JP2008058470A - Audio signal processor and audio signal reproduction system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reproduce an audio signal of high sound quality such that differences in bodily sensation effect due to individual differences in auditory characteristic are suppressed as much as possible by varying frequency characteristic of a high frequency range according to auditory characteristics of a user. <P>SOLUTION: An audio signal processor includes a frequency analyzer which analyzes frequency characteristics of an input audio signal; a correction amount predicting unit which detects the tendency of the frequency characteristics analyzed by the frequency analyzer, and predicts a correction amount for the high frequency range; an auditory parameter setter 41 which can set an auditory parameter based upon the auditory characteristics of the user; an auditory sensitivity corrector 40 which corrects the correction amount predicted by the correction amount predicting unit with the auditory parameter set by the auditory parameter setter 41; and an equalizer 33 which corrects the frequency characteristics of an audio signal output from a high-frequency-range interpolation processing unit 32 based upon the correction amount output from the auditory sensitivity corrector 40. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an audio signal processing apparatus that corrects high frequency components of a reproduced digital audio signal. The present invention also relates to an audio signal reproduction system including the audio signal reproduction device.

  In recent years, distribution of music data compressed in a format such as MP3 (MPEG1 audio layer 3) or WMA (Windows Media Audio, “Windows” is a registered trademark of Microsoft Corporation) has been rapidly spreading. In these types of music data, frequency components of a predetermined frequency or higher are removed in the compression process in order to avoid an increase in the amount of data and an increase in occupied bandwidth due to an excessively wide audio band. For example, in the MP3 format, frequency components of about 16 kHz or more are removed. Thus, music and the like reproduced based on music data from which frequency components of a certain value or more have been removed have poor sound quality because the high frequency components are cut off.

Therefore, Patent Documents 1 to 3 disclose configurations for preventing the above-described deterioration in sound quality. First, Patent Document 1 discloses a configuration in which a sound signal of about 20 kHz or more is generated and superimposed on an original audio signal. Patent Document 2 discloses a configuration in which an n-order harmonic component (for example, 20 kHz or more) outside the audible band is generated based on the high frequency component of the source audio signal and added to the original audio signal. Patent Document 3 discloses a configuration in which a high frequency component obtained by delaying random noise is added to an original audio signal.
Japanese Patent No. 2643349 Japanese Patent No. 3137289 Japanese Patent No. 3605706

  However, all of the configurations disclosed in Patent Documents 1 to 3 are configurations that simply interpolate a pseudo high frequency component in the original audio signal. In such a high frequency correction, since the interpolation characteristic is constant, a large difference occurs in the correction effect that can be experienced by the user. That is, there are individual differences in human auditory characteristics (such as audible frequency characteristics), and particularly with respect to high frequency components, there is a large difference in how sounds are heard depending on the age, sex, and living environment of the user. For example, women, young people, and people living in a quiet environment are said to have high ability to hear high-frequency components. Therefore, as disclosed in Patent Documents 1 to 3, when high-frequency correction is performed based on certain interpolation characteristics, a person who can feel comfortable sound quality by high-frequency correction and a person who can hardly feel the correction effect Will occur.

  In view of the above problems, the present invention provides a high-quality audio signal that suppresses the difference in the bodily sensation effect due to individual differences in auditory characteristics as much as possible by changing the high frequency characteristics in accordance with the user's auditory characteristics. An object of the present invention is to provide an audio signal correction apparatus capable of reproducing the sound.

  In order to achieve the above object, an audio signal processing apparatus of the present invention is provided. An audio signal processing apparatus including a high frequency correction unit including a high frequency interpolation processing unit for interpolating a high frequency component of an input audio signal, wherein the high frequency correction unit has a frequency characteristic of an input audio signal. Based on a frequency analysis unit to analyze, a tendency of the frequency characteristic analyzed by the frequency analysis unit, a correction amount prediction unit for predicting a high frequency correction amount and outputting a correction result, and a user's auditory characteristic An auditory parameter setting unit capable of setting an auditory parameter, an auditory sensitivity correction unit that corrects a correction result output from the correction amount prediction unit by an auditory parameter set in the auditory parameter setting unit, and the high-frequency interpolation An equalizer that corrects a frequency characteristic of the audio signal output from the processing unit, and the equalizer receives the frequency characteristic of the audio signal output from the high-frequency interpolation processing unit. It is corrected based on the correction result output from the sensitivity correction unit.

  The audio signal processing device of the present invention is an audio signal processing device including a high frequency correction unit including a high frequency interpolation processing unit that interpolates a high frequency component of an input audio signal, and the high frequency correction unit The frequency analysis unit that analyzes the frequency characteristics of the input audio signal, the frequency analysis result in the frequency analysis unit is subjected to regression analysis, the regression analysis result is derived, and the regression analysis result is reflected to the high frequency side A correction amount prediction unit that predicts a high-frequency correction amount and outputs a correction result; and an equalizer that corrects a frequency characteristic of an audio signal output from the high-frequency interpolation processing unit. Based on the correction result output from the correction amount prediction unit, the frequency characteristic of the audio signal is corrected.

