JP2024106008A - Digital voice processing device and digital voice processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a digital voice processing device capable of sufficiently compensating a higher-harmonic component included in a low frequency.

SOLUTION: A peripheral sample correction value generation unit 32 generates a peripheral sample correction value for correcting peripheral samples of a maximum sample and a minimum sample. A distortion correction value generation unit 33 generates a distortion correction value for correcting distortion correction objective samples, which are samples not corrected with the peripheral sample correction value, in an ascent sample section between a minimum sample located before a maximum sample and the maximum sample and a descent sample section between a maximum sample located after a maximum sample and a minimum sample. An addition/subtraction unit 37 adds/subtracts the peripheral sample correction value to/from the peripheral surface, and adds/subtracts the distortion correction value to/from the distortion correction objective sample.

SELECTED DRAWING: Figure 9

COPYRIGHT: (C)2024,JPO&INPIT

Description

The present invention relates to a digital audio processing device and a digital audio processing method for processing digital audio signals.

Patent document 1 describes a method of correcting the waveform of a digital audio signal by adding a correction value to samples located around a maximum sample having a maximum value of the waveform of the digital audio signal and subtracting a correction value from samples located around a minimum sample having a minimum value. Patent document 2 describes a method of correcting the waveform of a digital audio signal by adding a correction value to a maximum sample and subtracting a correction value from a minimum sample, in addition to the correction described in patent document 1.

Patent No. 3401171 Patent No. 3659489

The digital audio processing devices described in Patent Documents 1 and 2 cannot adequately compensate for harmonic components contained in low frequencies because the samples to be corrected are localized.

One or more embodiments aim to provide a digital audio processing device and a digital audio processing method that can adequately compensate for harmonic components contained in low frequencies.

A first aspect of one or more embodiments includes a maximum/minimum sample detection unit that detects maximum and minimum samples in samples constituting an input digital audio signal, and a first peripheral sample to be corrected, which includes one or more samples located before the maximum sample including the sample immediately before the maximum sample, one or more samples located after the maximum sample including the sample immediately after the maximum sample, which include one sample immediately after the maximum sample, one or more samples located before the minimum sample including the sample immediately before the minimum sample, which include one sample immediately before the minimum sample, which include one sample immediately after ... a peripheral sample correction value generation unit that multiplies a difference between each sample and the sample immediately preceding the first sample by a coefficient less than 1 to generate a first peripheral sample correction value for correcting the first peripheral sample, multiplies a difference between each sample of the second peripheral sample and the sample immediately preceding the second sample by a coefficient less than 1 to generate a second peripheral sample correction value for correcting the second peripheral sample, multiplies a difference between each sample of the third peripheral sample and the sample immediately preceding the third sample by a coefficient less than 1 to generate a third peripheral sample correction value for correcting the third peripheral sample, and multiplies a difference between each sample of the fourth peripheral sample and the sample immediately preceding the fourth sample by a coefficient less than 1 to generate a fourth peripheral sample correction value for correcting the fourth peripheral sample; one or more samples that are not corrected by the first or fourth surrounding sample correction value in an upward sample section between the minimum sample located before the maximum sample and the minimum sample are defined as first distortion correction target samples, and one or more samples that are not corrected by the second or third surrounding sample correction value in a downward sample section between the maximum sample and the minimum sample located after the maximum sample are defined as second distortion correction target samples, and a difference between each sample of the first distortion correction target samples and a sample adjacent to each sample is multiplied by a coefficient less than 1 to generate a first distortion correction value for correcting the first distortion correction target samples, and a difference between each sample of the second distortion correction target samples and a sample adjacent to each sample is multiplied by a coefficient less than 1 to generate a first distortion correction value for correcting the first distortion correction target samples, A digital audio processing device is provided that includes a distortion correction value generation unit that generates a second distortion correction value for correcting the second distortion correction target sample by multiplication, and an addition/subtraction unit that adds the first peripheral sample correction value to each sample of the first peripheral sample, adds the second peripheral sample correction value to each sample of the second peripheral sample, subtracts the third peripheral sample correction value from each sample of the third peripheral sample, subtracts the fourth peripheral sample correction value from each sample of the fourth peripheral sample, adds the first distortion correction value to the first distortion correction target sample or subtracts the first distortion correction value from the first distortion correction target sample, and adds the second distortion correction value to the second distortion correction target sample or subtracts the second distortion correction value from the second distortion correction target sample.

A second aspect of one or more embodiments is to detect a maximum sample and a minimum sample in samples constituting an input digital audio signal, and set one or more samples located before the maximum sample including the sample immediately before the maximum sample as a first peripheral sample to be corrected, one or more samples located after the maximum sample including the sample immediately after the maximum sample as a second peripheral sample to be corrected, one or more samples located before the minimum sample including the sample immediately before the minimum sample as a third peripheral sample to be corrected, and one or more samples located after the minimum sample including the sample immediately after the minimum sample as a fourth peripheral sample to be corrected, and the first peripheral sample multiplying a difference between each sample of the maximum and the sample immediately preceding the maximum sample by a coefficient less than 1 to generate a first peripheral sample correction value for correcting the first peripheral sample, multiplying a difference between each sample of the maximum and the sample immediately preceding the maximum sample by a coefficient less than 1 to generate a second peripheral sample correction value for correcting the second peripheral sample, multiplying a difference between each sample of the maximum and the sample immediately preceding the maximum sample by a coefficient less than 1 to generate a third peripheral sample correction value for correcting the third peripheral sample, multiplying a difference between each sample of the maximum and the sample immediately preceding the maximum sample by a coefficient less than 1 to generate a fourth peripheral sample correction value for correcting the fourth peripheral sample, One or more samples that are not corrected by the first or fourth surrounding sample correction value in an upward sample section between the minimum sample and the maximum sample located before the maximum sample are defined as first distortion correction target samples, and one or more samples that are not corrected by the second or third surrounding sample correction value in a downward sample section between the maximum sample and the minimum sample located after the maximum sample are defined as second distortion correction target samples, a difference between each of the first distortion correction target samples and a sample adjacent to each sample is multiplied by a coefficient less than 1 to generate a first distortion correction value for correcting the first distortion correction target samples, and a difference between each of the second distortion correction target samples and a sample adjacent to each sample is multiplied by a coefficient less than 1 to generate a first distortion correction value for correcting the first distortion correction target samples. A digital audio processing method is provided that generates a second distortion correction value for correcting the second distortion correction target sample by multiplying the sample by a coefficient less than 1/100, adds the first peripheral sample correction value to each sample of the first peripheral sample, adds the second peripheral sample correction value to each sample of the second peripheral sample, subtracts the third peripheral sample correction value from each sample of the third peripheral sample, subtracts the fourth peripheral sample correction value from each sample of the fourth peripheral sample, adds the first distortion correction value to the first distortion correction target sample or subtracts the first distortion correction value from the first distortion correction target sample, and adds the second distortion correction value to the second distortion correction target sample or subtracts the second distortion correction value from the second distortion correction target sample.

The digital audio processing device and digital audio processing method according to one or more embodiments can adequately compensate for harmonic components contained in low frequencies.

FIG. 1 is a block diagram illustrating a digital audio processing device according to one or more embodiments. FIG. 2 is a block diagram showing a specific example of the configuration of the voice waveform analysis unit 1 shown in FIG. FIG. 3 is a diagram showing an example of a waveform of an input digital audio signal. FIG. 4 is a diagram showing the operation of the maximum/minimum sample detection unit 12 shown in FIG. FIG. 5 is a diagram showing an example of grouping the number of extreme value intervals. FIG. 6 is a block diagram showing a specific example of the configuration of the correction target sample/correction coefficient setting unit 2 shown in FIG. FIG. 7 is a diagram showing an example of the correction target sample position pattern set in the correction target sample position pattern selection unit 21 shown in FIG. FIG. 8A is a diagram showing maximum and minimum sample coefficients and surrounding sample coefficients when the sampling frequencies are 44.1 kHz, 48 kHz, 88.2 kHz, and 96 kHz, which are set in the correction coefficient group selection unit 22 shown in FIG. FIG. 8B is a diagram showing the maximum and minimum sample coefficients and surrounding sample coefficients when the sampling frequencies are 176.4 kHz, 192 kHz, 352.8 kHz, and 384 kHz, which are set in the correction coefficient group selection unit 22 shown in FIG. FIG. 9 is a block diagram showing a specific example of the configuration of the voice waveform correction unit 3 shown in FIG. FIG. 10A is a diagram showing the operation of the voice waveform correction unit 3 to correct a maximum sample. FIG. 10B is a diagram showing the operation of the voice waveform correction unit 3 to correct the minimum sample. FIG. 11A is a diagram showing the operation of the voice waveform correction unit 3 to correct the sample immediately before and immediately after the maximum sample. FIG. 11B is a diagram showing the operation of the voice waveform correction unit 3 to correct the sample immediately before and immediately after the minimal sample. FIG. 12A is a diagram showing the operation of the voice waveform correction unit 3 to correct the samples one before, two before, one after, and two after the maximum sample. FIG. 12B is a diagram showing the operation of the voice waveform correction unit 3 to correct the samples one before, two before, one after, and two after the minimal sample. FIG. 13 is a diagram showing definitions of an ascending sample section, a descending sample section, the first and second intermediate samples, the first to fourth peripheral samples, and the first to eighth distortion correction samples. FIG. 14A is a diagram showing sample positions where the position constant by which the surrounding sample coefficients are multiplied is set to 1. FIG. 14B is a diagram showing sample positions where the position constant by which the surrounding sample coefficients are multiplied is set to 1/2. FIG. 15 is a diagram showing the distortion correction coefficients set in the correction coefficient group selection unit 22 shown in FIG. FIG. 16 is a diagram showing an example of the first distortion correction pattern. FIG. 17 is a diagram showing how the audio waveform correcting section 3 corrects each sample when the first distortion correction pattern shown in FIG. 16 is selected. FIG. 18 is a diagram showing an example of the second distortion correction pattern. FIG. 19 is a diagram showing how the audio waveform correcting section 3 corrects each sample when the second distortion correction pattern shown in FIG. 18 is selected. FIG. 20 is a diagram showing an example of the third distortion correction pattern. FIG. 21 is a diagram showing how the audio waveform correcting section 3 corrects each sample when the third distortion correction pattern shown in FIG. 20 is selected. FIG. 22 is a diagram showing an example of the fourth distortion correction pattern when the number of samples of the first and second distortion correction target samples is an even number. FIG. 23 is a diagram showing how the audio waveform correcting section 3 corrects each sample when the fourth distortion correction pattern shown in FIG. 22 is selected. FIG. 24 is a diagram showing an example of the fourth distortion correction pattern when the number of samples of the first and second distortion correction target samples is an odd number. FIG. 25 is a diagram showing how the audio waveform correcting section 3 corrects each sample when the fourth distortion correction pattern shown in FIG. 24 is selected. FIG. 26 is a block diagram illustrating an example configuration of a microcomputer that executes a digital audio processing program according to one or more embodiments of the present invention. FIG. 27 is a flowchart illustrating a digital audio processing method according to one or more embodiments, and a process that a digital audio processing program according to one or more embodiments causes a central processing unit to execute.

