JP2003060459A - Automatic gain control(agc) amplifier - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an AGC amplifier capable of keeping the intonation of voice, without omitting the head of the voice and suppressing noises, during the absence of the voice. SOLUTION: A voice level VP, in a voice section of an input signal S1, is calculated by a voice level calculation part 31, and a noise level NP in a non-voice section is calculated by a noise-level calculation part 32. The levels VP and NP are given to an S/N calculation part 33 for calculating S/N signal SN. Also, the level VP is given to a gain calculation part 34 to calculate a calculated gain VG, on the basis of the level difference between the level VP and a reference output level. The signal SN and the gain VG are given to a gain setting part 35, a setting gain AG according to each of the voice and non-voice sections is set, and the gain AG is given to the part 35 as a control signal, with respect to a variable gain amplifier 10. Accordingly, a voice communication which does not give a sense of incongruity can be made available, by setting the gain AG in the non-voice section according to the S/N.

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to a telephone line and a VoI.
The present invention relates to an automatic gain control amplifier (hereinafter referred to as "AGC amplifier") for outputting a voice signal at a predetermined level in P (Voice over Internet Protocol).

[0002]

2. Description of the Related Art Conventionally, as a technique in such a field,
For example, some documents were described in the following documents. Reference: Japanese Patent Laid-Open No. 7-235848

FIG. 2 shows a conventional AG disclosed in the above document.
It is a block diagram of a C amplifier. This AGC amplifier includes a variable gain amplifier (AMP) 1 that amplifies an input signal SI and outputs an output signal SO, a signal detector 2 that detects the presence or absence of the input signal SI, and an output signal SO that has a constant level. The gain control unit 3 controls the gain of the variable gain amplifier 1, and the storage unit 4 holds the gain value when the input signal SI is detected.

In this AGC amplifier, the gain of the variable gain amplifier 1 is sequentially controlled by the gain controller 3 so that the output signal SO becomes a constant level while the signal detector 2 detects the input signal SI. It Next, when the detection of the input signal SI by the signal detection unit 2 is stopped, the gain control unit 3 writes the gain value at that time into the storage unit 4, and the gain control unit 3 outputs the gain value to the variable gain amplifier 1. The control is performed so that the gain becomes the minimum value.

When the input signal SI is input again in this state, the signal detector 2 detects the input signal SI. As a result, the gain value held in the storage unit 4 is read by the gain control unit 3 and set as the initial value of the gain for the variable gain amplifier 1. After that, the gain of the variable gain amplifier 1 is sequentially controlled by the gain controller 3 so that the output signal SO has a constant level.

Therefore, it is possible to increase the gain tracking time constant in the gain control unit 3 to maintain the natural suppression of the voice and to significantly shorten the rise time of the output signal SO when the input signal SI is input. it can.

[0007]

However, the conventional AGC amplifier has the following problems. Since the gain of the variable gain amplifier 1 is switched every time voice is detected, the gain of the variable gain amplifier 1 is at the minimum value until voice is detected. For this reason, there is a problem that the beginning part of the voice becomes low level and the way of hearing becomes unnatural.

Further, in a telephone or the like, not only the voice of the speaker but also ambient noises are input as background noise, but the background noise alone is not detected as the input signal SI. For this reason, there is a problem in that the background noise also stops when the sound is stopped, and the sound becomes silent, which causes an uncomfortable feeling at the switching portion of the sound.

Unlike the above-mentioned AGC amplifier, the signal level in the voice section is measured, the level difference between the signal level and the reference level is calculated to set the gain value, and the gain value is set regardless of the presence or absence of voice. There is also an AGC amplifier adapted to amplify the input signal by using. In this case, if the ratio of noise in the input signal is large (low SN ratio), there is a problem that the noise when there is no voice becomes annoying.

The present invention solves the problems of the above-mentioned prior art, can maintain the voice intonation without cutting the beginning of the voice, and can suppress the noise when there is no voice. The present invention provides an AGC amplifier that can be used.

