JP2011186187A - Speech processor, speech processing method and speech processing program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To create a digital speech signal which is closer to original sound by suitably adding a high-frequency component according to an acquired digital speech signal. <P>SOLUTION: A speech processor 100 includes: a signal analysis section 124 which determines whether or not the high-frequency component of a predetermined frequency or more and a predetermined sound pressure or more is included in the digital speech signal; a correction value creation section 130 which creates a correction value for expanding an amplitude of the digital speech signal based on a coefficient that is different depending on whether or not the signal analysis section determines that the high-frequency component is included in the digital speech signal; and an addition section 134 for adding the correction value to the digital speech signal. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a sound processing apparatus, a sound processing method, and a sound processing program for analyzing a digital sound signal and processing the digital sound signal according to the analysis result.

  In recent years, due to advances in audio coding technology, it has become possible to reduce the file size while maintaining the sound quality of music recorded on CDs (Compact Discs) as much as possible. A large amount of music can be recorded and carried.

  However, the above-described speech coding technology cuts out high-frequency band speech signals that are not normally audible using human auditory characteristics, or thins out unacceptable sound data due to the masking effect. Compared to the original sound before the sound, it will be less stretched, spread, dynamic range and glossy. Therefore, a technique for improving the sound quality of a digital audio signal compressed by an audio encoding technique has been developed.

  For example, the inventor of the present invention sets a value obtained by multiplying a difference value between an extreme value sample of a digital audio signal and a sample immediately before the extreme value by a coefficient corresponding to the number of samples between the extreme values in the digital audio signal. A technique for adding a high-frequency component equal to or higher than a predetermined frequency to a digital audio signal by addition is proposed (for example, Patent Documents 1 and 2).

Japanese Patent No. 3401171 Japanese Patent No. 3659489

  The sound quality improvement processing for improving the sound quality by adding a high frequency component as in the techniques of Patent Documents 1 and 2 described above includes the CD standard, MPEG (Moving Picture Expert Group) -2, AAC (registered trademark). (Advanced Audio Coding), ATRAC (registered trademark) (Adaptive TRansform Acoustic Coding), MP3 (MPEG Audio Layer-3), WMA (Windows (registered trademark) Media Audio), etc. And can be applied in various scenes where decoding is performed. For this reason, the sound quality improvement processing may be performed on the digital audio signal that has already been subjected to the sound quality improvement processing.

  However, if a sound quality improvement process equivalent to a normal digital sound signal that has not been subjected to sound quality improvement processing is applied to a digital sound signal that has already been subjected to sound quality improvement processing, an excessive amount of high frequency components will be added, For example, the balance of the mid- and high-frequency ranges of the reproduced sound changes, resulting in a sound that is far from the original sound.

  In view of such a problem, the present invention provides a sound processing device, a sound processing method, and sound that can generate a digital sound signal closer to the original sound by appropriately adding a high frequency component according to the acquired digital sound signal. The purpose is to provide a processing program.

  In order to solve the above-described problems, a speech processing apparatus according to the present invention includes a signal analysis unit that determines whether or not a high-frequency component having a predetermined frequency or higher and a predetermined sound pressure or higher is included in a digital audio signal; A correction value generation unit that generates a correction value that expands the amplitude of the digital audio signal based on a coefficient that differs depending on whether or not the digital audio signal includes a high-frequency component, and digital audio And an addition unit for adding a correction value to the signal.

  The signal analysis unit has a smaller number of samples from an arbitrary extreme value to the next extreme value of the digital audio signal than a predetermined number, and is the larger extreme value between the arbitrary extreme value and the next extreme value. If the difference value between the extreme values, which is the difference between the local maximum value and any local extreme value or the next local extreme value, whichever is the smaller extreme value, exceeds a predetermined value, the digital audio signal It may be determined that a high frequency component is included.

  The signal analysis unit is configured to select any one of an extreme value and a next extreme value for a plurality of samples in which the number of samples from an arbitrary extreme value to the next extreme value of the digital audio signal is less than a predetermined number. The difference value between extreme values that is the difference between the maximum value that is the larger extreme value and the minimum value that is the smaller extreme value of any extreme value or the next extreme value exceeds the predetermined value. If the occupation ratio occupied by a certain sample exceeds a predetermined ratio, it may be determined that a high frequency component is included in the digital audio signal.

  The correction value generation unit is smaller than when the signal analysis unit determines that the digital audio signal does not include a high frequency component when the signal analysis unit determines that the digital audio signal includes a high frequency component. A correction value may be generated based on the coefficient.

  The correction value generation unit may generate a correction value based on different coefficients depending on the format of the digital audio signal.

  The signal analysis unit may determine a predetermined number and a predetermined value based on a format of the digital audio signal.

  The audio processing device includes a first conversion unit that upsamples a digital audio signal acquired by the audio processing device, and a digital audio signal after the addition unit adds a correction value before the first conversion unit upsamples the digital audio signal. And a second conversion unit that down-samples to the sampling frequency.

  The correction value generation unit multiplies the maximum difference value, which is the difference between the maximum value of the digital audio signal and the value of the sample one sample before the sample having the maximum value, by the coefficient, and the minimum of the digital audio signal. A correction value is generated by multiplying the minimum difference value, which is the difference between the value and the value of the sample one sample before the sample that is the minimum value, by a coefficient, and the correction value generated based on the maximum difference value is The correction value generated based on the minimum difference value may be associated with the sample having the minimum value so that the correction value generated based on the minimum difference value is subtracted from the minimum value so as to be added to the maximum value.

  The correction value generation unit multiplies the difference between the maximum value of the sample one sample before and one sample after the maximum value of the digital audio signal by the coefficient, and 1 of the minimum value of the digital audio signal. A correction value is generated by multiplying the difference between the value of each of the sample before the sample and the sample after one sample and the minimum value by a coefficient, and the difference between the value of the sample one sample before the maximum value and the maximum value is obtained. Corresponding to the sample one sample before the maximum value so that the correction value generated based on the sample value one sample before the maximum value is added, the value of the sample one sample after the maximum value and the maximum value The correction value generated based on the difference is associated with the sample after one sample of the maximum value so that it is added to the value of the sample after one sample of the maximum value, and one sample before the minimum value Corresponding to the sample one sample before the minimum value so that the correction value generated based on the difference between the sample value and the minimum value is subtracted from the value of the sample one sample before the minimum value, one sample of the minimum value The correction value generated based on the difference between the value of the subsequent sample and the minimum value may be associated with the sample after one sample of the minimum value so as to be subtracted from the value of the sample after one sample of the minimum value.

  In order to solve the above problems, the audio processing method of the present invention determines whether or not a high-frequency component having a predetermined frequency or higher and a predetermined sound pressure or higher is included in the digital audio signal, and the high-frequency component is included in the digital audio signal. A correction value that expands the amplitude of the digital audio signal is generated based on a coefficient that differs depending on whether or not the signal is included, and the correction value is added to the digital audio signal.

  In order to solve the above-described problem, the audio processing program of the present invention includes a step of determining whether a digital audio signal includes a high frequency component having a frequency equal to or higher than a predetermined frequency and higher than a predetermined sound pressure. Generating a correction value for enlarging the amplitude of the digital audio signal based on different coefficients depending on whether or not the signal contains a high frequency component; and adding the correction value to the digital audio signal; Is executed.

  As described above, according to the present invention, it is possible to generate a digital audio signal closer to the original sound by appropriately adding a high frequency component according to the acquired digital audio signal.