The following describes a digital audio processing device and a digital audio processing method according to one or more embodiments with reference to the accompanying drawings.

In FIG. 1, a digital audio processing device 100 according to one or more embodiments includes an audio waveform analysis unit 1, a correction target sample/correction coefficient setting unit 2, and an audio waveform correction unit 3. A digital audio signal with a predetermined quantization bit number and a predetermined sampling frequency is input to the audio waveform analysis unit 1 and the audio waveform correction unit 3 as an input digital audio signal to be corrected. Although not shown in FIG. 1, the input digital audio signal input to the audio waveform correction unit 3 is delayed by a delay device so that the audio waveform correction unit 3 corrects the input digital audio signal according to the audio waveform analyzed in the audio waveform analysis unit 1. This allows the timing of the analysis of the audio waveform by the audio waveform analysis unit 1 and the correction of the input digital audio signal by the audio waveform correction unit 3 to be matched.

Figure 2 shows a specific example configuration of the audio waveform analysis unit 1. As shown in Figure 2, the audio waveform analysis unit 1 has a sampling frequency acquisition unit 11, a maximum/minimum sample detection unit 12, and an extreme value interval group selection unit 13. The sampling frequency acquisition unit 11 acquires a sampling frequency from the input digital audio signal and supplies it to the correction target sample/correction coefficient setting unit 2. If the sampling frequency acquisition unit 11 cannot acquire the sampling frequency from the input digital audio signal, the user may manually input the sampling frequency.

The maximum and minimum sample detection unit 12 detects maximum samples having maximum values and minimum samples having minimum values in the input digital audio signal. The maximum and minimum sample detection unit 12 supplies the number of intervals (number of samples) from the minimum sample to the maximum sample and the number of intervals (number of samples) from the maximum sample to the minimum sample as the number of extreme value intervals to the extreme value interval group selection unit 13 and the correction target sample and correction coefficient setting unit 2. The maximum and minimum sample detection unit 12 supplies the correction target sample and correction coefficient setting unit 2 with a slope identification value indicating whether the waveform of the input digital audio signal has a positive slope from the minimum sample to the maximum sample or a negative slope from the maximum sample to the minimum sample.

The operation of the maximum and minimum sample detection unit 12 will be specifically described using as an example the waveform shown in Figure 3 input to the maximum and minimum sample detection unit 12. In Figure 3, S99 and S0 to S7 indicated by circles indicate samples of the input digital audio signal. As shown in Figure 4, the maximum and minimum sample detection unit 12 uses each of the samples S0 to S7 as a target sample and the sample immediately preceding each sample as a comparison sample, and compares the target sample with the comparison sample to determine whether the sample value is large or small. For samples S0 to S7, a large/small determination of small, large, large, large, small, small, small, large is obtained.

The maximum/minimum sample detection unit 12 detects whether the slope of the waveform is positive or negative based on the result of the large/small judgment. For example, if the slope of the waveform is positive, the maximum/minimum sample detection unit 12 outputs +1 as the slope discrimination value, and if the slope of the waveform is negative, the slope discrimination value is -1. The maximum/minimum sample detection unit 12 detects maximum samples having maximum values and minimum samples having minimum values based on the result of the large/small judgment. A sample of interest that is judged to be small when the large/small judgment changes from small to large is a minimum sample, and a sample of interest that is judged to be large when the large/small judgment changes from large to small is a maximum sample.

In the waveform shown in Figure 3, the hatched samples S0 and S6 are extremely small samples, and the hatched sample S3 is extremely large.

When the maximum/minimum sample detection unit 12 detects a maximum or minimum sample, the count value is set to 1 for the next sample, and the unit continues to count samples until it reaches a minimum sample and so on. As shown in FIG. 4, in the waveform shown in FIG. 3, the count value is 3 for sample S3, which is a maximum sample, and the count value is 3 for sample S6, which is a minimum sample. The maximum/minimum sample detection unit 12 outputs the count value thus calculated as the number of extreme value intervals.

The extreme value interval numbers 2 to 32 are referred to as extreme value intervals 2fs to 32fs, respectively. In FIG. 2, the extreme value interval group selection unit 13 supplies the extreme value interval group selected based on the input extreme value interval number to the correction target sample/correction coefficient setting unit 2.

The extreme value interval group selection unit 13 outputs extreme value interval groups obtained by grouping the number of extreme value intervals as shown in FIG. 5 as an example. In the example shown in FIG. 5, extreme value intervals of 2 fs to 5 fs are extreme value interval group G1, extreme value intervals of 6 fs to 8 fs are extreme value interval group G2, extreme value intervals of 9 fs to 11 fs are extreme value interval group G3, and extreme value intervals of 12 fs to 16 fs are extreme value interval group G4. Extreme value intervals of 17 fs to 20 fs are extreme value interval group G5, extreme value intervals of 21 fs to 24 fs are extreme value interval group G6, extreme value intervals of 25 fs to 28 fs are extreme value interval group G7, and extreme value intervals of 29 fs to 32 fs are extreme value interval group G8.

The digital audio processing device 100 excludes waveforms with an extreme value interval count exceeding 32 from the correction target. The digital audio processing device 100 corrects the waveform to compensate for the harmonic components of the input digital audio signal. Since compensating for the harmonic components of a waveform with a low frequency where the extreme value interval count exceeds 32 is ineffective and therefore not necessary, the upper limit of the extreme value interval to be corrected is set to 32 fs. The upper limit of the extreme value interval to be corrected is not limited to 32 fs, and a predetermined extreme value interval may be set as the upper limit. The extreme value interval group selection unit 13 may group the extreme value interval from 2 fs to the upper limit extreme value interval into a predetermined number of extreme value intervals.

Figure 6 shows a specific example of the configuration of the correction target sample/correction coefficient setting unit 2. As shown in Figure 6, the correction target sample/correction coefficient setting unit 2 has a correction target sample position pattern selection unit 21 and a correction coefficient group selection unit 22.

As an example, as shown in FIG. 7, the correction target sample position pattern selection unit 21 selects the correction target sample position pattern according to the input extreme value interval groups G1 to G8. In FIG. 7, S(-1) indicates the sample immediately before the maximum or minimum sample, and S(+1) indicates the sample immediately after the maximum or minimum sample. S(-2) indicates the sample two before the maximum or minimum sample, and S(+2) indicates the sample two after the maximum or minimum sample. In this way, if n is a number from 1 to 8, S(-n) indicates the sample n before the maximum or minimum sample, and S(+n) indicates the sample n after the maximum or minimum sample.

When the sample position pattern to be corrected in extreme value interval group G1 is S(-1), S(+1), it means that the samples one before and one after the maximum or minimum sample are samples to be corrected around the maximum or minimum sample (hereinafter, peripheral samples).When the sample position pattern to be corrected in extreme value interval group G2 is S(-1), S(-2), S(+1), S(+2), it means that the samples one before, two before, one after, and two after the maximum or minimum sample are peripheral samples to be corrected.

In this way, in the correction target sample position pattern shown in FIG. 7, the surrounding samples to be corrected that sandwich the maximum or minimum sample are set to be expanded by one sample each in the forward and backward directions as the extreme value interval group moves from G1 to G8. The correction target sample position pattern selection unit 21 selects the correction target sample position pattern that corresponds to the extreme value interval group.

The surrounding samples to be corrected may be limited to only the samples immediately before and after the maximum and minimum samples. For extreme value interval group G1 only, the samples immediately before and after the maximum and minimum samples may be the surrounding samples to be corrected, and for extreme value interval group G2 and onwards, the samples immediately before, two before, one after and two after the maximum and minimum samples may be the surrounding samples to be corrected. One or more samples located before and after the maximum sample, including the samples immediately before and one after the maximum sample, and one or more samples located before and after the minimum sample, including the samples immediately before and one after the minimum sample, may be the surrounding samples to be corrected.

The correction coefficient group selection unit 22 sets maximum and minimum sample coefficients less than 1 and surrounding sample coefficients less than 1 as shown in Figures 8A and 8B as one correction coefficient group. The maximum and minimum sample coefficients are used to generate correction values that correct the maximum and minimum samples, and the surrounding sample coefficients are used to generate correction values that correct the surrounding samples. The correction coefficient group selection unit 22 selects and outputs the maximum and minimum sample coefficients and surrounding sample coefficients that correspond to the sampling frequency and the extreme value interval of 2 fs to 32 fs.

In the example shown in FIG. 8A, a common set of coefficients is set for sampling frequencies of 44.1 kHz and 48 kHz, and a common set of coefficients is set for sampling frequencies of 88.2 kHz and 96 kHz. In the example shown in FIG. 8B, a common set of coefficients is set for sampling frequencies of 176.4 kHz and 192 kHz, and a common set of coefficients is set for sampling frequencies of 352.8 kHz and 384 kHz.

