JP2011071806A - Electronic device, and sound-volume control program for the same - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an electronic device for controlling amplitude of an input signal, or a sound-volume control program for the same. <P>SOLUTION: An input voice signal is partitioned in a frame unit to calculate an amplitude gain and to correct amplitude of the input signal according to the amplitude gain. An amplitude gain of an immediately preceding frame and that of a frame in processing are smoothly transited. The slowing-down or speeding-up of transition of the amplitude gains is changed according to voice characteristics of the immediately preceding frame and those of the frame in processing. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an electronic device that controls an output volume, and a volume control program for the electronic device.

  2. Description of the Related Art In recent years, electronic devices that can receive audio signals for telephone communication, radio waves, or television waves and output sound from an acoustic output device such as a speaker have become widespread. A mobile phone which is an example of such an electronic device receives a radio signal and extracts a voice signal modulated in the radio signal. By outputting the extracted audio signal as audio from a speaker of the mobile phone, for example, the user of the mobile phone can hear the audio signal.

  This sound signal is output from the sound output device with a certain amount of volume change due to, for example, a change in accent of the speaker who transmitted the sound signal. For this reason, when the change in volume is abrupt, for example, a sound with a large volume is suddenly output from the acoustic output device, which may make the user feel uncomfortable. Therefore, the input audio signal is buffered to detect whether the audio signal does not contain a large volume component. When a large volume component is included, the entire audio signal is normalized, and control is performed to reduce the volume over the entire buffered audio signal. An invention that prevents a situation in which a sound with a high sound volume is output from an acoustic output device by outputting a sound signal with a reduced sound volume is disclosed (for example, see Patent Document 1).

JP 2009-21834 A

  However, with the method of buffering the entire audio signal and then correcting the volume of the audio signal as described above, there is a great delay between the input of the audio signal and the output of the audio from the audio output device. Will occur. When making a call using a mobile phone, it is desirable that the delay from the input of an audio signal to the output be short.

  The present invention has been made to solve the above-described problems, and relates to an electronic device capable of controlling the amplitude of an input signal or a volume control program for an electronic device.

In order to achieve the object, an electronic device according to the present invention includes a voice input receiving unit that receives a voice input, a volume measuring unit that measures a voice input volume, a first time interval, and a first time interval. The sound received in the previous second time interval, the volume measuring means for measuring the input sound volume of the sound, the first time interval, and the second time interval before the first time interval. A storage means for storing each of the sound and the input volume of the sound; and a maximum volume of the input sound volume received in the first time interval stored by the storage means is equal to or higher than a predetermined volume In addition, the volume gain setting means for setting the volume gain of the first time interval according to the value of the input volume of the sound received during the first time interval, and at the start of the output of the sound received during the first time interval Volume gain for time interval 2 There is a volume gain transition function setting means for setting a volume gain transition function for transitioning to be the volume gain of the first time interval when the output of the sound received in the first time interval is ended, and the output volume of the sound is Volume control means for changing in accordance with a volume gain transition function and voice output means for outputting the sound received in the first time interval with the output volume changed by the volume control means.

In order to achieve the above object, the volume control program for an electronic device according to the present invention includes a sound received in a first time interval and a second time interval before the first time interval, and input of the sound. Storage means for storing each of the volume, and when the input volume of the sound received in the first time interval stored by the storage means is equal to or higher than a predetermined volume, the sound received in the first time interval The volume gain setting means for setting the volume gain of the first time interval according to the value of the input volume, and the volume gain of the second time interval at the start of the output of the sound accepted in the first time interval, A volume gain transition function setting means for setting a volume gain transition function for transitioning to be the volume gain of the first time section when the output of the sound received in the time section is completed, and the volume gain transition function In Flip and having and a volume control means vary.

  ADVANTAGE OF THE INVENTION According to this invention, the audio | voice apparatus which controls the output volume by dividing the input audio | voice signal into a time unit and performing an amplitude correction | amendment for every time unit, or the volume control of an electronic device A program is obtained.

The figure which shows the structure of the mobile telephone in this embodiment. The figure which shows the structure of the control part in this embodiment. The figure which shows the internal structure of the signal characteristic determination part in this embodiment. The figure which shows the structure of the sample unit gain calculation part in this embodiment. The figure which shows the structure of the input signal in this embodiment, a flame | frame, and a sub-frame. The figure which shows an example of the amplitude gain in this embodiment. The flowchart which shows the flow of a process of the volume control by the control part 100 of this embodiment. The figure which shows the internal structure of the mobile telephone in 2nd Embodiment. The flowchart which shows the flow of a process of the volume control part by the 2nd control part 200 of this embodiment.

  Embodiments of the present invention will be described below with reference to the drawings.

(Configuration of electronic equipment)
FIG. 1 is a diagram illustrating a configuration of a mobile phone which is an example of an electronic apparatus according to the present invention. FIG. 2 is a diagram showing a configuration of the control unit 100 provided in the mobile phone shown in FIG. Hereinafter, the operation of each component of the present invention will be described with reference to FIGS. 1 and 2 unless otherwise specified.

  The control unit 100 mounted on the mobile phone is configured by an electronic circuit such as a CPU (Central Processing Unit). The CPU executes processing according to a program stored in a ROM or RAM described later. Further, the CPU processes signals supplied from each circuit unit described above, generates various control signals, and supplies them to each circuit unit.

Through these processes, the CPU comprehensively controls the mobile phone. The storage unit 4 includes, for example, a flash memory element that is a nonvolatile memory that can be electrically rewritten and erased, an HDD (Hard Disc Drive), a ROM (Read Only Memory), a RAM (Random Access Memory), and the like. . The storage unit 4 stores various application programs executed by the CPU of the control unit 100, data groups, audio data, and the like.

  The antenna 1 transmits and receives radio signals to and from a radio base station (not shown). The antenna 1 inputs the received radio signal to the communication signal processing unit 2. Further, the radio signal input from the communication signal processing unit 2 is transmitted to the radio base station.

The communication signal processing unit 2 converts the radio signal input from the antenna 1 into an electric signal that can be processed by the control unit 100, and inputs the converted electric signal to the control unit 100. In addition, the electrical signal output from the control unit 100 is converted into a radio signal that can be transmitted from the antenna 1 and output to the antenna 1.

  The speaker 5 converts the sound output signal input from the control unit 100 into sound and outputs the sound.

  The microphone 6 converts the input voice into a voice input signal and outputs it to the control unit 100.

  The output volume control process described with reference to FIG. 2 is a process of receiving an input audio signal, processing it into an audio output signal, and outputting it. The input of the audio input signal may be input from, for example, the microphone 6 or may be read by the control unit 100 using the audio data stored in the storage unit 4 as an input. Further, the antenna 1 may be one that receives as a radio signal.

The audio input signal is digitally encoded for each predetermined audio signal unit (hereinafter simply referred to as a frame), and the audio input signal x [n] (n = 1, 2, 3,... N ) To the control unit 100 and processed into an audio output signal y [n]. The processed audio output signal is output as audio from the speaker 5. Note that the output destination of the audio output signal is not limited to the speaker 5. For example, the audio output signal may be converted into a radio signal via the communication signal processing unit 2 and the antenna 1 may transmit the audio output signal. . Alternatively, a sound output signal may be output to headphones or the like separately connected to the control unit 100.