It is explanatory drawing which showed the utilization condition of the audio processing apparatus. It is a functional block diagram for demonstrating the whole structure of a speech processing unit. It is explanatory drawing for demonstrating the relationship between the frequency corresponding to the number of samples from an arbitrary extreme value to the next extreme value, and the half-cycle digital audio signal including the arbitrary extreme value and the next extreme value. is there. It is explanatory drawing for demonstrating the process of a signal analysis part. It is explanatory drawing for demonstrating the process which a signal analysis part judges whether the high frequency component is contained in the digital audio | voice signal according to whether the occupation rate exceeds the predetermined ratio. It is explanatory drawing which shows an example of the format of a digital audio | voice signal, and a sample number threshold value and an extreme value level threshold value. It is explanatory drawing for demonstrating a coefficient table group and a coefficient table. It is explanatory drawing for demonstrating in more detail the process which adds a correction signal to a digital audio | voice signal. It is explanatory drawing for demonstrating the other process which adds a correction signal to a digital audio | voice signal. It is explanatory drawing for demonstrating the other process which adds a correction signal to a digital audio | voice signal. It is explanatory drawing for demonstrating the process of a 1st conversion part. It is the functional block diagram which showed the typical example of the computer (information processing apparatus) in which the sound quality improvement process by an audio processing apparatus is possible. It is the flowchart which showed the flow of the whole process of the audio | voice processing method. It is the flowchart which showed the flow of the whole process of the audio | voice processing method.

  Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. The dimensions, materials, and other specific numerical values shown in the embodiments are merely examples for facilitating the understanding of the invention, and do not limit the present invention unless otherwise specified. In the present specification and drawings, elements having substantially the same function and configuration are denoted by the same reference numerals, and redundant description is omitted, and elements not directly related to the present invention are not illustrated. To do.

(Speech processor 100)
FIG. 1 is an explanatory diagram for explaining a usage state of the speech processing apparatus 100. The audio processing apparatus 100 acquires a digital audio signal through broadcast waves from the broadcast station 102, through the communication network 106 from the content server 104, or directly from the storage medium 108, and adds a high frequency component to the digital audio signal. To improve the sound quality of digital audio signals. The user can listen to the improved digital audio signal directly from the audio processing device 100 or by transferring it to a playback device 110 such as a portable audio player or a mobile phone.

  In addition, the content server 104 may include the audio processing device 100. In this case, the audio signal to which the high frequency component is added by the audio processing device 100 of the content server 104 is transmitted via the communication network 106 to a personal computer or a portable computer. It is distributed to a playback device 110 such as an audio player or a mobile phone.

  Further, the playback device 110 such as a portable audio player or a mobile phone may include the sound processing device 100. In this case, the digital audio signal distributed from the content server 104 through the communication network 106 is reproduced with a high frequency component added by the audio processing device 100 of the reproduction device 110 such as a portable audio player or a mobile phone.

  Examples of digital audio signals that can be improved by the audio processing apparatus 100 include audio signals based on CD and DVD (Digital Versatile Disc) standards, MPEG-2, AAC (registered trademark), HE-AAC, and ATRAC (registered). (Trademark), MP3, WMA, and the like.

  Normally, when an audio signal is digitized, the audio signal is limited to a frequency component equal to or less than half the sampling frequency. Further, when the digitized audio signal (digital audio signal) passes through the communication network 106, it may be subjected to compression processing in order to reduce the communication load on the communication network 106. Therefore, a digital audio signal or a digital audio signal that has been digitized and further compressed does not have a high frequency component, and the reproducibility of the original sound is poor. Therefore, a sound quality improvement process for adding a high frequency component may be performed. The original sound is an audio signal before being digitized.

  This sound quality improvement process not only adds high-frequency components after digitization or compression processing in order to approximate the original sound, but also predicts that high-frequency components will be damaged by compression processing, for example, after compression processing This includes the case where a high frequency component is added to the original sound in advance so that an appropriate high frequency component is included.

  Therefore, the digital audio signal acquired by the audio processing device 100 may or may not be subjected to sound quality improvement processing. Under such circumstances, when the audio processing device 100 uniformly adds a high frequency component to the digital audio signal, if the high frequency component has already been added to the digital audio signal, the high frequency component is excessively added. In other words, the sound is far away from the original sound, for example, the balance of the mid-high range changes.

  The sound processing apparatus 100 according to the present embodiment appropriately adds a high-frequency component to the acquired digital sound signal according to whether sound quality improvement processing has already been performed, so that a digital sound signal closer to the original sound is generated. It becomes possible to do. Hereinafter, a detailed configuration of the sound processing apparatus 100 will be described.

  FIG. 2 is a functional block diagram for explaining the overall configuration of the speech processing apparatus 100. The audio processing device 100 includes a signal acquisition unit 120, an extreme value identification unit 122, a signal analysis unit 124, a table selection unit 126, a coefficient storage unit 128, a correction value generation unit 130, a delay unit 132, and an addition. Unit 134, signal output unit 136, first conversion unit 138, and second conversion unit 140.

  The sound processing apparatus 100 according to the present embodiment performs a selection process of selecting a coefficient table depending on whether or not a high frequency component is included in the acquired digital sound signal, and then performs correction using the selected coefficient table. Process. Hereinafter, the selection process and the correction process will be described in order.

(Selection process)
In the selection process, first, the signal acquisition unit 120 acquires a digital audio signal. Then, the signal acquisition unit 120 performs the acquisition of the acquired digital audio signal based on the header information of the digital audio signal and the information related to the standard of the digital audio signal such as CD, AAC (registered trademark), ATRAC (registered trademark), and MP3. Specify the number of quantization bits and sampling frequency. However, when the input digital audio signal is always a digital audio signal of the same format as in a CD player, the signal acquisition unit 120 does not have a function of specifying the number of quantization bits and the sampling frequency. Also good.

  The first conversion unit 138 upsamples the digital audio signal acquired by the signal acquisition unit 120 and outputs it to the extreme value specifying unit 122. Note that up-sampling is not essential, and the first converter 138 may not be included.

  The extreme value specifying unit 122 specifies a local maximum value and a local minimum value that are extreme values of the digital audio signal acquired by the signal acquisition unit 120. Specifically, the extreme value specifying unit 122 sequentially compares the sample values (sound pressure values of the samples) of the digital audio signal, and when the sample value changes from increasing or not increasing to decreasing, immediately before it starts decreasing. When the sample value is set to a maximum value and the sample value is changed from no decrease or increase / decrease to an increase, the sample value immediately before the sample value is increased is set to a minimum value. Then, the extreme value specifying unit 122 counts the number of samples from an arbitrary extreme value to the next extreme value, that is, the number of samples from the local maximum value to the local minimum value, or the number of samples from the local minimum value to the local maximum value. From this number of samples, the frequency corresponding to the part from the extreme value where the number of samples is counted to the next extreme value in the digital audio signal is known.

  FIG. 3 illustrates the relationship between the number of samples from an arbitrary extreme value to the next extreme value, and the frequency corresponding to the half-period digital audio signal including the arbitrary extreme value and the next extreme value. It is explanatory drawing of. 3A shows a case where the sampling frequency is 44.1 kHz, and FIG. 3B shows a case where the sampling frequency is 96 kHz. As shown in FIGS. 3 (a) and 3 (b), the smaller the number of samples from any extreme value to the next extreme value, the shorter the half-cycle digital audio signal, the shorter the waveform period, and the higher the frequency band. It can be said that it is a signal. For example, in the CD standard, since the sampling frequency is 44.1 kHz, the relationship shown in FIG. 3A is used, and when the number of samples from one extreme value to the next extreme value is 1, any extreme value And a half-cycle digital audio signal including the next extreme value is a signal included in the frequency band of 11.0525 to 22.050 kHz, and the number of samples from any extreme value to the next extreme value is 2. In this case, a half-cycle digital audio signal including an arbitrary extreme value and the next extreme value is a signal included in the frequency band of 7.35 to 11.025 kHz.