The values of the maximum and minimum sample coefficients and surrounding sample coefficients shown in Figures 8A and 8B are just an example, and the maximum and minimum sample coefficients and surrounding sample coefficients are not limited to the values shown in Figures 8A and 8B. With the exception of an extreme value interval of 2 fs at sampling frequencies of 176.4 kHz and 192 kHz, and an extreme value interval of 2 fs to 6 fs at sampling frequencies of 352.8 kHz and 384 kHz, the maximum and minimum sample coefficients and surrounding sample coefficients are generally set smaller as the extreme value interval becomes larger. However, the same coefficient may be set for multiple extreme value intervals.

When the correction coefficient group selection unit 22 selects and outputs the maximum/minimum sample coefficients and the surrounding sample coefficients based on the sampling frequency and the extreme value interval 2fs to 32fs, it determines whether the maximum/minimum sample coefficients and the surrounding sample coefficients are positive or negative based on the slope discrimination value as follows: When the correction coefficient group selection unit 22 determines that the maximum/minimum sample coefficients and the surrounding sample coefficients are positive, it outputs the maximum/minimum sample coefficients and the surrounding sample coefficients set in Figures 8A and 8B as is. When the correction coefficient group selection unit 22 determines that the maximum/minimum sample coefficients and the surrounding sample coefficients are negative, it multiplies the maximum/minimum sample coefficients and the surrounding sample coefficients set in Figures 8A and 8B by -1 and outputs them.

If the slope discrimination value is +1, the waveform moves from the minimum sample to the maximum sample. At this time, the correction coefficient group selection unit 22 sets the surrounding sample coefficient for generating a correction value for correcting the surrounding sample following the minimum sample to negative, and the surrounding sample coefficient for generating a correction value for correcting the surrounding sample before the maximum sample and the maximum and minimum sample coefficient for correcting the maximum sample to positive. If the slope discrimination value is -1, the waveform moves from the maximum sample to the minimum sample. At this time, the correction coefficient group selection unit 22 sets the surrounding sample coefficient for generating a correction value for correcting the surrounding sample following the maximum sample to positive, and the surrounding sample coefficient for generating a correction value for correcting the surrounding sample before the minimum sample and the maximum and minimum sample coefficient for correcting the minimum sample to negative.

Figure 9 shows a specific example of the configuration of the audio waveform correction unit 3. As shown in Figure 9, the audio waveform correction unit 3 has a maximum/minimum sample correction value generation unit 31, a surrounding sample correction value generation unit 32, a distortion correction value generation unit 33, and adders 34 to 36. The adders 34 to 36 form an addition/subtraction unit 37.

The input digital audio signal is input to the maximum/minimum sample correction value generation unit 31, the surrounding sample correction value generation unit 32, the distortion correction value generation unit 33, and the adder 34. The correction target sample position pattern output from the correction target sample position pattern selection unit 21 is input to the maximum/minimum sample correction value generation unit 31 and the surrounding sample correction value generation unit 32. The maximum/minimum sample coefficients output from the correction coefficient group selection unit 22 are input to the maximum/minimum sample correction value generation unit 31. The surrounding sample coefficients output from the correction coefficient group selection unit 22 are input to the surrounding sample correction value generation unit 32.

Figures 10A and 10B show the operation of the maximum/minimum sample correction value generator 31 generating a correction value for correcting the maximum sample Smax or the minimum sample Smin, and the adder 34 correcting the maximum sample Smax or the minimum sample Smin based on the generated correction value.

As shown in FIG. 10A, the maximum/minimum sample correction value generator 31 multiplies the difference Δ(-1) between the maximum sample Smax and the sample S(-1) immediately preceding the maximum sample Smax by the maximum/minimum sample coefficient to generate a maximum sample correction value Vaddmax for correcting the maximum sample Smax. Since the maximum/minimum sample coefficient is positive at this time, the adder 34 adds the maximum sample correction value Vaddmax, which is a positive value, to the maximum sample Smax in the input digital audio signal. As a result, the maximum sample Smax before correction, shown by the dashed line, is corrected to the maximum sample Smax', shown by the solid line, by adding the maximum sample correction value Vaddmax.

As shown in FIG. 10B, the maximum and minimum sample correction value generator 31 multiplies the difference Δ(-1) between the minimum sample Smin and the sample S(-1) immediately preceding the minimum sample Smin by the maximum and minimum sample coefficient to generate a minimum sample correction value Vsubmin for correcting the minimum sample Smin. Since the maximum and minimum sample coefficient is negative at this time, the adder 34 adds the minimum sample correction value Vsubmin, which is a negative value, to the minimum sample Smin in the input digital audio signal. As a result, the minimum sample Smin before correction, indicated by the dashed line, is corrected to the minimum sample Smin', indicated by the solid line, by subtracting the minimum sample correction value Vsubmin.

Specifically, the sampling frequency is 44.1 kHz, and the extreme value interval is 4 fs. The maximum/minimum sample correction value generator 31 generates the maximum sample correction value Vaddmax by multiplying the difference Δ(-1) shown in FIG. 10A by 1/8, and generates the minimum sample correction value Vsubmin by multiplying the difference Δ(-1) shown in FIG. 10B by 1/8. The adder 34 adds the maximum sample correction value Vaddmax to the maximum sample Smax, and subtracts the minimum sample correction value Vsubmin from the minimum sample Smin.

Figures 11A and 11B show a case where the sample S(-1) immediately before the maximum sample Smax or the sample S(+1) immediately after the minimum sample Smin are the surrounding samples to be corrected. Figures 11A and 11B show the operation of the surrounding sample correction value generator 32 generating correction values for correcting the samples S(-1) and S(+1), and the adder 35 correcting the samples S(-1) and S(+1) based on the generated correction values.

As shown in FIG. 11A, the surrounding sample correction value generator 32 multiplies the difference Δ(-1) between sample S(-1) and the maximum sample Smax, which is the sample immediately following it, by the surrounding sample coefficient to generate a correction value Vadd(-1) for correcting sample S(-1). Because the surrounding sample coefficient is positive at this time, the adder 35 adds the correction value Vadd(-1), which is a positive value, to sample S(-1) in the input digital audio signal. As a result, the uncorrected sample S(-1) shown by the dashed line is corrected to sample S(-1)' shown by the solid line by adding the correction value Vadd(-1).

The surrounding sample correction value generator 32 multiplies the difference Δ(+1) between sample S(+1) and the maximum sample Smax, which is the sample immediately preceding it, by the surrounding sample coefficient to generate a correction value Vadd(+1) for correcting sample S(+1). Since the surrounding sample coefficient is positive at this time, the adder 35 adds the correction value Vadd(+1), which is a positive value, to sample S(+1) in the input digital audio signal. As a result, the uncorrected sample S(+1) shown by the dashed line is corrected to sample S(+1)' shown by the solid line by adding the correction value Vadd(+1).

As shown in FIG. 11B, the surrounding sample correction value generator 32 multiplies the difference Δ(-1) between sample S(-1) and the minimum sample Smin, which is the sample immediately following sample, by the surrounding sample coefficient to generate a correction value Vsub(-1) for correcting sample S(-1). Because the surrounding sample coefficient at this time is negative, the adder 35 adds the negative correction value Vsub(-1) to sample S(-1) in the input digital audio signal. As a result, the uncorrected sample S(-1) shown by the dashed line is corrected to sample S(-1)' shown by the solid line by subtracting the correction value Vsub(-1).

The surrounding sample correction value generator 32 multiplies the difference Δ(+1) between sample S(+1) and the previous sample, the minimum sample Smin, by the surrounding sample coefficient to generate a correction value Vsub(+1) for correcting sample S(+1). Because the surrounding sample coefficient is negative at this time, the adder 35 adds the negative correction value Vsub(+1) to sample S(+1) in the input digital audio signal. As a result, the uncorrected sample S(+1) shown by the dashed line is corrected to sample S(+1)' shown by the solid line by subtracting the correction value Vsub(+1).

Specifically, take as an example a sampling frequency of 44.1 kHz and an extreme value interval of 4 fs. The surrounding sample correction value generation unit 32 generates a correction value Vadd(-1) by multiplying the difference Δ(-1) shown in Fig. 11A by 1/16, and generates a correction value Vadd(+1) by multiplying the difference Δ(+1) shown in Fig. 11A by 1/16. The surrounding sample correction value generation unit 32 generates a correction value Vsub(-1) by multiplying the difference Δ(-1) shown in Fig. 11B by 1/16, and generates a correction value Vsub(+1) by multiplying the difference Δ(+1) shown in Fig. 11B by 1/16. The adder 35 adds the correction values Vadd(-1) and Vadd(+1) to the samples S(-1) and S(+1) shown in FIG. 11A, respectively, and subtracts the correction values Vsub(-1) and Vsub(+1) from the samples S(-1) and S(+1) shown in FIG. 11B, respectively.

Incidentally, in a waveform with an extreme value interval of 2 fs, the sample S(-1) immediately before the maximum sample Smax and the sample S(+1) immediately after the minimum sample Smin are the same sample, and the sample S(+1) immediately after the maximum sample Smax and the sample S(-1) immediately before the minimum sample Smin are the same sample. Therefore, in a waveform with an extreme value interval of 2 fs, it is best to simply add the correction value Vadd(-1) to the sample S(-1) immediately before the maximum sample Smax to correct it to sample S(-1)', and add the correction value Vadd(+1) to the sample S(+1) immediately after the maximum sample Smax to correct it to sample S(+1)'.

Figures 12A and 12B show a case where the sample S(-1), the sample S(-2), the sample S(+1), and the sample S(+2) are the surrounding samples to be corrected. Figures 12A and 12B show the operation of the surrounding sample correction value generator 32 generating correction values for correcting the samples S(-1), S(-2), S(+1), and S(+2), and the adder 35 correcting the samples S(-1), S(-2), S(+1), and S(+2) based on the generated correction values.