FIG. 5 is a diagram illustrating an example of an audio signal processed by the electronic device according to the present invention. The audio input signal x [n] indicates an amplitude value when audio is sampled. The audio input signal x [n] has the same number of amplitude data as the number N of samples constituting the frame. A frame composed of N samples is divided into subframes composed of S samples by processing of a subframe dividing unit 102 described later. FIG. 5 shows an example in which N = 12, S = 4. In the volume control processing of the present invention described later, signal processing is performed for each subframe. By performing volume control for each subframe finely divided in the time domain, it is possible to accurately control the volume in the time domain even when an audio signal with a large volume change is input. . The audio input signal x [n] is expressed using 16-bit digital data having a range from, for example, −2 to the 15th power to +2 to the 15th power. However, the notation format of the audio input signal x [n] is not limited to this, and may be an audio signal or may be represented using digital data having a number of bits other than 16 bits. Alternatively, a floating (floating point) signal may be used. Note that the audio signal processing to be described later assumes that a frame composed of N samples is divided into subframes composed of S samples as an example, and the audio signal is processed in units of subframes. State. However, the audio signal processing in the present invention is not limited to this, and the audio signal processing may be performed in units of frames by omitting division into subframes. The audio signal processing described later will be described for the case where N = 12, for convenience. However, the audio signal processing in the present invention is not limited to this. Other values may be used for N, for example, 512, 1024, 2048 or the like that is a power of 2.

Hereinafter, each digital signal in this specification may be expressed using any of the formats described herein.

(First embodiment)
When outputting an audio signal of a mobile phone from a so-called acoustic output device such as the speaker 5, the frequency characteristic of the output audio depends on the output frequency characteristic of the acoustic output device.

When the output frequency characteristic of the acoustic output device is not flat, the output sound changes compared to the input sound. Therefore, the user of the mobile phone cannot experience the original sound quality of the input voice. For example, since an acoustic output device used for a mobile phone is configured to be small, frequency characteristics often deteriorate in a high frequency band of 4 kHz or higher.

  In order to prevent the deterioration of the output sound due to the output frequency characteristic of the acoustic device, a method of correcting the sound signal in advance according to the output frequency characteristic of the acoustic output device can be considered. In the acoustic output device used in the previous mobile phone, an audio signal in a band of 4 kHz or more that is attenuated by the acoustic output device is amplified. As a result, the frequency characteristic of the output sound approaches that of the input sound, and the user of the mobile phone can experience the original sound quality of the input sound.

  However, when such a method for correcting an audio signal is used, depending on the input audio signal, the signal is greatly amplified, which may cause amplitude saturation. As an audio signal that causes amplitude saturation, an audio signal that includes an unvoiced sound (a sound that exhaled from the lungs passes through the vocal cords without oscillating the vocal cords) in the case of making a call on the previous mobile phone can be considered. Since unvoiced sounds generally have frequency components concentrated at 4 kHz or higher, amplitude saturation occurs when an audio signal containing unvoiced sounds is corrected. Due to this amplitude saturation, the sound output device cannot correctly reproduce the input sound.

  Therefore, in the first embodiment, it is determined whether or not an unvoiced sound is included in the input sound signal. If the unvoiced sound is included, a process of reducing the amplitude value of the sound signal is performed. Even when an audio signal in a band of 4 kHz or higher is amplified in accordance with the output frequency characteristics of the acoustic output device by performing a process of reducing the amplitude value for a portion including an unvoiced sound in the audio signal. It is possible to prevent the audio signal from causing amplitude saturation.

Furthermore, when the volume correction is performed every unit time, a case where the degree of volume correction is large is considered. In this case, since the volume greatly changes in a short time for the user of the electronic device, the fluctuation of the volume is unnaturally felt. In the first embodiment, a limiter control gain is calculated for each subframe, and an amplitude gain for adjusting the amplitude for each sample is set based on the value of the limiter control gain. Here, the value of the amplitude gain is set so as to change while maintaining the differential continuity so as not to switch. Since the amplitude of each sample is corrected according to this amplitude gain, the user is not given an unnatural impression due to a switch-like change in the volume adjustment amount.

Further, when an unvoiced sound is suddenly generated from the silent state, the amplitude gain is set to change rapidly. As a result, the volume control can be accurately performed in time in anticipation of sudden amplitude saturation. On the other hand, when the voiced sound changes to the unvoiced sound, amplitude saturation does not occur suddenly, so the amplitude gain is changed gradually. As a result, the amplitude can be controlled without giving an unnatural impression to the user.

The configuration and role of each component shown in FIG. 2 will be described below. The control unit 100 includes a DC component control filter 101, a subframe division unit 10, a signal characteristic detection unit 103, a subframe unit gain calculation unit 104, a sample unit gain calculation unit 105, a volume control unit 106, and a frequency characteristic correction unit 107. Composed.

  The DC component removal filter 100 receives the audio input signal x [n], removes a DC component contained in the audio input signal x [n], and outputs a signal hp [n]. More specifically, the DC component removal filter 100 performs a high-pass filtering process on the audio input signal x [n]. As a result, the DC component removal filter 100 eliminates a low frequency component included in the audio input signal x [n], for example, a component lower than 50 Hz, and outputs the remaining signal as a signal hp [n]. For the high-pass filter processing, for example, a filter designed by an IIR Butterworth filter can be used. However, the process of removing the DC component is not limited to the method described here, and it is detected from the amplitude value of the audio input signal x [n] whether or not the DC component is included, and the DC component is deleted. A method may be used.

  The subframe dividing unit 102 receives a signal hp [n] composed of N samples and inputs several subframe signals sub_hpb composed of the signal hp [n] from S samples fewer than N. [S] (s = 1, 2, 3,... S) and output.

In the present invention, the value of the limiter control gain that changes the volume for each subframe is calculated by the process described later. Here, by performing volume control using the volume control signal calculated for each subframe, an effect that finer volume control can be performed can be obtained. That is, even when the input audio signal x [n] is a silent sound, which will be described later, or a signal with a sudden change in volume, the volume can be controlled following the change in volume.

  The signal characteristic detection unit 103 receives the subframe signal sub_hpb [s] and determines whether the subframe signal sub_hpb [s] includes voiced sound, unvoiced sound, or silence. Further, it is determined whether the subframe being processed is a location where the unvoiced sound is started or a portion where the unvoiced sound is ended. This detection result is output as sub_prm.

FIG. 3 shows a detailed configuration of the signal characteristic detection unit 103. The signal characteristic detection unit 103 includes a voiced / silent detection unit 1031 and a voiced / unvoiced sound detection unit 1032.

(Sound / silence detection operation)
The voice / silence detection unit 1031 determines whether the input subframe signal sub_hpb [s] is silent or voiced, and outputs the determination result as the voice / silence detection result sub_um. . The sound / silence detection unit 1031 sets sub_um = 1 when it is determined that the subframe signal sub_hpb [s] is sound, and sub_um = 0 when it is determined that the subframe signal sub_hpb [s] is sound. Is output.

The sound / silence detection unit 1031 includes a subframe amplitude maximum value detection unit 10311 and a sound / silence determination unit 10312.

Subframe amplitude maximum value detection section 10311 receives subframe signal sub_hpb [s] as an input, and detects the maximum value of the amplitude included in subframe signal sub_hpb [s]. The maximum value of the detected amplitude is output as the subframe amplitude maximum value sub_max. More specifically, the subframe amplitude maximum value detection unit 10311 converts the amplitude value of each sample included in the subframe signal sub_hpb [s] into an absolute value. The maximum value is detected from the absolute value of the amplitude. The detected amplitude value is output as the subframe amplitude maximum value sub_max.

Here, it has been described that the sub-frame amplitude maximum value detection unit 10311 detects and outputs the maximum value of the amplitude included in the sub-frame signal sub_hpb [s]. However, the value output from the subframe amplitude maximum value detection unit 10311 may be another value as long as it is a value used by the sound / silence determination unit 10312 to be described later to determine sound / silence. . For example, the subframe amplitude maximum value detection unit 10311 may output an average amplitude level of the subframe signal sub_hpb [s]. Moreover, you may output the average spectrum power of sub-frame signal sub_hpb [s].