[0011]

In order to solve the above-mentioned problems, a first invention of the present invention is, in an AGC amplifier, a voice detecting section for detecting a voice section in which voice exists in an input signal, A level calculator that calculates a voice level in a voice section and a noise level in a non-voice section of the input signal, and a signal-to-noise ratio (hereinafter, “S / N”) of the input signal based on the voice level and the noise level. Say)
An S / N calculating section for calculating a calculation gain, a gain calculating section for calculating a calculation gain based on a level difference between a reference output level of an output signal and the voice level, and the input based on the calculation gain and the S / N. The system includes a gain setting unit that generates a set gain for a voice section and a set gain for a non-voice section of a signal, and a variable gain amplifier that amplifies the input signal by the set gain and outputs the output signal.

According to the first invention, as described above, the AGC is used.
Since the amplifier is constructed, the following operations are performed. The gain calculator calculates the calculation gain based on the level difference between the voice level and the reference output level in the voice section of the input signal. The S / N calculator calculates the S / N based on the voice level and the noise level. The calculated gain and the S / N are given to the gain setting unit, and the set gain for the voice section and the set gain for the non-voice section are generated, respectively. The set gain is given to the variable gain amplifier, and the input signal is amplified according to the set gain and output as the output signal.

According to a second aspect of the present invention, the gain calculating section according to the first aspect calculates the calculation gain at a constant cycle so that the amount of change of the calculation gain in each cycle falls within a predetermined fluctuation range.
The value of the calculation gain is limited.

According to the second invention, the following operation is performed. The gain calculation unit calculates the calculation gain at a constant cycle. Then, when the variation amount of the calculation gain for each cycle exceeds the predetermined variation range, the variation amount is limited to fall within the predetermined range.

A third aspect of the present invention is the first or second aspect of the present invention, further comprising a hangover setting section for detecting that the input signal has changed from a voice section to a non-voice section and outputting a hangover signal. Further, the gain setting unit is configured to reduce a change in gain when the hangover signal is given and generate a set gain that is gently changed from a voice section to a non-voice section in a predetermined time. ing.

According to the third invention, the following operation is performed. When the input signal changes from the voice section to the non-voice section, the hangover setting unit outputs a hangover signal. When the hangover signal is given, the gain setting unit alleviates a change in gain at the time of transition from the voice section to the non-voice section, and generates a set gain that gently changes in a predetermined time.

In a fourth aspect based on the first to third aspects, a signal delay section for delaying an input signal by a predetermined time and giving it to a variable gain amplifier is provided. Furthermore, in order to match the timing of the input signal delayed by the signal delay unit,
While delaying the setting gain generated by the gain setting unit,
When the input signal changes from the non-voice section to the voice section, the delay time is used to gently change the set gain in advance so that a predetermined gain value is obtained at the start of the voice section, and then the variable gain amplifier is changed. A gain delay unit is provided.

According to the fourth invention, the following operation is performed. The input signal is delayed by the signal delay unit for a predetermined time and then applied to the variable gain amplifier. On the other hand, the set gain generated by the gain setting unit is delayed by the gain delay unit so as to match the timing of the input signal and is given to the variable gain amplifier. Further, when the input signal changes from the non-voice section to the voice section, the delay time is used in the gain delay unit so that the set gain becomes a predetermined value at the start of the voice section, so that the set gain is gradually changed in advance.

In a fifth aspect based on the first to fourth aspects, the gain setting section uses the calculated gain as the set gain as it is when the input signal is in the voice section, and when the input signal is in the non-voice section. The setting gain is set by reducing the calculation gain according to the S / N.

[0020]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS (First Embodiment) FIG. 1 is a configuration diagram of an AGC amplifier showing a first embodiment of the present invention. This AGC amplifier is a variable gain amplifier 10 that amplifies an input signal SI in a voice band (for example, 300 to 3400 Hz) digitized at a constant sampling frequency (for example, 8 kHz) and outputs an output signal SO, an input signal SI. A voice detection unit 20 for detecting the presence or absence of voice in the inside,
And a gain control unit 30 that controls the variable gain amplifier 10 so that the output signal SO has a predetermined level.

The voice detection unit 20 measures the level of the input signal SI at regular intervals (for example, 10 ms), compares it with a threshold value for voice detection, detects a voice section, and outputs the voice to the gain control unit 30. The detection signal VD is output. A value obtained by multiplying the noise level NP calculated by the gain control unit 30 by a fixed ratio (for example, 3 times) is used as the threshold value for voice detection.