The operation of the surrounding sample correction value generation unit 32 to correct the samples S(-1) and S(+1) that sandwich the maximum sample Smax, and the samples S(-1) and S(+1) that sandwich the minimum sample Smin, is as described in Figures 11A and 11B.

As shown in FIG. 12A, the surrounding sample correction value generation unit 32 multiplies the difference Δ(-2) between sample S(-2) and the sample immediately following it, sample S(-1), by the surrounding sample coefficient to generate a correction value Vadd(-2) for correcting sample S(-2). Because the surrounding sample coefficient at this time is positive, the adder 35 adds the correction value Vadd(-2), which is a positive value, to sample S(-2) in the input digital audio signal. As a result, the uncorrected sample S(-2) shown by the dashed line is corrected to sample S(-2)' shown by the solid line by adding the correction value Vadd(-2).

The surrounding sample correction value generator 32 also multiplies the difference Δ(+2) between sample S(+2) and the previous sample S(+1) by the surrounding sample coefficient to generate a correction value Vadd(+2) for correcting sample S(+2). Because the surrounding sample coefficient is positive at this time, the adder 35 adds the positive correction value Vadd(+2) to sample S(+2) in the input digital audio signal. As a result, the uncorrected sample S(+2) shown by the dashed line is corrected to sample S(+2)' shown by the solid line by adding the correction value Vadd(+2).

As shown in FIG. 12B, the surrounding sample correction value generator 32 multiplies the difference Δ(-2) between sample S(-2) and the sample immediately following it, sample S(-1), by the surrounding sample coefficient to generate a correction value Vsub(-2) for correcting sample S(-2). Because the surrounding sample coefficient at this time is negative, the adder 35 adds the negative correction value Vsub(-2) to sample S(-2) in the input digital audio signal. As a result, the uncorrected sample S(-2) shown by the dashed line is corrected to sample S(-2)' shown by the solid line by subtracting the correction value Vsub(-2).

The surrounding sample correction value generator 32 multiplies the difference Δ(+2) between sample S(+2) and the previous sample S(+1) by the surrounding sample coefficient to generate a correction value Vsub(+2) for correcting sample S(+2). Because the surrounding sample coefficient is negative at this time, the adder 35 adds the negative correction value Vsub(+2) to sample S(+2) in the input digital audio signal. As a result, the uncorrected sample S(+2) shown by the dashed line is corrected to sample S(+2)' shown by the solid line by subtracting the correction value Vsub(+2).

Specifically, take as an example a sampling frequency of 44.1 kHz and an extreme value interval of 6 fs. The surrounding sample correction value generation unit 32 generates a correction value Vadd(-2) by multiplying the difference Δ(-2) shown in Fig. 12A by 1/32, and generates a correction value Vadd(+2) by multiplying the difference Δ(+2) shown in Fig. 12A by 1/32. The surrounding sample correction value generation unit 32 generates a correction value Vsub(-2) by multiplying Δ(-2) shown in Fig. 12B by 1/32, and generates a correction value Vsub(+2) by multiplying the difference Δ(+2) shown in Fig. 12B by 1/32. The adder 35 adds the correction values Vadd(-2) and Vadd(+2) to the samples S(-2) and S(+2) shown in FIG. 12A, respectively, and subtracts the correction values Vsub(-2) and Vsub(+2) from the samples S(-2) and S(+2) shown in FIG. 12B, respectively.

As shown in Figure 13, the section between the minimum sample Smin located before the maximum sample Smax and the maximum sample Smax is called the rising sample section, and the section between the maximum sample Smax and the minimum sample Smin located after the maximum sample Smax is called the falling sample section. Figure 13 shows an example where the extreme value interval is 8 fs, with sample S8 being the maximum sample Smax and sample S16 being the minimum sample Smin.

One or more samples located before the maximum sample Smax, including the sample S(-1) immediately preceding the maximum sample Smax, are referred to as the first peripheral samples to be corrected, and one or more samples located after the maximum sample Smax, including the sample S(+1) immediately succeeding the maximum sample Smax, are referred to as the second peripheral samples to be corrected. One or more samples located before the minimum sample Smin, including the sample S(-1) immediately preceding the minimum sample Smin, are referred to as the third peripheral samples to be corrected, and one or more samples located after the minimum sample Smin, including the sample immediately succeeding the minimum sample Smin, are referred to as the fourth peripheral samples to be corrected.

As described above, the number of samples for the first to fourth surrounding samples is set according to the extreme value interval groups G1 to G8 as shown in FIG. 7.

The surrounding sample correction value generation unit 32 generates a first surrounding sample correction value for correcting the first surrounding sample by multiplying the difference between each sample of the first surrounding sample and the sample immediately following each sample by a surrounding sample coefficient. The surrounding sample correction value generation unit 32 generates a second surrounding sample correction value for correcting the second surrounding sample by multiplying the difference between each sample of the second surrounding sample and the sample immediately preceding each sample by a surrounding sample coefficient.

The peripheral sample correction value generating unit 32 generates a third peripheral sample correction value for correcting the third peripheral sample by multiplying the difference between each sample of the third peripheral sample and the sample immediately following each sample by a peripheral sample coefficient. The peripheral sample correction value generating unit 32 generates a fourth peripheral sample correction value for correcting the fourth peripheral sample by multiplying the difference between each sample of the fourth peripheral sample and the sample immediately preceding each sample by a peripheral sample coefficient.

At this time, the surrounding sample correction value generation unit 32 may generate the surrounding sample correction value by simply multiplying each difference by the surrounding sample coefficient set in Figures 8A and 8B, or may adjust the surrounding sample correction value by multiplying the surrounding sample coefficient by a predetermined constant. The surrounding sample correction value generation unit 32 adjusts the surrounding sample correction value, for example, as follows.

When the surrounding sample correction value generation unit 32 generates a surrounding sample correction value to be added to or subtracted from a sample close to the maximum sample Smax or minimum sample Smin, it multiplies each difference directly by the surrounding sample coefficient to generate the surrounding sample correction value. In this case, the surrounding sample coefficient is multiplied by a constant of 1. When the surrounding sample correction value generation unit 32 generates a surrounding sample correction value to be added to or subtracted from a sample far from the maximum sample Smax or minimum sample Smin, it multiplies each difference by a coefficient obtained by multiplying the surrounding sample coefficient by a constant less than 1 to generate the surrounding sample correction value. The constant by which the surrounding sample coefficient is multiplied is set according to the position of the sample, and is therefore referred to as a position constant.

As an example, the surrounding sample correction value generation unit 32 may set a position constant as shown in FIG. 14A and FIG. 14B, multiply the surrounding sample coefficient by the position constant, and generate the surrounding sample correction value. FIG. 14A shows sample positions where the position constant for each extreme value interval group G1 to G8 is 1. FIG. 14B shows sample positions where the position constant for each extreme value interval group G1 to G8 is 1/2.

When generating surrounding sample correction values for samples far from the maximum sample Smax or minimum sample Smin, the correction amount for distant samples can be intentionally made small by multiplying each difference by a coefficient obtained by multiplying the surrounding sample coefficient by a position constant less than 1 to generate the surrounding sample correction value. It is not essential for the surrounding sample correction value generation unit 32 to multiply the surrounding sample coefficient by the position constant.

In FIG. 9, the adder 34 of the addition/subtraction unit 37 adds the maximum sample correction value Vaddmax to the maximum sample Smax and subtracts the minimum sample correction value Vsubmin from the minimum sample Smin. The adder 35 of the addition/subtraction unit 37 adds the first surrounding sample correction value to each sample of the first surrounding sample and adds the second surrounding sample correction value to each sample of the second surrounding sample. The adder 35 subtracts the third surrounding sample correction value from each sample of the third surrounding sample and subtracts the fourth surrounding sample correction value from each sample of the fourth surrounding sample.

As can be seen from the above explanation, in the waveform correction of the input digital audio signal by the maximum/minimum sample correction value generator 31, the surrounding sample correction value generator 32, and the adders 34 and 35, the following samples are uncorrected samples. In the rising sample section where the sample value rises from the minimum sample Smin to the maximum sample Smax, the first intermediate sample located between the first and fourth surrounding samples shown in FIG. 13 is an uncorrected sample. In the falling sample section where the sample value falls from the maximum sample Smax to the minimum sample Smin, the second intermediate sample located between the second and third surrounding samples shown in FIG. 13 is an uncorrected sample.

The distortion correction value generating unit 33 generates a distortion correction value for reducing waveform distortion caused by the presence of uncorrected samples. The adder 36 further corrects the input digital audio signal by adding the distortion correction value to the input digital audio signal or subtracting the distortion correction value from the input digital audio signal. The distortion correction value generating unit 33 and the adder 36 further correct the input digital audio signal as described below, thereby further restoring the harmonic components of the original sound and generating an output digital audio signal that is closer to the original sound. By having the distortion correction value generating unit 33 and the adder 36, the harmonic components contained in the low frequency range can be sufficiently compensated for.

The correction coefficient group selection unit 22 further sets a distortion correction coefficient less than 1 as shown in FIG. 15 as another correction coefficient group. The correction coefficient group selection unit 22 selects and outputs a distortion correction coefficient corresponding to the sampling frequency and the extreme value interval of 4 fs to 32 fs. The numerical values of the distortion correction coefficients shown in FIG. 15 are only an example, and the distortion correction coefficients are not limited to the numerical values shown in FIG. 15. With the exception of the extreme value interval of 4 fs to 7 fs at sampling frequencies of 352.8 kHz and 384 kHz, the distortion correction coefficients are set smaller as the extreme value interval becomes larger. It is preferable to set the distortion correction coefficient to a value smaller than the maximum/minimum sample coefficients and the surrounding sample coefficients.

The distortion correction coefficients output from the correction coefficient group selection unit 22 are input to the distortion correction value generation unit 33. A distortion correction pattern signal corresponding to the distortion correction pattern selected by the user is input to the distortion correction value generation unit 33.