The voice / silence determination unit 10312 receives the value of the subframe amplitude maximum value sub_max output from the subframe amplitude maximum value detection unit 10311 as an input, and determines whether the subframe signal sub_hpb [s] is voiced or silent. Judgment is made. This determination result is output as a sound / silence determination result sub_um. More specifically, the sound / silence determination unit 10312 compares a predetermined threshold value α for sound / silence determination with the value of the subframe amplitude maximum value sub_max. If the comparison result is sub_max <α, the sound / silence determination unit 10312 does not include a sample with a large amplitude in the subframe signal sub_hpb [s], so that the subframe signal sub_hpb [s] It is determined that there is no sound. Then, the result determined to be silent is output as sub_um = 0. On the other hand, when the comparison result is sub_max> α, since the subframe signal sub_hpb [s] includes a sample having a large amplitude, it is determined that the subframe signal sub_hpb [s] is sound. Then, the result determined to be sound is output as sub_um = 1. In addition, the sound / silence determination method includes, for example, calculating the S / N ratio in each frequency band in the subframe signal sub_hpb [s], in addition to comparing the amplitude value described above with a threshold value. You may use the method compared with the threshold value (alpha) for sound / silence determination.

(Voiced / unvoiced sound detection operation)
The voiced / unvoiced sound detection unit 1032 receives the subframe signal sub_hpb [s] and the voiced / silent determination result sub_um as input, and whether the subframe signal sub_hpb [s] includes voiced sound or unvoiced sound Make a decision. Further, the subframe signal sub_hpb [s] to be processed is a signal that switches from voiced sound to unvoiced sound, a sound that switches from unvoiced sound to voiced sound, a signal that switches from silent to unvoiced sound, or from unvoiced sound to silent It is determined whether or not the signal is switched to. This determination result is output as voiced / unvoiced sound determination result sub_prm.

  The voiced / unvoiced sound detection unit 1032 includes a zero-cross number detection unit 10321 and a voiced / unvoiced sound determination unit 10322.

  The zero-cross number detection unit 10321 detects the number of zero-cross points included in the subframe signal sub_hpb [s] by using the subframe signal sub_hpb [s] and the sound / silence detection result sub_um as inputs. The number of detected zero cross points is output as the zero cross number sub_zc. In the present specification, the zero-cross point refers to a point where the polarity of the amplitude value is inverted when the amplitude value of a specific sample is compared with the amplitude value of the immediately preceding sample. Therefore, since the subframe signal sub_hpb [s] in which a number of zero cross points are included in the subframe is a fast-vibration signal, it is regarded as an audio signal containing a lot of energy in a high frequency range, that is, an unvoiced audio signal. be able to. On the other hand, when the number of zero cross points in the subframe is small, the subframe signal sub_hpb [s] is a signal that oscillates slowly, and thus can be regarded as a signal that does not include a frequency component in the high frequency range. That is, in the present invention, the detection of the zero cross point in the subframe signal sub_hpb [s] is used for analyzing the frequency component of the subframe signal sub_hpb [s].

In addition, when the value of the sound / silence detection result sub_um output from the sound / silence determination unit 10312 is sub_um = 0, it can be seen that the subframe signal sub_hpb [s] is a silence signal. In this case, the process of detecting the number of zero cross points is omitted, and the output is made as sub_zc = 0. On the other hand, when the value of the sound / silence detection result sub_um is sub_um = 1, the zero-cross number detection unit 10321 detects the number of zero-crosses included in the subframe signal sub_hpb [s] and outputs it as the zero-cross number sub_zc.

In the present embodiment, the zero cross number sub_zc is described as the number of zero cross points included in the subframe signal sub_hpb [s]. However, the operation of the present embodiment is not limited to this. For example, a value obtained by dividing the number of zero crosses included in the subframe signal sub_hpb [s] by the number of samples S of the subframe signal sub_hpb [s] may be used as the zero cross occurrence rate sub_zc instead of the zero cross number sub_zc. By using this value, even when the number of samples S of the subframe signal sub_hpb [s] changes, the zero-cross number detection unit 10321 outputs a value to the voiced / unvoiced sound detection unit 10322 described later by the same processing. In addition, voiced / unvoiced sound can be detected.

Further, in the present embodiment, it has been described that the zero-cross number detection unit 10321 detects the number of zero-cross points included in the subframe signal sub_hpb [s]. However, the configuration of the present embodiment is not limited to this, and the zero cross number detection unit 10321 may be a feature amount that determines whether or not the audio signal includes a signal of a predetermined frequency band. A feature amount may be detected. For example, the sub-frame signal sub_hpb [s] may be subjected to spectrum analysis to detect high-frequency component energy. Alternatively, the maximum value of the autocorrelation of the LPC prediction residual may be detected.

In the present embodiment, the zero cross number detection unit 10321 is used as a means for detecting a high frequency component of 4 kHz or more as an example. However, since the frequency band detected by the zero-cross number detection unit 10321 may be set according to the frequency characteristics of the acoustic output device, the frequency band is not limited to a high frequency of 4 kHz or more, and an arbitrary frequency component may be detected.

  The voiced / unvoiced sound determination unit 10322 receives the voiced / silent detection result sub_um and the number of zero crosses sub_zc, and determines how the signal characteristics such as voiced / unvoiced / silent are switched between the subframe being processed and the immediately preceding subframe. judge. This determination result is output as a signal characteristic determination result sub_prm. More specifically, the voiced / unvoiced sound determination unit 10322 holds the detection result of the signal characteristic in the immediately preceding subframe as the immediately preceding subframe signal characteristic determination result sub_prm1, and the sounded / silent detection result sub_um, the number of zero crosses sub_zc As input. Voiced / unvoiced sound determination section 10322 receives the previous three values as input, updates and outputs signal characteristic determination result sub_prm and immediately preceding subframe signal characteristic determination result sub_prm1. Note that a voiced / unvoiced sound threshold β, which will be described later, uses a predetermined value to determine whether the subframe being processed is an unvoiced sound, a voiced sound or a silent sound. Note that when determining the signal characteristics using Equation 1, sub_zc = β may be treated as being equal to either sub_zc <β or sub_zc> β.

  When the voiced / unvoiced sound determination unit 10322 determines that both the immediately preceding subframe and the subframe being processed are silent, the value of the signal characteristic determination result sub_prm is set to sub_prm = 0, and the immediately preceding subframe signal characteristic determination result sub_prm1. Is output as sub_prm1 = 0.

When the voiced / unvoiced sound determination unit 10322 determines that the immediately preceding subframe is a voiced sound and the subframe being processed is silent, the value of the signal characteristic determination result sub_prm = 1, sub_prm = 1, and the immediately preceding subframe signal The value of the characteristic determination result sub_prm1 is output as sub_prm1 = 0.

When the voiced / unvoiced sound determination unit 10322 determines that the immediately preceding subframe is an unvoiced sound and the subframe being processed is silent, the value of the signal characteristic determination result sub_prm = 2, sub_prm = 2, and the immediately preceding subframe signal characteristic The value of the determination result sub_prm1 is output as sub_prm1 = 0.

When the voiced / unvoiced sound determination unit 10322 determines that the immediately preceding subframe is silent and the subframe being processed is a voiced sound, the value of the signal characteristic determination result sub_prm is 3, sub_prm = 3, and the immediately preceding subframe signal. The value of the characteristic determination result sub_prm1 is output as sub_prm1 = 1.

When the voiced / unvoiced sound determination unit 10322 determines that both the immediately preceding subframe and the subframe being processed are voiced sounds, the value of the signal characteristic determination result sub_prm is 4, sub_prm = 4, and the immediately preceding subframe signal characteristic determination result. The value of sub_prm1 is output as sub_prm1 = 1.