  The signal analysis unit 124 includes, for example, a high-frequency component equal to or higher than a predetermined frequency and equal to or higher than a predetermined sound pressure in the digital audio signal for a predetermined period after the signal acquisition unit 120 starts acquiring the digital audio signal. Determine whether or not. The predetermined period may be a preset period, or may be a period for one piece of music as long as the digital audio signal is an audio signal that can be divided into predetermined time units such as music. The digital audio signal may include a noise signal generated at the time of AD conversion or transmission, and this noise signal also includes a frequency component of a predetermined frequency or higher. Since this noise signal has a low sound pressure, in the present embodiment, the high frequency component is distinguished from the noise signal as a frequency component having a predetermined frequency or higher and a predetermined sound pressure or higher. The predetermined sound pressure is, for example, an intermediate sound pressure between the average sound pressure of the noise signal and the average sound pressure of the high frequency component of the digital audio signal excluding the noise signal, or a predetermined multiple of the average sound pressure of the noise signal. Or sound pressure.

  Specifically, the signal analysis unit 124 counts the number of samples from an arbitrary extreme value of the digital audio signal to the next extreme value counted by the extreme value specifying unit 122 less than a predetermined number (hereinafter referred to as a sample number threshold value). , The level of the local maximum that is the larger of any extreme and the next extreme, and the local minimum that is the smaller of either the extreme and the next extreme The difference value between the extreme values, which is the difference between the value level (the difference between any extreme value and the next extreme value, or the difference between the next extreme value and the arbitrary extreme value) is a predetermined value (hereinafter, extreme value) If it exceeds the level threshold), it is determined that the digital audio signal contains a high frequency component.

  The sample number threshold is a threshold for determining whether or not a half-cycle digital audio signal including an arbitrary extreme value and the next extreme value is an audio signal having a predetermined frequency or higher, and is acquired digitally. It is determined by the audio signal standard, sampling frequency, and the like. If the digital audio signal acquired like a CD player is always a digital audio signal having the same sampling frequency, the sample number threshold may always be the same value. When digital audio signals with various sampling frequencies and various standards are acquired as in a personal computer, the signal analysis unit 124 sets a sample number threshold according to the sampling frequency and standards. Specifically, the signal analysis unit 124 detects the sampling frequency of the acquired digital audio signal based on the header information and information on the standard, and the number of samples to which the frequency component higher than the frequency component included in the sampling frequency corresponds. Is the sample number threshold.

  The extreme level threshold is a threshold for determining whether or not a half-cycle digital audio signal including an arbitrary extreme value and the next extreme value is equal to or higher than a predetermined sound pressure. This is a threshold value for exclusion.

  FIG. 4 is an explanatory diagram for explaining the processing of the signal analysis unit 124. FIG. 4A shows the frequency spectrum of the digital audio signal 150 that has been subjected to the sound quality improvement process and the frequency spectrum of the digital audio signal 152 that has not been subjected to the sound quality improvement process superimposed, and FIG. FIG. 4C shows only the frequency spectrum of the digital audio signal 152 that has not been subjected to the sound quality improvement process, and FIG. 4C shows only the frequency spectrum of the digital audio signal 152 that has not been subjected to the sound quality improvement process.

  As shown in FIG. 4A, the digital audio signal 150 that has been subjected to the sound quality improvement process is a high frequency component that is not included in the digital audio signal 152 that has not been subjected to the sound quality improvement process (FIG. 4A). (Indicated by hatching).

  Therefore, for example, when the frequency component 154 in the range of the frequency f1 or higher (substantially in the range of the frequencies f1 to f2) is the sound pressure higher than the predetermined sound pressure, the frequency component of the digital signal 152 shown in FIG. Instead, it can be determined that the high-frequency component is included only in the digital audio signal 150 that has been subjected to the sound quality improvement processing shown in FIG. In other words, the number of samples from any extreme value of the digital audio signal to the next extreme value is less than the sample number threshold, and the difference value between extreme values that is the difference between the arbitrary extreme value and the next extreme value is the extreme value level. When the threshold value is exceeded, the digital audio signal has a frequency component that should not be included unless sound quality improvement processing is performed, and is a digital sound signal that has been subjected to sound quality improvement processing in the past. I can judge.

  Here, the signal analysis unit 124 has a predetermined number of samples in which the number of samples from any extreme value of the digital audio signal to the next extreme value is less than the sample number threshold and the difference value between extreme values exceeds the extreme level threshold. If at least one digital audio signal is included, it is determined that the digital audio signal includes a high frequency component and has been subjected to sound quality improvement processing. However, the signal analysis unit 124 is not limited to such a case. For example, if the number of such samples exceeds a predetermined threshold, the digital audio signal includes a high-frequency component and is subjected to sound quality improvement processing. The digital audio signal may be determined.

  When the number of samples from any extreme value to the next extreme value is less than the sample number threshold value and the difference value between extreme values exceeds the extreme value level threshold value, the signal analysis unit 124 adds the digital audio signal to a predetermined frequency and a predetermined frequency. It is determined that a high frequency component higher than the sound pressure is included. The signal analysis unit 124 determines whether or not a high frequency component is included in the acquired digital audio signal, the sample number threshold for specifying the frequency band, the number of samples from an arbitrary extreme value to the next extreme value, and the like. The speech processing apparatus 100 determines whether the high frequency component is a noise signal or not, and the comparison between the extreme value difference value and the extreme value level threshold for identifying whether the high frequency component is a noise signal. It is possible to perform sound quality improvement processing while suppressing an increase in load.

  Further, the signal analyzer 124 compares the difference value between the extreme values and the extreme value level threshold value, so that the signal analyzer 124 can detect the noise signal even though the digital audio signal has not been subjected to the sound quality improvement processing yet. It is possible to avoid a situation in which it is erroneously determined that the sound quality improvement processing is performed due to the influence of the above.

  Next, another example of a method for determining whether or not a digital audio signal is a digital audio signal that has already undergone sound quality improvement processing will be described. In this example, the signal analysis unit 124, for example, for a digital audio signal acquired by the signal acquisition unit 120 in a predetermined period, the whole of a plurality of samples in which the number of samples from any extreme value to the next extreme value is less than the sample number threshold. On the other hand, if the occupation ratio occupied by the samples whose difference value between extreme values exceeds the extreme level threshold exceeds a predetermined ratio, it is determined that a high frequency component is included in the digital audio signal. Here, the occupation ratio refers to the ratio of the number of samples in which the difference value between extreme values exceeds the extreme level threshold with respect to the total number of a plurality of samples whose number of samples is smaller than the sample number threshold.

  FIG. 5 is a diagram for explaining processing performed by the signal analysis unit 124 to determine whether or not a high frequency component is included in a digital audio signal depending on whether or not the occupation ratio exceeds a predetermined ratio. FIG. Similar to FIG. 4A, FIG. 5A superimposes the frequency spectrum of the digital audio signal 156 that has been subjected to the sound quality improvement process and the frequency spectrum of the digital audio signal 158 that has not been subjected to the sound quality improvement process. 5B shows only the frequency spectrum of the digital audio signal 156 that has been subjected to the sound quality improvement processing, and FIG. 5C shows only the frequency spectrum of the digital audio signal 158 that has not been subjected to the sound quality improvement processing. Is shown.