  In addition, the audio signal reproduction system of the present invention includes a high frequency correction unit including a high frequency interpolation processing unit that interpolates a high frequency component of an input audio signal, and an audio signal output from the high frequency correction unit. An audio signal reproduction system comprising an output terminal capable of outputting to the system and an operation unit capable of inputting various operation commands in the system, wherein the high frequency correction unit analyzes the frequency characteristics of the input audio signal A frequency analysis unit, a correction amount prediction unit that detects a tendency of the frequency characteristic analyzed by the frequency analysis unit, predicts a high-frequency correction amount, and outputs a correction result; and an auditory parameter based on a user's auditory characteristic An auditory parameter setting unit that can set a correction result output from the correction amount prediction unit using an auditory parameter that is set in the auditory parameter setting unit; An equalizer that corrects a frequency characteristic of the audio signal output from the interpolation processing unit; and the equalizer outputs the frequency characteristic of the audio signal output from the high-frequency interpolation processing unit from the auditory sensitivity correction unit. Correction is performed based on the correction result.

  The present invention can reproduce a high-quality sound signal that matches the auditory characteristics of the user.

  In the audio signal processing device according to the present invention, the frequency analysis unit subtracts the left channel and right channel audio signals from the input stereo audio signal and extracts a difference signal; and the difference signal A fast Fourier transform processing unit for performing frequency analysis by performing a fast Fourier transform process on the difference signal output from the extraction unit; and an envelope detection unit for performing peak detection of the frequency analysis result output from the fast Fourier transform processing unit. It is good also as a structure. With this configuration, since high-frequency correction is predicted from the frequency envelope of the audio signal, dynamic sound quality control according to the state of music (musical tone, etc.) can be performed instead of conventional tone control. Therefore, it is possible to output a natural sound without a sense of incongruity.

  The correction amount prediction unit may include a regression analysis unit that performs a regression analysis of the frequency analysis result in the frequency analysis unit. With this configuration, the tune of the input music data is grasped, and the inclination of the regression line is corrected according to the tune to be grasped, so that it is possible to perform equalization adapted to music information (music tune, rhythm, etc.). High frequency interpolation with comfortable sound quality can be realized. In addition, by combining interpolation frequency prediction by regression analysis and user's auditory characteristics, equalization that is suitable for music information is possible, not just high-frequency emphasis equalization, and high frequency with more natural and comfortable sound quality. Interpolation can be realized.

  The correction amount prediction unit may further include a filter that removes high-frequency components of fluctuations in the regression analysis result in the regression analysis unit. With this configuration, the change of the coefficients a and b of the regression line (primary formula) in the regression analysis unit can be smoothly changed according to the tone, and the slope of the regression line can also be smoothly changed according to the tone. . Therefore, it is possible to prevent noise that occurs when the regression line changes abruptly according to the tone of music.

  The audio signal processing device or the audio signal reproduction system of the present invention includes a normal mode and an audio measurement mode capable of measuring a user's audio characteristics, and the audio parameter setting unit is set to a predetermined frequency in advance. A frequency oscillating unit that generates a pure sound while changing the level, and a memory that records a frequency and a level when the operating unit is operated when a pure sound is generated in the frequency oscillating unit. When the system is shifted to the auditory measurement mode, the frequency oscillating unit generates the pure sound while changing the level, outputs the generated pure sound from the output terminal, and the pure sound is output. Sometimes when the operation unit is operated by the user, the frequency and level of the pure sound at that time are recorded in the memory, and the frequency recorded in the memory It may set the auditory parameters based on micro-level information. With this configuration, optimum high frequency characteristics can be obtained for each user. Therefore, the user can enjoy music with comfortable sound quality.

(Embodiment 1)
[1. overall structure〕
FIG. 1 is a block diagram showing the configuration of the audio signal processing apparatus according to the first embodiment.

  In FIG. 1, an audio signal processing apparatus 1 corresponds to, for example, a personal computer or a portable music player that can reproduce music data in the MP3 format or the like. In this embodiment, a personal computer is taken as an example. The audio signal processing device 1 includes an information medium 11, an audio data processing unit 12, a high frequency correction unit 13, a digital interface (hereinafter referred to as a digital IF) 14, an operation unit 15, and a control unit 16. For convenience, FIG. 1 shows only the components necessary for the description of the present embodiment, and actually includes components other than those shown in the figure, such as a central processing unit (CPU) and a semiconductor memory. Note that the music data that can be played back by the audio signal processing apparatus 1 is not limited to the MP3 format, but may be music data composed of at least a digital signal such as a WMA format or a linear PCM format. In the following description, the MP3 format will be taken up and described.

  The information medium 11 is composed of a medium capable of writing or reading at least digital data, such as a hard disk drive or a semiconductor memory. The information medium 11 stores music data compressed in MP3 format.

  The audio data processing unit 12 reads out music data recorded on the information medium 11, demodulates it into a PCM (Pulse Code Modulation) signal, and outputs it.

  The high frequency correction unit 13 interpolates a high frequency component of about 16 kHz or more with respect to the PCM signal output from the audio data processing unit 12, and processes the frequency characteristic according to the auditory characteristic of the user. . A detailed configuration of the high frequency correction unit 13 will be described later.