  When the voiced / unvoiced sound determination unit 10322 determines that the immediately preceding subframe is an unvoiced sound and the subframe being processed is a voiced sound, the value of the signal characteristic determination result sub_prm = 5, the previous subframe signal The value of the characteristic determination result sub_prm1 is output as sub_prm1 = 1.

  When the voiced / unvoiced sound determination unit 10322 determines that the immediately preceding subframe is silent and the subframe being processed is unvoiced, the value of the signal characteristic determination result sub_prm is set to sub_prm = 6, and the immediately preceding subframe signal characteristic The value of the determination result sub_prm1 is output as sub_prm1 = 2.

  When the voiced / unvoiced sound determination unit 10322 determines that the immediately preceding subframe is a voiced sound and the subframe being processed is an unvoiced sound, the value of the signal characteristic determination result sub_prm = sub_prm = 7, the immediately preceding subframe signal The value of the characteristic determination result sub_prm1 is output as sub_prm1 = 2.

  The voiced / unvoiced sound determination unit 10322 sets the value of the signal characteristic determination result sub_prm to sub_prm = 8, the previous subframe signal, when it is determined that the immediately preceding subframe is an unvoiced sound and the subframe being processed is an unvoiced sound. The value of the characteristic determination result sub_prm1 is output as sub_prm1 = 2.

(Volume control operation)
The subframe unit gain calculation unit 104 receives the signal characteristic determination result sub_prm and calculates a subframe unit gain sub_pow, which is a reference value for performing volume control, for each subframe. More specifically, the value of the subframe unit gain sub_pow is set to sub_pow = f (a) (a = 0, 1, 2,... 8) according to the value of the signal characteristic determination result sub_prm. f (a) is a value set in advance corresponding to each of the eight signal characteristic determination results sub_prm. For example, when the value of the signal characteristic determination result sub_prm is sub_prm = 6, the value of the subframe unit gain sub_pow is set as sub_pow = f (6).

  Here, when the value of the signal characteristic determination result sub_prm is any of sub_prm == 6, sub_prm == 7, sub_prm == 8, that is, when it is determined that the subframe being processed is an unvoiced sound, The audio signal of the subframe being processed includes a frequency component of 4 kHz or higher. As described above, when performing a process of amplifying a frequency component of 4 kHz or more in order to correct the output frequency characteristic of the acoustic output device, there is a possibility that an audio signal including an unvoiced sound will cause amplitude saturation. Therefore, in order to avoid a situation where the output audio signal causes amplitude saturation, when it is determined that the subframe being processed is an unvoiced sound, the value of the subframe unit gain sub_pow includes the value of the subframe being processed. A value that attenuates the output volume is set. For example, a value such that 0 <sub_pow <1 is set as the value of the subframe unit gain sub_pow.

Furthermore, the value of the subframe unit gain sub_pow may be set so that 0 <f (6) <f (7) <f (8) <1. The case where sub_prm = 8 is a case where it is determined that the subframe being processed is an unvoiced sound and that the immediately preceding subframe is also determined to be an unvoiced sound. In such a case, by setting f (8) to a value close to 1, the amplitude value of the immediately preceding subframe and the amplitude value of the subframe being processed are close to each other. Thereby, the user who hears the sound output from the output acoustic device can hear natural sound with little change in volume. On the other hand, the case of sub_prm (6) is a case where it is determined that the subframe being processed is an unvoiced sound and that the immediately preceding subframe is determined to be silent. In such a case, by setting f (6) to a value close to 0, the amplitude value of the subframe being processed is greatly reduced. As a result, even when a silent sound is suddenly output from the silent state, it is possible to quickly follow the change in volume and perform a process of reducing the amplitude value, thereby preventing a situation where amplitude saturation occurs.

On the other hand, when the value of the signal characteristic determination result sub_prm is a value other than sub_prm == 6, sub_prm == 7, sub_prm == 8, that is, when the subframe being processed is determined to be silent or voiced. Is less likely to cause amplitude saturation even when a process of amplifying a frequency component of 4 kHz or higher is performed to correct the output frequency characteristics of the acoustic output device. Therefore, when it is determined that the subframe being processed is silent or voiced, the value of the subframe unit gain sub_pow is set as, for example, sub_pow = 1.

  The sample unit gain calculator 105 receives the signal characteristic determination result sub_prm and the subframe unit gain sub_pow, and controls the sample unit gain sm_pow [s] for controlling the amplitude value of each sample constituting the subframe signal sub_hpb [s]. ] (S = 1, 2, 3,... S). The sample unit gain calculation unit 105 includes a subframe change amount calculation unit 1051, a window function adaptation unit 1052, and a delay unit 1053.

The sample unit gain calculation unit 105 outputs a sample unit gain sm_pow [s] that smoothly transitions so as to maintain differential continuity for each sample based on the value of the subframe unit gain sub_pow [s] by a process described later. To do. Thereby, it is possible to perform volume control while maintaining a natural volume change. The sample unit gain calculation unit 105 changes the transition speed of the sample unit gain sm_pow [s] according to the signal characteristic determination result sub_prm.

  The subframe change amount calculation unit 1051 receives the subframe unit gain sub_pow and the immediately preceding subframe limiter control gain limit_pow1 and outputs the limiter control gain limit_pow. The limiter control gain limit_pow is a value used when the window function adaptation unit 1053 described later calculates the sample unit gain sm_pow [s]. The subframe change amount calculation unit 1051 sets the value of the limiter control gain limit_pow to a value close to the limiter control gain limit_pow1 of the immediately preceding subframe and outputs the set value. As a result, the amplitude value of the immediately preceding subframe and the amplitude value of the subframe being processed can be changed within a range that is naturally felt by the user. More specifically, the subframe change amount calculation unit 1051 compares the difference in value between the limiter control gain limit_pow1 and the subframe unit gain sub_pow of the immediately preceding subframe. When the difference between the two values falls below a predetermined threshold γ, that is, when | sub_pow−limit_pow1 | <1 / γ, the value of the limiter control gain limit_pow is set to be limit_pow = sub_pow. On the other hand, if the difference between the two values exceeds the threshold γ, that is, (sub_pow−limit_pow1)> 1 / γ, the value of the limiter control gain limit_pow is set as limit_pow = limit_pow1 × γ. If (sub_pow-limit_pow1) <1 / γ, the limiter control gain limit_pow is set as limit_pow = limit_pow1 × (1 / γ). When | sub_pow−limit_pow1 | = 1 / γ, any of the above-described values may be set as the value of the limiter control gain limit_pow. Here, the threshold γ is felt as a natural amount of change when the user listens to the sound output from the acoustic output device when the difference in amplitude between the immediately preceding subframe and the subframe being processed changes. Set the limit value. For example, when the time during which the subframe signal sub_hpb [s] is output is 4 ms, the threshold γ is set to a value within the range of 1.25 to 1.50. Note that the value of the threshold γ is not limited to the value described here.

  The delay unit 1052 holds the limiter control gain limit_pow in the immediately preceding subframe as an input, and outputs it as the limiter control gain limit_pow1 in the immediately preceding subframe. When the control unit 100 starts the sound volume control process, that is, when the value of the limiter control gain limit_pow1 of the immediately preceding subframe does not exist, the value of the limiter control gain limit_pow1 of the immediately preceding subframe is output as limit_pow1 = 1. .