As shown in FIG. 13, one or more first intermediate samples that are not corrected by the first or fourth surrounding sample correction value in the rising sample section are referred to as first distortion correction target samples, and one or more second intermediate samples that are not corrected by the second or third surrounding sample correction value in the falling sample section are referred to as second distortion correction target samples.

For example, if the extreme value interval is 4 fs, the central sample in the rising sample section is the first sample to be corrected for distortion, and the central sample in the falling sample section is the second sample to be corrected for distortion. If the extreme value interval is 8 fs, the central three samples in the rising sample section are the first sample to be corrected for distortion, and the central three samples in the falling sample section are the second sample to be corrected for distortion.

The distortion correction value generation unit 33 generates first and second distortion correction values for correcting at least the first and second distortion correction target samples as distortion correction target samples. As shown in FIG. 13, it is preferable that the distortion correction value generation unit 33 further sets the first and fourth peripheral samples as third and fourth distortion correction target samples, respectively, and sets the second and third peripheral samples as fifth and sixth distortion correction target samples, respectively. In this case, the distortion correction value generation unit 33 generates third to sixth distortion correction values for correcting the third to sixth distortion correction target samples.

As shown in FIG. 13, it is preferable that the distortion correction value generation unit 33 further sets the maximum sample Smax as the seventh distortion correction target sample and the minimum sample Smin as the eighth distortion correction target sample. In this case, the distortion correction value generation unit 33 generates seventh and eighth distortion correction values that correct the seventh and eighth distortion correction target samples.

The first distortion correction pattern will be described using Figures 16 and 17. Figures 16 and 17 show an example in which all of the first to eighth distortion correction target samples are distortion correction target samples, using an extreme value interval of 8 fs as an example. The distortion correction value generation unit 33 multiplies the difference between each sample of the first distortion correction target sample and the sample immediately following each sample by a distortion correction coefficient to generate a first distortion correction value. The distortion correction value generation unit 33 multiplies the difference between each sample of the second distortion correction target sample and the sample immediately following each sample by a distortion correction coefficient to generate a second distortion correction value.

The distortion correction value generation unit 33 multiplies the difference between each of the third to sixth distortion correction target samples and the sample immediately following each sample by a distortion correction coefficient to generate the third to sixth distortion correction values, respectively. The distortion correction value generation unit 33 multiplies the difference between the seventh distortion correction target sample and the sample immediately preceding it by the distortion correction coefficient to generate the seventh distortion correction value, and multiplies the difference between the eighth distortion correction target sample and the sample immediately preceding it by the distortion correction coefficient to generate the eighth distortion correction value.

The adder 36 of the addition/subtraction unit 37 adds the first, third, and fourth distortion compensation values to the first, third, and fourth distortion compensation target samples, respectively, and subtracts the second, fifth, and sixth distortion compensation values from the second, fifth, and sixth distortion compensation target samples, respectively. The adder 36 adds the seventh distortion compensation value to the seventh distortion compensation target sample, and subtracts the eighth distortion compensation value from the eighth distortion compensation target sample.

Figure 16 conceptually illustrates waveform correction of an input digital audio signal by only the distortion correction value generator 33 and the adder 36, without considering waveform correction of the input digital audio signal by the maximum/minimum sample correction value generator 31, the surrounding sample correction value generator 32, and the adders 34 and 35. As shown in Figure 16, when the distortion correction value generator 33 and the adder 36 correct the input digital audio signal with the first distortion correction pattern, the sample values of each sample from sample S1 to sample S8 (maximum sample Smax) are corrected to increase. When the distortion correction value generator 33 and the adder 36 correct the input digital audio signal with the first distortion correction pattern, the sample values of each sample from sample S9 to sample S16 (minimum sample Smin) are corrected to decrease.

Figure 17 shows how each sample is corrected by the maximum/minimum sample correction value generator 31, the surrounding sample correction value generator 32, the distortion correction value generator 33, and the adders 34 to 36 when the first distortion correction pattern is selected. In Figure 17 and Figures 19, 21, 23, and 25 described below, the first and second intermediate samples are samples that were previously uncorrected. In Figure 17, an upward arrow indicates the addition of a maximum/minimum sample correction value, surrounding sample correction value, or distortion correction value, and a downward arrow indicates the subtraction of a maximum/minimum sample correction value, surrounding sample correction value, or distortion correction value.

If the arrows for correcting maximum/minimum samples or surrounding samples and for correcting distortion correction values point in the same direction, each sample is corrected with a correction value that is the sum of the maximum/minimum sample correction value or surrounding sample correction value and the distortion correction value. If the arrows for correcting maximum/minimum samples or surrounding samples and for correcting distortion correction values point in opposite directions, each sample is corrected with a correction value that is the difference between the maximum/minimum sample correction value or surrounding sample correction value and the distortion correction value.

The first distortion correction pattern is expected to reduce waveform distortion, restore the harmonic components of the original sound, and improve the sound response by speeding up the sound's rise time.

The second distortion correction pattern will be explained using Figures 18 and 19. Figures 18 and 19 also show an example in which the extreme value interval is 8 fs and all of the first to eighth distortion correction target samples are distortion correction target samples. In explaining the second distortion correction pattern, explanations that overlap with the explanation of the first distortion correction pattern may be omitted.

The distortion correction value generating unit 33 multiplies the difference between each sample of the first distortion correction target sample and the sample immediately preceding each sample by a distortion correction coefficient to generate a first distortion correction value. The distortion correction value generating unit 33 multiplies the difference between each sample of the second distortion correction target sample and the sample immediately preceding each sample by a distortion correction coefficient to generate a second distortion correction value. The distortion correction value generating unit 33 multiplies the difference between each sample of the third distortion correction target sample and the sample immediately following each sample by a distortion correction coefficient to generate a third distortion correction value. The distortion correction value generating unit 33 multiplies the difference between each sample of the fourth distortion correction target sample and the sample immediately preceding each sample by a distortion correction coefficient to generate a fourth distortion correction value.

The distortion correction value generating unit 33 multiplies the difference between each sample of the fifth distortion correction target sample and the sample immediately preceding each sample by a distortion correction coefficient to generate a fifth distortion correction value. The distortion correction value generating unit 33 multiplies the difference between each sample of the sixth distortion correction target sample and the sample immediately following each sample by a distortion correction coefficient to generate a sixth distortion correction value. The distortion correction value generating unit 33 multiplies the difference between the seventh distortion correction target sample and the sample immediately preceding it by the distortion correction coefficient to generate a seventh distortion correction value, and multiplies the difference between the eighth distortion correction target sample and the sample immediately preceding it by the distortion correction coefficient to generate an eighth distortion correction value.

The adder 36 adds the first and third distortion compensation values to the first and third distortion compensation target samples, respectively, and subtracts the fourth distortion compensation value from the fourth distortion compensation target sample. The adder 36 adds the second and fifth distortion compensation values to the second and fifth distortion compensation target samples, respectively, and subtracts the sixth distortion compensation value from the sixth distortion compensation target sample. The adder 36 adds the seventh distortion compensation value to the seventh distortion compensation target sample, and subtracts the eighth distortion compensation value from the eighth distortion compensation target sample.

As shown in FIG. 18, when the distortion correction value generator 33 and the adder 36 correct the input digital audio signal with the second distortion correction pattern, the sample values of each sample from sample S3 to sample S8 (maximum sample Smax) are corrected to increase, and the sample values of each sample from sample S1 to sample S2 are corrected to decrease. When the distortion correction value generator 33 and the adder 36 correct the input digital audio signal with the second distortion correction pattern, the sample values of each sample from sample S9 to sample S13 are corrected to increase, and the sample values of each sample from sample S14 to sample S16 (minimum sample Smin) are corrected to decrease.

Figure 19 shows how each sample is corrected by the maximum/minimum sample correction value generator 31, the surrounding sample correction value generator 32, the distortion correction value generator 33, and the adders 34 to 36 when the second distortion correction pattern is selected.

The second distortion correction pattern is expected to reduce waveform distortion and restore the harmonic components of the original sound, as well as increase the overall sense of volume of the song, making the song feel more powerful.

The third distortion correction pattern will be explained using Figures 20 and 21. Figures 20 and 21 also show an example in which the extreme value interval is 8 fs and all of the first to eighth distortion correction target samples are distortion correction target samples. In explaining the third distortion correction pattern, explanations that overlap with the explanation of the first distortion correction pattern may be omitted.

The distortion correction value generating unit 33 multiplies the difference between each sample of the first distortion correction target sample and the sample immediately following each sample by a distortion correction coefficient to generate a first distortion correction value. The distortion correction value generating unit 33 multiplies the difference between each sample of the second distortion correction target sample and the sample immediately following each sample by a distortion correction coefficient to generate a second distortion correction value. The distortion correction value generating unit 33 multiplies the difference between each sample of the third distortion correction target sample and the sample immediately following each sample by a distortion correction coefficient to generate a third distortion correction value. The distortion correction value generating unit 33 multiplies the difference between each sample of the fourth distortion correction target sample and the sample immediately preceding each sample by a distortion correction coefficient to generate a fourth distortion correction value.

The distortion correction value generating unit 33 multiplies the difference between each sample of the fifth distortion correction target sample and the sample immediately preceding each sample by a distortion correction coefficient to generate a fifth distortion correction value. The distortion correction value generating unit 33 multiplies the difference between each sample of the sixth distortion correction target sample and the sample immediately following each sample by a distortion correction coefficient to generate a sixth distortion correction value. The distortion correction value generating unit 33 multiplies the difference between the seventh distortion correction target sample and the sample immediately preceding it by the distortion correction coefficient to generate a seventh distortion correction value, and multiplies the difference between the eighth distortion correction target sample and the sample immediately preceding it by the distortion correction coefficient to generate an eighth distortion correction value.

