JP3039051B2 - Adaptive noise suppression device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an adaptive noise suppressor incorporated in a video camera or the like.

[0002]

2. Description of the Related Art In recent years, adaptive noise suppressors have been frequently used for the purpose of collecting voice in a noise environment and extracting a weak voice signal buried in noise.

A conventional adaptive noise suppression apparatus will be described below.
The input adaptive noise suppression device will be described with reference to the drawings.

As shown in FIG. 5, a conventional adaptive noise suppressor applies a main microphone 51, a reference microphone 52, a digital filter 53, and a subtractor 5 to a noise source 50.
4, a filter control unit 55.

[0005] With respect to the adaptive noise suppressor composed of the above components, the relationship and operation of each component will be described below. First, as shown in FIG. 5, both a main microphone 51 for sound pickup and a reference microphone 52 for noise pickup use omnidirectional microphones, and the reference microphone 52 has a noise level so as not to be affected by a desired sound signal. It is located near the source 50. The output signal from the reference microphone 52 is
The output signal of the digital filter 53 is input to the digital filter 53, and the output signal of the digital filter 53 is subtracted by the subtractor 54 from the output signal of the main microphone 51 that collects noise-superimposed voice.
The filter control unit 55 estimates the impulse response of the acoustic transmission system from the noise source 50 to the main microphone 51 based on the algorithm of the learning identification method, and corrects the filter coefficient of the digital filter 53.

However, if a desired audio signal is mixed into the reference microphone 52, an error in the estimation of the transmission system occurs, so that a normal speech section is detected and the estimation operation of the transmission system is stopped while the desired audio is being input. Is added.

Next, a conventional example of a two-input adaptive noise suppressor for detecting a voice section will be described with reference to FIG.

[0008] As shown in FIG. 6, the adaptive noise suppressor applies a main microphone 61, a reference microphone 62, a digital filter 63, a subtractor 64, a filter control unit 65, and a voice section detection unit 66 to a noise source 60. It consists of.

With respect to the adaptive noise suppressor composed of the above components, the relationship and operation of each component will be described below. First, as shown in FIG. 6, the voice section detection unit 66 measures the power of the output signal of the main microphone 61, compares the power with an appropriately set threshold value, and the power exceeds the threshold value. In this case, it is determined to be a voice section. The filter control unit 65 receives the output signal of the subtractor 64, the output signal of the reference microphone 62, and the determination result of the voice section detection unit 66 as inputs, and updates the filter coefficient of the digital filter 63 when it is determined that the section is a noise section, If it is determined to be a voice section, updating of the filter coefficient is stopped.

[0010] In the adaptive noise suppressor configured as described above, the output of the digital filter is adjusted to be the optimum predicted value for the noise component included in the output signal of the main microphone, so that it is superimposed on the desired sound. Noise can be suppressed.

In the learning identification method, the filter coefficient of the digital filter for realizing the estimated impulse response of the transmission system is updated by (Equation 1).

[0012]

(Equation 1)

In equation (1), N is the number of taps of the digital filter 63, hj (i) is the filter coefficient of the j-th tap at time i, s (i) is the output signal of the subtractor, and x (i) is The output signal from the reference microphone, α, is the error correction gain. The error correction gain α needs to satisfy 0 <α ≦ 1 in order to perform a stable adaptive operation.
Α in the conventional adaptive noise suppression device is a fixed constant.

[0014]

However, in the above conventional configuration, since the error correction gain α is fixed, the SN ratio (the ratio of the power of the noise to the power of the desired voice signal)
Is smaller, the convergence time of the adaptive noise suppressor is S
It is longer than when the N ratio is large. However, increasing α decreases the convergence time but reduces the noise suppression effect,
Conversely, when α is reduced, the noise suppression effect increases, but the convergence time increases.

In particular, when only noise generated from a mechanical system housed in a video camera housing is to be suppressed, generally all sounds coming from the surroundings need to be regarded as desired audio signals. The level fluctuation of the audio signal is severe. Furthermore, the magnitude of the noise differs depending on the operation state of various video cameras, such as a state in which the video cassette tape is only running, a state in which auto focus is operating, and a state in which the zooming function of the optical system is operating. . As described above, when the SN ratio fluctuates greatly, there is a problem that the adaptive noise suppression device using the fixed α cannot obtain a stable noise suppression effect.