  The window function adaptation unit 1053 receives the limiter control gain limit_pow, the previous subframe limiter control gain limit_pow1, and the signal characteristic determination result sub_prm, and calculates the window function window [s] according to Equation 1. Further, the sample unit gain sm_pow [s] is calculated according to the window function window [s]. Then, the value of the sample unit gain sm_pow [n] obtained by collecting the sample unit gain sm_pow [s] obtained for each subframe for each frame is output. Further, the window function adaptation unit 1053 changes the transition speed of the window function window [s] according to the signal characteristic determination result sub_prm. More specifically, first, the window function adaptation unit 1053 reads the value of the signal characteristic determination result sub_prm. Then, the window function adaptation unit 1053 sets the values of the smoothing parameters (XGate, YGate) according to the value of the signal characteristic determination result sub_prm. For the setting of the smoothing parameters (XGate, YGate), for example, the window function adaptation unit 1053 holds table data that associates the value of the signal characteristic determination result sub_prm with the value of the smoothing parameter (XGate, YGate). This is done by setting a value according to the data. The smoothing parameter XGate is a parameter that determines the slow speed of the change of the window function window [s] with respect to the time direction, and a value of 0 <XGate ≦ 1 is used. The smoothing parameter YGate is a parameter that determines the slow speed of the change in the window function window [s] with respect to the gain direction.

Hereinafter, the limiter control gain limit_pow, the limiter control gain limit_pow1 of the immediately preceding subframe, and the signal characteristic determination result sub_prm are input, and the window function window [s], the sample unit gain sm_pow [s], and each subframe are calculated. Equation 1 is shown as an equation for calculating the sample unit gain sm_pow [n] in which the sample unit gain sm_pow [s] is summarized for each frame.

  First, the smoothing parameters (XGate, YGate) are input, and the window function window [s] is calculated from Equation 1 (a, b, c). Next, the sample unit gain sm_pow [s] is calculated from Equation 1 (d) using the calculated window function window [s] as an input. The sample unit gain sm_pow [s] is calculated as a function of transition from the limiter control gain limit_pow1 of the immediately preceding frame to the limiter control gain limit_pow. Finally, the sample unit gain sm_pow [s] calculated for each subframe is combined to calculate the sample unit gain sm_pow [n].

FIG. 6 shows how the sample unit gain sm_pow [s] changes according to the values of the smoothing parameters (XGate, YGate). FIG. 6A shows the sample unit gain sm_pow [s] when the values of the smoothing parameters (XGate, YGate) are set to (XGate, YGate) = (0.5, 1.0). . FIG. 6B shows the sample unit gain sm_pow [s] when the values of the smoothing parameters (XGate, YGate) are set to (XGate, YGate) = (0.25, 1.0). . It can be seen that the value of the sample unit gain sm_pow [s] shown in FIG. 6B transitions at a faster speed in the time direction than the sample unit gain sm_pow [s] shown in FIG. FIG. 6C shows the sample unit gain sm_pow [s] when the values of the smoothing parameters (XGate, YGate) are set to (XGate, YGate) = (0.5, 2.0). . FIG. 6D shows the sample unit gain sm_pow [s] when the values of the smoothing parameters (XGate, YGate) are set to (XGate, YGate) = (0.5, 3.0). . It can be seen that the sample unit gain sm_pow [s] shown in FIG. 6D transitions at a faster speed in the gain direction than the sample unit gain sm_pow [s] shown in FIG.

  For example, when the value of the signal characteristic determination result sub_prm is sub_prm = 6, that is, when the immediately preceding subframe is silent and the subframe being processed is unvoiced, it is expected that the amplitude value is switched in a switch manner. Is done. In such a case, the smoothing parameters (XGate, YGate) are set to large values and the transition of the sample unit gain sm_pow [s] is advanced so that the control following the switching of the amplitude value is performed. Can do. On the other hand, when the value of the signal characteristic determination result sub_prm is sub_prm = 7, that is, when it is determined that the immediately preceding subframe is voiced sound and the subframe being processed is unvoiced sound, the amplitude value gradually increases. The situation is expected to change. In such a case, the value of the smoothing parameter (XGate, YGate) is set to a small value, and the amplitude value is gradually controlled by delaying the transition of the sample unit gain sm_pow [s].

When the value of the signal characteristic determination result sub_prm is other than this, that is, when sub_prm = 1, 2, 3, 4, 5, 8, the values of the smoothing parameters (XGate, YGate) are set to the values of sub_prm = 6 and sub_prm = 7. It may be set to an intermediate value compared to the case. Thereby, the volume of the sound output from the acoustic output device can be naturally changed for the user.

  In this embodiment, Expression 1 is given as an example of a function for the window function adaptation unit 1053 to calculate the sample unit gain sm_pow [n]. However, the method of calculating the sample unit gain sm_pow [n] used by the present invention is not limited to this, and the sample unit gain sm_pow [n] may change the slow transition of the value while maintaining differential continuity. For example, you may calculate using another function.

  The volume control unit 106 receives the signal hp [n] and the sample unit gain sm_pow [n], changes the amplitude value included in the signal hp [n], and outputs this as the pre-process signal d [n]. To do. More specifically, for example, the pre-process signal d [n] is calculated by integrating the signal hp [n] and the sample unit gain sm_pow [n]. As a result, the amplitude of the pre-processed signal d [n] changes along with the sample unit gain sm_pow [n]. Here, it has been described that the volume control unit 106 calculates the pre-process signal d [n] by integration of the signal hp [n] and the sample unit gain sm_pow [n] as an example. However, the operation of the volume control unit 106 used in the present invention is not limited to this. For example, the signal hp [n] may be multiplied by a power of the sample unit gain sm_pow [n], or The pre-processed signal d [n] may be output by weighting the amplitude of the signal hp [n] according to the sample unit gain sm_pow [n] by another method.

The frequency characteristic correction unit 107 receives the pre-processed signal d [n], corrects the pre-processed signal d [n] according to the frequency characteristic of the sound output device, and outputs it as an output signal y [n]. For example, when the acoustic output device outputs an output signal y [n] having a frequency of 4 kHz or higher, the frequency characteristic is such that the output volume is attenuated as compared with the output of the output signal y [n] having a frequency of 4 kHz or lower. Consider having it. In this case, the frequency characteristic correction unit 107 amplifies a frequency component of 4 kHz or more included in the pre-process signal d [n] and outputs the amplified signal as an output signal y [n]. As a result, the attenuation of the frequency component of 4 kHz or more by the acoustic output device and the amplification of the frequency component of 4 kHz or more performed by the frequency characteristic correction unit 107 cancel each other. Thereby, the user who hears the sound output from the sound output device can experience the original sound quality of the input sound. More specifically, the frequency characteristic correction unit 107 performs, for example, FFT (Finite Fourier transform) processing on the pre-processed signal d [n] and orthogonally transforms it into a frequency domain signal. Frequency characteristics are added to or subtracted from the orthogonally transformed frequency domain signal in accordance with the acoustic output device. For example, IFFT (InverseFinit F) is applied to a frequency domain signal subjected to frequency characteristic addition / subtraction.
(orientor transform) processing and the like, and orthogonal transform into a signal in the time domain. The orthogonally transformed time domain signal is output as an output signal y [n]. In the present embodiment, as an example of the operation of the frequency characteristic correcting unit 107, the operation of correcting the frequency characteristic by converting the frequency of the preprocess signal d [n] has been described. However, the operation of the frequency characteristic correction unit 107 in the present invention is not limited to this, and for example, the orthogonal transformation of the preprocessed signal d [n] to the frequency domain is omitted, and the preprocessed signal d [n] in the time domain is omitted. Similar effects of the present invention can also be obtained by applying an IIR (Infinite Impulse Response) filter or FIR (Finite Impulse Response) filter processing.

(Flow of volume control processing by the control unit)
FIG. 7 is a flowchart illustrating a substantial processing flow when the input signal x [n] input to the control unit 100 is output as the output signal y [n]. Below, the process which the control part 100 performs is described along FIG.