  The digital IF 14 is an interface that can output the PCM signal that has been subjected to high-frequency correction in the high-frequency correction unit 13 to the outside of the apparatus. The digital IF 14 is configured by a USB (Universal Serial Bus) interface, for example, and includes a port compliant with the USB standard. It is assumed that driver software that can control the operation of a USB-compatible device such as the DAC 21 is incorporated in the audio signal processing apparatus 1 in advance.

  The operation unit 15 can input various commands in the audio signal processing device 1. For example, when the audio signal processing device 1 is configured by a personal computer, it corresponds to a keyboard or a mouse, and is configured by a portable music player. In this case, a playback button capable of inputting a command for reproducing music data and a stop button capable of inputting a command for stopping the reproduction of music data are included.

  The control unit 16 controls the operation of at least the audio data processing unit 12 and the high frequency correction unit 13 based on the contents operated by the operation unit 15.

  In addition, a headphone 22 is connected to the audio signal processing device 1 via a DAC 21. The DAC (Digital-Analog Converter) 21 converts the PCM signal output from the digital IF 14 into an analog audio signal, and is configured as a unit independent of the audio signal processing device 1 and the headphones 22. The DAC 21 is provided with a USB terminal conforming to the USB standard, and can be connected to the digital IF 14 via a cable. Further, the DAC 21 includes an analog terminal (for example, a stereo mini jack) to which headphones 22 corresponding to analog audio signals can be connected, and the headphones 22 can be connected.

  When listening to music data with the audio signal processing apparatus 1 having the above configuration, first, the headphones 22 are connected to the analog terminal of the DAC 21, and the USB terminal of the DAC 21 is connected to the digital IF 14. Then, it is detected that the DAC 21 is connected in the music signal processing apparatus 1, and the audio signal processing apparatus 1 is ready to output music data (PCM signal) from the digital IF 14.

  Next, when the operation unit 15 is operated by the user and a music data playback command is input, the control unit 16 outputs the playback command to the audio data processing unit 12. The audio data processing unit 12 reads out desired music data from the music data recorded on the information medium 11 based on the input reproduction command and demodulates it into a PCM signal. The demodulated PCM signal is output to the high frequency correction unit 13. When the PCM signal is compressed into the MP3 format, the high frequency component of about 16 kHz or more is removed, so that the demodulated PCM signal also has the high frequency component of 16 kHz or more removed.

  The high frequency correction unit 13 adds a high frequency component of about 16 kHz or more to the input PCM signal and corrects the frequency characteristic according to the user's auditory characteristics. A specific frequency characteristic correction operation will be described later.

  The PCM signal output from the high frequency correction unit 13 is output to the DAC 21 via the digital IF 14. The DAC 21 generates an analog audio signal based on the input PCM signal and outputs it to the headphones 22. Thereby, a sound is output from the headphones 22, and the user can listen to the output sound.

[2. Configuration of high-frequency correction unit 13]
FIG. 2 is a block diagram showing the configuration of the high frequency correction unit 13. As shown in FIG. 2, the high frequency correction unit 13 includes an input terminal 31, a high frequency interpolation processing unit 32, an equalizer 33, an output terminal 34, a difference signal extraction unit 35, a fast Fourier transform processing unit (hereinafter referred to as an FFT processing unit). 36), envelope detection unit 37, regression analysis unit 38, LPF 39, auditory sensitivity correction unit 40, and auditory parameter setting unit 41.

  The input terminal 31 is a terminal to which a PCM signal output from the audio data processing unit 12 is input.

  The high frequency interpolation processing unit 32 interpolates high frequency components for the PCM signal input to the input terminal 31. In the present embodiment, since the PCM signal input to the input terminal 31 has a signal with a frequency of about 16 kHz or more removed, the high-frequency interpolation processing unit 32 interpolates a signal with a frequency of about 16 kHz or more. The specific configuration and operation of the high-frequency interpolation processing unit 32 are the same as the configuration and operation disclosed in, for example, the description of the frequency interpolation unit in Japanese Unexamined Patent Application Publication No. 2002-73096. Omitted.

  The equalizer 33 corrects the frequency characteristic of the PCM signal output from the high frequency interpolation processing unit 32 based on the auditory characteristic parameter output from the auditory sensitivity correction unit 40. In the present embodiment, the equalizer 33 is configured by an IIR filter (IIR: Infinite duration Impluse Response), and the transfer function is as shown in (Equation 1).

  The output terminal 34 is a terminal that outputs the PCM signal output from the equalizer 33 to the outside (digital IF 14 in FIG. 1).

  The difference signal extraction unit 35 subtracts the left channel (Lch) signal and the right channel (Rch) signal from the stereo PCM signal input to the input terminal 31 to extract a difference signal. Thereby, a reverberant signal can be extracted. In the present embodiment, the input PCM signal is a stereo signal, but may be a monaural signal. In the case of a monaural signal, a difference signal (reverberation component) cannot be extracted even if subtraction processing is performed. In this case, frequency analysis processing is directly performed.

