JP2015032933A - Low-frequency compensation device and low-frequency compensation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To present bass enhancement effects to a maximum while suppressing sound quality deterioration in middle and low sound ranges by weighting a harmonic tone signal in accordance with a level or a fluctuation amount of a low-frequency signal.SOLUTION: A low-frequency compensation device comprises: band dividing means 2 and 3 for band-dividing an input signal into a low-frequency signal and a high-frequency signal; harmonic tone signal generation means 4 for generating a first harmonic tone signal on the basis of the low-frequency signal; level detection signal generation means 5 for generating a level detection signal on the basis of a signal level of the low-frequency signal; level fluctuation amount detection means 6 for generating a level fluctuation amount detection signal indicating a level fluctuation amount of the low-frequency signal; multiplication means 7 for multiplying the first harmonic tone signal by the level detection signal and the level fluctuation amount detection signal to generate a second harmonic tone signal; and addition means 13 for adding the high-frequency signal, the low-frequency signal and the second harmonic tone signal to generate an output signal.

Description

  The present invention relates to a low-frequency complement device and a low-frequency complement method, and more specifically, by generating a harmonic signal based on a low-frequency signal, and adding the generated harmonic signal to a high-frequency signal and a low-frequency signal, The present invention relates to a low-frequency interpolation apparatus and a low-frequency interpolation method characterized by performing low-frequency interpolation processing.

  In a general speaker, the reproducible frequency band differs depending on the performance of the speaker. For this reason, the sound quality of the audio signal output from the speaker is greatly influenced by the frequency band that can be reproduced by the speaker. For example, when music is played back with a speaker that is inferior in low-frequency playback capability, low-frequency audio signals are not sufficiently played back, so there is a tendency that the low-frequency power is lacking and the sound is not sharp. In addition, the frequency of the low sound reproduction limit in the speaker is referred to as the lowest resonance frequency.

  Nowadays, as a technique for complementing bass (basis below the lowest resonance frequency) that cannot be sufficiently reproduced by a speaker, a complementation method using missing psychological psychology is known. In the complementary method using missing fundamentals, the harmonic component of the target frequency is added to the audio signal and output by the speaker, so that the sound of that frequency is not actually included in the output sound of the speaker. Even if it exists, a listener can be made to perceive that the sound of the frequency is included.

  As an example of a complementary method using missing fundamentals, a low frequency signal is extracted from a sound source signal output from a sound source to generate a harmonic signal, and the harmonic signal level is weighted according to the volume setting of the playback device. In the above, a method has been proposed in which a harmonic signal is added to a sound source signal and output by a speaker (see Patent Document 1).

JP 2012-147179 A

  However, when the harmonic signal is weighted according to the volume and then added to the sound source signal, the low frequency interpolation is performed by the harmonic signal, while depending on the effect (weighting amount) of the complement, There was a problem of causing deterioration.

  Specifically, the mid-bass may be unnecessarily enhanced with the addition of a harmonic signal. For this reason, compared to the low frequency sound reproduced by a loudspeaker capable of reproducing heavy bass, the low frequency of the output sound that has been supplemented by the low frequency with the harmonic signal is inferior in output sound responsiveness and is not tightened. -There was a tendency to be perceived as an extended sound.

  On the other hand, if the overtone signal is not complemented, the output sound in the frequency band that cannot be reproduced by the speaker is not reproduced, so that there is a possibility that the sharpness of the sound is lost as described above. For this reason, the low-frequency interpolation method using the harmonic signal is one of effective interpolation methods. Therefore, in the complementary processing using the harmonic signal, there has been a demand for a method for suppressing deterioration in sound quality while realizing the bass enhancement effect by the harmonic signal.

  The present invention has been made in view of the above-described problems, and a low-frequency complementing device capable of complementing low frequencies below the lowest resonance frequency that cannot be sufficiently reproduced by a speaker using a harmonic signal, and A low-frequency interpolation method that weights overtone signals according to the level and amount of fluctuation of the low-frequency signal contained in the sound source, thereby maximizing the bass enhancement effect while suppressing sound quality degradation in the mid-low range. It is an object of the present invention to provide a low-frequency complement device and a low-frequency complement method that can be exhibited.

  In order to solve the above-described problems, a low-frequency complement apparatus according to the present invention includes a band dividing unit that divides an input signal into a low-frequency signal and a high-frequency signal, and the low-frequency band that is divided by the band dividing unit. Harmonic signal generating means for generating a first harmonic signal having a predetermined signal level output at a frequency several times the predetermined low frequency with reference to a predetermined low frequency of the signal, and a band by the band dividing means The absolute value of the signal level of the divided low-frequency signal is calculated and then the DC component is cut to perform level detection signal generation means for generating a level detection signal, and the band divided by the band dividing means Level fluctuation amount detecting means for generating a level fluctuation amount detection signal indicating the level fluctuation amount of the low frequency band signal based on the signal level fluctuation of the low frequency band signal, and the level overtone signal with respect to the first harmonic signal. Multiplication means for generating a second overtone signal by multiplying the level detection signal and the level variation detection signal, the high frequency signal and the low frequency signal band-divided by the band division means, and the multiplication means And adding means for adding the second overtone signal generated in step (1) to generate an output signal.

  Further, the low-frequency interpolation method according to the present invention includes a band dividing step in which the band dividing unit divides the input signal into a low-frequency signal and a high-frequency signal, and the low-frequency signal divided in the band dividing step. A harmonic signal generation step in which the harmonic signal generation means generates a first harmonic signal having a predetermined signal level output at a frequency several times the predetermined low frequency with reference to a predetermined low frequency, and the band division A level detection signal generation step in which a level detection signal generation means generates a level detection signal by performing a DC component cut process after calculating the absolute value of the signal level of the low-frequency signal divided in the step; Based on the signal level fluctuation of the low frequency signal band-divided in the band division step, the level fluctuation amount detecting means is configured to detect the level fluctuation amount of the low frequency signal. A level fluctuation amount detection step for generating a level fluctuation amount detection signal to be indicated; and by multiplying the first harmonic signal by the level detection signal and the level fluctuation amount detection signal, the multiplying means outputs the second harmonic signal. A multiplying step to be generated; the high-frequency signal and the low-frequency signal band-divided in the band-dividing step; and the second overtone signal generated in the multiplying step are added, and an adding unit outputs an output signal. And an adding step to be generated.