The adder 36 adds the first and third distortion compensation values to the first and third distortion compensation target samples, respectively, and subtracts the fourth distortion compensation value from the fourth distortion compensation target sample. The adder 36 subtracts the second and sixth distortion compensation values from the second and sixth distortion compensation target samples, respectively, and adds the fifth distortion compensation value to the fifth distortion compensation target sample. The adder 36 adds the seventh distortion compensation value to the seventh distortion compensation target sample, and subtracts the eighth distortion compensation value from the eighth distortion compensation target sample.

As shown in FIG. 20, when the distortion correction value generator 33 and the adder 36 correct the input digital audio signal with the third distortion correction pattern, the sample values of each sample from sample S3 to sample S8 (maximum sample Smax) are corrected to increase, and the sample values of each sample from sample S1 to sample S2 are corrected to decrease. When the distortion correction value generator 33 and the adder 36 correct the input digital audio signal with the third distortion correction pattern, the sample values of each sample from sample S9 to sample S10 are corrected to increase, and the sample values of each sample from sample S11 to sample S16 (minimum sample Smin) are corrected to decrease.

Figure 21 shows how each sample is corrected by the maximum/minimum sample correction value generator 31, the surrounding sample correction value generator 32, the distortion correction value generator 33, and the adders 34 to 36 when the third distortion correction pattern is selected.

The third distortion correction pattern is expected to reduce waveform distortion, restore the harmonic components of the original sound, and extend the overtones, making the entire song sound clearer.

The fourth distortion correction pattern will be explained using Figures 22 to 25. Figures 22 and 23 show an example in which the extreme value interval is 7 fs, and Figures 24 and 25 show an example in which the extreme value interval is 8 fs, and all of the first to eighth distortion correction target samples are distortion correction target samples. In explaining the fourth distortion correction pattern, explanations that overlap with the explanation of the first distortion correction pattern may be omitted.

In the extreme value interval 7 fs shown in FIG. 22 and FIG. 23, the number of samples of the first distortion correction target samples is an even number, and the number of samples of the second distortion correction target samples is an even number. The distortion correction value generating unit 33 multiplies the difference between each sample of the half number of samples on the first peripheral sample side of the first distortion correction target samples and the sample immediately following each sample by a distortion correction coefficient to generate a first distortion correction value for the half number of samples on the first peripheral sample side. Also, the distortion correction value generating unit 33 multiplies the difference between each sample of the half number of samples on the fourth peripheral sample side of the first distortion correction target samples and the sample immediately preceding each sample by a distortion correction coefficient to generate a first distortion correction value for the half number of samples on the fourth peripheral sample side.

The distortion correction value generating unit 33 multiplies the difference between each sample of the half number of samples on the second peripheral sample side of the second distortion correction target samples and the sample immediately preceding each sample by a distortion correction coefficient to generate a second distortion correction value for the half number of samples on the second peripheral sample side. The distortion correction value generating unit 33 also multiplies the difference between each sample of the half number of samples on the third peripheral sample side of the second distortion correction target samples and the sample immediately following each sample by a distortion correction coefficient to generate a second distortion correction value for the half number of samples on the third peripheral sample side.

The distortion correction value generating unit 33 multiplies the difference between each sample of the third distortion correction target sample and the sample immediately following each sample by a distortion correction coefficient to generate a third distortion correction value, and multiplies the difference between each sample of the fourth distortion correction target sample and the sample immediately preceding each sample by a distortion correction coefficient to generate a fourth distortion correction value. The distortion correction value generating unit 33 multiplies the difference between each sample of the fifth distortion correction target sample and the sample immediately preceding each sample by a distortion correction coefficient to generate a fifth distortion correction value, and multiplies the difference between each sample of the sixth distortion correction target sample and the sample immediately following each sample by a distortion correction coefficient to generate a sixth distortion correction value.

The distortion correction value generation unit 33 multiplies the difference between the seventh distortion correction target sample and the sample immediately preceding it by a distortion correction coefficient to generate a seventh distortion correction value, and multiplies the difference between the eighth distortion correction target sample and the sample immediately preceding it by a distortion correction coefficient to generate an eighth distortion correction value.

The adder 36 adds a first distortion compensation value to each sample of half the number of samples on the first peripheral sample side, and subtracts the first distortion compensation value from each sample of half the number of samples on the fourth peripheral sample side. The adder 36 adds a second distortion compensation value to each sample of half the number of samples on the second peripheral sample side, and subtracts the second distortion compensation value from each sample of half the number of samples on the third peripheral sample side.

The adder 36 adds the third distortion compensation value to the third distortion compensation target sample and subtracts the fourth distortion compensation value from the fourth distortion compensation target sample. The adder 36 adds the fifth distortion compensation value to the fifth distortion compensation target sample and subtracts the sixth distortion compensation value from the sixth distortion compensation target sample. The adder 36 adds the seventh distortion compensation value to the seventh distortion compensation target sample and subtracts the eighth distortion compensation value from the eighth distortion compensation target sample.

Figure 23 shows how each sample is corrected by the maximum/minimum sample correction value generator 31, the surrounding sample correction value generator 32, the distortion correction value generator 33, and the adders 34 to 36 when the fourth distortion correction pattern is selected with an extreme value interval of 7 fs.

In the extreme value interval 8 fs shown in FIG. 24 and FIG. 25, the number of samples of the first distortion correction target sample is an odd number, and the number of samples of the second distortion correction target sample is an odd number. The distortion correction value generating unit 33 multiplies the difference between each sample of the samples located between the first central sample located in the center of the first distortion correction target sample and the first peripheral sample and the sample immediately following each sample by a distortion correction coefficient to generate a first distortion correction value for the sample on the first peripheral sample side. Also, the distortion correction value generating unit 33 multiplies the difference between each sample of the samples located between the first central sample and the fourth peripheral sample and the sample immediately preceding each sample by a distortion correction coefficient to generate a first distortion correction value for the sample on the fourth peripheral sample side.

The distortion correction value generating unit 33 multiplies the difference between each sample of the samples located between the second central sample located in the center of the second distortion correction target sample and the second peripheral sample and the sample immediately preceding each sample by a distortion correction coefficient to generate a second distortion correction value for the samples on the second peripheral sample side. The distortion correction value generating unit 33 multiplies the difference between each sample of the samples located between the second central sample and the third peripheral sample and the sample immediately following each sample by a distortion correction coefficient to generate a second distortion correction value for the samples on the third peripheral sample side.

The distortion correction value generator 33 generates the third to eighth distortion correction values in the same manner as when the number of samples in the first and second distortion correction target samples is an even number.

The adder 36 adds a first distortion compensation value to each sample located between the first central sample and the first peripheral sample, and subtracts the first distortion compensation value from each sample located between the first central sample and the fourth peripheral sample. The adder 36 adds a second distortion compensation value to each sample located between the second central sample and the second peripheral sample, and subtracts the second distortion compensation value from each sample located between the second central sample and the third peripheral sample. The adder 36 adds/subtracts the third to eighth distortion compensation values to/from the third to eighth distortion compensation target samples, in the same way as when the number of samples in the first and second distortion compensation target samples is an even number.

Figure 25 shows how each sample is corrected by the maximum/minimum sample correction value generator 31, the surrounding sample correction value generator 32, the distortion correction value generator 33, and the adders 34 to 36 when the fourth distortion correction pattern is selected with an extreme value interval of 8 fs. As can be seen from Figures 24 and 25, when the number of samples in the first and second distortion correction target samples is odd, no distortion correction value is added or subtracted from the first and second central samples, and the original sample values are maintained.

The fourth distortion correction pattern is expected to reduce waveform distortion, restore the harmonic components of the original sound, and make the entire song sound more cohesive.

Figure 22 shows a case where the rising sample section in which the sample value rises from the minimum sample Smin to the maximum sample Smax, and the falling sample section in which the sample value falls from the maximum sample Smax to the minimum sample Smin, both have an extreme value interval of 7 fs. Therefore, the number of samples in the first and second distortion correction target samples is an even number. Figure 24 shows a case where the rising sample section in which the sample value rises from the minimum sample Smin to the maximum sample Smax, and the falling sample section in which the sample value falls from the maximum sample Smax to the minimum sample Smin, both have an extreme value interval of 8 fs. Therefore, the number of samples in the first and second distortion correction target samples is an odd number.

There are cases where the number of samples in the first distortion correction target sample is even and the number of samples in the second distortion correction target sample is odd, and cases where the number of samples in the first distortion correction target sample is odd and the number of samples in the second distortion correction target sample is even. Even in such cases, the distortion correction value generation unit 33 generates the first distortion correction value as described above depending on whether the number of samples in the first distortion correction target sample is even or odd, and generates the second distortion correction value as described above depending on whether the number of samples in the second distortion correction target sample is even or odd.

The adder 36 adds or subtracts the first distortion compensation value to the first distortion compensation target sample as described above depending on whether the number of samples in the first distortion compensation target sample is even or odd. The adder 36 adds or subtracts the second distortion compensation value to the second distortion compensation target sample as described above depending on whether the number of samples in the second distortion compensation target sample is even or odd.

In the digital audio processing device 100 described above, the maximum/minimum sample correction value generating unit 31 and the adder 34 in the audio waveform correction unit 3 may be omitted, and the audio waveform correction unit 3 may be configured to have an addition/subtraction unit 37 consisting of a peripheral sample correction value generating unit 32, a distortion correction value generating unit 33, and adders 35 and 36.

The surrounding sample correction value generation unit 32 generates first to fourth surrounding sample correction values. The distortion correction value generation unit 33 multiplies the difference between each sample of the first distortion correction target sample and a sample adjacent to each sample by a distortion correction coefficient to generate a first distortion correction value for correcting the first distortion correction target sample. The distortion correction value generation unit 33 multiplies the difference between each sample of the second distortion correction target sample and a sample adjacent to each sample by a distortion correction coefficient to generate a second distortion correction value for correcting the second distortion correction target sample. The distortion correction value generation unit 33 generates at least the first and second distortion correction values.