  The FFT processing unit 36 uses a fast Fourier transform (FFT) technique for the PCM signal output from the difference signal extraction unit 35 to convert the spectrum distribution of the PCM signal into a set of complex numbers (specifically, Is expressed as a set of two values representing the real part and imaginary part of the complex number), and frequency analysis is performed. The specific operation of the FFT processing unit 36 is the same as the configuration and operation disclosed in, for example, Japanese Unexamined Patent Application Publication Nos. 2002-175092 and 2002-171588 (Japanese Patent No. 3713200). Detailed description will be omitted.

  The envelope detection unit 37 detects only the peak of each bin in the result of frequency analysis performed by the FFT processing unit 36, and performs detection processing. That is, the result analyzed by the FFT processing unit 36 is that the X axis is the frequency, whereas the envelope detection unit 37 performs the waveform processing considering the X axis as the time axis. Therefore, the information output from the envelope detection unit 37 is information representing the sound quality tendency of the frequency (high frequency energy gradient).

  The regression analysis unit 38 performs regression analysis of information output from the envelope detection unit 37. For example, if the FFT accuracy in the FFT processing unit 36 is 2048 points, regression analysis is performed on 2048 pieces of data. There are many regression analysis methods such as linear regression and exponential regression, but any method may be used for analysis. However, assuming that the regression analysis process is performed in real time, it is preferable to use “linear regression analysis” with a small amount of calculation. In the present embodiment, linear regression analysis is used, and as a result, coefficients a and b, which are parameters necessary for the linear expression (y = ax + b), are output. In addition, the smaller the number of target data when performing regression analysis, for example, 256 is preferable because the burden on the CPU (central processing unit) in the audio signal processing device 1 can be reduced.

  The LPF 39 removes high frequency components in the coefficients a and b output from the regression analysis unit 38. Since the analysis processing by FFT is performed in a relatively short cycle, if the analysis result output from the regression analysis unit 38 is used for control as it is, the slope of the regression line changes abruptly in accordance with the change in tune, A sense of incongruity such as noise will occur. Therefore, by passing the coefficients a and b through the LPF 39, the changes of the coefficients a and b can be smoothly changed according to the tune, and the inclination of the regression line can also be changed smoothly according to the tune. In the present embodiment, the LPF 39 can be composed of a second-order IIR filter, and the transfer function is as shown in (Expression 2).

  The auditory sensitivity correction unit 40 corrects the coefficients a and b output from the LPF 39 based on the auditory parameter obtained from the auditory parameter setting unit 41.

  The auditory parameter setting unit 41 has a function of measuring a user's auditory characteristics. The measurement result of the auditory characteristic is stored as a table of three parameters Q, f, and gain. Q represents the sharpness of the characteristic line (see FIG. 7) in the high frequency emphasis characteristic. f represents the peak frequency. The method for setting the auditory parameter will be described in detail later.

  The frequency analysis unit includes a difference signal extraction unit 35, an FFT processing unit 36, and an envelope detection unit 37. The correction amount prediction unit includes a regression analysis unit 38 and an LPF 39.

  The operation will be described below.

  In FIG. 2, the PCM signal output from the audio data processing unit 12 is input to the high frequency interpolation processing unit 32 and the difference signal extraction unit 35 via the input terminal 31.

  The difference signal extraction unit 35 performs a subtraction process between the Lch signal and the Rch signal in the input stereo PCM signal to extract a difference signal (reverberation component). Next, the difference signal output from the difference signal extraction unit 35 is input to the FFT processing unit 36, where FFT processing is performed on the difference signal, and frequency analysis is performed. The result of the frequency analysis is input to the envelope detector 37. The envelope detection unit 37 detects the peak of each bin of the input analysis result and performs detection processing. The information obtained by the detection process is information that represents the sound quality tendency of the frequency. Through the above processing, the frequency characteristics of the input PCM signal can be analyzed.

  Next, the regression analysis unit 38 performs a regression analysis based on the information output from the envelope detection unit 37. FIG. 3 is a graph showing the result of frequency analysis of an input PCM signal by FFT, where the X axis represents frequency and the Y axis represents gain. When regression analysis (in this embodiment, linear regression analysis) is performed on the PCM signal shown in FIG. 3, a linear analysis result as shown in FIG. 4 is obtained. In general, since the gain of the voice tends to decrease as the frequency increases, the result of the regression analysis also has a sloped characteristic as shown in FIG.

  As shown in FIG. 4, since the coefficients a and b of the linear expression are determined based on the result of regression analysis, the degree of high frequency emphasis after high frequency interpolation can be predicted. FIG. 5 is a graph obtained by enlarging the maximum value of the X axis to 25 kHz based on the graph shown in FIG. As shown in FIG. 5, the frequency component of about 16 kHz or more is removed from the input PCM signal. FIG. 6 shows a graph obtained by adding a predicted amount to the result of the regression analysis based on the characteristics shown in FIG. That is, in FIG. 6, the characteristics up to 16 kHz are the results of regression analysis based on the characteristics shown in FIG. 5, and the characteristics above 16 kHz are the characteristics predicted based on the results of the regression analysis. As shown in FIG. 6, based on the linear coefficients a and b calculated by the regression analysis, the regression line up to 16 kHz (linear model) is extended to the high frequency side to predict the characteristics of 16 kHz or higher. Can do. An arithmetic expression for correcting the coefficients a and b is as shown in (Equation 3).