  In the low-frequency complement apparatus or low-frequency complement method according to the present invention, not only the level detection signal generated based on the absolute value of the signal level of the low-frequency signal but also the low-frequency range for the generated first harmonic signal. The second overtone signal is generated by multiplying the level variation detection signal generated based on the signal level variation of the signal. Therefore, not only the signal level of the harmonic signal is adjusted according to the value of the signal level of the low frequency signal, but also the signal level of the harmonic signal is adjusted according to the fluctuation amount of the signal level of the low frequency signal. Will be.

  Therefore, according to the low-frequency interpolation apparatus or the low-frequency interpolation method according to the present invention, when the level fluctuation amount of the low-frequency signal is large, it is possible to improve the low-frequency range complementation effect by increasing the weighting specific gravity. On the other hand, if the level fluctuation amount of the low-frequency signal is small, it is possible to prevent unnecessary supplementary processing from being performed in the mid-high range by the signal output of the overtone signal by weakening the specific gravity of the weighting, It becomes possible to reduce deterioration of sound quality.

  Further, in the above-described low-frequency complementing device, the level fluctuation amount detecting means is detected by the maximum value detecting means for detecting the maximum value of the low-frequency signal band-divided by the band dividing means, and the maximum value detecting means. A maximum value hold means for generating a maximum value hold signal by holding the maximum value for a predetermined time, and performing a differentiation process on the maximum value hold signal generated by the maximum value hold means, And a differential processing means for generating the level fluctuation amount detection signal indicating the level fluctuation amount of the low frequency signal.

  Further, in the low-frequency interpolation method described above, in the level fluctuation amount detection step, a maximum value detection step in which a maximum value detection unit detects a maximum value of the low-frequency signal band-divided in the band division step, A maximum value hold step in which the maximum value hold means generates a maximum value hold signal by holding the maximum value detected in the maximum value detection step for a predetermined time, and the maximum value hold generated in the maximum value hold step. The differential processing means may include a differential processing step of generating the level fluctuation amount detection signal indicating the level fluctuation amount of the low-frequency signal by performing differential processing on the signal.

  In the level fluctuation detection means of the low frequency complementing apparatus or the level fluctuation detecting step of the low frequency complementing method according to the present invention, the signal level is detected by detecting the maximum value of the low frequency signal and holding the maximum value. It is possible to detect the fluctuation state of the maximum value of. By holding the maximum value in this manner, the relative change state of the signal level can be obtained as the maximum value hold signal even when the signal level of the low-frequency signal is obtained by the amplitude value.

  Furthermore, by differentiating the maximum value hold signal obtained in this way, the level fluctuation amount of the low frequency signal can be obtained as a level fluctuation amount detection signal, and this level fluctuation amount detection signal is as described above. In addition, by multiplying the first harmonic signal together with the level detection signal, the signal level of the first harmonic signal can be adjusted according to the amount of fluctuation of the signal level of the low frequency signal.

  Further, in the above-described low-frequency complementing device, the level fluctuation amount detection unit includes a gain adjustment unit that performs gain adjustment so that a signal level of the level fluctuation amount detection signal is higher than zero. The multiplying unit may multiply the first harmonic signal by the level detection signal and a level variation detection signal whose gain is adjusted by the gain adjusting unit.

  Further, in the above-described low-frequency interpolation method, the level variation detection step is a gain adjustment step in which gain adjustment means performs gain adjustment so that a signal level of the level variation detection signal is higher than zero. And in the multiplication step, the multiplication means multiplies the first harmonic signal by the level detection signal and the level variation detection signal that has been gain-adjusted in the gain adjustment step. There may be.

  The level fluctuation amount detection signal generated in the above-described low-frequency complementing apparatus and low-frequency complementing method according to the present invention is a signal indicating the level fluctuation amount of the low-frequency signal, and is zero when the signal level does not fluctuate. It becomes the value of (0). Therefore, even if the signal level of the low frequency signal is large or small, if the signal level does not fluctuate, the low frequency signal is not supplemented by the second overtone signal, and overall compensation is performed. The effect is reduced.

  For this reason, even if the signal level of the low-frequency signal does not fluctuate by performing gain adjustment on the level fluctuation amount detection signal so that the signal level becomes a value higher than zero (0), It is possible to achieve low-frequency complementation by using a harmonic signal, and it is possible to maintain a steady minimum low-frequency complementation effect. Furthermore, if there is a significant change in the signal level of the low frequency signal, it will be possible to achieve a greater complementary effect with the second overtone signal, thus realizing both the bass enhancement effect and the suppression of sound quality degradation. It becomes possible to do.

  Further, the low-frequency complementing device described above performs a filtering process using a high-pass filter on the second harmonic signal generated by the multiplication unit, so that the frequency of the second harmonic signal increases as the frequency increases. Filter means for gradually reducing the signal level output, wherein the adding means adds the high-frequency signal, the low-frequency signal, and the second overtone signal filtered by the filter means, and outputs the output signal. May be generated.

  Furthermore, the low-frequency interpolation method described above performs filtering using a high-pass filter on the second harmonic signal generated in the multiplication step, so that the filter means increases the frequency as the frequency increases. A filtering process step for gradually reducing the signal level output of the second harmonic signal, wherein the adding means is filtered in the high-frequency signal, the low-frequency signal, and the filtering process step; The output signal may be generated by adding the second overtone signal.

  The harmonic signal generated in the low frequency complementing apparatus and low frequency complementing method according to the present invention has a predetermined signal level output at a frequency several times higher than a predetermined low frequency of the low frequency signal. For this reason, the complementary effect by the second harmonic signal becomes too strong at a high frequency where the multiple becomes large, and there is a possibility that the listener may feel uncomfortable in terms of hearing.

  Therefore, it is possible to prevent the listener from giving a sense of incongruity by performing the filtering process so as to gradually reduce the signal level output of the second overtone signal as the frequency increases. .