The gain control unit 30 includes a voice level calculation unit 31.
And a noise level calculator 32. The voice level calculation unit 31 averages the signal level of the input signal SI when the voice detection signal VD determines the voice section, and averages the signal level over a constant period (for example, 100 ms), and at regular intervals (for example, 10 ms). The audio level VP is output. Similarly, the noise level calculation unit 32, based on the input signal SI when it is determined that the non-voice section,
The noise level NP is calculated. Voice level VP
Is given to the S / N calculator 33 and the gain calculator 34, and the noise level NP is given to the S / N calculator 33 and the voice detector 20.

The S / N calculation unit 33 uses a fixed period (for example,
Every 10 ms), S / N is calculated based on the voice level VP and the noise level NP, and the S / N signal S is sent to the gain setting unit 35.
It outputs N.

The gain calculator 34 is a voice level calculator 31.
The calculation gain VG is calculated and output based on the level difference between the audio level VP given from the above and the reference output level VS (for example, -20 dBm0) of the output signal SO. The calculation gain VG is limited so as not to exceed a preset maximum value and is given to the gain setting unit 35.

The gain setting section 35 calculates the calculated gain VG and S / N.
Based on the signal SN, the set gain AG for the variable gain amplifier 10 is output corresponding to the voice section and the non-voice section. That is, in the voice section, the calculated gain VG calculated by the gain calculator 34 is output as it is as the set gain AG. On the other hand, in the non-voice section, the calculation gain VG is reduced according to the value of the S / N signal SN and is output as the set gain AG.

FIG. 3 is a signal waveform diagram showing the operation of FIG. The operation of FIG. 1 will be described below with reference to FIG.

The input signal SI is input to the variable gain amplifier 10 and also to the voice detector 20 and the gain controller 30 at the same time.

In the voice detection unit 20, the level of the input signal SI is judged by a voice detection threshold value set based on the noise level NP, and the voice detection signal VD is output. For example, in the voice sections T1, T3, T5 of FIG. 3, the input signal SI exceeds the threshold and the voice detection signal VD becomes "1".
In addition, the non-voice sections T2 and T4 in which the input signal SI is less than or equal to the threshold value.
Then, the voice detection signal VD becomes "0".

When the voice detection signal VD is "1", the level of the input signal SI is calculated by the voice level calculation unit 31, averaged over a certain period, and the voice level VP is output every fixed period. When the voice detection signal VD is “0”, the noise level calculator 32 similarly calculates the level of the input signal SI and outputs the noise level NP.

The voice level VP output from the voice level calculation unit 31 is given to the gain calculation unit 34, and the output signal SO
The calculation gain VG is calculated by the following equation based on the level difference from the reference output level VS of. VG (dB) = VS (dBm) −VP (dBm) For example, if VP = −30 dBm and VS = −20 dBm0, then VG = −20 dBm − (− 30 dBm) = 10 dB. Actually, the calculation gain VG is calculated as follows using the true value instead of the logarithmic (dB) calculation. VG = reference output level / audio level = 0.1 / 0.0316 = 3.16

The noise level NP output from the noise level calculation unit 32 is given to the S / N calculation unit 33 together with the voice level VP output from the voice level calculation unit 31, and S / N is calculated by the following equation at regular intervals. The N signal SN is calculated. SN (dB) = VP (dBm) -NP (dBm)

Voice detection signals VD, S / of the voice detection unit 20
The S / N signal SN of the N calculator 33 and the calculated gain VG of the gain calculator 34 are given to the gain setting unit 35, and the actual set gain AG for the variable gain amplifier 10 is set as follows. The voice section T in which the voice detection signal VD is "1"
1, T3, T5, the calculation gain VG of the gain calculation unit 34
Is output as it is as the set gain AG.

Non-voice section T in which the voice detection signal VD is "0"
At 2 and T4, the set gain AG is set as follows according to the value of the S / N signal SN of the S / N calculator 33. SN ≧ 24 dB ・ ・ ・ AG = VG 18 dB <SN <24 dB ・ ・ ・ AG = (3/4) × VG SN ≦ 18 dB ・ ・ ・ AG = (1/2) × VG

For example, in the non-voice section T2, the S / N signal S
When N is 24 dB or more, the calculation gain VG is the same as the set gain A.
This is the case where it is output as G, and the non-voice section T4 is S /
The case where the N signal SN is 18 dB or less and 1/2 of the calculated gain VG is output as the set gain AG is shown.