  First, the input signal x [n] input to the control unit 100 is input to the DC component removal filter unit 101 (step 1001). The DC component removal filter unit 101 removes the DC component included in the input signal x [n] and outputs the signal hp [n] obtained by extracting only the audio signal to the subframe division unit 102 and the volume control unit 106. .

The signal hp [n] is input to the subframe dividing unit 102 (step 1002). The subframe dividing unit 102 divides the signal hp [n] composed of N samples into the subframe signal sub_hpb [s] composed of S samples, and outputs it to the signal characteristic detection unit 103.

The subframe signal sub_hpb [s] is input to the subframe amplitude maximum value detection unit 10311 (step 1003). The subframe amplitude maximum value detection unit 10311 extracts a sample having the maximum amplitude value from the samples in the subframe signal sub_hpb [s], and uses the amplitude value as the subframe maximum amplitude value sub_max. The sound is output to the silence determination unit 10312.

The subframe maximum amplitude value sub_max is input to the sound / silence determination unit 10312 (step 1004). The voice / silence determination unit 10312 compares the subframe maximum amplitude value sub_max with the voice / silence threshold α to determine whether the subframe signal sub_hpb [s] is voiced or silent.

If it is determined that the subframe signal sub_hpb [s] is sound (“sound” in step 1004), the subframe signal sub_hpb [s] is received by the zero cross number detection unit 10321 that receives the input of sub_hpb [s]. The zero cross point included in the frame is detected (step 1005). When the zero cross point is detected, the number of zero cross points included in the subframe signal sub_hpb [s] is output to the voiced / unvoiced sound determination unit 10322 as the zero cross number sub_zc.

On the other hand, if it is determined that the subframe signal sub_hpb [s] is silent (“silence” in step 1004), detection of the zero cross point by the zero cross number detection unit 10321 is omitted, and the value of the zero cross number sub_zc is sub_zc = 0. Is output to the voiced / unvoiced sound determination unit 10322.

The zero-cross number sub_zc is input to the voiced / unvoiced sound determination unit 10322 (step 1006). Voiced / unvoiced sound determination section 10322 compares the number of zero crosses sub_zc and threshold β to determine whether subframe signal sub_hpb [s] is an unvoiced sound.

In step 1004 and step 1006, it is determined whether sub_hpb [s] is silent, voiced, or unvoiced, and this determination result is a signal characteristic determination result sub_prm. The data is output to the window function adaptation unit 1052.

The signal characteristic determination result sub_prm is input to the subframe unit gain calculation unit 104. When the subframe signal sub_hpb [s] is silent or voiced (“silence” in step 1004 and “voiced” in step 1006), the subframe unit gain calculation unit 104 performs subframe unit gain. The value of sub_pow is set to sub_pow = 1 and is output to subframe unit gain calculation section 1051 (step 1008).

On the other hand, when the subframe signal sub_hpb [s] is an unvoiced sound (“unvoiced sound” in step 1006), the value of the subframe unit gain sub_pow is set to a value satisfying sub_pow <1, and the subframe unit gain calculation unit 1051 (Step 1007). Here, as described in the section of the operation of the subframe unit gain calculation unit 104, the value of the subframe unit gain sub_pow is whether the immediately preceding subframe is silent, voiced, or unvoiced. You may change according to.

The subframe unit gain sub_pow is input to the subframe change amount calculation unit 1051 (step 1009). The subframe change amount calculation unit 1051 compares the subframe unit gain sub_pow with the limiter control gain limit_pow1 of the immediately preceding frame. When the difference between the subframe unit gain sub_pow and the limiter control gain limit_pow1 of the immediately preceding frame is large, the subframe change amount calculation unit 1051 sets the limiter control gain limit_pow to obtain a natural volume change based on the threshold γ. Reset and output to the window function adaptation unit 1052.

  The limiter control gain limit_pow is input to the window function adaptation unit 1052. The window function adaptation unit 1052 determines whether the immediately preceding subframe signal is sound or silence from the signal characteristic determination result sub_prm. If this is silence (“silence” in step 1010), the window function adaptation unit 1052 quickly changes from the value of the previous limiter control gain limit_pow1 to the value of the limiter control gain limit_pow of the frame being processed. sm_pow [s] is calculated (step 1011).

On the other hand, if the immediately preceding subframe signal is sound (“sound” in step 1010), the window function adaptation unit 1052 determines the value of the limiter control gain limit_pow of the frame being processed from the value of the immediately previous limiter control gain limit_pow1. The sample unit gain sm_pow [s] that gradually transitions to is calculated (step 1012). The window function adaptation unit 1052 combines the calculated sample unit gains sm_pow [s] to generate a sample unit gain sm_pow [n] (step 1013). The generated sample unit gain sm_pow [n] is output to the volume control unit 106.

  The sample unit gain sm_pow [n] is input to the volume control unit 106 (step 1014). The volume control unit 106 changes the value of the amplitude included in the signal hp [n] according to the sample unit gain sm_pow [n], and outputs this to the frequency characteristic correction unit 107 as the pre-processed signal d [n]. To do.

  The pre-process signal d [n] is input to the frequency characteristic correction unit 107 (step 1015). The frequency characteristic correction unit 107 corrects the preprocessed signal d [n] according to the frequency characteristic of the acoustic output device connected to the control unit 100, and outputs the corrected signal as an output signal y [n].

  Through the series of processes described above, the control unit 100 determines whether or not an unvoiced sound is included in the input sound signal. If the unvoiced sound is included, the control unit 100 performs a process of reducing the amplitude value of the sound signal. This is a case where an audio signal in a band of 4 kHz or more is amplified in accordance with the output frequency characteristic of the acoustic output device by performing a process of reducing the amplitude value for a portion including an unvoiced sound in the audio signal. However, it is possible to prevent the audio signal from causing amplitude saturation.

Furthermore, the control unit 100 calculates the limiter control gain limit_pow for each subframe, and calculates the sample unit gain sm_pow [n] so as to smoothly change the limiter control gain limit_pow. Since the amplitude of each sample is corrected according to the sample unit gain sm_pow [n], the user is not given an unnatural impression due to a sudden change in volume.

Further, the control unit 100 abruptly changes the sample unit gain sm_pow [n] when an unvoiced sound is suddenly generated from the silent state. As a result, sudden amplitude saturation can be prevented and volume control can be performed with high accuracy in terms of time. On the other hand, when changing from voiced sound to unvoiced sound, amplitude saturation does not occur suddenly, so the sample unit gain sm_pow [n] is gradually changed. As a result, the amplitude can be controlled without giving an unnatural impression to the user.

(Second embodiment)
FIG. 8 is a diagram illustrating a configuration of the control unit 100 and the second control unit 200 connected to the control unit 100, which are built in the electronic apparatus according to the second embodiment of the present invention.

The output signal y [n] output from the control unit 100 is a signal in which the amplitude of the unvoiced sound part included in the input signal x [n] is attenuated. Therefore, a situation in which amplitude saturation caused by amplification of unvoiced sound occurs is prevented. By the way, the input signal x [n] generates, for example, the accent of the speaker who generated the input signal x [n], the strength of music that has become the original signal for generating the input signal x [n], or the input signal. Due to changes in the state of the device, etc., there will be a large difference in the amplitude. The output signal y [n] output from the control unit 100 still includes a large difference in amplitude. When such a signal with a large change in amplitude is output from the sound output device, for example, a sudden loud sound may be output from the sound output device from a silent state, which may give an unpleasant impression to the user of the mobile phone, or the volume. There is a possibility that the user of the mobile phone cannot hear the voice because the small voice is output from the acoustic output device.

