JPH11187479A - Audio signal processing unit - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain an arbitrary attack and recovery characteristics with gain control of an audio signal. SOLUTION: An output of a detector DET and a control output value SET are compared in the steps [11]-[16]. When [DET>SET], the output of a recovery counter INC is cleared [INC=0], and the output value SET is incremented to [SET+2<n> ]. Furthermore, [DET=SET] is discriminated in the step [17]. The recovery counter is driven in steps [18]-[23], an output value INC is incremented by 1, and the incremented output value is compared with a value DEC of a recovery coefficient, and the processing in the steps [18]-[23] is repeated until the result [INC=DEC] is obtained. As an output value SET of a control counter, a preceding value is held and outputted during this time interval. Then when [INC=DEC], the output INC is cleared [INC=0] and the output value SET is decremented to [SET-1].

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an audio signal processing device suitable for use in, for example, a sound collection device, a noise reduction device, a sound quality improvement device, and the like.

[0002]

2. Description of the Related Art For example, in a sound pickup device incorporating a microphone or connecting a microphone to input a sound signal, an audio signal processing device for controlling the level of the sound signal such as an AGC circuit or a limiter circuit is used. I have. Also,
For example, an audio signal processing device that controls the level of an audio signal is also used as a noise reduction device for reducing wind noise and mechanical noise in a VCR integrated video camera device.

Further, an audio signal processing device for controlling the level of an audio signal is also used in an apparatus for improving sound quality (such as expansion and compression of a dynamic range) when an audio signal is reproduced by a speaker or the like. In various other devices that handle audio signals, an audio signal processing device for controlling the level of the audio signal is frequently used.

[0004]

FIG. 6 shows a configuration of an apparatus for reducing, for example, wind noise, as an example of the above-described audio signal processing apparatus. In FIG. 6, microphones 61L and 61R are two microphones provided close to each other, for example, a stereo microphone of a VTR integrated video camera.

The left (L) and right (R) sound signals picked up by the microphones 61L and 61R are converted into an analog-to-digital (ADC) circuit 64 through amplifiers 62L and 62R and an automatic gain control (AGC) circuit 63, respectively. Supplied to The left and right audio signals digitized by the AD conversion circuit 64 are respectively delayed (DL) circuits 65L and 65L.
The signal is supplied to an arithmetic circuit (Σ) 66L, 66R described later through R.

The left and right audio signals from the AD conversion circuit 64 are supplied to an arithmetic circuit (Σ) 67, for example, where the right (R) audio signal is subtracted from the left (L) audio signal. The subtracted output from the arithmetic circuit 67 is, for example, a first low-pass filter (LPF) 68 of digital signal processing having a pass band of several kHz or less of the digital audio signal.
Supplied to

That is, in the low-pass filter 68,
The low-frequency component of the difference between the audio signals collected by the microphones 61L and 61R is extracted. Here, the low-frequency component of this difference corresponds to the wind sound signal of the collected sound signal. Then, the low-frequency component of this difference is supplied to arithmetic circuits 66L and 66R through a gain control amplifier 69 described later,
It is subtracted from the original left and right audio signals from the delay circuits 65L and 65R.

Further, a second low-pass filter (LPF) 70 for digital signal processing in which the subtraction output of the arithmetic circuit 67 has a pass band of, for example, several tens Hz or less of the digital audio signal.
Supplied to The output signal of the low-pass filter 70 is supplied to a detection circuit (DET) 72 through an amplifier 71, and the signal is converted into an absolute value. The absolute value signal is supplied to a gain control amplifier 69 through a circuit 73 for providing an attack characteristic and a recovery characteristic described later.

As a result, the above-described arithmetic circuits 66L, 66
In R, the low-frequency component of the difference corresponding to the wind signal of the signal collected by the microphones 61L and 61R is subtracted, and the level of the subtracted wind signal is further reduced to the lower-frequency component of the wind signal. Is controlled according to the absolute value. Only when a wind sound signal is generated by this control, the wind sound signal is removed according to the level.