  In the low-frequency complement apparatus or low-frequency complement method according to the present invention, not only the level detection signal generated based on the absolute value of the signal level of the low-frequency signal but also the low-frequency range for the generated first harmonic signal. The second overtone signal is generated by multiplying the level variation detection signal generated based on the signal level variation of the signal. Therefore, not only the signal level of the harmonic signal is adjusted according to the value of the signal level of the low frequency signal, but also the signal level of the harmonic signal is adjusted according to the fluctuation amount of the signal level of the low frequency signal. Will be.

  Therefore, according to the low-frequency interpolation apparatus or the low-frequency interpolation method according to the present invention, when the level fluctuation amount of the low-frequency signal is large, it is possible to improve the low-frequency range complementation effect by increasing the weighting specific gravity. On the other hand, if the level fluctuation amount of the low-frequency signal is small, it is possible to prevent unnecessary supplementary processing from being performed in the mid-high range by the signal output of the overtone signal by weakening the specific gravity of the weighting, It becomes possible to reduce deterioration of sound quality.

It is the block diagram which showed schematic structure of the low frequency complementation apparatus which concerns on embodiment. It is the flowchart which showed the outline of the processing content in the low frequency complementation apparatus which concerns on embodiment. (A) shows the filter characteristics of the high pass filter of the first HPF unit and the low pass filter of the first LPF unit, and (b) shows the low pass filter of the second HPF unit and the low pass filter of the second LPF unit. It is the figure which showed the filter characteristic with a band pass filter. (A) is the block diagram which showed schematic structure of the level detection part. (B) is a block diagram showing a schematic configuration of a level variation detection unit. It is the flowchart which showed the process of each function part in a level fluctuation amount detection part. It is the 1st figure which showed the operation example which detects the level fluctuation amount of a low frequency signal in a level fluctuation amount detection part for every functional part. It is the 2nd figure which showed the operation example which detects the level fluctuation amount of a low frequency signal in a level fluctuation amount detection part for every functional part. It is the 3rd figure which showed the operation example which detects the level fluctuation amount of a low frequency signal in a level fluctuation amount detection part for every functional part. (A) is the figure which showed the low frequency signal after passing the 1st LPF part, and the level detection signal produced | generated by the level detection part, (b) is the harmonic signal produced | generated by the edge detection part. And a level detection signal generated by the level detection unit. FIG. 9A is a diagram showing a weighted harmonic signal obtained by multiplying the harmonic signal shown in FIG. 9B by the level detection signal shown in FIG. 9B in the multiplication unit. The frequency characteristics of the band signal are shown. (A) is the figure which showed the frequency characteristic of the weighted harmonic overtone signal shown to Fig.10 (a), (b) is the band restriction | limiting of a low region and a high region by the 2nd LPF part and the 2nd HPF part. It is the figure which showed the frequency characteristic of the overtone signal. (A) is a diagram illustrating an input signal, (b) is a diagram illustrating a low-frequency signal output from the first LPF unit, and (c) is a harmonic signal output from the edge detection unit. FIG. (A) is the figure which showed the level detection signal produced | generated in a level detection part, (b) and (c) were the figures which showed the level fluctuation amount detection signal produced | generated in a level fluctuation amount detection part. is there. (A)-(c) has shown the weighted harmonic signal. FIGS. 14A to 14C are weighted harmonic signals shown in FIGS. 14A to 14C, and the harmonic signals after the filter processing by the second LPF unit and the second HPF unit are performed. FIG.

  Hereinafter, an example of a low-frequency interpolation device according to the present invention will be shown and described in detail with reference to the drawings. FIG. 1 is a block diagram showing a schematic configuration of a low frequency complement device according to the present embodiment. FIG. 2 is a flowchart showing an outline of processing contents in the low-frequency complement device 1. The processing content of the low-frequency interpolation device 1 will be described according to the processing procedure of the flowchart.

  As shown in FIG. 1, the low-frequency interpolation device 1 includes a first HPF unit (band division unit) 2, a first LPF unit (band division unit) 3, an edge detection unit (overtone signal generation unit) 4, and a level detection unit. (Level detection signal generation means) 5, level fluctuation amount detection section (level fluctuation amount detection means) 6, multiplication section (multiplication means) 7, phase inversion section 8, first amplification section 9, and second LPF section (Filter means) 10, second HPF section 11, second amplification section 12, and addition section (addition means) 13.

  First, an input signal input from the sound source to the low frequency interpolation device 1 is input to the first HPF unit 2 and the first LPF unit 3. The first HPF unit 2 includes a high-pass filter, and the first LPF unit 3 includes a low-pass filter. The input signal input to the first HPF unit 2 passes through a high-pass filter, so that only the high band is extracted and a high-frequency signal is generated. Further, the input signal input to the first LPF unit 3 passes through the low-pass filter, so that only the low band is extracted and a low-frequency signal is generated.

  FIG. 3A is a diagram illustrating filter characteristics of the high pass filter of the first HPF unit 2 and the low pass filter of the first LPF unit 3. As an example, the high-pass filter and the low-pass filter shown in FIG. 3A are third-order Butterworth filters, and the cutoff frequency is 60 Hz. The input sound source signal is divided into a high frequency signal and a low frequency signal by the first HPF unit 2 and the first LPF unit 3 (step S.1 in FIG. 2).

  The high frequency signal output from the first HPF unit 2 is output to the adding unit 13. The low frequency signal output from the first LPF unit 3 is output to the second amplifying unit 12, the edge detecting unit 4, the level detecting unit 5, and the level fluctuation amount detecting unit 6, respectively.

  Next, in the low frequency interpolation device 1, a process of generating a harmonic signal in the edge detection unit 4, generating a level detection signal in the level detection unit 5, and generating a level variation detection signal in the level variation detection unit 6. (Step S.2 in FIG. 2). More detailed processing contents will be described next.

  The edge detection unit 4 has a role of obtaining an overtone signal (first overtone signal) by obtaining an impulse train from the input low-frequency signal. When the low frequency signal is input to the edge detection unit 4, the edge detection unit 4 detects the position (timing) at which the low frequency signal changes from the negative side to the positive side based on the amplitude change of the low frequency signal. An impulse train signal is generated based on the detection position (detection timing). The edge detection unit 4 according to the present embodiment generates an impulse train having an amplitude of 1. The impulse train generated by the edge detection unit 4 in this way is output to the multiplication unit 7 as a harmonic signal (first harmonic signal).