As described above, the AG of the first embodiment is
The C amplifier has a gain setting unit 35 that controls the set gain AG in the non-voice section according to the S / N (signal-to-noise ratio) value of the input signal SI. As a result, when the S / N is high, the noise is not offensive to the ears, so by maintaining the same gain as that of the voice section even in the non-voice section, the beginning portion of the voice is not cut, and the intonation of the voice is maintained and the voice does not feel uncomfortable. Can communicate. Further, when the S / N is low, there is an advantage that the noise in the non-voice section can be suppressed by reducing the set gain AG according to the value of the S / N.

(Second Embodiment) FIG. 4 shows a second embodiment of the present invention.
3 is a configuration diagram of a gain control unit showing the embodiment of FIG. This gain control unit 30A is used in place of the gain control unit 30 in FIG. 1, and elements common to those in FIG. 1 are designated by common reference numerals.

The gain control section 30A has a voice level calculation section 31, a noise level calculation section 32, an S / N calculation section 33, and a gain setting section 35, which are the same as those in FIG. Different gain calculation units 34A are provided, and a gain storage unit 36 is added.

The gain calculator 34A includes a gain calculator 34a,
Gain limiter 34b, output reference value VS, and gain limit value LG
have.

The gain calculator 34a is a voice level calculator 3
The calculation gain VG 'is calculated based on the audio level VP given from 1 and the output reference value VS which is the reference output level of the output signal SO. The calculated gain VG 'is provided to the gain limiting unit 34b.

The gain limiter 34b refers to the calculated gain VG 'given from the gain calculator 34a and the previous calculated gain VG0 stored in the gain memory 36, and the difference exceeds the gain limit value LG. A new calculation gain VG is calculated so as not to exist. The calculated gain VG obtained by the gain limiting unit 34b is given to the gain setting unit 35 and stored in the gain storage unit 36 for the next calculation.

Next, the operation of the gain calculator 34A will be described. In the gain calculation unit 34a, the audio level VP given from the audio level calculation unit 31 and the output reference value VS
Based on, the calculated gain VG ′ is calculated by the following equation,
It is given to the gain limiting unit 34b. VG '= VS-VP

In the gain limiter 34b, the gain calculator 3
4a, the calculated gain VG ′ and the gain storage unit 36.
The previous calculated gain VG0 stored in is compared.
Then, a new calculation gain VG is calculated as follows so that the difference does not exceed the gain limit value LG (for example, 6 dB). VG '> VG0 + LG ... VG = VG + LG VG0-LG≤VG'≤VG0 + LG ... VG = VG' VG '<VG0-LG ... VG = VG-LG However, the new calculation gain VG is 1 or less. In the case, VG = 1
And The calculated new calculated gain VG is output to the gain setting unit 35 and is given to the gain storage unit 36.
The contents of the gain storage unit 36 are rewritten with a new calculation gain VG. Other operations are the same as in FIG.

As described above, the gain control unit 30A of the second embodiment has the gain storage unit 36 for storing the previous calculated gain VG, and the previous calculated gain VG0 and the calculated gain VG calculated this time. The gain limiter 34b that calculates a new calculated gain VG so that the difference between the gain limit value LG and
have. As a result, when the level fluctuation of the input signal SI in the voice section is large, the followability is lowered, so that the feeling of intonation of the voice is not impaired, and there is an advantage that it is possible to perform voice communication without discomfort.

(Third Embodiment) FIG. 5 shows a third embodiment of the present invention.
3 is a configuration diagram of a gain control unit showing the embodiment of FIG. This gain control unit 30B is used in place of the gain control unit 30 in FIG. 1, and elements common to those in FIG. 1 are designated by common reference numerals.

The gain control section 30B has a voice level calculation section 31, a noise level calculation section 32, an S / N calculation section 33 and a gain calculation section 34, which are the same as those in FIG. A different gain setting unit 35A is provided, and a hangover setting unit 37 is added.

The hangover setting section 37 detects the timing of change of the input signal SI from the voice section to the non-voice section based on the voice detection signal VD, and gradually sets the set gain AG over a certain remaining time. The hangover signal HO for changing is output to the gain setting unit 35A.