  Also in FIG. 5A, the frequency spectrum 156 of the digital audio signal subjected to the sound quality improvement process is a high frequency component not included in the frequency spectrum 158 not subjected to the sound quality improvement process (FIG. 5A). (Indicated by hatching). However, the frequency band of the high frequency component by the sound quality improvement processing is narrower than the frequency band (frequencies f3 to f4) serving as a criterion, and the frequencies f3 to f4 that can be specified by the number of samples from an arbitrary extreme value to the next extreme value. When the range includes a high-frequency component by the sound quality improvement process and a frequency component other than the high-frequency component by the sound quality improvement process, the sound quality is simply compared with the extreme level threshold described with reference to FIG. It may not be possible to determine whether or not improvement processing has been performed. In the following description regarding FIG. 5, the sampling frequency of the digital audio signal is 44.1 kHz, and the range of frequencies f3 to f4 is 11.0525 kHz to 22.050 kHz. In addition, the sample number threshold is 2.

  As shown in FIG. 5, a frequency component 160 indicated by a half-cycle digital audio signal including a certain extreme value and the next extreme value is a frequency spectrum of the digital audio signal subjected to the sound quality improvement processing in FIG. If the frequency spectrum 158 that is not subjected to the sound quality improvement process in FIG. 5C is also included in the frequency spectrum 158, the sound quality has already been improved only because the digital audio signal has a frequency component included in the range of the frequencies f3 to f4. It cannot be determined whether or not improvement processing has been performed. That is, even if the number of samples from an arbitrary extreme value to the next extreme value is “1”, which is smaller than the sample number threshold, a half-cycle digital audio signal including the extreme value and the samples up to the next extreme value is obtained. It can only be specified that the frequency component it has is a frequency component in the range of 11.0525 kHz to 22.050 kHz, and is the frequency component of the digital audio signal of that half cycle added by the sound quality improvement processing? Cannot judge whether or not.

  Therefore, the signal analysis unit 124 has a plurality of samples in which the number of samples from an arbitrary extreme value to the next extreme value is smaller than the sample number threshold, that is, a range of the frequency f3 or more (substantially a range of frequencies f3 to f4). For a plurality of included samples, it is determined whether or not the difference value between extreme values exceeds an extreme value level threshold. Then, the signal analysis unit 124, for example, for an entire sample in which the number of samples from an arbitrary extreme value to the next extreme value is smaller than the sample number threshold in a predetermined period, the difference value between extreme values exceeds the extreme level threshold. The occupancy occupied by the sample is derived. Specifically, the signal analysis unit 124 determines that the difference value between extreme values exceeds the extreme value level threshold among digital audio signals having a half cycle in which the number of samples from an arbitrary extreme value to the next extreme value is less than two. The occupancy rate of the half-cycle digital audio signal is derived.

  As can be seen by comparing the areas indicated by cross-hatching in FIGS. 5B and 5C, in the frequency components of the frequencies f3 to f4 where the number of samples from any extreme value to the next extreme value is less than the sample number threshold. The ratio at which the difference value between extreme values exceeds the extreme value level threshold is higher in the case of a digital audio signal including a high frequency component by sound quality improvement processing shown in FIG. Therefore, the signal analysis unit 124 compares the occupancy rate of the samples whose difference value between extreme values exceeds the extreme value level threshold value with a predetermined ratio (for example, 50%), thereby improving the sound quality of the digital audio signal. It is determined whether or not processing has been performed.

  Thus, depending on the combination of the sampling frequency of the digital audio signal and the frequency band of the high frequency component, whether or not the high frequency component is included in the digital audio signal of a certain half cycle can be determined by the sample of the half cycle. Even if it cannot be determined, the signal analysis unit 124 of the present embodiment analyzes the sample over a plurality of periods, so that the sound quality improvement processing has already been performed on the digital audio signals of the plurality of periods. It is possible to reliably determine whether or not.

  Next, a description will be given of a means for the signal analysis unit 124 to determine the sample number threshold value and the extreme value level threshold value based on the format of the digital audio signal.

  FIG. 6 is an explanatory diagram showing an example of the relationship between the format of the digital audio signal, the sample number threshold value, and the extreme value level threshold value. As shown in FIG. 6, the signal analysis unit 124 selects from any extreme value to the next extreme value depending on the format of the digital audio signal (for example, AAC (registered trademark), HE-AAC, MP3) and the bit rate. A sample number threshold value and an extreme value level threshold value are determined. In FIG. 6, when the sound pressure corresponding to the extreme value level threshold value is AD-converted, the extreme value level threshold value is several times (for example, 128) the LSB (Least Significant Bit) which is the quantization unit of AD conversion. It is shown in what.

  As shown in FIG. 6, the frequency band that is removed by digitization or compression processing is determined by the format of the digital audio signal, and the frequency band that may be added by the sound quality improvement processing is determined. Therefore, if the format and bit rate of the digital audio signal are determined, the frequency of the digital audio signal is derived based on the sampling frequency and the number of samples from an arbitrary extreme value to the next extreme value. The sound pressure level of the noise signal included in the digital audio signal is also determined by the number of quantization bits, the bit rate, and the like. Therefore, the signal analysis unit 124 can reliably determine whether the sound quality improvement processing has already been performed by determining the sample number threshold value and the extreme value level threshold value based on the format.

  The table selection unit 126 selects one coefficient table from among a plurality of coefficient tables for selecting a coefficient according to whether or not the signal analysis unit 124 determines that a high frequency component is included in the digital audio signal. select. At this time, when the digital audio signal includes a high frequency component, the table selection unit 126 selects a coefficient table that includes a smaller coefficient than when the digital audio signal does not include a high frequency component. A correction value generation unit 130, which will be described later, generates a correction value based on the coefficients included in the coefficient table.

  If the digital audio signal contains high frequency components, it can be determined that the sound quality improvement processing has already been performed. The table selection unit 126 suppresses the correction amount by selecting a coefficient table in which a smaller coefficient is included in the digital audio signal subjected to the sound quality improvement process, and reliably avoids an excessive sound quality improvement process.

  The coefficient storage unit 128 includes a RAM (Random Access Memory), an EEPROM, a nonvolatile RAM, a flash memory, an HDD (Hard Disk Drive), and the like, and a first coefficient table group and a second coefficient table group are stored in advance. Yes.

  FIG. 7 is an explanatory diagram for explaining a coefficient table group and a coefficient table. As shown in FIGS. 7A and 7B, the coefficient table group (the first coefficient table group 166 and the second coefficient table group 168) includes the coefficient of the coefficient table A and the coefficient table in the first coefficient table group 166. B coefficient,..., For example, for each format (standard) of the digital audio signal, as shown in the second coefficient table group 168 by the columns of the coefficient table A ′, the coefficient table B ′ coefficient,. In addition, a coefficient corresponding to the number of samples is set. Here, for ease of understanding, the coefficients of the coefficient tables A, B, A ′, B ′,... Are shown in columns of the coefficient table group, but more specifically, the coefficient tables A, B, In each of A ′, B ′,..., The number of samples and the coefficient are associated one-to-one.