  FIG. 4A is a block diagram showing a schematic configuration of the level detection unit 5. As shown in FIG. 4A, the level detection unit 5 includes an absolute value detection unit 21 and a DC cut unit 22. The level detection unit 5 calculates the absolute value of the low-frequency signal input by the absolute value detection unit 21, and then the DC cut unit 22 performs a DC cut component cut process. The signal that has been cut by the DC cut unit 22 is output to the multiplication unit 7 as a level detection signal.

  FIG. 4B is a block diagram illustrating a schematic configuration of the level variation detection unit 6. FIG. 5 is a flowchart showing the processing of each functional unit in the level variation detection unit 6. Further, FIGS. 6 to 8 are diagrams illustrating an example of an operation for detecting the level variation amount of the low frequency signal by the level variation amount detection unit 6 for each functional unit.

  The level fluctuation amount detection unit 6 includes a monaural conversion unit 31, a maximum value detection unit (maximum value detection unit) 32, a maximum value hold unit (maximum value hold unit) 33, a smoothing unit 34, and a decibel conversion unit 35. The first level limiting unit 36, the differential processing unit (differential processing unit) 37, the second level limiting unit 38, the linear conversion unit 39, the fluctuation amount extracting unit 40, and the gain processing unit (gain adjusting unit) 41. And have.

  When the low-frequency signal input from the first LPF unit 3 is a plurality of channels, the monaural conversion unit 31 divides the signal of each channel by the number of channels and then adds the monaural signal (monaural conversion). ) (Step S.11 in FIG. 5). FIG. 6A shows an example of a low-frequency signal that is monauralized by the monaural converter 31. For example, when the input signal is a two-channel stereo signal (a signal composed of an L channel and an R channel), the amplitude of each channel signal is divided by two and then added to obtain a stereo signal. Performs monaural processing. The low frequency signal converted to monaural by the monaural converter 31 is output to the maximum value detector 32.

  After obtaining the absolute value of the amplitude, the maximum value detection unit 32 performs processing for obtaining the maximum value for each predetermined sample (step S.12 in FIG. 5). FIG. 6B shows an example of a low-frequency signal for which maximum value detection has been performed by the maximum value detection unit 32. The maximum value detection unit 32 according to the present embodiment obtains a maximum value (maximum value within one frame) for every 32 samples. In the maximum value detection unit 32 according to the present embodiment, since the overlap processing is not performed, the sampling frequency is 1/32. The low frequency signal subjected to the maximum value processing in the maximum value detection unit 32 is output to the maximum value holding unit 33.

  The maximum value holding unit 33 has a role of holding (holding) the maximum value of the input low frequency signal by a predetermined number of samples (predetermined number of frames) (step S.13 in FIG. 5). FIG. 6C shows an example of a low frequency signal that has been subjected to the maximum value hold processing in the maximum value hold unit 33. As described above, the maximum value detection unit 32 obtains the maximum value for 32 samples (predetermined samples are set to 32 samples). Since the maximum value hold unit 33 performs overlap processing, the sampling frequency of the low frequency signal input from the maximum value detection unit 32 does not change. By holding the maximum value for a predetermined sample in the maximum value holding unit 33, it is possible to remove fine fluctuations in the maximum value. The low frequency signal that has been subjected to the maximum value hold processing in the maximum value hold unit 33 is output to the smoothing unit 34.

  The smoothing unit 34 has a role of performing a smoothing process on the input low frequency signal (step S.14 in FIG. 5). FIG. 6D illustrates an example of a low frequency signal that has been subjected to the smoothing process in the smoothing unit 34. Specifically, in the smoothing unit 34, it is possible to remove even finer vibrations by performing low-pass filter processing on the input low-frequency signal. As described above, the smoothing process is performed on the input signal to obtain the envelope in the smoothing unit 34. The envelope obtained by the smoothing unit 34 is output to the decibel conversion unit 35.

  The decibel conversion unit 35 has a role of converting the output level of the signal from the amplitude value to the decibel value by converting the input envelope into a logarithm (step S.15 in FIG. 5). FIG. 7A shows an example of a low-frequency signal subjected to decibel conversion in the decibel conversion unit 35. Since it is known that a human senses a change in volume logarithmically, the decibel converter 35 converts the signal level into a decibel value by converting the envelope into a logarithm.

  Specifically, the decibel conversion unit 35 uses the input envelope as the input envelope and the output as the decibel signal [dB] of the output signal, thereby reducing the decibel converted into decibels based on the following Equation 1. A range signal can be obtained.

output [dB] = 20 * log ₁₀ (input) Expression 1
The low frequency signal obtained by the decibel conversion unit 35 is output to the first level limiting unit 36.

  The first level limiting unit 36 has a role of preventing an extreme signal level change by performing level limiting (signal level limiting) using a predetermined lower limit value (step S.16 in FIG. 5). . FIG. 7B shows an example of a low-frequency signal that has been subjected to level restriction processing in the first level restriction unit 36. When the change in the signal level of the input low-frequency signal is extremely large, for example, when the signal level changes greatly from a silent state (a state where the signal level is close to zero (0)) to near full scale, The level fluctuation amount detection signal generated in the level fluctuation amount detector 6 also fluctuates greatly and rapidly. For this reason, in order to suppress an extreme change in signal level due to such a large and rapid fluctuation, the level is limited at a predetermined lower limit value. By limiting the signal level below the predetermined lower limit value to the lower limit value, it is possible to prevent an extreme change in signal level. In the first level limiting unit 36 according to the present embodiment, for example, −20 [dB] is set as the lower limit value. Due to this limitation, a signal level of −20 [dB] or less becomes −20 [dB]. The low frequency signal whose level is limited by the first level limiting unit 36 is output to the differentiation processing unit 37.