The left and right audio signals from the arithmetic circuits 66L and 66R are respectively converted into delay (DL) circuits 74.
L and 74R are supplied to the arithmetic circuits (Σ) 75L and 75R. The left and right audio signals from the arithmetic circuits 66L and 66R are supplied to the arithmetic circuit (Σ) 76 to add the left and right audio signals, and the added output has a third pass band having a predetermined pass band corresponding to the wind noise signal. Low-pass filter (LPF)
77.

Thus, in the low-pass filter 77,
A signal corresponding to the in-phase component of the wind noise signal is extracted. A signal corresponding to the in-phase component is supplied to the gain control amplifier 78. The gain control amplifier 78 is controlled by a signal from the circuit 73. And this gain control amplifier 7
8 are supplied to the arithmetic circuits 75L and 75R,
It is subtracted from the original left and right audio signals through the delay circuits 74L and 74R. Further, output signals of the arithmetic circuits 75L and 75R are taken out to terminals 79L and 79R.

That is, these operation circuits 75L, 75R
Then, the in-phase component of the wind noise signal in the signals collected by the microphones 61L and 61R is subtracted. Then, at the time of this subtraction, the level to be subtracted is controlled according to the absolute value of the wind sound signal, so that the wind sound signal is removed according to the level only when the wind sound signal is generated, The wind noise signal can be further reduced without attenuating the original audio signal to be collected.

By the way, in this device, the above-mentioned circuit 7
In No. 3, for example, predetermined attack characteristics and recovery characteristics as shown in FIG. 5 are provided. That is, for example, at the beginning of a wind noise signal, an attack characteristic with a quick response is provided to improve the response to a gust or the like. Also, when the level of the wind noise signal is reduced, if the response of the control is made faster, the change point can be easily recognized by the audibility, so that a recovery characteristic for slowing the response is provided for the purpose of solving the problem.

However, in order to provide such an attack characteristic and a recovery characteristic, the conventional means requires a complicated and large-scale circuit configuration. In the past, such attack and recovery characteristics were empirically required, but if they were to be changed, for example, conventional means would require a significant design change, and such changes could not be made. It was extremely difficult. Furthermore, it has been difficult to perform so-called adaptive processing according to an input signal or the like.

This application has been made in view of the above points, and the problem to be solved is that the conventional means requires a complicated and large-scale circuit configuration.
It is difficult to change the attack characteristics and the recovery characteristics, and it is also difficult to perform so-called adaptive processing.

[0016]

For this reason, in the present invention, a control signal is formed by using a counter means, and a value serving as a unit of counting on the up side and the down side of the counter means can be arbitrarily set. According to this, arbitrary attack characteristics and recovery characteristics can be obtained with a simple configuration, and these characteristics can be easily changed arbitrarily, and so-called adaptive processing can be performed. it can.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS That is, the present invention detects the level of an audio signal, compares the detected level to form a control signal, and uses the formed control signal to perform feedforward or feedback. In an audio signal processing device for controlling the level of an audio signal, a control signal is formed using a counter means, and an attack characteristic of the level control is changed by changing a value used as a unit of counting on an up side and a down side of the counter means. And recovery characteristics.

[0018]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below with reference to the drawings.
FIG. 1 is a flowchart showing an example of processing when the audio signal processing device of the present invention is formed by a feedforward circuit as shown in FIG. 2A, for example. The level control by such a feedforward circuit is applied, for example, to wind noise reduction processing.

That is, in FIG. 2A, the signal input supplied to the terminal 1 is taken out to the terminal 3 as a signal output through the operation means 2. Further, a signal input from the terminal 1 is supplied to the detection means 4, and this detection output is supplied to a comparison / control means 5 described later. The control output of the comparison / control means 5 is supplied to the operation means 2 through the LOG conversion means 6, and the control output is fed back to the comparison / control means 5.

In this circuit, for example, when the object is an audio signal, the detection means 4 is represented by a signal of +/- sign. A peak value and an average value are obtained, and a filtering process or the like for extracting a target frequency band is performed. For example, in the case of using the above-described wind noise reduction processing, a processing of converting a difference component between the collected left and right audio signals to an absolute value after performing a low-pass filter processing is performed.