  Next, high-frequency components in the coefficients a and b of the linear expression, which are the results of regression analysis performed by the regression analysis unit 38, are removed. Thereby, since the change of the coefficients a and b can be smoothly changed according to the tune, the generation of noise can be suppressed. The coefficients a and b output from the LPF 39 are input to the auditory sensitivity correction unit 40.

  Note that the correction of the frequency characteristics based on the regression analysis as described above is executed in real time.

  On the other hand, the auditory parameter setting unit 41 sets an auditory parameter that matches the auditory characteristics of the user. The auditory parameters set by the auditory parameter setting unit 41 may be arbitrarily input by the user, or may be determined by measurement based on a predetermined auditory measurement method. An example of the auditory measurement method will be described later.

  An example of characteristics based on auditory parameters is shown in FIG. The auditory parameter is composed of Q (sharpness of characteristics shown in FIG. 7), f (peak frequency), and gain so that the gain around a predetermined frequency is slightly emphasized as shown in FIG. That is, the characteristic curve shown in FIG. 7 is determined by these three auditory parameters.

  The set three auditory parameters are output to the auditory sensitivity correction unit 40. The auditory sensitivity correction unit 40 corrects the regression analysis result (see FIG. 6) expressed by the linear expression output from the LPF 39 based on the auditory parameter output from the auditory parameter setting unit 41. The corrected frequency characteristic is shown by the solid line in FIG. In FIG. 8, the dotted line characteristic is the characteristic before correction in the auditory sensitivity correction unit 40 (that is, the characteristic shown in FIG. 6). Further, the corrected frequency characteristics shown in FIG. 8 are Q = 0.707, f = 16 kHz, and the gain is 12 dB. The auditory parameter corrected by the auditory sensitivity correction unit 40 is output to the equalizer 33.

  The equalizer 33 corrects the PCM signal output from the high-frequency interpolation processing unit 32 based on the auditory parameter output from the auditory sensitivity correction unit 40. Specifically, in order to correct the frequency characteristic based on the characteristic shown in FIG. 8, the level around 16 kHz is corrected to be emphasized as shown in FIG. FIG. 9 shows the frequency characteristics of the PCM signal output from the equalizer 33, and shows that a frequency of about 16 kHz or more is added by the interpolation process.

  The PCM signal output from the equalizer 33 is output to the digital IF 14 (FIG. 1) via the output terminal 34.

[3. Auditory parameter setting method)
Next, an example of an auditory parameter setting method will be described.

  In the present embodiment, the user has a configuration capable of measuring his / her own auditory characteristics. Since the auditory measurement function is arbitrarily executed by the user, the auditory measurement function can be shifted to the auditory measurement mode by operating the operation unit 15. Specifically, it is possible to shift to the audio measurement mode by executing software capable of performing audio measurement.

  In order to perform auditory measurement, the audio signal processing apparatus 1 includes a frequency oscillating unit 17 as shown in FIG. When the frequency oscillating unit 17 shifts to the auditory measurement mode, the frequency oscillating unit 17 sequentially generates pure sounds (sin waves) having a plurality of frequencies under the control of the control unit 16. The generated pure sound is output to the digital IF 14.

  The operation will be described below.

  In FIG. 1, when performing auditory measurement, the DAC 21 needs to be connected to the digital IF 14 of the audio signal processing apparatus 1, and the headphones 22 need to be connected to the DAC 21. Next, the operation unit 15 is operated to shift the audio signal processing device 1 to the auditory measurement mode. Specifically, the auditory measurement software is activated. When the music data recorded on the information medium 11 is being played back when the auditory measurement mode is entered, the control unit 16 causes the audio data processing unit 12 to play back the music data. Control to stop.

  Next, when auditory measurement is started, the control unit 16 controls the frequency oscillation unit 17 to generate a pure sound having a predetermined frequency. Specifically, first, a continuous pure sound having a frequency of 10 kHz and a gain of −50 dB is generated. Thereafter, the gain is increased by 5 dB every second while the frequency is kept constant. The pure sound generated in this way is output to the DAC 21 via the digital IF 14, converted into an analog signal, and then output to the headphones 22. Therefore, the headphone 22 outputs a sound with a constant frequency and a level gradually increasing from a minimum.

  The user listens to the pure sound output from the headphones 22 and operates the operation unit 15 at the time when the user hears the pure sound. In the present embodiment, the timing at which a sound is heard can be determined by operating the ENTER key of the keyboard attached to the personal computer at the timing when the pure sound is heard. Note that the operation unit 15 operated at the time of auditory measurement is not limited to a keyboard, and a dedicated push button may be provided. In that case, a push button may be provided on the cable of the headphones 22.