  Further, the first coefficient table group 166 shown in FIG. 7A is a digital sound signal that has not been subjected to the sound quality improvement process, and the second coefficient table group 168 shown in FIG. This is used for digital audio signals subjected to. In the present embodiment, the coefficients of the coefficient tables (coefficient table A ′, coefficient table B ′,...) Of the second coefficient table group 168 are the corresponding coefficient tables (coefficient tables) of the first coefficient table group 166. Unlike the coefficients of A, coefficient table B,..., Here, half of the coefficients of the corresponding coefficient tables (coefficient table A, coefficient table B,...) Of the first coefficient table group 166 are used. It is a value. The coefficient table A and the coefficient table A ′ are coefficient tables used when the sampling frequency of the acquired digital audio signal is, for example, 44.1 kHz. The coefficient table B and the coefficient table B ′ are, for example, the acquired digital audio signal. It is a coefficient table used when a sampling frequency is 96 kHz, for example. As shown in FIG. 7, the coefficient tables are made different depending on whether or not the sound quality improvement processing has already been performed for digital audio signals having the same sampling frequency.

  Also, the first coefficient table group 166 and the second coefficient table group 168 are combined into one table, and for example, an identification information item (column) corresponding to whether or not sound quality improvement processing is performed is added to the table. The table selection unit 126 may select a coefficient table based on the identification information in addition to the format of the digital audio signal.

  The table selection unit 126 first selects one coefficient table group from the first coefficient table group 166 and the second coefficient table group 168 depending on whether or not a high frequency component is included in the digital audio signal, Further, using the selected one coefficient table group, one coefficient table corresponding to the format of the digital audio signal is selected. The correction value generation unit 130 described later generates a correction value using the coefficient table selected by the table selection unit 126.

  Here, in the coefficient tables A and B of the first coefficient table group 166 and the coefficient tables A ′ and B ′ of the second coefficient table group 168 in FIGS. 7A and 7B, the coefficient value increases as the number of samples increases. Is small for the following reasons. That is, when the number of samples from an arbitrary extreme value to the next extreme value is large, the frequency of the digital audio signal is low. Therefore, for example, even if filtering has already been performed with a 22.1 kHz low-pass filter (LPF), the harmonics of the low-frequency frequency components remain without being suppressed. Therefore, a sufficiently high sound quality can be maintained without adding a large high-frequency component, and the coefficient can be small.

  On the other hand, when the number of samples from any extreme value to the next extreme value is small, the frequency of the digital audio signal is high. Therefore, for example, if filtering has already been performed with a 22.1 kHz low-pass filter, the harmonics of the high-frequency frequency components are almost reduced. Therefore, the sound quality cannot be improved unless sufficient high-frequency components are added, so the coefficient needs to be large.

  Therefore, the table selection unit 126 calculates the coefficients of the coefficient tables A and B of the first coefficient table group 166 and the coefficients of the second coefficient table group 168 associated with the coefficients according to the number of samples from an arbitrary extreme value to the next extreme value. Using the tables A ′ and B ′, a coefficient table is selected so that an appropriate correction amount is obtained according to the number of samples.

  As described above, the table selection unit 126 associates a coefficient having a generally smaller value than the first coefficient table group 166 used when the sound quality improvement processing has already been performed on the digital audio signal and has not yet been performed. The second coefficient table group 168 including the coefficient table is selected, and one coefficient table is further selected from the coefficient table group. With this configuration, the audio processing apparatus 100 can perform appropriate correction according to the frequency band of the digital audio signal and whether or not the sound quality improvement processing has already been performed.

  By the selection process described above, an appropriate coefficient table is selected according to whether or not a high frequency component is included. Subsequently, the voice processing apparatus 100 performs correction processing using the selected coefficient table. In the correction process, the signal acquisition unit 120 acquires the digital audio signal again, and the processes of the first conversion unit 138 and the extreme value specifying unit 122 are performed. The signal acquisition unit 120 may temporarily hold the digital audio signal in a buffer unit (not shown) without acquiring the digital audio signal again. The processing from the signal acquisition unit 120 to the extreme value identification unit 122 is substantially the same as the selection processing except that the digital audio signal up-sampled by the first conversion unit 138 is also output to the delay unit 132, so the description will be made. The correction process will be described from the process of the correction value generation unit 130 omitted.

(Correction process)
In the correction process, the correction value generation unit 130 generates a correction value that increases the amplitude of the digital audio signal according to the coefficient selected from the coefficient table. Specifically, the correction value generation unit 130 determines the maximum difference that is the difference between the maximum value of the digital audio signal and the value of the sample one sample before the sample having the maximum value (hereinafter referred to as the maximum value sample). The value is multiplied by a coefficient corresponding to the number of samples in the coefficient table selected by the table selection unit 126 to generate a correction value to be added to the sample of the maximum value, and the minimum value of the digital audio signal and the minimum value thereof Multiplying the minimum difference value, which is the difference between the sample that has become the sample (hereinafter referred to as the minimum value sample) and the value of the previous sample, by a coefficient corresponding to the number of samples in the coefficient table selected by the table selection unit 126. To generate a correction value to be subtracted from the sample of the minimum value.

  For example, if the sampling frequency of the digital audio signal acquired by the signal acquisition unit 120 is 44.1 kHz and the signal analysis unit 124 determines that the sound quality improvement processing has not been performed in the past, the table selection unit 126 selects the coefficient table A. The correction value generation unit 130 obtains the number of samples from the minimum value sample one sample before the maximum value sample to be corrected to the maximum value sample to be corrected from the extreme value specifying unit 122 and stored in the coefficient storage unit 128. One coefficient corresponding to the number of samples is selected from the coefficient table A, and a correction value to be added to the sample of the maximum value is generated by multiplying the maximum difference value by the coefficient. In other words, when the coefficient table A is selected by the table selection unit 126, as shown in FIG. 7A, the correction value generation unit 130 is 1/2 if the number of samples is 1 to 5, and the number of samples is 6. If it is ~ 9, it is 1/4, if the number of samples is 10-14, 1/8, and if the number of samples is 15 or more, 1/16 is multiplied by the maximum difference value to add to the maximum value sample. A correction value is generated.

  Note that when the signal analysis unit 124 determines that the digital audio signal to be corrected has been subjected to sound quality improvement processing in the past, the table selection unit 126 selects the coefficient table A ′, so that correction is performed when the number of samples is the same. The value generation unit 130 generates a correction value using a smaller coefficient than when the signal analysis unit 124 determines that the digital audio signal has not been subjected to sound quality improvement processing in the past.

  Similarly, the correction value generation unit 130 obtains the number of samples from the maximum value sample one sample before the minimum value sample to be corrected to the minimum value sample to be corrected from the extreme value specifying unit 122, and One coefficient corresponding to the number of samples is selected from among them, and the coefficient is multiplied by the minimum difference value to generate a correction value to be added to the minimum value sample. In other words, when the coefficient table A is selected by the table selection unit 122, as shown in FIG. 7A, the correction value generation unit 130 is 1/2 if the number of samples is 1 to 5, and the number of samples is 6. Correction to add to the minimum value sample by multiplying the minimum difference value by 1/4 if it is ~ 9, 1/8 if the number of samples is 10-14, or 1/16 if the number of samples is 15 or more. Generate a value.

  Even in the generation of the correction value of the minimum value sample, when the signal analysis unit 124 determines that the sound quality improvement processing has been performed on the digital audio signal to be corrected in the past, the table selection unit 126 selects the coefficient table A ′. When the number of samples is the same, the correction value generation unit 130 generates a correction value using a smaller coefficient than when the signal analysis unit 124 determines that the digital audio signal has not been subjected to sound quality improvement processing in the past. It will be.

  Then, the correction value generation unit 130 associates the correction value generated based on the maximum difference value with the maximum value sample so that the correction value is added to the maximum value, and subtracts the correction value generated based on the minimum difference value from the minimum value. Associating with the local minimum sample.