In this circuit, the comparison / control means 5 performs the following processing shown in FIG.

That is, in FIG. 1, step [11]
Then, for example, the value DET (0000 _{hex to} 7fff _hex ) of the 16-bit detection output from the detection means 4 is compared with the value of the control output SET (0000 _{hex to} 7fff _hex ) of the control counter, also described later, of 16 bits. Then, [DET> SET] is determined. And [DET
> SET] (y), the output value INC of the recovery counter described later is cleared in step [12] [INC =
0].

[0023] The maximum value of the value SET control signal in step [13], is compared with the example (7fff _hex -2 ^{n) (n} is an arbitrary integer), wherein when the value SET is not the maximum value, i.e. [SET In the case of <(7ff _hex −2 ⁿ )], (y) enters the attack mode. In this attack mode, the output value SET of the control counter is incremented to [SET + 2 ⁿ ] in step [14]. Further, this value SET is output as a control output, and the operation is terminated.

This operation is performed, for example, at a sampling period of an audio signal (for example, 1/48 kHz, 1 / 44.1 kHz).
z, 1/32 kHz). The operation of incrementing the output value SET of the control counter is continued until a condition [DET = SET] or [SET = 7fff _hex ] described later is satisfied.

That is, in step [13], [SET ≧
(7fff-2 ⁿ ) _hex ], (n) is set to [SET = 7ffff _hex ] in step [15]. Then, in step [16], this output value [SET = 7fff
_hex ] is held at the value SET, this value SET is output as a control output, and the operation is terminated.

In step [11], [DET> SE
If [T] is not (n), then [DET = SET] is determined in step [17]. And [DET = S
ET], (y) is the value of the previous value S in step [16].
ET is held as it is, this value SET is output as a control output, and the operation is terminated.

Next, [DET <SET] is determined. That is, when [DET = SET] is not satisfied in step [17], [DET <SET] is determined in (n). When [DET <SET] is determined, step [18]
Drives the above-described recovery counter, for example, the 16-bit output value INC is incremented by +1 [INC = I
NC + 1]. In step [19], the value I
The NC is compared with a 16-bit recovery coefficient value DEC.

In step [19], [INC <DE
C], (y) is the previous value SE in step [16].
T is held as it is, this value SET is output as a control output, and the operation is terminated. Thus, the operation in which the output value INC of the above-mentioned recovery counter is incremented by +1 is performed by [INC = DEC] in step [19].
The output value SET of the control counter during this period is output while the previous value is held as it is.

Then, in step [19], [INC = DE
C], that is, when [INC <DEC] is not satisfied (n), the output value INC of the recovery counter is cleared [INC = 0] in step [20]. Further, in step [21], the value SET is the minimum value of the control signal,
For example, it is compared with (0000 _hex ). And the value SET
Is not the minimum value, that is, [SET> 000
0 _hex ], the recovery mode is set (y).

In this recovery mode, step [2]
In [2], the output value SET of the control counter is decremented to [SET-1]. Further, this value SET is output as a control output, and the operation is terminated. The operation of decrementing the output value SET of the control counter is repeated until the value SET becomes the minimum value of the control signal, that is, [SET ≦ 0000 _hex ] in step [21].

Further, in step [21], [SET ≦ 00
[00 _hex ], (n) is set to [SET = 0000 _hex ] in step [23], and this output value [SET = 0000 _hex ] is held at the value SET in step [16]. Then, this value SET is output as the control output, and the operation ends.

Thus, the value SET of the control output extracted from the control counter always follows the value DET of the detection output from the detection means 4. The time constant is a time determined by n during an attack, for example, a sampling period of 1
/ 32 kHz, n = 1, 500 msec. At the time of recovery, for example, when the sampling period is 1/32 kHz, DEC = 10
_{In dec} , 10 sec. Converges.