The present invention solves the above-mentioned conventional problems. An adaptive noise suppression apparatus used in a video camera or the like which can shorten the convergence time of an adaptive operation and obtain an optimum noise suppression effect. The purpose is to provide.

[0017]

In order to achieve the above object, an adaptive noise suppressor according to the present invention varies an error correction gain α according to an S / N ratio. The noise suppression amount is obtained from the output of the reference microphone and the output of the subtractor, and the error correction gain α is made variable in accordance with the noise suppression amount.

[0018]

According to the present invention, the error correction gain α is reduced when the SN ratio is large, and α is increased when the SN ratio is small, so that the convergence time is kept constant even when the SN ratio changes. can do.

When the noise suppression amount is large, α is reduced, and when the noise suppression amount is small, α is increased, so that a stable noise suppression effect can be obtained even when the SN ratio varies.

[0020]

【Example】

(Embodiment 1) An embodiment of the present invention will be described below with reference to the drawings.

As shown in FIG. 1, the adaptive noise suppressor of the first embodiment has a noise source 1 generated in a video camera 10.
1, the main microphone 12 and the reference microphone 1
3, a digital filter 14, a subtracter 15, a filter controller 16, an SN ratio calculator 17, a corrected gain calculator 18, and the like.

With respect to the adaptive noise suppressor composed of the above components, the relationship and operation of each component will be described with reference to FIG.

As shown in FIG. 1, the adaptive noise suppressor comprises:
Reference microphone 1 for noise pickup shown in FIG.
In this embodiment, a close-talking microphone is used, and the microphone 3 is disposed near the noise source 11 in the housing of the video camera 10. Since the reference microphone 13 uses a close-talking type, it has low sensitivity to a desired voice signal from a distance,
Only noise generated from the noise source 11 can be collected. The output signal from the reference microphone 13 is input to the digital filter 14, and the output signal of the digital filter 14 is subtracted by the subtracter 15 from the output signal from the main microphone 12 that picks up the voice on which noise is superimposed. SN
The ratio calculator 17 calculates an SN ratio R from the ratio of the power of the output signal from the main microphone 12 to the power of the output signal from the reference microphone 13. Modified gain calculator 18
Calculates an error correction gain α according to the SN ratio in accordance with R and an appropriately set rule. FIG. 2 shows an example of a method for calculating α. In FIG. 2, R _max and R _min are the maximum value and the minimum value of the SN ratio that can be assumed in the operation state of the video camera 10, respectively. Α _max is the maximum value of α set for R _min . The corrected gain calculator 18 receives R as the S / N ratio and calculates α according to the relationship shown in FIG. The filter control unit 16 substitutes the calculated α into (Equation 1) and updates the filter coefficient of the digital filter 14.

As described above, the adaptive noise suppressor according to the first embodiment of the present invention reduces α when the SN ratio is large, and increases α when the SN ratio is small. , The change in the convergence time due to the fluctuation of However, as described above, when the value of α is determined only by the SN ratio, there is a problem that the noise suppression effect is reduced when the SN ratio is small compared to when the SN ratio is large.

(Embodiment 2) A second embodiment of the present invention will be described below with reference to the drawings.

As shown in FIG. 3, the adaptive noise suppressor applies a main microphone 32, a reference microphone 33, a digital filter 34, a subtractor 35, and a filter control unit 36 to a noise source 31 generated in a video camera 30. , A noise suppression amount calculation unit 37, a corrected gain calculation unit 38, and the like.

With respect to the adaptive noise suppression device comprising the above components, the relationship between the components and the operation will be described with reference to FIG.

As shown in FIG. 3, in the adaptive noise suppressing apparatus according to the second embodiment of the present invention, a close-talking microphone is used as a reference microphone 33 for noise pickup, and a noise source in the housing of the video camera 30 is used. 31 is provided in the vicinity. An output signal from the reference microphone 33 is input to a digital filter 34, and the digital filter 34
Is subtracted by the subtractor 35 from the output signal from the main microphone 32 that collects the voice on which the noise is superimposed. The noise suppression amount calculation unit 37 calculates the noise suppression amount E defined by (Equation 2).

