And a coding step of coding envelope data of the frequency data of the deleted frequency band of the plurality of channels based on the grouping.   6. The audio signal encoding method according to claim 4, wherein the intensity of the frequency data is calculated by a power of an absolute value of the frequency data.

Images