That is, in this apparatus, for example, in the case of the above-described wind noise reduction processing, the gain control by the operating means 2 is performed by increasing the amount of reduction in the + direction of the value SET and by decreasing the value of-
The amount of reduction is small in the direction. As a result, when wind noise is generated, an instantaneous operation is performed by an attack, and in a subsequent recovery, a gradual recovery operation is performed, so that natural control can be performed in terms of audibility.

Further, in this device, by changing the value of n and the value DEC, a time constant determined by the value of n during the attack and a time constant determined by the value DEC during the recovery are arbitrarily determined. be able to. That is, it is possible to arbitrarily and easily change the time constant during the attack and the recovery.

Therefore, in this apparatus, the control signal is formed by using the counter means, and the unit value of the counting on the up side and the down side of the counter means can be arbitrarily set. Arbitrary attack characteristics and recovery characteristics can be obtained, and these characteristics can be arbitrarily changed.

According to the present invention, the conventional means requires a complicated and large-scale circuit configuration, and it is difficult to change the attack characteristics and the recovery characteristics. It can be resolved in the following.

FIG. 3 is a flowchart showing an example of processing when the audio signal processing apparatus of the present invention is formed by a feedback circuit as shown in FIG. 2B, for example. The level control by such a feedback circuit is applied to, for example, processing of an AGC circuit and a limiter circuit.

That is, in FIG. 2B, the signal input supplied to the terminal 1 is taken out to the terminal 3 as a signal output through the operating means 2 as a gain control circuit. A signal output from the operation means 2 is supplied to a detection means 4, and the detection output is supplied to a comparison / control means 5, which will be described later, and is compared with a target (threshold) value from an arbitrary terminal 7 to output a control output. It is formed. And this comparison
The control output of the control means 5 is supplied to the operation means 2 through the LOG conversion means 6.

In this circuit, for example, when the object is an audio signal, the signal is represented by a signal of +/- sign. A peak value and an average value are obtained, and a filtering process or the like for extracting a target frequency band is performed. For example, in the case of using the above-described AGC circuit processing, a peak detection process is performed after converting the larger one of the collected left and right audio signals into an absolute value.

In this circuit, the comparison / control means 5 performs the processing shown in FIG. 3 described below.

That is, in FIG. 3, step [31]
Then, for example, a detection output value DET (0000 _{hex to} 7fff _hex ) of, for example, 16 bits from the detection means 4 and a target value of 16 bits (TARGET value = 0000)
_{hex to} 7fff _hex ) and [DET> TA
RGET] is determined. And [DET> TARGE
T] (y), the output value INC of the recovery counter described later is cleared [INC = 0] in step [32].

In step [33], for example, a control output value SET (0
000 _{hex to} 7fff _hex ) is compared with the minimum value of the control signal, for example, (2 ⁿ ) (n is an arbitrary integer), where the value S
When ET is not the minimum value, that is, when [SET> 2 ⁿ ] (y), the attack mode is set. In this attack mode, the output value SET of the above-described control counter is decremented to [SET-2 ⁿ ] in step [34]. Further, this value SET is output as a control output, and the operation is terminated.

This operation is performed, for example, at the sampling period of an audio signal (for example, 1/48 kHz, 1 / 44.1 kHz).
z, 1/32 kHz). The operation of incrementing the output value SET of the control counter is described later [DET = TARGET]
Until the condition is satisfied.

That is, in step [33], [SET ≦ 2
when it becomes ^n] (n) at step [35] [SE
T = 0]. Then, in step [36], this output value [SET = 0] is held at the value SET, and this value S
ET is output as the control output, and the operation ends.

In step [31], [DET> TAR
GET] (n), a further step [3]
7], [DET = TARGET] is determined. When [DET = TARGET] (y), the previous value SET is held as it is in step [36], this value SET is output as a control output, and the operation is terminated.

Next, [DET <TARGET] is determined. That is, in step [37], [DET = TARGE
T], it is determined that [DET <TARGET] at (n). When [DET <TARGET] is determined, the above-described recovery counter is driven in step [38], and for example, the 16-bit output value INC is incremented by +1 [INC = INC + 1]. Further, in step [39], the value INC is compared with the value DEC of the 16-bit recovery coefficient.