The addition/subtraction unit 37 adds the first peripheral sample correction value to each sample of the first peripheral sample, and adds the second peripheral sample correction value to each sample of the second peripheral sample. The addition/subtraction unit 37 subtracts the third peripheral sample correction value from each sample of the third peripheral sample, and subtracts the fourth peripheral sample correction value from each sample of the fourth peripheral sample. The addition/subtraction unit 37 adds the first distortion correction value to the first distortion correction target sample or subtracts the first distortion correction value from the first distortion correction target sample, and adds the second distortion correction value to the second distortion correction target sample or subtracts the second distortion correction value from the second distortion correction target sample.

In the digital audio processing device 100, the audio waveform correction unit 3 preferably has a maximum/minimum sample correction value generation unit 31, and the addition/subtraction unit 37 preferably has adders 34-36. The maximum/minimum sample correction value generation unit 31 generates maximum sample correction values and minimum sample correction values. The addition/subtraction unit 37 adds the maximum sample correction value to the maximum sample and subtracts the minimum sample correction value from the minimum sample.

It is preferable that the distortion correction value generation unit 33 generates third to sixth distortion correction values in addition to the first and second distortion correction values. It is preferable that the addition/subtraction unit 37 adds or subtracts the third to sixth distortion correction values to the third to sixth distortion correction target samples. The third to sixth distortion correction target samples are the first, fourth, second, and third surrounding samples, respectively.

It is preferable that the distortion correction value generation unit 33 generates a seventh and an eighth distortion correction value in addition to the first to sixth distortion correction values. It is preferable that the addition/subtraction unit 37 adds the seventh distortion correction value to the seventh distortion correction target sample and subtracts the eighth distortion correction value from the eighth distortion correction target sample. The seventh distortion correction target sample is a maximum sample, and the eighth distortion correction target sample is a minimum sample.

As described above, the digital audio processing device 100 executes the following digital audio processing method. The digital audio processing device 100 detects maximum and minimum samples in the samples that constitute the input digital audio signal. The digital audio processing device 100 generates first to fourth surrounding sample correction values and first and second distortion correction values.

The digital audio processing device 100 adds the first and second peripheral sample correction values to each sample of the first and second peripheral samples, respectively, and subtracts the third and fourth peripheral sample correction values from each sample of the third and fourth peripheral samples, respectively. The digital audio processing device 100 adds the first distortion correction value to the first distortion correction target sample or subtracts the first distortion correction value from the first distortion correction target sample, and adds the second distortion correction value to the second distortion correction target sample or subtracts the second distortion correction value from the second distortion correction target sample.

The digital audio processing device 100 shown in FIG. 1 can be realized by a central processing unit (CPU) of a microcomputer executing a digital audio processing program. In FIG. 26, the CPU 50, main memory 55, and storage medium 60 are connected by a bus. The storage medium 60 is any non-transitory storage medium such as a hard disk drive, an optical disk, or a semiconductor memory. The storage medium 60 stores a digital audio processing program. The digital audio processing program may be transmitted from an external server via a communication line such as the Internet and stored in the storage medium 60.

The CPU 50 loads the digital audio processing program stored in the storage medium 60 into the main memory 55. The CPU 50 executes each command written in the digital audio processing program loaded into the main memory 55, thereby performing the process shown in FIG. 27.

The flowchart shown in FIG. 27 will be used to explain the processing that the digital audio processing program according to one or more embodiments causes the CPU 50 to execute. When the processing starts, the CPU 50 analyzes the audio waveform of the input digital audio signal in step S01. Step S01 includes a step of detecting extremely maximum samples and extremely minimum samples in the samples that make up the input digital audio signal. In step S02, the CPU 50 determines the sample position to be corrected and selects a group of correction coefficients.

The CPU 50 generates maximum and minimum sample correction values in step S03, generates surrounding sample correction values in step S04, and generates a distortion correction value in step S05. Step S03 includes a step of generating a maximum sample correction value and a minimum sample correction value. Step S04 includes a step of generating first to fourth surrounding sample correction values. Step S05 includes a step of generating first and second distortion correction values.

In step S06, the CPU 50 adds or subtracts maximum and minimum sample correction values, peripheral sample correction values, and distortion correction values to the input digital audio signal. Step S06 includes a step of adding the first peripheral sample correction value to the first peripheral sample, and a step of adding the second peripheral sample correction value to the second peripheral sample. Step S06 includes a step of subtracting the third peripheral sample correction value from the third peripheral sample, and a step of subtracting the fourth peripheral sample correction value from the fourth peripheral sample. Step S06 includes a step of adding the first distortion correction value to the first distortion correction target sample or subtracting the first distortion correction value from the first distortion correction target sample, and a step of adding the second distortion correction value to the second distortion correction target sample or subtracting the second distortion correction value from the second distortion correction target sample.

In step S07, the CPU 50 determines whether the input of the digital audio signal has ended. If the input of the digital audio signal has not ended (NO), the CPU 50 repeats the processing of steps S01 to S07. If the input of the digital audio signal has ended (YES), the CPU 50 ends the processing.

The present invention is not limited to the embodiment described above, and various modifications are possible without departing from the spirit of the present invention.

REFERENCE SIGNS LIST 1 Audio waveform analysis section 2 Correction target sample/correction coefficient setting section 3 Audio waveform correction section 11 Sampling frequency acquisition section 12 Maximum/minimum sample detection section 13 Extreme value interval group selection section 21 Correction target sample position pattern selection section 22 Correction coefficient group selection section 31 Maximum/minimum sample correction value generation section 32 Surrounding sample correction value generation section 33 Distortion correction value generation section 34 to 36 Adder 37 Addition/subtraction section 50 Central processing unit (CPU)
55 Main memory 60 Storage medium 100 Digital audio processing device

Claims (8)