  The differentiation processing unit 37 has a role of obtaining a fluctuation amount in the signal level of the low frequency signal (step S.17 in FIG. 5). FIG. 7C shows an example of the amount of fluctuation of the low-frequency signal obtained by the differentiation processing unit 37. Specifically, by performing a high-pass filter process on the input low-frequency signal, it is possible to determine the amount of fluctuation in the signal level [dB] of the input low-frequency signal. Therefore, when the obtained variation amount value is large, it can be determined that the signal level variation is large, and when the variation amount value is small, it can be determined that the signal level variation is small. . The fluctuation amount of the low frequency signal obtained in the differentiation processing unit 37 is output to the second level limiting unit 38.

  The second level limiting unit 38 has a role of extracting only the positive fluctuation amount by setting the lower limit value of the inputted fluctuation amount to zero (0) (step S.18 in FIG. 5). ). FIG. 8A shows an example of the amount of fluctuation that has been extracted on the positive side in the second level limiting unit 38. The signal from which only the positive fluctuation amount is extracted in the second level limiting unit 38 is output to the linear conversion unit 39.

  The linear conversion unit 39 has a role of converting the input fluctuation amount signal into a linear value (step S.19 in FIG. 5). FIG. 8B shows an example of a low-frequency signal converted into a linear value by the linear converter 39. Since the fluctuation amount obtained by the second level limiting unit 38 is in decibels, the linear conversion unit 39 converts the fluctuation amount into an amplitude. The conversion from decibel to amplitude can be realized by performing a reverse process of the process using the equation 1 performed in the decibel conversion unit 35. The low frequency signal converted into a linear value (amplitude conversion) in the linear conversion unit 39 is output to the fluctuation amount extraction unit 40.

The fluctuation amount extraction unit 40 has a role of subtracting 1 from the amplitude value converted into a linear value by the linear conversion unit 39 (step S.20 in FIG. 5). FIG. 8C shows an example of a signal obtained by subtracting the amplitude value by 1 in the fluctuation amount extraction unit 40. The conversion of the linear value in the linear conversion unit 39 is performed by the transformation of Expression 1 as described above. Specifically, Equation 2 can be obtained by modifying Equation 1 as follows. In Equations 1 and 2, input represents an amplitude (linear value), and output represents a decibel signal [dB].
output [dB] = 20 * log ₁₀ (input) Expression 1
= Log ₁₀ (input) ²⁰

From this equation 1, (input) ²⁰ = 10 ^(output) ... Equation 2 is obtained. In this equation 2, when output is zero (0) [dB], that is, when the variation is zero (0). The value of “10 to the power of zero (0)” on the right side of Equation 2 is “1”. That is, when the fluctuation amount is zero (0), the amplitude (input value) becomes 1. Therefore, the fluctuation amount extraction unit 40 performs adjustment so that the amplitude is zero (0) when the fluctuation amount is zero (0) by subtracting 1 from the obtained amplitude value (linear value). Do. The signal subjected to the subtraction of 1 in the fluctuation amount extraction unit 40 is output to the gain processing unit 41.

  The gain processing unit 41 has a role of setting a lower limit value in amplitude (step S.21 in FIG. 5). FIG. 8D shows an example of a signal for which a lower limit value has been set for the amplitude in the gain processing unit 41.

  By subtracting 1 in the fluctuation amount extraction unit 40, the amplitude becomes zero (0) when the fluctuation amount of the low-frequency signal is zero (0). For this reason, also in the signal input to the gain processing unit 41, the amplitude value becomes zero (0) when the variation amount of the low frequency signal is zero (0). However, if the amount of fluctuation of the low-frequency signal is zero (0), it is impossible to adjust the amount of fluctuation of the overtone signal, and it becomes difficult to constantly achieve the minimum low-frequency complementing effect. . For this reason, the gain processing unit 41 sets a lower limit value of the amplitude between zero (0) and 1, thereby maintaining a steady minimum low-frequency complementing effect, and a bass enhancement effect and sound quality deterioration suppression. To achieve both.

  In order to effectively increase / decrease the complementary effect at the location where the fluctuation amount of the low frequency signal is large (the level fluctuation is large), the gain processing unit 41 first adds gain to the signal input from the fluctuation amount extraction unit 40. It is also possible to use the above-described method of setting the lower limit value after performing overall gain adjustment. The signal for which the lower limit value has been set by the gain processing unit 41 is output to the multiplication unit 7 as a level fluctuation amount detection signal.

  As described above, the multiplication unit 7 generates the harmonic signal (first harmonic signal) generated by the edge detection unit 4, the level detection signal generated by the level detection unit 5, and the level fluctuation amount detection unit 6. The level fluctuation amount detection signal is input. The multiplication unit 7 has a role of multiplying the harmonic signal by the level detection signal and the level variation detection signal (step S.3 in FIG. 2). The multiplication unit 7 can multiply the harmonic signal (first harmonic signal) by the signal level of the low-frequency signal by multiplying the harmonic signal by the level detection signal. Furthermore, by multiplying the harmonic signal by the level variation detection signal, it is possible to perform weighting according to the level variation of the low frequency signal. The harmonic signal weighted in the multiplication unit 7 in this way is output to the phase inverting unit 8 as a “weighted harmonic signal (second harmonic signal)”.

  The phase inversion unit 8 performs phase inversion of the weighted harmonic signal input from the multiplication unit 7 (step S.4 in FIG. 2). The weighted harmonic signal whose phase has been inverted by the phase inverting unit 8 is output to the first amplifying unit 9. The first amplifying unit 9 performs an amplification process on the weighted harmonic signal whose phase has been inverted. With this amplification process, the final gain adjustment is performed on the weighted harmonic signal (step S.4 in FIG. 2). The weighted harmonic overtone signal that has been amplified in the first amplification unit 9 is output to the second LPF unit 10.

  The weighted overtone signal (second overtone signal) input to the second LPF unit 10 is subjected to low frequency adjustment in the second LPF unit 10, and thereafter, high frequency adjustment is performed in the second HPF unit 11 ( Step S.5 in FIG. The second LPF unit 10 includes a low-pass filter, and the second HPF unit 11 includes a high-pass filter. FIG. 3B is a diagram illustrating an example of filter characteristics of the low-pass filter and the high-pass filter. The low-pass filter according to the present embodiment is a fourth-order Butterworth filter, and 60 Hz is set as the cutoff frequency. On the other hand, the high-pass filter is a second-order Butterworth filter, and 60 Hz is set as the cutoff frequency.