On the other hand, the gain setting section 35A determines the input signal SI
Corresponding to the voice section and the non-voice section of the variable gain amplifier 1
The set gain AG for 0 is output. That is,
In the voice section, the calculated gain V calculated by the gain calculator 34
G is output as it is as the set gain AG. on the other hand,
In the non-voice section, the calculation gain VG is reduced according to the value of the S / N signal SN to calculate the target set gain AG, and the target value is gradually increased over a certain period of time based on the hangover signal HO. The set gain AG is output so that

FIG. 6 is a signal waveform diagram showing the operation of FIG. The operation of FIG. 5 will be described below with reference to FIG. The input signal SI of FIG. 6 is simultaneously input to the variable gain amplifier 10, the voice detection unit 20, and the gain control unit 30B.
The voice detection unit 20 determines the level of the input signal SI, and in the voice sections T1, T3, T5, the voice detection signal VD
Becomes "1", and the voice detection signal VD becomes "0" in the non-voice sections T2 and T4.

When the voice detection signal VD is "1", the level of the input signal SI is calculated by the voice level calculation unit 31 and the voice level VP is output at regular intervals. Also,
When the voice detection signal VD is “0”, the noise level calculation unit 3
The level of the input signal SI is calculated by 2, and the noise level NP is output at regular intervals.

The voice level VP is given to the gain calculator 34, and the calculated gain VG is calculated based on the level difference from the reference output level VS. Further, the noise level NP is given to the S / N calculation unit 33 together with the voice level VP, and the S / N signal SN is calculated at every constant cycle.

Further, the voice detection signal VD is given to the hangover setting section 37, and the change of the input signal SI from the voice section to the non-voice section is detected. When the input signal SI changes from the voice section to the non-voice section, the hangover signal HO is sent from the hangover setting unit 37 to the gain setting unit 35A.
Is output.

The voice detection signal VD, the S / N signal SN, the calculation gain VG and the hangover signal HO are supplied to the gain setting section 3.
5 to the variable gain amplifier 10 from the gain setting section 35.
The actual set gain AG for is set.

In the voice sections T1, T3 and T5, the calculated gain VG is set as it is as the set gain AG. In the non-voice section T2, since the S / N signal SN is 24 dB or more, the calculation gain VG is set as it is as the set gain AG and output.

On the other hand, since the S / N signal SN is 18 dB or less in the non-voice section T4, it is 1/2 of the calculation gain VG.
Is set as the set gain AG. At this time, the set gain AG from the gain setting unit 35 to the variable gain amplifier 10 is
Instead of being output as it is, it is linearly changed over a predetermined time (hangover time HOT).

For example, when the final set gain AG in the non-voice section T4 is VG / 2 and the elapsed time from the start of the non-voice section T4 is tn, the variable gain amplifier 10 is provided in this non-voice section T4. Set gain AG (tn)
Is expressed by the following equation. When tn ≦ HOT: AG (tn) = VG− (VG / 2)
× (tn / HOT) When tn> HOT: AG (tn) = VG / 2

As described above, the gain control section 30B of the third embodiment detects the change of the input signal SI from the voice section to the non-voice section and outputs the hangover signal HO, and the hangover setting section 37. And the hangover signal H
It has a gain setting unit 35A that linearly changes the set gain AG in the non-voice section according to O over a predetermined time. As a result, there is an advantage that the change in the noise level when the input signal SI changes from the voice section to the non-voice section becomes gradual, and the discomfort caused by the level change during conversation can be reduced.

(Fourth Embodiment) FIG. 7 shows a fourth embodiment of the present invention.
3 is a configuration diagram of a gain control unit showing the embodiment of FIG. This gain control section 30C is used in place of the gain control section 30 in FIG. 1, and elements common to those in FIGS. 4 and 5 are designated by common reference numerals.

The gain control section 30C has a voice level calculation section 31, a noise level calculation section 32, an S / N calculation section 33, a gain calculation section 34A and a gain storage section 36 similar to those in FIG. It is configured to have a similar gain setting unit 35A and hangover setting unit 37.

The operation of the gain control section 30C is a combination of the operations of FIGS. 4 and 5, and the maximum change amount of the set gain AG when the input signal SI changes from the voice section to the non-voice section is limited. At the same time, the gain gradually changes due to the hangover setting and shifts to the target set gain AG. As a result, an AGC amplifier having the same advantages as those of the first, second and third embodiments can be obtained.

(Fifth Embodiment) FIG. 8 shows the fifth embodiment of the present invention.
2 is a configuration diagram of the AGC amplifier showing the embodiment of FIG. 1, in which elements common to those in FIG. 1 are designated by common reference numerals.