  Next, the value of the pure sound gain (for example, −20 dB) at the time when the operation unit 15 is operated is recorded in the memory included in the frequency oscillation unit 17. That is, the lowest level (audible gain) at which a 10 kHz sound can be heard is recorded in the memory. Therefore, a person who has an excellent ability to hear a 10 kHz signal has a low audible gain value.

  Next, the control unit 16 controls the frequency oscillating unit 17 to generate a continuous pure sound of 12 kHz. The frequency oscillating unit 17 generates a pure sound having a frequency of 12 kHz and a gain of −50 dB, Thereafter, the gain is increased by 5 dB every second. As described above, the user operates the operation unit 15 at a timing when a pure sound is heard from the headphones 22. The gain of pure sound (for example, −14 dB) when the operation unit 15 is operated is recorded in the memory in the frequency oscillation unit 17.

  Thereafter, as described above, measurement is performed with pure sounds of 14 kHz, 16 kHz, 18 kHz, and 20 kHz.

  The level of the pure sound output from the frequency oscillating unit 17 is gradually increased from −50 dB to 0 dB. If the operating unit 15 is not operated by the user even though 0 dB is output, It is determined that the user cannot hear the sound having the frequency of, and the value of the audible gain is not recorded in the memory. That is, the result is “impossible to measure”.

  Thereby, the audible gain at each frequency of 12 to 20 kHz is recorded in the memory. An example of the measurement results is shown in (Table 1).

  The person to be measured who gave the results shown in (Table 1) can hear low-frequency sounds well, but it is difficult to hear high-frequency sounds. Will be.

  Auditory parameters are set based on the result of the auditory measurement obtained as described above. Specifically, as shown in FIG. 7, three auditory parameters of Q (sharpness of characteristics shown in FIG. 7), f (peak frequency), and gain are set so that the gain around 16 kHz is slightly emphasized. That is, the characteristic curve shown in FIG. 7 is determined by these three auditory parameters. The frequency characteristics shown in FIG. 7 are Q = 0.707, f = 16 kHz, and gain = 12 dB.

  The set three auditory parameters are output to the auditory sensitivity correction unit 40. The auditory sensitivity correction unit 40 corrects the regression analysis result (see FIG. 6) expressed by the linear expression output from the LPF 39 based on the auditory parameter output from the auditory parameter setting unit 41. FIG. 8 shows the frequency characteristics after correction (solid line). In FIG. 8, the dotted line characteristic is the characteristic before correction in the auditory sensitivity correction unit 40 (that is, the characteristic shown in FIG. 6). The auditory parameter corrected by the auditory sensitivity correction unit 40 is output to the equalizer 33.

  The equalizer 33 corrects the PCM signal output from the high-frequency interpolation processing unit 32 based on the auditory parameter output from the auditory sensitivity correction unit 40. Specifically, in order to correct the frequency characteristics based on the characteristics shown in FIG. 8, the level around 16 kHz is corrected to be emphasized.

  The PCM signal output from the equalizer 33 is output to the digital IF 14 (FIG. 1) via the output terminal 34.

[4. Effects of the embodiment, etc.]
As described above, according to the present embodiment, the auditory parameter setting unit 41 sets the auditory parameter according to the user's auditory characteristics, and corrects the high frequency of the PCM signal in the equalizer 33 based on the auditory parameter. Therefore, optimum high frequency characteristics can be obtained for each user. Therefore, the user can enjoy music with comfortable sound quality.

  In addition, the regression analysis unit 38 grasps the tune of the input music data, and corrects the slope of the regression line according to the tune to be grasped, so that equalization adapted to music information (music tune, rhythm, etc.) is possible. Thus, high-frequency interpolation with natural and comfortable sound quality can be realized.

  In addition, by combining interpolation frequency prediction by regression analysis and user's auditory characteristics, equalization that is suitable for music information is possible, not just high-frequency emphasis equalization, and high frequency with more natural and comfortable sound quality. Interpolation can be realized.

  That is, generally, when listening to music, the human brain listens while predicting the frequency gradient of the music information, so that the frequency is corrected in accordance with the user's auditory characteristics as in this embodiment. In this way, the user can recognize the correction feeling moderately and enjoy music with pleasant sound quality. For example, in the case of a musical instrument that emphasizes the high range, such as a cymbal, if the sound continues to some extent, it becomes possible to recognize comfort such as reverberation. That is, a pleasant sound is effective for a sound source that continues continuously. In the present embodiment, a sound with a feeling of reverberation can be created by performing frequency enhancement on such a sound. Therefore, sound quality that is natural and easy to hear can be realized by correcting the high frequency sound as predicted by the regression analysis. Furthermore, a clear sound can be realized by correcting a higher frequency than expected.

  In addition, the reverberation of sound has a proportional relationship with the amount of reverberant sound. Therefore, in this embodiment, since the difference signal between Lch and Rch extracted by the difference signal extraction unit 35 is a reverberant sound, the amount of reverberation is predicted by measuring the amount of the difference signal. be able to. Then, by performing the high frequency correction prediction based on the envelope of the difference signal, the high frequency correction can be performed without deteriorating the feeling of sound.