  The cut-off frequency and filter characteristics of the low-pass filter are determined in consideration of the reproduction capability of the speaker. In the second LPF unit 10, by performing a filtering process on the weighted harmonic signal using a low-pass filter, the signal level at a frequency below the reproduction capability of the speaker is limited. Further, the cut-off frequency and filter characteristics of the high-pass filter are determined to values that do not cause a sense of incongruity in the mid-high range due to the complementary processing using the harmonic signal in consideration of the listener's audibility. In the second HPF unit 11, it is possible to generate a harmonic signal having no sense of incongruity on hearing by performing a filtering process on the weighted harmonic signal using a high-pass filter. The harmonic signal filtered by the second LPF unit 10 and the second HPF unit 11 is output to the adding unit 13 as a harmonic signal that has been weighted and band-limited (filtered second harmonic signal).

  On the other hand, it is a low-frequency signal filtered by the first LPF unit 3 and is not input to the edge detection unit 4, the level detection unit 5, and the level fluctuation amount detection unit 6, but is input to the second amplification unit 12. The low frequency signal is gain-adjusted in the second amplifying unit 12. The gain adjusted in the second amplifying unit 12 includes a high-frequency signal output from the first HPF unit 2 to the adding unit 13 and weighted and band-limited harmonics output from the second HPF unit 11 to the adding unit 13. It is set in consideration of the gain with the signal (filtered second overtone signal). The low frequency signal whose gain is adjusted in the second amplifying unit 12 is output to the adding unit 13.

  The adding unit 13 includes a high frequency signal input from the first HPF unit 2, a low frequency signal input from the second amplifying unit 12, and a weighted and band limited harmonic signal (filter) input from the second HPF unit 11. 2), the output signal is generated (step S.6 in FIG. 2). The output signal generated by the adding process of the adding unit 13 is output to a speaker or the like (not shown).

  The low-frequency interpolation processing of the low-frequency interpolation device that has been used conventionally is characterized in that the harmonic signal is weighted according to the signal level (absolute value, squared value, etc.) of the harmonic signal. . However, in the low frequency complement device 1 according to the present embodiment, the level fluctuation amount detection unit 6 detects not only the signal level of the harmonic signal but also the level fluctuation amount in the signal level of the low frequency signal, and the detected level fluctuation. The level variation detection unit 6 performs weighting processing on the harmonic signal according to the amount. For this reason, effective weighting is performed according to the signal level of the low-frequency signal and the amount of fluctuation of the signal level by increasing / decreasing the level fluctuation amount detection signal output from the level fluctuation amount detector 6 or setting the lower limit value. Is possible.

  Next, as a specific example of operation, the level fluctuation amount detection unit 6 performs weighting based on the level fluctuation amount and the signal state of the output signal when weighting based on the level fluctuation amount is not performed. An example of each operation will be described by showing signal states of output signals in the case of breakage. In either case, the input signal (audio signal) to be input is a 50 Hz sine wave.

[When weighting based on level fluctuation is not performed]
When weighting based on the level fluctuation amount is not performed, the level fluctuation amount detection unit 6 does not perform the process of outputting the level fluctuation amount detection signal to the multiplication unit 7. Therefore, the first HPF unit 2 and the first LPF unit 3 divide the input signal into a high-frequency signal and a low-frequency signal, and then generate a harmonic signal generated by the edge detection unit 4 by the level detection unit 5. The overtone signal, the high frequency signal, and the low frequency signal that have been multiplied by the level detection signal are added by the adder 13 and output as an output signal.

  FIG. 9A shows a low-frequency signal (first LPF unit output) after passing through the first LPF unit 3 and a level detection signal generated by the level detection unit 5, and FIG. 2 shows a harmonic signal (edge detection unit output) generated by the edge detection unit 4 and a level detection signal generated by the level detection unit 5. Here, the DC component cut processing in the DC cut unit 22 of the level detection unit 5 is performed using a first-order Butterworth filter, and the cut-off frequency is set to 20 Hz.

  As shown in FIG. 9B, the harmonic signal output from the edge detection unit is an impulse train having an amplitude of 1, and the output of this impulse train is a sine wave of a low frequency signal output from the first LPF unit 3. In (see FIG. 9A), it is generated at a position that changes from the negative side to the positive side. Further, the level detection signal shown in FIG. 9B is offset to the negative side with the output position of the impulse train as a reference by the DC component cut processing in the DC cut unit 22.

  FIG. 10A shows a weighted harmonic signal weighted by multiplying the harmonic signal (impulse train) shown in FIG. 9B by the level detection signal shown in FIG. Is shown. Since the weighted harmonic signal has the signal level (amplitude) of the impulse train in FIG. 9B, the signal level (amplitude value) of the output position of the impulse train is the signal level (amplitude value) of the level detection signal. To be equal to.

  FIG. 10 (b) shows the frequency characteristics of the low frequency signal, and FIG. 11 (a) shows the frequency characteristics of the weighted harmonic signal shown in FIG. 10 (a). As already described, since the input signal is a sine wave of 50 Hz, the frequency output of the low frequency signal is only 50 Hz as shown in FIG. However, the weighted harmonic signal shown in FIG. 11 (a) is subjected to signal level detection at an integer multiple frequency with 50 Hz as a reference. Specifically, as shown in FIG. 11 (a), a signal output (overtone) of the same level as the signal level of 50 Hz at a frequency of 100 Hz, 150 Hz, 200 Hz, 250 Hz,. Is generated.

  Further, as shown in FIG. 11A, the signal obtained by multiplying the harmonic detection signal by the level detection signal produces the same signal level at all the integral multiples of 50 Hz. This may cause a sense of incongruity to the listener. For this reason, as described above, the second LPF unit 10 and the second HPF unit 11 perform the band limitation adjustment of the low band and the high band.

  FIG. 11B is a diagram showing the frequency characteristics of a harmonic signal (weighted and band-limited harmonic signal) that has been subjected to band limitation of the low frequency band and the high frequency band by the second LPF unit 10 and the second HPF unit 11. is there. As shown in FIG. 11 (b), the band-limited harmonic signal has a lower harmonic signal level as it becomes higher. Thus, by gradually reducing the high-frequency signal level, it is possible to obtain a harmonic signal that does not cause a sense of incongruity in hearing. Further, as shown in FIGS. 10B, 11A, and 11B, the signal level of 50 Hz is normalized and expressed to be 0 [dB] for convenience of explanation.