Here, in step [39], [INC <DE
C] (y), the previous value SE in step [36]
T is held as it is, this value SET is output as a control output, and the operation is terminated. Thus, the operation in which the output value INC of the recovery counter is incremented by +1 is performed in step [39] at [INC = DEC].
The output value SET of the control counter during this period is output while the previous value is held as it is.

Then, in step [39], [INC = DE
C], that is, when [INC <DEC] is not satisfied (n), the output value INC of the recovery counter is cleared [INC = 0] in step [40]. Further, in step [41], the value SET is the maximum value of the control signal,
For example, it is compared with (7ffff _hex ). And the value SET
Is not the maximum value, that is, [SET <7ff
f _hex ], (y) enters the recovery mode.

In this recovery mode, step [4]
In [2], the output value SET of the control counter is incremented to [SET + 1]. Further, this value SET is output as a control output, and the operation is terminated. The operation in which the output value SET of this control counter is incremented is
This is repeated until the value SET becomes the maximum value of the control signal, that is, [SET ≧ 7fff _hex ] in step [41].

Further, in step [41], [SET ≧ 7f
(ff _hex )], (n) is set to [SET = 7fff _hex ] in step [43], and the output value [SET = 7fff _hex ] is held at the value SET in step [36]. Then, this value SET is output as the control output, and the operation ends.

Thus, the control output value SET taken out of the control counter is controlled such that the detection output value DET from the detection means 4 always follows the target value (TARGET). The time constant is a time determined by n during an attack, for example, a sampling period of 1/32 kHz.
z, n = 1 and 500 msec. At the time of recovery, for example, when the sampling period is 1/32 kHz, DEC = 10 _dec and 1
0 sec. Converges.

That is, in this apparatus, for example, AGC
When applied to the circuit, the gain control by the operation means 2 is performed in the minus direction of the value SET, the small gain, and the plus value of the value SET.
If the gain is large in the direction, the signal output obtained at the terminal 3 in FIG. 2B operates instantaneously with an attack when the value DET exceeds the value TARGET, and in the subsequent recovery,
Perform the recovery operation gradually.

Further, in this apparatus, by changing the value of n and the value DEC, a time constant determined by the value of n during an attack and a time constant determined by the value DEC during a recovery are arbitrarily determined. be able to. That is, it is possible to arbitrarily and easily change the time constant during the attack and the recovery.

Therefore, also in this apparatus, the control signal is formed by using the counter means, and the value serving as the unit of counting on the up side and the down side of the counter means can be arbitrarily set, thereby achieving a simple configuration. Can obtain an arbitrary attack characteristic and a recovery characteristic, and can arbitrarily change these characteristics.

As a result, the conventional means requires a complicated and large-scale circuit configuration, and it is difficult to change the attack characteristics and the recovery characteristics. It can be resolved in the following.

Further, in this device, the value n (n is an integer) on the attack side and the value DEC on the recovery side are coefficients for determining respective time constants, and can be optimized depending on the purpose of processing at the time of circuit design. is there. Generally, these time constants are settled in the above-described time. However, depending on the processing, there are cases where it is better to make them adaptively variable. In particular, in the recovery operation, it is sometimes necessary to make the recovery time variable according to certain information instead of always having a recovery time of several seconds.

FIG. 4 shows a configuration in which the recovery time is adaptively changed according to the attack operation time. That is, in FIG. 4, for example, the detection value DET from the detection means 4 supplied to the input terminal 11 is taken out to the output terminal 13 of the control output SET through the comparison / control means 12, and the detection value DET from the input terminal 11 is attacked. It is supplied to the time measuring means 14.

The attack time measuring means 14 counts the number of audio signal samples after entering the attack mode, for example. The count value is supplied to the recovery coefficient control means 15, and the value DEC formed by the recovery coefficient control means 15 is controlled according to the count value. Further, the value DEC formed by the recovery coefficient control means 15 is supplied to the comparison / control means 12.