This AGC amplifier is provided with a gain control section 30D having a slightly different configuration in place of the gain control section 30 of the AGC amplifier shown in FIG. 1, and a signal delay for delaying the input signal SI and giving it to the variable gain amplifier 10. The part 40 is added. In this AGC amplifier, the digitized input signal SI is divided into frames of, for example, 10 ms and processed.

The gain control section 30D has a gain delay section 38 in addition to a voice level calculation section 31, a noise level calculation section 32, an S / N calculation section 33, a gain calculation section 34 and a gain setting section 35 which are the same as those in FIG. have.

The gain delay unit 38 delays the set gain AG output from the gain setting unit 35 by a predetermined time (for example, 10 ms for one frame) to obtain the delay set gain AGD.
Is output. The gain delay unit 38 immediately sets the set gain AG when changing from the non-voice section to the voice section.
Instead of switching between, the delay time for one frame is used to output the delay setting gain AGD that gradually rises over the time of the one frame. The delay setting gain AGD is provided as a gain control signal for the variable gain amplifier 10.

On the other hand, the signal delay section 40 delays the input signal SI and outputs the delayed input signal SID in order to match the timing of the delay setting gain AGD given to the variable gain amplifier 10 and the timing of the input signal SI. is there. The delay time in the signal delay unit 40 is, for example, the delay time (10
ms) and the delay time in the gain delay unit 38 (10 m
s), and is set to 2 frames (20 ms).

FIG. 9 is a signal waveform diagram showing the operation of FIG. The operation of FIG. 8 will be described below with reference to FIG.

In the voice detecting section 20, the input signal SI is divided into frames every 10 ms, the average level is calculated, the presence or absence of voice is judged by the threshold value for voice detection, and the voice detection signal VD is output.

Further, the voice level calculator 31 calculates the level of the input signal SI in the voice section for each frame, and outputs the voice level VP. Similarly, the noise level calculator 32 calculates the level of the input signal SI in the non-voice section and outputs the noise level NP.

The voice level VP is given to the gain calculator 34, and the calculated gain VG is calculated based on the level difference from the reference output level VS. The noise level NP is given to the S / N calculator 33 together with the voice level VP, and the S / N signal SN is calculated for each frame. Voice detection signal VD, S /
The N signal SN and the calculated gain VG are given to the gain setting unit 35, and the actual set gain AG for the variable gain amplifier 10 is given.
Is set.

The set gain AG is given to the gain delay unit 38, and after being delayed by one frame, is output to the variable gain amplifier 10 as a delay set gain AGD. When changing from the non-voice section to the voice section, as shown by the arrow in FIG. 9, the delay time for one frame is used, and the delay setting gain AGD that gradually rises over the time of the one frame is used.
Is output.

On the other hand, on the input side of the variable gain amplifier 10,
The signal delay unit 40 gives the delayed input signal SID delayed by two frames. As a result, in the variable gain amplifier 10, the timings of the delay input signal SID and the delay setting gain AGD match, and the delay output signal SOD amplified by the setting gain AG corresponding to the input signal SI is obtained.

As described above, the AG of the fifth embodiment is
The C amplifier delays the setting gain AG and delays the setting gain A
It has a gain delay unit 38 that outputs GD and a signal delay unit 40 that delays the input signal SI and outputs the delayed input signal SID accordingly. Therefore, it becomes possible for the gain delay unit 38 to generate the delay setting gain AGD that gradually changes using the delay time.

As described above, the AGC amplifier of the fifth embodiment has the same advantages as those of the first embodiment, and in addition, changes in the noise level when the input signal SI changes from the non-voice section to the voice section. Has an advantage that the discomfort caused by the change in level during conversation can be further reduced.

(Sixth Embodiment) FIG. 10 is a block diagram of a gain controller showing a sixth embodiment of the present invention. The gain control section 30E is used in place of the gain control section 30D in FIG. 8, and elements common to those in FIGS. 4 and 8 are designated by common reference numerals.

The gain control section 30E has a voice level calculation section 31, a noise level calculation section 32, an S / N calculation section 33, a gain setting section 35 and a gain delay section 38 similar to those in FIG. It is configured to have a similar gain calculation unit 34A and gain storage unit 36.