Therefore, in the second embodiment, the second control unit 200 is further connected to the control unit 100, and the amplitude of the output signal y [n] is normalized by the second control unit. Then, the normalized output signal y [n] is output to the acoustic output device as the output signal z [n]. By normalizing the amplitude, it is possible to output an easily audible audio signal with a constant volume.

Furthermore, when an audio signal that suddenly generates audio from a silent state is input, the change in amplitude gain is also abrupt. As a result, it is possible to avoid a situation in which a large sound volume is suddenly output from the silent state, and to perform sound volume control with high accuracy in time. On the other hand, when an audio signal that continues the audio is input, the change in the amplitude gain is made gradual. As a result, the amplitude can be controlled without giving an unnatural impression to the user.

Hereinafter, the configuration and role of each component of the second control unit 200 shown in FIG. 3 will be described. The control unit 200 includes a sound / silence detection unit 201, a frame amplitude maximum value detection unit 202, a frame unit gain calculation unit 203, a second sample unit gain calculation unit 204, and a volume normalization unit 205.

The voice / silence detection unit 201 receives the signal hp [n] output from the DC component control filter 101 and determines whether the signal hp [n] is voiced or silent. Also, using information on whether the signal hp [n] of the immediately preceding frame is sounded or silent, information indicating the switching between sounded and silenced between the immediately preceding frame and the frame being processed is used as sounded. -Output as silence detection result fr_um. The voice / silence detection unit 201 performs the same process as the voice / silence detection unit 1031 that receives the subframe signal sub_hpb [s] described in the first embodiment, to generate the voice / silence of hp [n]. To determine.

In the following, the sound / silence detection result fr_um and the sound / silence detection result fr_um1 of the immediately preceding frame are calculated depending on whether the frame being processed is sound / silence or whether the immediately preceding frame is sound / silence. Indicates conditions.

  When it is determined that both the immediately preceding frame and the frame being processed are silent, the value of the sound / silence detection result fr_mum is 0, and the value of the sound / silence detection result fr_mum1 of the immediately previous frame is fr_mum1 = 0. And

  When it is determined that the previous frame is silent and the frame being processed is voiced, the value of the voiced / silent detection result fr_um = 1 is set to fr_mum = 1, and the value of the voiced / silent detection result fr_mum1 of the immediately preceding frame is set. Is fr_um1 = 1.

  When it is determined that the immediately preceding frame is sound and the frame being processed is silent, the value of the sound / silence detection result fr_mum is set to fr_mum = 2, and the value of the sound / silence detection result fr_mum1 of the immediately preceding frame is determined. Is fr_um1 = 0.

  When it is determined that both the previous frame and the frame being processed are sound, the value of the sound / silence detection result fr_mum is 3, and the value of the sound / silence detection result fr_mum1 of the immediately previous frame is fr_mum1 = Set to 1.

The frame amplitude maximum value detection unit 202 receives the sound / silence detection result fr_um and the output signal y [n], and detects the maximum value of the amplitude included in the output signal y [n]. The maximum value of the detected amplitude is output as the frame amplitude maximum value fr_max. Further, when the value of the input sound / silence detection result fr_um is fr_um = 0, the maximum value detection operation may be omitted and the value of the frame amplitude maximum value fr_max may be output as fr_max = 0. . Further, it has been stated that the frame amplitude maximum value detection unit 202 detects and outputs the maximum value of the amplitude included in the signal hp [n]. However, the value output by the frame amplitude maximum value detection unit 203 is not limited to this, and the average amplitude level of the signal hp [n] may be output. Moreover, you may output the average spectral power of signal hp [n].

The frame unit gain calculation unit 203 receives the frame amplitude maximum value fr_max output from the frame amplitude maximum value detection unit 202 and the sound / silence detection result fr_um output from the sound / silence detection unit 201, and performs volume control. A frame unit gain fr_pow, which is a reference value for performing, is calculated for each frame. More specifically, the value of the frame unit gain fr_pow is set to fr_pow = f (b) (b = 0, 1, 2, 3) according to the value of the sound / silence detection result fr_um. Here, in the case where the value of the sound / silence detection result fr_um is either fr_um == 0 or fr_um == 2, that is, when it is determined that the frame being processed is silent, There is no need to perform control to increase or decrease the amplitude. Therefore, when it is determined that the frame being processed is silent, the value of the frame unit gain fr_pow is set to fr_pow = 1. On the other hand, if the value of the sound / silence detection result fr_um is either fr_um == 1 or fr_um == 3, that is, if it is determined that the frame being processed is sound, the frame being processed is It is necessary to normalize the amplitude so that the sound volume becomes constant when output from the sound output device. Therefore, when it is determined that the frame being processed is sound, the value of the frame unit gain fr_pow is set to fr_pow = (OUT_LEVEL / fr_max). Here, the volume normalization target level OUT_LEVEL is an amplitude level that is an output target. A predetermined value may be used as the volume normalization target level OUT_LEVEL, or a value determined by the user of the mobile phone as necessary.

  The second sample unit gain calculation unit 204 receives the frame unit gain fr_pow and the sound / silence detection result fr_um output from the frame unit gain calculation unit 203 as input, and calculates the amplitude of each sample constituting the output signal y [n]. A second sample unit gain auto_gain [n] for controlling the value is output. The second sample unit gain calculation unit 204 receives the signal characteristic determination result sub_prm and the subframe unit gain sub_pow and outputs the second sample by the same process as the sample unit gain calculation unit 105 that outputs the sample unit gain sm_pow [s]. The unit gain auto_gain [n] is output. The above-described sample unit gain calculation unit 105 changes the slow speed of the transition of the sample unit gain sm_pow [s] according to the signal characteristic determination result sub_prm. On the other hand, the second sample unit gain calculation unit 204 changes the slowness of the transition of the second sample unit gain auto_gain [n] according to the sound / silence detection result fr_um. For example, when the value of the voice / silence detection result fr_um is fr_um == 1, that is, the previous frame is silent and the frame being processed is voiced, the change from silent to voice is switched. It is conceivable that the situation will happen suddenly. In such a case, the volume can be corrected with good follow-up even when the volume of the switch is switched by advancing the transition of the second sample unit gain auto_gain [n]. On the other hand, when the value of the voiced / silent detection result fr_um is fr_um == 3, that is, when both the immediately preceding frame and the frame being processed are both voiced, the volume may change relatively slowly. Conceivable. In such a case, by delaying the transition of the second sample unit gain auto_gain [n], the volume can be corrected with a natural correction amount for the user of the mobile phone.

  The volume normalization unit 205 receives the output signal y [n] and the sample unit gain auto_gain [n] and changes the amplitude value of each sample of the output signal y [n]. More specifically, the volume normalization unit 205 calculates and outputs the output signal z [n] from the integration of the output signal y [n] and the sample unit gain auto_gain [n]. For example, when the sample unit gain auto_gain [n] is a value smaller than 1, the output signal z [n] is attenuated to a value having a smaller amplitude than the output signal y [n]. On the other hand, for example, when the sample unit gain auto_gain [n] is a value larger than 1, the output signal z [n] is amplified to a value having a larger amplitude than the output signal y [n]. The output signal z [n] along the volume normalization target level OUT_LEVEL can be obtained by manipulating the amplitude value by the volume normal lower part 205.

(Flow of volume control processing by second control unit)
FIG. 9 is a flowchart describing a substantial process flow when the output signal y [n] input to the second control unit 200 is output as the output signal z [n]. Below, the process which the 2nd control part 200 performs is described along FIG.

  First, the output signal y [n] input to the second control unit 200 is input to the frame amplitude maximum value detection unit 202 (step 2001). The frame amplitude maximum value detection unit extracts a sample having the maximum amplitude value from the samples in the output signal y [n], and sets the amplitude value as the frame maximum amplitude value fr_max to the frame unit gain calculation unit 203. Is output.