Thus, for example, by increasing the value DEC as the number of attack counts increases, the recovery time is shortened for the impulse fluctuation and small amplitude fluctuation of the input detection value DET, and the recovery time is reduced for long time fluctuation and large amplitude fluctuation. On the other hand, the recovery time can be lengthened, and further, control without a sense of incongruity can be realized.

In the above-described apparatus, since the comparison / control means 5 and 12 obtain the control output value SET by addition and subtraction by the counter means, the change is a linear change. The level change of the audio signal to be performed is also a linear change. However, such a linear change in level cannot be heard as a uniform change in human hearing. This is because the hearing of the human ear is proportional to the logarithm of the level change.

Therefore, in the above-described apparatus, the control output value SET is set to L
The OG conversion means 6 performs conversion using the following equation. CO
NT = K · log (SET) Here, K is a constant, and therefore, by controlling the operation means 2 using the control output CONT, a level change proportional to the logarithm is given, and a smoothness suitable for human hearing is given. A variable operation can be performed.

In the above-mentioned apparatus, the operating means 2 controls the level of an incoming signal based on the supplied control output value CONT. Therefore, this operation means 2
Is configured by, for example, a multiplier, and performs level control by multiplying an input signal value by a control output value CONT.

Thus, according to the above-described audio signal processing apparatus, the level of the audio signal is detected, the detected level is compared, and a control signal is formed. In a device for controlling the level of an audio signal by feedback, a control signal is formed by using a counter means, and an attack characteristic and an attack characteristic of a level control by changing a value as a unit of counting on an up side and a down side of the counter means. By providing the recovery characteristic, an arbitrary attack characteristic and a recovery characteristic can be obtained with a simple configuration, and these characteristics can be arbitrarily changed.

[0064]

Therefore, according to the present invention, the control signal is formed by using the counter means, and the value as the unit of counting on the up side and the down side of the counter means can be arbitrarily set. Any attack characteristics and recovery characteristics can be obtained with a simple configuration, and these characteristics can be arbitrarily changed, and so-called adaptive processing can be performed.

Thus, the conventional means requires a complicated and large-scale circuit configuration, it is difficult to change the attack characteristics and the recovery characteristics, and it is also difficult to perform the so-called adaptive processing. According to the present invention, these problems can be easily solved.

Further, according to the present invention, a LOG conversion means is provided to convert a change in the control output into a logarithmic change, thereby enabling a level change corresponding to the sense of the human ear. is there.

[Brief description of the drawings]

FIG. 1 is a flowchart for explaining the operation of an example of an audio signal processing device to which the present invention is applied;

FIG. 2 is a configuration diagram of an example of a feedforward circuit and a feedback circuit that constitute an audio signal processing device to which the present invention is applied;

FIG. 3 is a flowchart for explaining the operation of another example of the audio signal processing device to which the present invention is applied;

FIG. 4 is a configuration diagram of an example of a circuit for adapting an audio signal processing device to which the present invention is applied;

FIG. 5 is a diagram for explaining an attack and a recovery time.

FIG. 6 is a configuration diagram of an example of an apparatus for reducing wind noise to which the present invention is applied.

[Explanation of symbols]

1 ... signal input terminal, 2 ... operation means, 3 ... signal output terminal,
4 detection means, 5 comparison / control means, 6 LOG conversion means

Claims (4)

[Claims] 1. A level of an audio signal is detected, a control signal is formed by comparing the detected level, and the level of the audio signal is controlled by feedforward or feedback using the formed control signal. In the audio signal processing apparatus, the control signal is formed by using a counter means, and the attack and recovery characteristics of the level control are changed by changing the unit values of the counting on the up side and the down side of the counter means. An audio signal processing device comprising: 2. The audio signal processing device according to claim 1, wherein a value serving as a unit of counting on an up side and a down side of said counter means is arbitrarily determined. 3. The audio signal processing apparatus according to claim 1, wherein the values of the counting unit on the up side and the down side of the counter means are respectively changed adaptively. 4. The audio signal processing apparatus according to claim 1, wherein the control signal is formed by supplying the counted output value of said counter means to log conversion means. .