The operation of the gain control section 30E is a combination of the operations of FIGS. 8 and 4, and the maximum change amount of the set gain AG when the input signal SI changes from the voice section to the non-voice section is limited. To be done. Further, when the input signal SI changes from the non-voice section to the voice section, the gain gradually changes and shifts to the target set gain AG. This allows
A with similar advantages to the first, second and fifth embodiments
A GC amplifier is obtained.

(Seventh Embodiment) FIG. 11 is a block diagram of a gain controller showing a seventh embodiment of the present invention. This gain control unit 30F is used in place of the gain control unit 30D in FIG. 8, and elements common to those in FIGS. 5 and 8 are designated by common reference numerals.

The gain control section 30F has a voice level calculation section 31, a noise level calculation section 32, an S / N calculation section 33, a gain calculation section 34 and a gain delay section 38 similar to those in FIG. It is configured to have a similar gain setting unit 35A and hangover setting unit 37.

The operation of the gain control unit 30F is a combination of the operations of FIGS. 8 and 5, and when the input signal SI changes from the voice section to the non-voice section, the gain is moderated by the hangover setting. It changes and shifts to the target set gain AG. Further, when the input signal SI changes from the non-voice section to the voice section, the gain gradually changes and shifts to the target set gain AG. As a result, an AGC amplifier having the same advantages as those of the first, third and fifth embodiments can be obtained.

(Eighth Embodiment) FIG. 12 is a block diagram of a gain controller showing an eighth embodiment of the present invention. This gain control unit 30G is used in place of the gain control unit 30D in FIG. 8, and elements common to those in FIGS. 7 and 8 are designated by common reference numerals.

The gain control section 30G has a voice level calculation section 31, a noise level calculation section 32, an S / N calculation section 33, and a gain delay section 38 similar to those in FIG. 34A, gain storage unit 36, gain setting unit 3
5A and a hangover setting unit 37.

The operation of the gain control section 30G is a combination of the operations of FIGS. 8 and 7, and the maximum change amount of the set gain AG when the input signal SI changes from the voice section to the non-voice section is limited. At the same time, the gain gradually changes due to the hangover setting, and the target gain AG is set. Further, when the input signal SI changes from the non-voice section to the voice section, the gain gradually changes and shifts to the target set gain AG. As a result, an AGC amplifier having the same advantages as those of the first, fourth and fifth embodiments can be obtained.

The present invention is not limited to the above embodiment, and various modifications can be made. Examples of this modification include the following. (A) Although the input signal SI has been described as a digitized audio signal, it can be similarly applied to an analog signal.

(B) The processing of the voice level calculation section 31, the noise level calculation section 32, etc. may be controlled by a program using a processing device such as a microprocessor or a digital signal processor.

(C) The calculated gain VG calculated by the gain calculation section 34 or the like is smoothed by using a low pass filter or the like,
You may make it give to the gain setting part 35 grade | etc.,. As a result, an averaged gain value can be obtained, and a voice signal with no discomfort can be obtained.

(D) In the gain setting section 35 and the like, the setting gain AG in the non-voice section is switched depending on the value of the S / N signal SN. However, the S / N signal SN and the calculation gain V
The set gain AG can be switched more finely according to the value of G. This makes it possible to further reduce the discomfort caused by noise in the non-voice section.

[0086]

As described in detail above, according to the first aspect of the present invention, the calculation gain calculated based on the level difference between the reference output level and the voice level and the S / N ratio of the input signal are used. , And a gain setting unit for controlling the variable gain amplifier by generating a set gain for the voice section and a set gain for the non-voice section of the input signal, respectively. This allows
When the S / N is high, the noise is not offensive to the ears, so by maintaining the same gain as in the voice section even in the non-voice section, the beginning portion of the voice is not cut, and voice intonation can be maintained without discomfort. It will be possible.

According to the second aspect of the invention, there is provided the gain calculating section for calculating the calculation gain in a constant cycle and keeping the variation amount of the calculation gain in each cycle within a predetermined fluctuation range. As a result, when the level fluctuation of the input signal is large, it is possible to reduce the followability and further maintain the intonation of the voice.

According to the third aspect of the invention, the hangover setting section is provided for detecting that the input signal has changed from the voice section to the non-voice section. As a result, in the gain setting unit, the change in the gain is relaxed when the hangover signal is given, and the set gain gently changed from the voice section to the non-voice section in a predetermined time is generated. Therefore, the background noise remains even when the voice signal disappears, and voice communication without discomfort becomes possible.