  On the other hand, the signal hp [n] input to the second control unit 200 is input to the voice / silence detection unit 201 (step 2002). The voice / silence detection unit 201 determines whether the signal hp [n] is voiced or silent from the amplitude value of the signal hp [n]. The voice / silence detection unit 201 determines whether the immediately preceding frame is voiced or silent, and the voiced / silent detection result fr_um that determines whether the frame being processed is voiced or silent. This is output to the frame unit gain calculation unit 203.

  The voiced / silent detection result fr_um and the maximum frame amplitude value fr_max are input to the frame unit gain calculation unit 203. The frame unit gain calculation unit 203 determines whether the input signal y [n] is sound or sound from the sound / silence detection result fr_um (step 2002). When it is determined that the input signal y [n] is sound (“sound” in step 2002), the frame unit gain calculation unit 203 outputs a frame unit gain fr_pow in which the frame amplitude maximum value fr_max is standardized ( Step 2003).

  On the other hand, when it is determined that the input signal y [n] is silent (“silence” in step 2002), the frame unit gain calculation unit 203 outputs the value of the frame unit gain fr_pow = 1 as fr_pow = 1 (step 2004). .

  The frame unit gain fr_pow and the sound / silence detection result fr_um are input to the second sample unit gain calculation unit 204. The second sample unit gain calculation unit 204 determines whether the input signal y [n] of the immediately preceding frame is sound or sound from the sound / silence detection result fr_um (step 2005). When it is determined that the input signal y [n] of the immediately preceding frame is silent (“silence” in step 2005), the second sample unit gain calculating unit 204 calculates the frame being processed from the frame unit gain fr_pow1 of the immediately preceding frame. A sample unit gain auto_gain [n] that makes a transition to the frame unit gain fr_pow is calculated (step 2006).

  On the other hand, when it is determined that the input signal y [n] of the immediately preceding frame is sound (“sound” in step 2005), the second sample unit gain calculator 204 determines the frame unit gain fr_pow1 of the immediately previous frame. Sample unit gain auto_gain [n] that gradually transitions from to the frame unit gain fr_pow of the frame being processed (step 2007).

  The sample unit gain auto_gain [n] is input to the volume normalization unit 205 (step 2008). The volume normalization unit 205 changes the amplitude value included in the output signal y [n] according to the amplitude gain auto_gain [n], and outputs this as the output signal z [n].

  Through the series of processes described above, the second control unit 200 normalizes the amplitude value of the output signal y [n] output from the control unit 100 according to the volume normalization target level OUT_LEVEL, and normalizes the output signal z [n ] Is output. By normalizing the amplitude, it is possible to output an easily audible audio signal with a constant volume.

Furthermore, when an output signal y [n] that suddenly generates sound from a silent state is input, the amplitude gain auto_gain [n] also changes rapidly. As a result, it is possible to avoid a situation in which a large sound volume is suddenly output from the silent state, and to perform sound volume control with high accuracy in time. On the other hand, when the output signal y [n] that continues the voice is input, the change of the amplitude gain auto_gain [n] is made gradual. As a result, the amplitude can be controlled without giving an unnatural impression to the user.

In addition, this invention is not limited to the said embodiment, As long as it is a range which does not deviate from the summary of invention, you may deform | transform and embody a component. Various inventions may be formed by proper combinations of a plurality of constituent elements disclosed in the above embodiments. For example, some components may be deleted from all the components shown in each embodiment.

DESCRIPTION OF SYMBOLS 100 Control part 101 DC component control filter 102 Sub-frame division | segmentation part 103 Signal characteristic detection part 1031 Signal characteristic detection part 10311 Sub-frame amplitude maximum value detection part 10312 Voice / silence determination part 1032 Voiced / unvoiced sound detection part 10321 Zero cross number detection part 10322 Voiced / unvoiced sound determination unit 104 Subframe unit gain calculation unit 105 Sample unit gain calculation unit 1051 Subframe change amount calculation unit 1052 Window function adaptation unit 1053 Delay unit 106 Volume control unit 107 Frequency characteristic correction unit 200 Second control unit 201 Sound / Silence Detection Unit 202 Frame Amplitude Maximum Value Detection Unit 203 Frame Unit Gain Calculation Unit 204 Second Sample Unit Gain Calculation Unit 205 Volume Normalization Unit

Claims (5)

Voice input receiving means for receiving voice input;
Volume measuring means for measuring the input volume and frequency component of the sound received by the sound input receiving means;
When the frequency component included in the predetermined frequency band out of the sound received during the predetermined time interval exceeds the predetermined energy, the volume according to the value of the input volume of the audio received during the predetermined time interval Volume gain setting means for setting the gain;
Audio output means for outputting the sound received during the predetermined time interval at a volume corresponding to the volume gain set by the volume gain setting means;
An electronic device comprising: Voice input receiving means for receiving voice input;
Volume measuring means for measuring the input volume and frequency component of the sound received by the sound input receiving means;
When the frequency component included in the predetermined frequency band out of the sound received during the predetermined time interval exceeds the predetermined energy, the volume according to the value of the input volume of the audio received during the predetermined time interval Volume gain setting means for setting the gain;
The volume gain transition function
At the end of the output of the sound received during the first time interval, the sound volume gain setting means sets the volume gain value for the sound received during the first time interval,
At the start of the output of the sound received during the first time interval, the sound volume gain setting means sets the value of the volume gain set by the sound volume received during the second time interval before the first time interval. Volume gain transition function setting means for setting to transition,
Voice output means for outputting the voice received in the first time interval at a volume corresponding to the volume gain transition function;
An electronic device comprising: The volume gain transition function setting means includes:
Set the first volume gain transition function when the input volume of the sound received in the first time interval and the second time interval is less than or equal to a predetermined volume,
The first volume gain transition function when the audio input volume received during the first time interval is greater than or equal to the predetermined volume and the audio input volume received during the second time interval is less than or equal to the predetermined volume. Compared to, set a second volume gain transition function that suddenly transitions to the volume gain value of the first time interval,
When the input volume of the sound received in the second time interval is equal to or higher than the predetermined volume, the volume gain value gradually changes in the first time interval compared to the first volume gain transition function. The electronic device according to claim 1, wherein a third volume gain transition function is set. The volume gain transition function setting means includes:
Set the first volume gain transition function when the input volume of the sound received in the first time interval and the second time interval is less than or equal to a predetermined volume,
The audio received during the first time interval has a frequency component included in the predetermined frequency band in the spectrum distribution exceeding the predetermined energy, and the input volume of the audio received during the second time interval is Set the second volume gain transition function that transitions to the value of the volume gain set in the first time interval abruptly compared to the first volume gain transition function when the volume is below a predetermined volume,
The audio received during the first time interval has a frequency component included in the predetermined frequency band in the spectrum distribution exceeding the predetermined energy, and the input volume of the audio received during the second time interval is Setting a third volume gain transition function that transitions to the value of the volume gain set in the first time interval more slowly than the first volume gain transition function when the volume is equal to or higher than a predetermined volume. The electronic apparatus according to claim 1, wherein the electronic apparatus is an electronic device. A voice input receiving means for receiving a voice signal input;
Sound volume measuring means for acquiring the input sound volume and frequency component value of the sound signal received by the sound input receiving means;
When the frequency component included in the predetermined frequency band out of the sound received during the predetermined time interval exceeds the predetermined energy, the volume according to the value of the input volume of the audio received during the predetermined time interval Volume gain setting means for setting the gain;
Output volume control means for setting the output volume of the sound received during the predetermined time interval to an output volume corresponding to the volume gain set by the volume gain setting means;
A volume control program for an electronic device, comprising:

Images