  Further, the correction value generation unit 130 adds the correction value generated from the maximum difference value to the corresponding maximum value sample, and the correction value generated from the minimum difference value becomes the minimum value of the corresponding digital audio signal. A correction signal with a correction value is generated so as to be subtracted from the sample.

  The adder 134 adds the correction signal generated by the correction value generator 130 to the digital audio signal. In the present embodiment, the correction value is added to the digital audio signal by adding the correction signal to the digital audio signal.

As a result, as shown in the following two equations, the adding unit 134 adds the maximum difference value multiplied by the coefficient to the maximum value, and subtracts the minimum difference value multiplied by the coefficient from the minimum value. Here, the maximum value before adding the correction signal is Vmax, the minimum value before adding the correction signal is Vmin, the maximum value after adding the correction signal is V'max, and the minimum value after adding the correction signal V′min, the maximum difference value is dl0, the minimum difference value is ds0, and the coefficient selected by the correction value generation unit 130 based on the number of samples from the coefficients selected by the table selection unit 126 is Amax and Amin. Then, the maximum value and the minimum value after adding the correction signal are expressed as the following Expression 1 and Expression 2, respectively.
V′max = Vmax + Amax × dl0 (Formula 1)
V′min = Vmin−Amin × ds0 (Formula 2)

  FIG. 8 is an explanatory diagram for explaining the process of adding the correction signal to the digital audio signal in more detail. Note that Amax · dl0 = Δdl0 and Amin · ds0 = Δds0. When the digital audio signal 170 as shown in FIG. 8A is acquired, the correction value generation unit 130 multiplies the maximum difference value dl0 and the minimum difference value ds0 by coefficients Amax and Amin, respectively, and corrects Δdl0 and Δds0. And a correction signal 172 as shown in FIG. 8B is generated. When the adding unit 134 adds the correction signal 172 to the digital audio signal, the corrected digital audio signal 174 increases in amplitude between the maximum value and the minimum value, as indicated by the white arrow in FIG. .

  As described above, a digital audio signal from which high-frequency components have been cut may have poor sound expansion, spread, dynamic range, glossiness, and the like. As described above, the correction value generation unit 130 and the addition unit 134 of the present embodiment complement the high-frequency component that has been cut by simple processing for increasing the absolute value of the maximum value and the minimum value of the digital audio signal. Therefore, a digital audio signal closer to the original sound can be generated.

  Next, another example of the correction value generated by the correction value generation unit 130 will be described. The correction value generation unit 130 multiplies the difference between the maximum value of the sample one sample before and one sample after the maximum value of the digital audio signal by the coefficient of the coefficient table selected by the table selection unit 126. At the same time, the correction value is obtained by multiplying the difference between the value of the sample one sample before and the sample after the sample of the minimum value of the digital audio signal by the coefficient of the coefficient table selected by the table selection unit 126. Generate.

  Then, the correction value generation unit 130 maximizes the correction value generated based on the difference between the sample value one sample before the maximum value and the maximum value to the value of the sample one sample before the maximum value. The correction value generated based on the difference between the sample value one sample after the maximum value and the maximum value in correspondence with the sample one sample before the value is added to the value of the sample one sample after the maximum value. The correction value generated based on the difference between the value of the sample one sample before the minimum value and the value of the minimum value is subtracted from the value of the sample one sample before the minimum value. In this way, the correction value generated based on the difference between the sample value one sample after the minimum value and the minimum value is associated with the sample one sample before the minimum value and the value of the sample one sample after the minimum value is Subtracted Correspond to the sample after one sample of sea urchin minimum value.

  Here, the correction value is generated for the sample one sample before the extreme value and the sample one sample after the extreme value. However, the correction value is not limited to such a case, and two or more before the extreme value and two or more after the extreme value. Correction values may also be generated for samples. In this case, the local maximum and local minimum difference values are the difference between the value of the sample for which the correction value is to be generated and the value of the sample closer to the extreme value than that sample. Hereinafter, such processing of the correction value generation unit 130 will be described with reference to FIGS. 9 and 10.

  Also in FIG. 9, as in FIG. 8A, the sampling frequency of the digital audio signal acquired by the signal acquisition unit 120 is 44.1 kHz, and the signal analysis unit 124 performs a sound quality improvement process on the digital audio signal in the past. If it is determined that it is not, the table selection unit 126 selects the coefficient table A. When the digital audio signal 170 as illustrated in FIG. 9A is acquired, the correction value generation unit 130 determines the correction target from the minimum value sample immediately before the maximum value sample to which the correction target local maximum sample corresponds. The number of samples up to the maximum value sample corresponding to the local maximum sample is acquired from the extreme value specifying unit 122. Then, the correction value generation unit 130 obtains one coefficient corresponding to the number of samples from the coefficient table A stored in the coefficient storage unit 128, and the difference between the maximum local sample value and the maximum value (maximum) As shown in FIG. 9B, correction values Δdl1 and Δdl2 are generated by multiplying dl1 and dl2 which are neighboring difference values) by their coefficients.

  Similarly, the correction value generation unit 130 determines the number of samples from the maximum value sample immediately before the maximum value sample to which the correction target local minimum sample corresponds to the local minimum sample to which the correction target local minimum sample corresponds. One coefficient corresponding to the number of samples is selected from the coefficient table A acquired from the specifying unit 122 and stored in the coefficient storage unit 128, and the difference between the local minimum sample value and the local minimum value (local minimum difference) Value) ds1 and ds2 are multiplied by the coefficients to generate correction values Δds1 and Δds2, as shown in FIG. 9B.

  Then, the correction value generation unit 130 sets Δdl1 to a sample one sample before the maximum value sample (sample that has reached the maximum value), Δdl2 sets a sample one sample after the maximum value sample, and Δds1 sets a minimum value sample (minimum value). The correction signal 176 is generated by associating the sample one sample before the sample) and the sample one sample after the minimum value sample with Δds2, and the adding unit 134 adds the correction signal to the digital audio signal. To do. Then, the correction values Δdl1 and Δdl2 generated based on the local maximum difference values dl1 and dl2 are added to the values of the corresponding local maximum samples, and the correction values Δds1 and Δds2 generated based on the local minimum difference values ds1 and ds2 are obtained. , Subtracted from the value of the local minimum sample. As a result, the waveform of the corrected digital audio signal 178 approaches a rectangular wave, as indicated by a white arrow in FIG. 9C.

  The correction value generation unit 130 and the addition unit 134 of the present embodiment are corrected by a simple process of increasing the absolute value of the local maximum sample value and the local minimum sample value of the digital audio signal so as to approximate the rectangular wave. A digital audio signal closer to the original sound can be generated.

  Further, a correction value may be generated and corrected for all of the maximum value sample, the minimum value sample, the local maximum sample, and the local minimum sample. Assume that the local maximum difference value and the local maximum difference value before and after the local maximum sample are equal to dl3, and the local minimum difference value and the local minimum difference value before and after the local minimum sample are equal to ds3. In this case, when the digital audio signal 170 as shown in FIG. 10A is acquired, the correction value generation unit 130 corrects the maximum value sample and the front and rear local maximum samples as shown in FIG. 10B. The correction signal 180 is generated by associating the value Δdl3 with the correction value Δds3 in association with the minimum value sample and the front and rear minimum neighboring samples, and the adding unit 134 adds the correction signal 180 to the digital audio signal. As a result of the processing of the correction value generation unit 130 and the addition unit 134, the corrected digital audio signal 182 indicated by the white arrow in FIG. 10C has the correction effect described with reference to FIGS. Therefore, it is possible to reproduce a sound close to the original sound.