According to the fourth aspect of the present invention, there is provided the signal delay section for delaying the input signal by a predetermined time and the gain delay section for delaying the set gain so as to match the timing of the input signal. Accordingly, when the input signal changes from the non-voice section to the voice section, the delay time can be used to gently change the set gain in advance so that the gain value becomes a predetermined gain value at the start of the voice section. Therefore, the output signal gradually increases at the transition from the silence state of the non-voice section to the voice section,
In addition, voice communication without discomfort becomes possible.

According to the fifth aspect, the calculation gain is used as it is as the set gain in the voice section, and the calculation gain is reduced in the non-voice section according to the magnitude of the S / N to set the set gain. It constitutes the setting section. Therefore, S /
By reducing the set gain according to the value of N, noise in the non-voice section when the S / N is low can be suppressed, and voice communication can be performed without annoyance.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an AGC amplifier showing a first embodiment of the present invention.

FIG. 2 is a configuration diagram of a conventional AGC amplifier.

FIG. 3 is a signal waveform diagram showing the operation of FIG.

FIG. 4 is a configuration diagram of a gain control section showing a second embodiment of the present invention.

FIG. 5 is a configuration diagram of a gain control section showing a third embodiment of the present invention.

FIG. 6 is a signal waveform diagram showing the operation of FIG.

FIG. 7 is a configuration diagram of a gain control section showing a fourth embodiment of the present invention.

FIG. 8 is a configuration diagram of an AGC amplifier showing a fifth embodiment of the present invention.

9 is a signal waveform diagram showing the operation of FIG.

FIG. 10 is a configuration diagram of a gain control section showing a sixth embodiment of the present invention.

FIG. 11 is a configuration diagram of a gain control section showing a seventh embodiment of the present invention.

FIG. 12 is a configuration diagram of a gain control section showing an eighth embodiment of the present invention.

[Explanation of symbols]

10 Variable gain amplifier 20 Voice detector 30 gain control section 31 Voice level calculator 32 Noise level calculator 33 S / N calculator 34 Gain calculator 35 Gain Setting Section 36 Gain memory 37 Hangover setting section 38 Gain Delay Unit 40 signal delay unit

Claims (5)

[Claims] 1. A voice detection unit that detects a voice section in which voice exists in an input signal, a level calculation unit that calculates a voice level in a voice section of the input signal and a noise level in a non-voice section, and the voice level. And a signal-to-noise ratio calculation unit that calculates a signal-to-noise ratio of the input signal based on the noise level, and a gain calculation unit that calculates a calculation gain based on the level difference between the reference output level of the output signal and the voice level. A gain setting unit that generates a set gain for a voice section and a set gain for a non-voice section of the input signal based on the calculated gain and the signal-to-noise ratio, and amplifies the input signal by the set gain. And a variable gain amplifier that outputs the output signal. 2. The gain calculating section calculates the calculation gain in a constant cycle, and limits the value of the calculation gain so that a change amount of the calculation gain in each cycle falls within a predetermined variation range. The automatic gain control amplifier according to claim 1, wherein: 3. A hangover setting unit that outputs a hangover signal upon detecting that the input signal has changed from a voice section to a non-voice section, wherein the gain setting section is provided with the hangover signal. The automatic gain according to claim 1 or 2, characterized in that the change in gain is sometimes relaxed to generate the set gain that is gradually changed from a voice section to a non-voice section in a predetermined time. Control amplifier. 4. A signal delay unit that delays the input signal by a predetermined time and applies the variable gain amplifier to the signal delay unit, and a gain setting unit that generates the signal to match the timing of the input signal delayed by the signal delay unit. When the input signal changes from the non-voice section to the voice section while delaying the set gain, the delay time is utilized to set the set gain in advance so that the gain value becomes a predetermined gain value at the start of the voice section. 4. An automatic gain control amplifier according to claim 1, 2 or 3, further comprising: a gain delay unit that is gradually changed and is applied to the variable gain amplifier. 5. The gain setting unit uses the calculated gain as the set gain as it is when the input signal is in a voice section, and uses the calculated gain as the signal-to-noise when the input signal is in a non-voice section. 5. The automatic gain control amplifier according to claim 1, wherein the set gain is set by reducing the ratio according to the magnitude of the ratio.