  In addition, there are individual differences in high-frequency sensitivity, and even if the degree of high-frequency emphasis is the same, it may be felt that the high-frequency emphasis is excessive or high-frequency emphasis is not noticed. Therefore, in this embodiment, the user's auditory frequency characteristics are measured, and the high frequency correction amount is suppressed for a user with high sensitivity in the high frequency range, and the normal for the user with low sensitivity in the high frequency range. By performing high-frequency correction (equalizing) that is higher than the correction, it is possible to achieve sound quality that matches individual differences in high-frequency sensitivity.

  In the present embodiment, since high-frequency correction is predicted from the frequency envelope of the PCM signal, dynamic sound quality control according to the state of music (musical tone, etc.) is performed instead of conventional tone control. Can do. Therefore, it is possible to output a natural sound without a sense of incongruity.

  In the present embodiment, the LPF 39 is arranged so that the change of the coefficients a and b of the regression line (primary expression) in the regression analysis unit 38 can be smoothly changed according to the tone of the curve, and the slope of the regression line Can be changed smoothly according to the tune. Therefore, it is possible to prevent noise that occurs when the regression line changes abruptly according to the tone of music.

  Further, the high-frequency interpolation according to the present embodiment can realize the correction effect more strongly in an audio output device whose reproduction frequency band is widely covered up to a high frequency such as headphones.

  The frequency generated from the frequency oscillating unit 17 at the time of auditory measurement is not limited to the range of 10 to 20 kHz shown in (Table 1), and may be measured with other values. Further, the number of samples for auditory measurement is not limited to six shown in (Table 1), and may be seven or more. By increasing the number of samples, the auditory sensitivity can be measured more finely, and the high-frequency correction can be performed more finely.

  Further, the auditory measurement method shown in the present embodiment is an example, and a configuration in which auditory parameters are set by other methods may be used. For example, an auditory parameter corresponding to a combination of information such as age, sex, and living environment is tabulated in advance as a specified value, and the user inputs information such as his / her age and gender to the input information. A method may be used in which optimal hearing parameters are selected from a table. Alternatively, the method may be such that the user inputs the real number of the auditory parameter.

  The auditory parameter measured by the above measuring method may be recorded in a memory for each user. With this configuration, if you share a computer with multiple users, you only need to read your own auditory parameters when you listen to the music, so each time you listen to the music, you make an auditory measurement Eliminates the need for improved usability.

  In the present embodiment, the high-frequency correction of music data in the MP3 format has been described. However, high-frequency correction can be similarly performed for music data in a format other than the MP3 format. In particular, it is possible to achieve high sound quality in music data from which high frequency components have been removed by compression.

  In addition, the information medium in the present embodiment is not indispensable. For example, when reproducing music data acquired from the network, streaming playback of an audio stream acquired from the network, or receiving digital radio broadcasting, the information medium can be obtained. The present embodiment is also effective when reproducing the recorded audio data. In particular, it is possible to achieve high sound quality in music data from which high frequency components have been removed by compression.

  In the present embodiment, the audio signal processing device 1 (personal computer) includes a configuration that performs high frequency correction based on auditory characteristics, such as the high frequency correction unit 13 and the frequency oscillation unit 17. It is good also as a structure with which external apparatuses, such as are equipped. In the case of such a configuration, an external device is provided with an operation unit that is operated by the user during auditory measurement. With this configuration, it is not necessary to install software for performing high frequency correction on a personal computer, so that the usability can be improved. Moreover, it can connect and use not only a personal computer but various audio equipment.

  The DAC 21 is configured as a unit independent of the headphones 22, but may be configured to be incorporated in the headphones 22. In that case, the headphone 22 includes a USB terminal (male pin) instead of the stereo mini plug.

  The DAC 21 is configured as a unit independent of the audio signal processing device 1, but may be configured to be incorporated in the audio signal processing device 1. In this case, the digital IF 14 is configured by an analog IF such as a stereo mini jack. For example, when the audio signal processing device 1 is included in a portable music player, the audio signal processing device 1 includes a DAC 21 and a stereo mini jack as an audio output terminal. Even in such a configuration, it is possible to realize high-frequency correction based on the auditory characteristic that is a feature of the present embodiment.

  The audio signal processing apparatus of the present invention is useful for a device capable of reproducing music signals and music data. For example, it is useful for amplifiers, speakers, headphones, personal computers, mobile phones, portable music players, and the like.

The block diagram which shows the structure of the audio | voice signal processing apparatus in Embodiment 1 of this invention. FIG. 3 is a block diagram showing a configuration of a high frequency correction unit in the first embodiment. Characteristic diagram of input signal in linear regression calculation in embodiment 1 Characteristic diagram of regression line in linear regression calculation in embodiment 1 Characteristic diagram of input signal in linear regression calculation in embodiment 1 Characteristic diagram of regression line after high-frequency interpolation in primary regression calculation in Embodiment 1 Characteristic diagram showing an example of frequency emphasis characteristics based on auditory measurement results in the first embodiment The characteristic view which shows the regression line which added the auditory measurement result in Embodiment 1 Characteristic diagram showing frequency characteristics after high-frequency interpolation in the first embodiment