[When weighting based on level fluctuation amount]
When weighting based on the level fluctuation amount is performed, the level fluctuation amount detection signal generated in the level fluctuation amount detection unit 6 is output to the multiplication unit 7. Therefore, the first HPF unit 2 and the first LPF unit 3 divide the input signal into a high-frequency signal and a low-frequency signal, and then generate a harmonic signal generated by the edge detection unit 4 by the level detection unit 5. The added level detection signal is multiplied by the level fluctuation amount detection signal generated by the level fluctuation amount detection unit 6, and the overtone signal multiplied by the multiplication process, the high frequency signal, and the low frequency signal are added by the addition unit 13. And output as an output signal.

  FIG. 12A shows an input signal, and FIG. 12B shows a low frequency signal output from the first LPF unit 3. The band of the input signal is limited by the low-pass filter of the first LPF unit 3, so that the state shown in FIG. Next, FIG. 12C shows a harmonic signal s generated by the edge detection unit 4. The edge detection unit 4 generates a harmonic overtone signal s composed of an impulse train having an amplitude of 1, based on the low frequency signal.

FIG. 13A shows the level detection signal w ₁ generated in the level detection unit 5. In the level detection unit 5, the absolute value detection unit 21 detects the absolute value of the input signal, and the DC cut unit 22 performs DC cut processing. FIG. 13B and FIG. 13C show level fluctuation amount detection signals generated by the level fluctuation amount detection unit 6. However, FIG. 13B shows the level variation detection signal w _2a when the gain adjustment in the gain processing unit 41 is not performed, and FIG. 13C shows the gain of 0.3 amplitude in the gain processing unit 41. adjustment indicates the level variation detected signal w _2b when done.

Next, the weighting harmonic signal generated in the multiplying unit 7, described above, the harmonic signal s of FIG. 12 (c), the level detection signal _{w 1} in FIG. 13 (a), FIG. 13 (b) (c) The level fluctuation amount detection signals w _2a and w _2b will be described.

14A to 14C show weighted harmonic signals. More specifically, FIG. 14A shows a weighted harmonic signal generated by the multiplication unit 7 by multiplication of the harmonic signal s and the level detection signal w ₁ . FIG. 14B shows a weighted harmonic signal generated by multiplication of the harmonic signal s, the level detection signal w _1, and the level variation detection signal w _{2 a} in the multiplier 7. Further, FIG. 14 (c), the multiplication unit 7 shows the weighted harmonic signal generated by multiplying the harmonic signal s and the level detection signal w ₁ and level variation detection signal w _2b.

  15 (a) to 15 (c) show harmonic signals (FIG. 14 (a) to FIG. 14 (c)) obtained by filtering the weighted harmonic signals shown in FIGS. 14 (a) to 14 (c) by the second LPF unit 10 and the second HPF unit 11. A weighted and band-limited harmonic signal).

  The weighted harmonic signal shown in FIG. 14A and the weighted and band-limited harmonic signal shown in FIG. 15A are not weighted by the level variation detection signal, and the level (amplitude) of the low frequency signal is not. This is equivalent to a signal weighted according to only the absolute value or the square of. The weighted harmonic signal shown in FIG. 14A and the weighted and band-limited harmonic signal shown in FIG. 15A are weighted only by the level detection signal, so regardless of the fluctuation amount of the low frequency signal. A harmonic signal having a relatively large level (amplitude) is output. For this reason, it is possible to sufficiently add the acoustic effect / acoustic correction to the low frequency range of the output signal, but on the other hand, there is a possibility that excessive correction may be performed in the mid / low frequency range. There is a risk of sound quality degradation. The actual effect amount due to low-frequency interpolation depends on the final gain adjustment, but it has been difficult to achieve both the bass enhancement effect and the suppression of sound quality deterioration.

  On the other hand, the weighted harmonic signal shown in FIG. 14B and the weighted and band-limited harmonic signal shown in FIG. 15B are weighted by the level variation detection signal, but the gain processing unit 41 2 shows a signal that has not been subjected to gain adjustment for the level fluctuation amount detection signal.

  The weighted harmonic signal shown in FIG. 14 (b) and the weighted and band-limited harmonic signal shown in FIG. 15 (b) are weighted by the level fluctuation amount detection signal. In response to such a large low-frequency sound, a harmonic signal is output only at a location where there is a large change in the low-frequency level. For this reason, while the overtone signal subjected to weighting and band limitation has a sharp signal output, the overtone signal is not constantly output, so that the complementary effect of the entire signal tends to decrease.

  Further, the weighted harmonic signal shown in FIG. 14C and the weighted and band-limited harmonic signal shown in FIG. 15C are weighted by the level variation detection signal, and the gain processing unit 41 is further weighted. FIG. 5 shows a signal that has been subjected to gain adjustment for the level variation detection signal.

  In the weighted harmonic signal shown in FIG. 14C and the weighted and band-limited harmonic signal shown in FIG. 15C, the harmonic signal is weighted by the gain-adjusted level variation detection signal. In addition, a particularly large complementary signal can be obtained at a location where the level of the low range has changed significantly, and a harmonic signal corresponding to the level detection signal can be obtained at other locations. Therefore, it is possible to produce the maximum complementary effect when there is a large change in the low level while maintaining the minimum complementary effect constantly, and to achieve both the bass enhancement effect and the suppression of sound quality deterioration Is possible.

  As described above, in the low-frequency complement device 1 according to the present embodiment, it is possible to compensate for the low frequency below the lowest resonance frequency that cannot be sufficiently reproduced by the speaker by using the harmonic signal.

  Furthermore, the low frequency complementing device 1 weights the harmonic overtone signal in accordance with the signal level of the low frequency signal included in the sound source and the amount of level fluctuation. It is possible to exert as much as possible. For example, the amount of low-frequency interpolation is constantly kept to a minimum according to the level of the low-frequency signal, while the amount of low-frequency fluctuation is large in response to low-frequency sounds such as bass drums included in the sound source. In such a case, a sharp bass can be reproduced from the speaker by instantaneously amplifying the effect amount.