  In this embodiment, the correction value generation unit 130 generates the correction signal, and the addition unit 134 adds the correction signal to the digital audio signal. However, the present invention is not limited to this, and the correction value generation unit 130 corrects the correction signal. May be output to the adding unit 134, and the adding unit 134 may adjust the sign and timing of the correction value to add the correction value to the digital audio signal.

  Note that the delay unit 132 delays the digital audio signal input from the signal acquisition unit 120 by the total processing time in each functional unit from the first conversion unit 138 to the second conversion unit 140, and the first conversion unit It synchronizes with the digital audio signal that has passed through the functional units from 138 to the second conversion unit 140.

  As described above, the high-frequency component is included in the digital audio signal that has been subjected to the sound quality improvement processing. The signal analysis unit 124 performs frequency analysis of the digital audio signal and determines whether or not sound quality improvement processing has been performed by identifying the presence or absence of high frequency components. And the table selection part 126 selects the suitable coefficient according to whether the sound quality improvement process is already performed, and the correction value production | generation part 130 produces | generates a correction signal. With this configuration, the sound processing apparatus 100 performs an appropriate sound quality improvement process depending on whether or not the acquired digital sound signal has already been subjected to the sound quality improvement process, maintains a sense of balance in the middle and high frequencies, and further increases the original sound. A digital audio signal close to can be generated.

  Next, operations of the first conversion unit 138 and the second conversion unit 140 will be described. The 1st conversion part 138 can change the sampling frequency, and can ensure the area | region which adds a high frequency component.

  The first conversion unit 138 upsamples the digital audio signal acquired by the signal acquisition unit 120 and outputs it to the extreme value specifying unit 122, the correction value generation unit 130, and the delay unit 132. The second conversion unit 140 downsamples the digital audio signal after the addition unit 134 adds the correction signal to a sampling frequency before being upsampled by the first conversion unit 138.

  FIG. 11 is an explanatory diagram for explaining processing of the first conversion unit 138. When a digital audio signal having a sampling frequency of fs shown in FIG. 11A is added to the digital audio signal by a correction process with a frequency component that is 1/2 or more of the sampling frequency, aliasing noise occurs. There is a case.

  Therefore, as shown in FIG. 11B, the first converter 138 upsamples the sampling frequency of the digital audio signal by a factor of two. Then, as shown in FIG. 11C, a high frequency component (indicated by hatching in FIG. 11C) is added to this signal. Thereafter, the second conversion unit 140 can suppress the occurrence of aliasing noise by down-sampling the corrected digital audio signal to set the sampling frequency to the original sampling frequency.

(Computer 200)
The voice processing apparatus 100 described above can be realized using a computer. Hereinafter, an example in which the speech processing apparatus 100 is realized using a computer will be described.

  FIG. 12 is a functional block diagram showing a typical example of a computer (information processing apparatus) 200 that can perform sound quality improvement processing by the sound processing apparatus 100. The computer 200 includes a central processing unit 210, a temporary storage device 212, an external storage device 214, an acquisition unit 216, and an output unit 218.

  A central processing unit (CPU) 210 controls the entire computer 200 with programs and applications in the temporary storage device 212 and the external storage device 214. The temporary storage device 212 includes a RAM, an EEPROM, a nonvolatile RAM, and the like, and temporarily stores digital audio signals and the like processed by the central processing unit 210. The external storage device 214 includes a flash memory, an HDD, and the like, and stores a program processed by the central processing unit 210. The acquisition unit 216 acquires a digital audio signal and temporarily stores it in the temporary storage device 212. The output unit 218 outputs the digital audio signal corrected by the computer 200 to the playback device 110 or the like.

  The central processing unit 210 executes the program to execute an extreme value specifying unit 122, a signal analysis unit 124, a table selection unit 126, a correction value generation unit 130, a delay unit 132, an addition unit 134, It functions as a first conversion unit 138 and a second conversion unit 140. Accordingly, in the present embodiment, the computer 200 determines whether or not the digital audio signal includes a high frequency component having a predetermined frequency or higher and a predetermined sound pressure or higher, and the digital audio signal includes the high frequency component. Also provided is an audio processing program that executes a step of generating a correction value that expands the amplitude of the digital audio signal and a step of adding the correction value to the digital audio signal based on different coefficients depending on whether or not Is done. In addition, this program may be read from a storage medium and loaded into a computer, or may be loaded into the computer 200 via the communication network 106.

(Audio processing method)
Next, an audio processing method for analyzing a digital audio signal using the above-described audio processing apparatus 100 and correcting the digital audio signal using the analysis result will be described.

  13 and 14 are flowcharts showing the overall flow of the voice processing method. Here, the voice processing method is roughly divided into two, the process of selecting a coefficient table will be described using FIG. 13, and the correction process based on the selected coefficient table will be described using FIG.

  When the coefficient table selection process is started, first, counting for a predetermined period (not shown) is started. Next, as shown in FIG. 13, the signal acquisition unit 120 acquires a digital audio signal (YES in S300), and the extreme value specifying unit 122 is the local maximum that is the extreme value of the digital audio signal acquired by the signal acquisition unit 120. When the value and the minimum value are specified (YES in S302), the extreme value specifying unit 122 counts the number of samples from any extreme value to the next extreme value (S304), and the difference between the maximum value and the minimum value A difference value between extreme values (maximum difference value and minimum difference value) is derived (S306).

  Subsequently, the signal analysis unit 124 determines whether or not the number of samples from an arbitrary extreme value to the next extreme value is smaller than the sample number threshold (S308). When the number of samples from any extreme value to the next extreme value is smaller than the sample number threshold (YES in S308), the signal analysis unit 124 indicates that the difference value between extreme values, which is the difference between the maximum value and the minimum value, is an extreme value. It is determined whether or not the level threshold is exceeded (S310). If the difference value between the extreme values exceeds the extreme level threshold (YES in S310), it is determined that the digital audio signal includes a high frequency component, and the table selection unit 126 determines from the second coefficient table group 168. A coefficient table is selected based on the format of the digital signal (S312). Therefore, when a coefficient table is selected from the second coefficient table group 168, the coefficient table selection process ends without waiting for the elapse of a predetermined period.

  When the signal acquisition unit 120 has not acquired a digital audio signal (NO in S300), the extreme value specifying unit 122 specifies the maximum value and the minimum value that are the extreme values of the digital audio signal acquired by the signal acquisition unit 120. Until the number of samples from any extreme value to the next extreme value is greater than or equal to the sample number threshold (NO in S308), or the difference value between extreme values is less than or equal to the extreme level threshold (NO in S302) In step S310, the table selection unit 126 determines whether a predetermined period has elapsed (step S314). If the predetermined period has not elapsed (NO in S314), the process is repeated from the signal acquisition step (S300). When the predetermined period has elapsed (YES in S314), the table selection unit 126 selects a coefficient table from the first coefficient table group 166 based on the format of the digital signal (S316). Thus, if the number of samples and the difference value between extreme values do not satisfy the predetermined condition within the predetermined period, the coefficient table is selected from the first coefficient table group 166.

  When the table selection unit 126 finishes selecting the coefficient table in the second coefficient table group selection step (S312) or the first coefficient selection step (S316), the process proceeds to the correction process shown in FIG.