[0029]

(Equation 2)

In the definition equation (Equation 2) of E, P _R is the power of the output signal from the reference microphone 33, and P _S
Is the power of the difference signal between the output signal from the main microphone 32 and the output signal from the subtractor 35. The value of the noise suppression amount E decreases as the residual noise included in the output of the subtractor 35 increases, so that the degree of noise suppression can be known from the magnitude of E. The correction gain calculator 38 calculates an error correction gain α according to the noise suppression amount according to E and an appropriately set rule. FIG. 4 shows an example of a method for calculating α. FIG.
Residual noise E _MAX is included in the output of the subtractor at 0
Is the value of E when. The corrected gain calculator 38 receives the noise suppression amount E as input and calculates α according to the relationship shown in FIG. The filter control unit 36 substitutes the calculated α into (Equation 1) and updates the filter coefficient of the digital filter 34.

As described above, the adaptive noise suppressor according to the second embodiment of the present invention reduces α when the amount of noise suppression is large, and increases α when the amount of noise suppression is small. A stable noise suppression effect can be obtained regardless of the S / N ratio.

[0032]

As is clear from the above embodiments, the present invention prevents the increase in the convergence time of the adaptive operation due to the decrease in the SN ratio by making the error correction gain α variable according to the SN ratio. be able to.

Further, the error correction gain α according to the noise suppression amount
Is variable, it is possible to realize an excellent adaptive noise suppression device, such as being able to obtain a stable noise suppression effect without being affected by fluctuations in the SN ratio.

[Brief description of the drawings]

FIG. 1 is a block diagram of an adaptive noise suppression apparatus according to a first embodiment of the present invention.

FIG. 2 is a characteristic diagram of an SN ratio and an error correction gain in the embodiment.

FIG. 3 is a block diagram of an adaptive noise suppressor according to a second embodiment of the present invention;

FIG. 4 is a characteristic diagram of a noise suppression amount and an error correction gain in the embodiment.

FIG. 5 is a block diagram of a conventional adaptive noise suppression device.

FIG. 6 is a block diagram of another example of the conventional adaptive noise suppression apparatus.

[Explanation of symbols]

 10, 30 Video camera 11, 31 Noise source 12, 32 Main microphone 13, 33 Reference microphone (close-talking type) 14, 34 Digital filter 15, 35 Subtractor 16, 36 Filter control unit 17 SN ratio calculation unit 37 Noise suppression amount Calculation unit 18, 38 Corrected gain calculation unit

──────────────────────────────────────────────────続 き Continuation of front page (56) References JP-A-3-295400 (JP, A) JP-A-53-113530 (JP, A) JP-A-3-230370 (JP, A) JP-A 57-113 141112 (JP, A) JP-A-62-135019 (JP, A) JP-A-64-48750 (JP, U) (58) Fields investigated (Int. Cl. ⁷ , DB name) G10K 11/178 H04N 5 / 225 H04R 1/02 107 G11B 33/08 G10L 3/02 301 G10L 9/00

Claims (2)

(57) [Claims] A main microphone for sound pickup disposed on a surface of a video camera housing; a close-talking microphone disposed near a noise source generated in the video camera housing; A digital filter for inputting an output signal from the microphone, a subtracter for subtracting an output signal of the digital filter from an output signal from the main microphone, and a signal from the main microphone for the power of the output signal from the close-talking microphone. An SN ratio calculator for calculating an SN ratio from a power ratio of the output signal;
A correction gain calculator for calculating an error correction gain based on an output of the ratio calculator and an appropriately set rule; an error correction gain determined by the correction gain calculator; an output signal of the subtractor; And a filter control unit which receives an output signal from a type microphone, estimates an impulse response of an acoustic transmission system from the noise source to the main microphone, and updates a filter coefficient of the digital filter. An adaptive noise suppression device for a video camera, comprising: 2. A main microphone for sound pickup disposed on a surface of a video camera housing, a close-talking microphone disposed near a noise source generated in the video camera housing, and the close-talking microphone. A digital filter for inputting an output signal from a microphone, a subtracter for subtracting an output signal of the digital filter from an output signal from the main microphone, an output signal from the main microphone, and an output signal from the close-talking microphone And an output signal of the subtractor, and calculates a noise suppression amount, and calculates an error correction gain based on an output of the noise suppression amount calculation unit and an appropriately set rule. A correction gain calculator, an error correction gain determined by the correction gain calculator, an output signal of the subtractor, and an output signal from the close-talking microphone; Estimating the impulse response of an acoustic transmission system from the noise source to the main microphone,
An adaptive noise suppression device for a video camera, comprising: a filter control unit configured to update a filter coefficient of the digital filter.