  The preferred embodiments of the low-frequency complementing apparatus according to the present invention have been described above with reference to the accompanying drawings, showing the low-frequency complementing apparatus 1 as an example, but the present invention is not limited to such examples. Absent. It will be apparent to those skilled in the art that various changes and modifications can be made within the scope of the claims and belong to the technical scope of the present invention.

DESCRIPTION OF SYMBOLS 1 ... Low-frequency supplement apparatus 2 ... 1st HPF part (band-division means)
3 ... 1st LPF section (band dividing means)
4 Edge detection unit (overtone signal generation means)
5: Level detection unit (level detection signal generation means)
6: Level fluctuation amount detection unit (level fluctuation amount detection means)
7: Multiplication unit (multiplication means)
8: Phase inversion unit 9: First amplification unit 10: Second LPF unit (filter means)
11 ... 2nd HPF part 12 ... 2nd amplification part 13 ... Addition part (addition means)
21 ... Absolute value detection unit 22 ... DC cut unit 31 ... Monaural conversion unit 32 ... Maximum value detection unit (maximum value detection means)
33 ... Maximum value holding section (maximum value holding means)
34 ... smoothing unit 35 ... decibel conversion unit 36 ... first level limiting unit 37 ... differentiation processing unit (differentiation processing means)
38 ... 2nd level restriction part 39 ... Linear conversion part 40 ... Variation amount extraction part 41 ... Gain processing part (gain adjustment means)

Claims (8)

Band dividing means for dividing an input signal into a low-frequency signal and a high-frequency signal;
A first harmonic signal having a predetermined signal level output is generated at a frequency several times the predetermined low frequency with reference to a predetermined low frequency of the low frequency signal band-divided by the band dividing means. Overtone signal generating means;
Level detection signal generating means for generating a level detection signal by performing a DC component cut process after calculating the absolute value of the signal level of the low frequency band band-divided by the band dividing means;
Level fluctuation amount detection means for generating a level fluctuation amount detection signal indicating the level fluctuation amount of the low frequency signal based on the signal level fluctuation of the low frequency signal band-divided by the band division means;
Multiplication means for generating a second harmonic signal by multiplying the first harmonic signal by the level detection signal and the level variation detection signal;
Adding means for adding the high-frequency signal and the low-frequency signal band-divided by the band-dividing unit and the second overtone signal generated by the multiplying unit to generate an output signal. A low-frequency interpolation device that is characterized. The level fluctuation amount detecting means includes
Maximum value detecting means for detecting a maximum value of the low frequency signal band-divided by the band dividing means;
Maximum value hold means for generating a maximum value hold signal by holding the maximum value detected by the maximum value detection means for a predetermined time; and
Differential processing means for generating the level fluctuation amount detection signal indicating the level fluctuation amount of the low-frequency signal by performing differentiation processing on the maximum value hold signal generated by the maximum value holding means;
The low-frequency interpolation apparatus according to claim 1, comprising: The level fluctuation amount detection means has gain adjustment means for performing gain adjustment so that the signal level becomes a value higher than zero with respect to the level fluctuation amount detection signal,
3. The multiplication means according to claim 2, wherein the first harmonic signal is multiplied by the level detection signal and a level variation detection signal whose gain is adjusted by the gain adjustment means. Low-frequency interpolation device. A filter that gradually reduces the signal level output of the second overtone signal as the frequency increases by performing a filtering process on the second overtone signal generated by the multiplication means using a high-pass filter. Having means,
The said addition means adds the said high frequency signal, the said low frequency signal, and the 2nd harmonic signal filtered by the said filter means, and produces | generates the said output signal. The Claim 1 thru | or 3 characterized by the above-mentioned. The low-frequency complement device according to any one of the above. A band dividing step in which the band dividing means divides the input signal into a low-frequency signal and a high-frequency signal;
The harmonic signal generating means includes a first signal level output at a frequency several times the predetermined low frequency with reference to a predetermined low frequency of the low frequency signal divided in the band dividing step. A harmonic signal generation step for generating a harmonic signal;
Level detection signal generation step in which the level detection signal generation means generates a level detection signal by performing a DC component cut process after calculating the absolute value of the signal level of the low-frequency signal band-divided in the band division step. When,
Level fluctuation amount detection in which the level fluctuation amount detection means generates a level fluctuation amount detection signal indicating the level fluctuation amount of the low band signal based on the signal level fluctuation of the low band signal band-divided in the band division step. Steps,
A multiplication step in which a multiplying unit generates a second harmonic signal by multiplying the first harmonic signal by the level detection signal and the level variation detection signal;
An addition step in which the addition means generates an output signal by adding the high-frequency signal and the low-frequency signal band-divided in the band-division step and the second overtone signal generated in the multiplication step. A low-frequency interpolation method characterized by comprising: In the level variation detection step,
A maximum value detecting step in which a maximum value detecting means detects a maximum value of the low frequency signal band-divided in the band dividing step;
A maximum value hold step in which a maximum value hold means generates a maximum value hold signal by holding the maximum value detected in the maximum value detection step for a predetermined time; and
Differentiation processing for generating the level fluctuation amount detection signal indicating the level fluctuation amount of the low-frequency signal by performing differentiation processing on the maximum value hold signal generated in the maximum value holding step. Steps,
The low-frequency interpolation method according to claim 5, wherein: The level fluctuation amount detection step includes a gain adjustment step in which gain adjustment means performs gain adjustment so that a signal level of the level fluctuation amount detection signal is higher than zero.
In the multiplication step, the multiplication means multiplies the first harmonic signal by the level detection signal and a level variation detection signal for which gain adjustment has been performed in the gain adjustment step. Item 7. The low-frequency interpolation method according to Item 6. By performing a filtering process using a high-pass filter on the second harmonic signal generated in the multiplication step, the filter means gradually outputs the signal level output of the second harmonic signal as the frequency increases. Having a filtering step that reduces to
In the adding step, the adding means adds the high frequency signal, the low frequency signal, and the second overtone signal filtered in the filtering step to generate the output signal. The low-frequency interpolation method according to any one of claims 5 to 7.

Images