  As shown in FIG. 14, in the correction process, when the signal acquisition unit 120 acquires a digital audio signal (YES in S300), the extreme value specifying unit 122 is the extreme value of the digital audio signal acquired by the signal acquisition unit 120. The signal acquisition step (S300) is repeated until the maximum value and the minimum value are specified (NO in S302). When the maximum value and the minimum value, which are the extreme values of the digital audio signal acquired by the signal acquisition unit 120, are specified (YES in S302), the extreme value specifying unit 122 counts the number of samples from the extreme value to the extreme value ( S304). The digital audio signal in FIG. 14 and the digital audio signal in FIG. 13 are similar signals.

  Then, the correction value generation unit 130 uses the coefficient table selected by the table selection unit 126 in the second coefficient table group selection step (S312) or the first coefficient selection step (S316), and the number of samples of the digital audio signal. A coefficient corresponding to 1 is selected (S330).

  The correction value generation unit 130 multiplies the maximum difference value, which is the difference between the maximum value of the digital audio signal and the value of the sample one sample before the maximum value sample, by the coefficient of the coefficient table selected by the table selection unit 126. Thus, a correction value to be added to the maximum value sample is generated, and the table selection unit 126 sets the difference value between the minimum value of the digital audio signal and the value of the sample one sample before the minimum value sample. Is multiplied by the coefficient of the selected coefficient table to generate a correction value to be subtracted from the sample of the minimum value (S332), and the maximum value is added so that the correction value generated based on the maximum difference value is added to the maximum value. In association with the sample, the correction value generated based on the minimum difference value is associated with the minimum value sample so as to be subtracted from the minimum value (S334).

  Subsequently, in the addition unit 134, the correction value generation unit 130 generates the correction value generated from the maximum difference value from the minimum difference value to the maximum value of the digital audio signal associated with the correction value itself. A correction signal in which the correction value is arranged so as to be synchronized with the minimum value of the digital audio signal associated with the correction value is generated (S336). The adder 134 adds the digital audio signal and the correction signal (S338), and the signal output unit 136 outputs the digital audio signal added with the correction signal (S340), and a signal acquisition step (S300 shown in FIG. 14). Return to).

  Even with the audio processing method described above, a digital audio signal that has been subjected to sound quality improvement processing can be distinguished from a normal digital audio signal, and correction processing is performed using coefficients corresponding to each, so that the digital audio signal closer to the original sound is used. Can be generated.

  As mentioned above, although preferred embodiment of this invention was described referring an accompanying drawing, it cannot be overemphasized that this invention is not limited to this embodiment. 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 these are naturally within the technical scope of the present invention. Is done.

  Note that each step in the voice processing method of the present specification does not necessarily have to be processed in time series in the order described in the flowchart, and may include parallel or subroutine processing.

  The present invention can be used in an audio processing apparatus, an audio processing method, and an audio processing program that analyze a digital audio signal and process the digital audio signal according to the analysis result.

DESCRIPTION OF SYMBOLS 100 ... Voice processing apparatus 120 ... Signal acquisition part 124 ... Signal analysis part 126 ... Table selection part 128 ... Coefficient memory | storage part 130 ... Correction value generation part 134 ... Addition part 136 ... 1st conversion part 138 ... 2nd conversion part 200 ... Computer

Claims (11)

A signal analysis unit for determining whether or not a high frequency component of a predetermined frequency or higher and a predetermined sound pressure or higher is included in the digital audio signal;
Correction that generates a correction value that expands the amplitude of the digital audio signal based on a different coefficient depending on whether or not the signal analysis unit determines that the high-frequency component is included in the digital audio signal A value generator,
An adder for adding the correction value to the digital audio signal;
An audio processing apparatus comprising:   The signal analysis unit has a number of samples from an arbitrary extreme value of the digital audio signal to the next extreme value, which is smaller than a predetermined number, and the larger extreme value between the arbitrary extreme value and the next extreme value. A difference value between extreme values that is a difference between a maximum value that is a value and a minimum value that is the smaller one of the arbitrary extreme value and the next extreme value exceeds a predetermined value The audio processing apparatus according to claim 1, wherein the digital audio signal is determined to include the high frequency component.   The signal analysis unit includes the arbitrary extreme value and the next extreme value for a plurality of samples in which the number of samples from an arbitrary extreme value to the next extreme value of the digital audio signal is less than a predetermined number. The difference value between the extreme values, which is the difference between the maximum value that is the larger extreme value, and the local minimum value that is the smaller of the arbitrary extreme value and the next extreme value, The audio processing according to claim 1, wherein if the occupation ratio occupied by samples exceeding a predetermined value exceeds a predetermined ratio, it is determined that the high-frequency component is included in the digital audio signal. apparatus.   When the signal analysis unit determines that the high frequency component is included in the digital audio signal, the correction value generation unit does not include the high frequency component in the digital audio signal. The speech processing apparatus according to any one of claims 1 to 3, wherein the correction value is generated based on a smaller coefficient than the case where the determination is made.   5. The audio processing apparatus according to claim 1, wherein the correction value generation unit generates the correction value based on a coefficient that differs according to a format of the digital audio signal. 6.   The audio processing apparatus according to claim 2, wherein the signal analysis unit determines the predetermined number and the predetermined value based on a format of the digital audio signal. A first converter for upsampling a digital audio signal acquired by the audio processing device;
A second conversion unit that downsamples the digital audio signal after the addition unit adds the correction value to a sampling frequency before being upsampled by the first conversion unit;
The speech processing apparatus according to claim 1, further comprising:   The correction value generation unit multiplies the coefficient by a maximum difference value that is a difference between the maximum value of the digital audio signal and the value of the sample one sample before the sample that has reached the maximum value, and the digital audio signal The correction value is generated by multiplying the minimum difference value, which is the difference between the minimum value of the signal and the value of the sample one sample before the sample having the minimum value, by the coefficient, and based on the maximum difference value The correction value generated in association with the sample having the maximum value so that the correction value is added to the maximum value, and the correction value generated based on the minimum difference value is subtracted from the minimum value. The speech processing apparatus according to claim 1, wherein the speech processing apparatus is associated with a sample having a minimum value. The correction value generation unit
The difference between the respective values of the sample one sample before and one sample after the maximum value of the digital audio signal and the maximum value is multiplied by the coefficient, and one sample before the minimum value of the digital audio signal. The correction value is generated by multiplying the difference between each value of the sample and the sample after one sample and the minimum value by the coefficient,
One sample before the maximum value so that a correction value generated based on the difference between the sample value one sample before the maximum value and the maximum value is added to the value of the sample one sample before the maximum value. Map to the sample
After one sample of the maximum value, a correction value generated based on the difference between the sample value after one sample of the maximum value and the maximum value is added to the value of the sample after one sample of the maximum value. Map to the sample
One sample before the minimum value so that a correction value generated based on the difference between the value of the sample one sample before the minimum value and the minimum value is subtracted from the value of the sample one sample before the minimum value. Map to the sample
After one sample of the minimum value, the correction value generated based on the difference between the sample value after one sample of the minimum value and the minimum value is subtracted from the value of the sample after one sample of the minimum value. The voice processing apparatus according to claim 1, wherein the voice processing apparatus is associated with the sample. Determine whether the digital audio signal contains high frequency components above a specified frequency and above a specified sound pressure,
Based on different coefficients depending on whether the high frequency component is included in the digital audio signal, generating a correction value that expands the amplitude of the digital audio signal,
An audio processing method comprising adding the correction value to the digital audio signal. On the computer,
Determining whether the digital audio signal includes a high frequency component of a predetermined frequency or higher and a predetermined sound pressure or higher;
Generating a correction value for enlarging the amplitude of the digital audio signal based on different coefficients depending on whether the high frequency component is included in the digital audio signal;
Adding the correction value to the digital audio signal;
A voice processing program characterized by causing

Images