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Audio | voice signal processing apparatus 11 Information medium 12 Audio | voice data processing part 13 High region correction | amendment part 14 Digital IF
15 Operation unit 16 Control unit 17 Frequency oscillation unit 21 DAC
22 Headphones 32 High-frequency interpolation processing unit 33 Equalizer 35 Difference signal extraction unit 36 FFT processing unit 37 Envelope detection unit 38 Regression analysis unit 39 LPF
40 Auditory sensitivity correction unit 41 Auditory parameter setting unit

Claims (8)

An audio signal processing device including a high frequency correction unit including a high frequency interpolation processing unit for interpolating a high frequency component of an input audio signal,
The high-frequency correction unit is
A frequency analyzer that analyzes the frequency characteristics of the input audio signal;
Detecting a tendency of the frequency characteristic analyzed by the frequency analysis unit, predicting a correction amount in a high range, and outputting a correction result;
An auditory parameter setting unit capable of setting an auditory parameter based on the user's auditory characteristics;
An auditory sensitivity correction unit that corrects a correction result output from the correction amount prediction unit using an auditory parameter set in the auditory parameter setting unit;
An equalizer that corrects a frequency characteristic of an audio signal output from the high-frequency interpolation processing unit;
The equalizer corrects a frequency characteristic of an audio signal based on a correction result output from the auditory sensitivity correction unit. An audio signal processing device including a high frequency correction unit including a high frequency interpolation processing unit for interpolating a high frequency component of an input audio signal,
The high-frequency correction unit is
A frequency analyzer that analyzes the frequency characteristics of the input audio signal;
A correction amount prediction unit that performs regression analysis of the frequency analysis result in the frequency analysis unit to derive a regression analysis result, reflects the regression analysis result to a high frequency side, predicts a high frequency correction amount, and outputs a correction result When,
An equalizer that corrects a frequency characteristic of an audio signal output from the high-frequency interpolation processing unit;
The equalizer corrects a frequency characteristic of an audio signal based on a correction result output from the correction amount prediction unit. The frequency analysis unit
In the input stereo audio signal, a difference signal extraction unit that subtracts the audio signal of the left channel and the right channel and extracts a difference signal;
A fast Fourier transform processing unit for performing a frequency analysis by performing a fast Fourier transform on the difference signal output from the difference signal extraction unit;
The audio signal processing device according to claim 1, further comprising an envelope detection unit that detects a peak of a frequency analysis result output from the fast Fourier transform processing unit. The correction amount prediction unit
The audio signal processing apparatus according to claim 1, further comprising a regression analysis unit that performs regression analysis on a frequency analysis result in the frequency analysis unit. The correction amount prediction unit
The audio signal processing apparatus according to claim 2, further comprising a filter that removes a high-frequency component of fluctuation of the regression analysis result in the regression analysis unit. It has a normal mode and an audio measurement mode that can measure the user's audio characteristics,
The auditory parameter setting unit
A frequency oscillating unit that generates a pure sound set at a predetermined frequency in advance while changing the level;
A memory for recording a frequency and a level when the operation unit is operated when generating a pure sound in the frequency oscillating unit;
When the system enters the auditory measurement mode,
The frequency oscillator generates the pure sound while changing the level,
The generated pure sound is output from the output terminal,
When the operation unit is operated by a user when the pure sound is being output, the frequency and level of the pure sound at that time are recorded in the memory,
Set auditory parameters based on frequency and level information recorded in the memory;
The audio signal processing apparatus according to claim 1. A high-frequency correction unit having a high-frequency interpolation processing unit for interpolating a high-frequency component of the input audio signal;
An output terminal capable of outputting the audio signal output from the high frequency correction unit to the outside;
An audio signal reproduction system including an operation unit capable of inputting various operation commands in the system,
The high-frequency correction unit is
A frequency analyzer that analyzes the frequency characteristics of the input audio signal;
Detecting a tendency of the frequency characteristic analyzed by the frequency analysis unit, predicting a correction amount in a high range, and outputting a correction result;
An auditory parameter setting unit capable of setting an auditory parameter based on the user's auditory characteristics;
An auditory sensitivity correction unit that corrects a correction result output from the correction amount prediction unit using an auditory parameter set in the auditory parameter setting unit;
An equalizer that corrects a frequency characteristic of an audio signal output from the high-frequency interpolation processing unit;
The equalizer corrects frequency characteristics of an audio signal based on a correction result output from the auditory sensitivity correction unit. It has a normal mode and an audio measurement mode that can measure the user's audio characteristics,
The auditory parameter setting unit
A frequency oscillating unit that generates a pure sound set at a predetermined frequency in advance while changing the level;
A memory for recording a frequency and a level when the operation unit is operated when generating a pure sound in the frequency oscillating unit;
When the system enters the auditory measurement mode,
The frequency oscillator generates the pure sound while changing the level,
The generated pure sound is output from the output terminal,
When the operation unit is operated by a user when the pure sound is being output, the frequency and level of the pure sound at that time are recorded in the memory,
Set auditory parameters based on frequency and level information recorded in the memory;
The audio signal reproduction system according to claim 7.

Images