a maximum/minimum sample detector for detecting maximum and minimum samples in samples constituting an input digital audio signal;
One or more samples located before the maximum sample including the sample immediately before the maximum sample are defined as first peripheral samples to be corrected, one or more samples located after the maximum sample including the sample immediately after the maximum sample are defined as second peripheral samples to be corrected, one or more samples located before the minimum sample including the sample immediately before the minimum sample are defined as third peripheral samples to be corrected, and one or more samples located after the minimum sample including the sample immediately after the minimum sample are defined as fourth peripheral samples to be corrected, and a difference between each of the first peripheral samples and the sample immediately after each of the first peripheral samples is multiplied by a coefficient less than 1. a peripheral sample correction value generation unit that multiplies a difference between each of the second peripheral samples and the sample immediately preceding the second peripheral sample by a coefficient less than 1 to generate a first peripheral sample correction value for correcting the first peripheral sample, multiplies a difference between each of the second peripheral samples and the sample immediately preceding the second peripheral sample by a coefficient less than 1 to generate a third peripheral sample correction value for correcting the third peripheral sample, and multiplies a difference between each of the fourth peripheral samples and the sample immediately preceding the fourth peripheral sample by a coefficient less than 1 to generate a fourth peripheral sample correction value for correcting the fourth peripheral sample;
a distortion correction value generation unit that defines one or more samples that are not corrected by the first or fourth surrounding sample correction value in an upward sample section between the minimum sample located before the maximum sample and the minimum sample as first distortion correction target samples, and defines one or more samples that are not corrected by the second or third surrounding sample correction value in a downward sample section between the maximum sample and the minimum sample located after the maximum sample as second distortion correction target samples, multiplies a difference between each of the first distortion correction target samples and a sample adjacent to each sample by a coefficient less than 1 to generate a first distortion correction value for correcting the first distortion correction target sample, and multiplies a difference between each of the second distortion correction target samples and a sample adjacent to each sample by a coefficient less than 1 to generate a second distortion correction value for correcting the second distortion correction target sample;
an adding/subtracting unit that adds the first peripheral sample correction value to each sample of the first peripheral sample, adds the second peripheral sample correction value to each sample of the second peripheral sample, subtracts the third peripheral sample correction value from each sample of the third peripheral sample, subtracts the fourth peripheral sample correction value from each sample of the fourth peripheral sample, adds the first distortion correction value to the first distortion correction target sample or subtracts the first distortion correction value from the first distortion correction target sample, and adds the second distortion correction value to the second distortion correction target sample or subtracts the second distortion correction value from the second distortion correction target sample;
A digital audio processing device comprising: The distortion correction value generation unit
the first and fourth surrounding samples are respectively third and fourth distortion correction target samples, and the second and third surrounding samples are respectively fifth and sixth distortion correction target samples;
multiplying a difference between each sample of the first distortion correction target samples and a sample immediately following each sample by a coefficient less than 1 to generate the first distortion correction value;
multiplying a difference between each sample of the second distortion correction target samples and a sample immediately following each sample by a coefficient less than 1 to generate the second distortion correction value;
multiplying a difference between each of the third to sixth distortion correction target samples and the sample immediately following each of the third to sixth distortion correction samples by a coefficient less than 1 to generate third to sixth distortion correction values, respectively;
The addition/subtraction unit is
adding the first, third, and fourth distortion compensation values to the first, third, and fourth distortion compensation target samples, respectively;
2 . The digital audio processing device according to claim 1 , wherein the second, fifth, and sixth distortion compensation values are subtracted from the second, fifth, and sixth distortion compensation target samples, respectively. The distortion correction value generation unit
the first and fourth surrounding samples are respectively third and fourth distortion correction target samples, and the second and third surrounding samples are respectively fifth and sixth distortion correction target samples;
multiplying a difference between each sample of the first distortion correction target samples and a sample immediately following each sample by a coefficient less than 1 to generate the first distortion correction value;
multiplying a difference between each sample of the second distortion correction target samples and a sample immediately before the sample by a coefficient less than 1 to generate the second distortion correction value;
multiplying a difference between each sample of the third distortion correction target samples and a sample immediately following each sample by a coefficient less than 1 to generate a third distortion correction value;
multiplying a difference between each sample of the fourth distortion correction target samples and a sample immediately preceding each sample by a coefficient less than 1 to generate a fourth distortion correction value;
multiplying a difference between each sample of the fifth distortion correction target samples and a sample immediately preceding each sample by a coefficient less than 1 to generate a fifth distortion correction value;
multiplying a difference between each sample of the sixth distortion correction target samples and a sample immediately following each sample by a coefficient less than 1 to generate a sixth distortion correction value;
The addition/subtraction unit is
adding the first and third distortion compensation values to the first and third distortion compensation target samples, respectively;
subtracting the fourth distortion correction value from the fourth distortion correction target sample;
adding the second and fifth distortion compensation values to the second and fifth distortion compensation target samples, respectively;
The digital audio processing device according to claim 1 , wherein the sixth distortion compensation value is subtracted from the sixth distortion compensation target sample. The distortion correction value generation unit
the first and fourth surrounding samples are respectively third and fourth distortion correction target samples, and the second and third surrounding samples are respectively fifth and sixth distortion correction target samples;
multiplying a difference between each sample of the first distortion correction target samples and a sample immediately following each sample by a coefficient less than 1 to generate the first distortion correction value;
multiplying a difference between each sample of the second distortion correction target samples and a sample immediately following each sample by a coefficient less than 1 to generate the second distortion correction value;
multiplying a difference between each sample of the third distortion correction target samples and a sample immediately following each sample by a coefficient less than 1 to generate a third distortion correction value;
multiplying a difference between each sample of the fourth distortion correction target samples and a sample immediately preceding each sample by a coefficient less than 1 to generate a fourth distortion correction value;
multiplying a difference between each sample of the fifth distortion correction target samples and a sample immediately preceding each sample by a coefficient less than 1 to generate a fifth distortion correction value;
multiplying a difference between each sample of the sixth distortion correction target samples and a sample immediately following each sample by a coefficient less than 1 to generate a sixth distortion correction value;
The addition/subtraction unit is
adding the first and third distortion compensation values to the first and third distortion compensation target samples, respectively;
subtracting the fourth distortion correction value from the fourth distortion correction target sample;
subtracting the second and sixth distortion compensation values from the second and sixth distortion compensation target samples, respectively;
The digital audio processing device according to claim 1 , further comprising: adding the fifth distortion compensation value to the fifth distortion compensation target sample. The distortion correction value generation unit
the first and fourth surrounding samples are respectively third and fourth distortion correction target samples, and the second and third surrounding samples are respectively fifth and sixth distortion correction target samples;
when the number of samples of the first distortion correction target samples is an even number, multiplying a difference between each sample of half the number of samples on the first peripheral sample side among the first distortion correction target samples and a sample immediately following each sample by a coefficient less than 1, and multiplying a difference between each sample of half the number of samples on the fourth peripheral sample side among the first distortion correction target samples and a sample immediately preceding each sample by a coefficient less than 1 to generate the first distortion correction value;
when the number of samples of the first distortion correction target samples is odd, multiplying a difference between each sample located between a first central sample located at the center of the first distortion correction target samples and the first peripheral sample and a sample immediately following each sample by a coefficient less than 1, and multiplying a difference between each sample located between the first central sample and the fourth peripheral sample and a sample immediately preceding each sample by a coefficient less than 1 to generate the first distortion correction value;
when the number of samples of the second distortion correction target samples is an even number, multiplying a difference between each sample of half the number of samples on the second peripheral sample side among the second distortion correction target samples and a sample immediately preceding each sample by a coefficient less than 1, and multiplying a difference between each sample of half the number of samples on the third peripheral sample side and a sample immediately following each sample by a coefficient less than 1 to generate the second distortion correction value;
when the number of samples of the second distortion correction target samples is odd, multiplying a difference between each sample located between a second central sample located at the center of the second distortion correction target samples and the second peripheral sample and a sample immediately preceding each sample by a coefficient less than 1, and multiplying a difference between each sample located between the second central sample and the third peripheral sample and a sample immediately following each sample by a coefficient less than 1 to generate the second distortion correction value;
multiplying a difference between each sample of the third distortion correction target samples and a sample immediately following each sample by a coefficient less than 1 to generate a third distortion correction value;
multiplying a difference between each sample of the fourth distortion correction target samples and a sample immediately preceding each sample by a coefficient less than 1 to generate a fourth distortion correction value;
multiplying a difference between each sample of the fifth distortion correction target samples and a sample immediately preceding each sample by a coefficient less than 1 to generate a fifth distortion correction value;
multiplying a difference between each sample of the sixth distortion correction target samples and a sample immediately following each sample by a coefficient less than 1 to generate a sixth distortion correction value;
The addition/subtraction unit is
when the number of samples of the first distortion correction target samples is an even number, adding the first distortion correction value to each sample of half the number of samples on the first peripheral sample side, and subtracting the first distortion correction value from each sample of half the number of samples on the fourth peripheral sample side;
when the number of samples of the first distortion correction target samples is odd, adding the first distortion correction value to each sample located between the first central sample and the first peripheral sample, and subtracting the first distortion correction value from each sample located between the first central sample and the fourth peripheral sample;
when the number of samples of the second distortion correction target samples is an even number, adding the second distortion correction value to each sample of half the number of samples on the second peripheral sample side, and subtracting the second distortion correction value from each sample of half the number of samples on the third peripheral sample side;
when the number of samples to be subjected to the second distortion correction is an odd number, adding the second distortion correction value to each sample located between the second central sample and the second peripheral sample, and subtracting the second distortion correction value from each sample located between the second central sample and the third peripheral sample;
adding the third distortion correction value to the third distortion correction target sample;
subtracting the fourth distortion correction value from the fourth distortion correction target sample;
adding the fifth distortion correction value to the fifth distortion correction target sample;
The digital audio processing device according to claim 1 , wherein the sixth distortion compensation value is subtracted from the sixth distortion compensation target sample. a maximum/minimum sample correction value generation unit that multiplies a difference between the maximum sample and a sample immediately preceding the maximum sample by a coefficient less than 1 to generate a maximum sample correction value for correcting the maximum sample, and multiplies a difference between the minimum sample and a sample immediately preceding the minimum sample by a coefficient less than 1 to generate a minimum sample correction value for correcting the minimum sample,
The addition/subtraction unit is
adding the maximum sample correction value to the maximum sample;
6. The digital audio processing device according to claim 1, wherein the minimal sample correction value is subtracted from the minimal sample. a maximum/minimum sample correction value generation unit that multiplies a difference between the maximum sample and a sample immediately preceding the maximum sample by a coefficient less than 1 to generate a maximum sample correction value for correcting the maximum sample, and multiplies a difference between the minimum sample and a sample immediately preceding the minimum sample by a coefficient less than 1 to generate a minimum sample correction value for correcting the minimum sample,
The distortion correction value generation unit
The maximum sample is a seventh sample to be corrected for distortion, and the minimum sample is an eighth sample to be corrected for distortion,
multiplying a difference between the maximum sample and a sample immediately preceding the maximum sample by a coefficient less than 1 to generate a seventh distortion correction value;
multiplying a difference between the minimum sample and a sample immediately preceding the minimum sample by a coefficient less than 1 to generate an eighth distortion correction value;
The addition/subtraction unit is
adding the seventh distortion correction value to the seventh distortion correction target sample;
6. The digital audio processing device according to claim 2, further comprising: subtracting the eighth distortion compensation value from the eighth distortion compensation target sample. Detecting maximum and minimum samples in the samples constituting the input digital audio signal;
one or more samples located before the maximum sample including the sample immediately before the maximum sample are defined as first peripheral samples to be corrected, one or more samples located after the maximum sample including the sample immediately after the maximum sample are defined as second peripheral samples to be corrected, one or more samples located before the minimum sample including the sample immediately before the minimum sample are defined as third peripheral samples to be corrected, and one or more samples located after the minimum sample including the sample immediately after the minimum sample are defined as fourth peripheral samples to be corrected,
multiplying a difference between each of the first surrounding samples and a sample immediately following the first surrounding sample by a coefficient less than 1 to generate a first surrounding sample correction value for correcting the first surrounding sample;
multiplying a difference between each of the second surrounding samples and a sample immediately preceding the second surrounding sample by a coefficient less than 1 to generate a second surrounding sample correction value for correcting the second surrounding sample;
multiplying a difference between each of the third neighboring samples and a sample immediately following the third neighboring sample by a coefficient less than 1 to generate a third neighboring sample correction value for correcting the third neighboring sample;
multiplying a difference between each of the fourth surrounding samples and a sample immediately preceding the fourth surrounding sample by a coefficient less than 1 to generate a fourth surrounding sample correction value for correcting the fourth surrounding sample;
one or more samples that are not corrected by the first or fourth surrounding sample correction value in an ascending sample section between the minimum sample and the maximum sample located before the maximum sample are defined as first distortion correction target samples, and one or more samples that are not corrected by the second or third surrounding sample correction value in a descending sample section between the maximum sample and the minimum sample located after the maximum sample are defined as second distortion correction target samples,
multiplying a difference between each sample of the first distortion correction target sample and a sample adjacent to the sample by a coefficient less than 1 to generate a first distortion correction value for correcting the first distortion correction target sample;
multiplying a difference between each sample of the second distortion correction target samples and a sample adjacent to the sample by a coefficient less than 1 to generate a second distortion correction value for correcting the second distortion correction target samples;
adding the first surrounding sample correction value to each sample of the first surrounding samples;
adding the second surrounding sample correction value to each sample of the second surrounding samples;
subtracting the third surrounding sample correction value from each sample of the third surrounding samples;
subtracting the fourth surrounding sample correction value from each sample of the fourth surrounding sample;
adding the first distortion compensation value to the first distortion compensation target sample or subtracting the first distortion compensation value from the first distortion compensation target sample;
a second distortion compensation value being added to the second distortion compensation target sample or a second distortion compensation value being subtracted from the second distortion compensation target sample.

Images