JP2002095084A - Directivity reception system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To largely improve directional characteristics when external noises deteriorating the articulation of a hearing aid are eliminated by the direction characteristics of a microphone. SOLUTION: A reception array 5 composed of microphones 1, 2, 3 and 4 having a plurality of channels is arranged on the frame of a spectacle type fixture or a supporting beam for a headphone type fixture. Outputs from each microphone are wavelet-transformed 6 and amplitude and phase spectra are obtained, and the next operation is conducted at every frequency band. Difference is expanded 11 by multiplying phase difference among each channel by an arbitrary coefficient, and amplitude and phase among each channel are interpolated and the number of the channels is multiplied 16 in arbitrary size. These outputs are multiplied 21 by an arbitrary weighting function and phase adjust addition 26 is conducted, and the result is used as a conventional amplifier input by wavelet inverse transformation 27. Phase difference among each channel is expanded, and directional width is reduced by multiplying the number of the channels, and external-noise elimination performance is improved largely.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION The present invention relates to hearing aids.

[0002]

2. Description of the Related Art External noise is one of the factors that reduce the clarity of a hearing aid. However, in a conventional hearing aid, the input / output characteristics and frequency characteristics of a hearing-impaired ear are compensated by the characteristics of the hearing aid, and the signal-to-noise ratio is reduced. The mainstream is to prevent further deterioration. Some amplifiers are divided into frequency bands to suppress noise by controlling the degree of amplification in the noise band. However, there is a problem when the amplifier overlaps with the signal band.
In this case, noise can be reduced by the directivity of the microphone and the array, but the sharpness of the directivity is not sufficient because the size of the device is limited as compared with the wavelength.

[0003]

When external noise that degrades the clarity of a hearing aid is to be removed by the directional characteristics of a microphone, a sharp directional characteristic is obtained by the conventional method due to the limitation on the size of the microphone array. I can't. The present invention significantly improves this directional characteristic.

[0004]

A receiving array comprising microphones of a plurality of channels is arranged on a frame of an eyeglass-type wearing device or a support beam of a headphone-type wearing device.
The output of each microphone is wavelet transformed to obtain an amplitude spectrum and a phase spectrum, and the following operation is performed for each frequency band. The difference between the channels is expanded by multiplying the phase difference between the channels by an arbitrary coefficient, and further interpolating the amplitude and phase between the channels to calculate the interpolation channel output. Multiply. These outputs are multiplied by an arbitrary weighting function, subjected to phasing addition (or simply added when a beam is formed in the front direction), and subjected to inverse wavelet transform to obtain a conventional amplifier input.

[0005]

FIG. 1 is a system diagram showing an embodiment of the present invention. M ₁ , M ₂ , M _i, and M _m (i = 1, 2,... M) of ₁ , ₂ , 3, and 4 are m-channel microphones, which become d on the frame of the eyeglass-type wearing device or the support beam of the headphones. The microphone arrays are arranged at horizontal distance intervals and constitute five microphone arrays.

F _i (t) (i = 1, 2,..., M) is the output of each microphone, and the m-channel frequency spectrum F _i (ω) (i = 1,
2,..., M). A of 7, 8, 9 and 10
_i (ω) (i = 1, 2,..., m) and α _i (ω) (i
= 1, 2,... M) are F _i (ω) (i = 1,
2,..., M). Ω is an angular frequency. At 11, the following operations are performed on these outputs for each frequency band. _{α i (ω) (i =} 1,2, ..., m) determine the phase difference between channels than to enlarge the phase difference between channels is multiplied by the arbitrary coefficient K _p to the difference, new phase spectrum β _i
(Ω) (i = 1, 2,..., M). Frequency spectrum _{G i (ω) (i =} 1,2, ..., m) , and the 12,
13, 14, and 15 show the radiation spectrum A
_i (ω) (i = 1, 2,..., m) and the phase spectrum β _i (ω) (i = 1, 2,..., m). Next 16
Then, for each frequency band, A _i (ω) (i = 1, 2, 2)
.., M) and β _i (ω) (i = 1, 2,... M), by interpolating an arbitrary number of amplitudes and phases between channels,
Out a total spectrum of the interpolation channel, new frequency spectra by multiplying the number of channels including the interpolation channel arbitrary size _{n = K c m H k (} ω) (k = 1,2, ...,
n). 17, 18, 19 and 20 show the amplitude spectrum B _k (ω) (k = 1, 2,..., N) and the phase spectrum γ _k (ω) (k = 1, 2,... N).
Next, at 21, the amplitude B _k (ω) of each channel (k = 1,
2,..., N) have an arbitrary weighting function W _k (k = 1, 2,.
n) to perform shading to obtain a new frequency spectrum J _k (ω) (k = 1, 2,..., n). 2
2, 23, 24 and 25 have their radiation spectra C _k
(Ω) (k = 1, 2,..., N) and phase spectrum γ
_k (ω) (k = 1, 2,..., n). In step 26, these inputs are subjected to phasing addition in an arbitrary direction (or simply adding when forming a beam in the front direction) to obtain a directional output J (ω). At 27, J (ω) is inversely wavelet transformed to a directional output j (t), which is used as a conventional amplifier input.

A calculation example is shown below for a linear array. The array length is al, the microphone interval is d, the number of microphone channels is m, the sound wave arrival direction is measured from the front of the array, θ, the phase difference expansion coefficient is K _p , the channel multiplication coefficient is K _c , and the wavelength is λ. From the above, the microphone interval d is d = al / (m-1), and the number n of multiplied channels is n =
K _c m, multiplication channel spacing s is _{s = al · K p / (} n-
1). The phase spectrum β _i (ω) is given by equation (1). β _i (ω) = K _p α _i (ω) − (α ₁ (ω) + α _m (ω)) (K _p −1) / 2 (1) Addition output when the phasing direction is 0 degrees Is as follows depending on each condition. 1. In the case of a non-directional microphone R (θ) = R
(Θ). 2. In the case of a unidirectional microphone R (θ) = R (θ) _c R (θ) _c = R (θ) _o (1 + cos (θ)) / 2 (3) 3. of the secondary sound pressure gradient microphone Case R (θ) = R (θ) _g R (θ) _g = R (θ) _o ((1 + cos (θ)) / 2) sin (πa cos (θ) / λ) / sin (πa / λ) (4) where a is the distance before and after the sound pressure gradient microphone. 4 When Shading is not Performed When W _k = 1, R in Expressions (2), (3) and (4)
(Θ) _o is given by equation (5). _{R (θ) o = sin (} πn al K p sin (θ) / (n-1) / λ) / s in (π al K p sin (θ) / (n-1) / λ) / n ...... (5)

[0008]

The calculation results of the directional characteristics are shown in FIG. 2 and subsequent figures.
FIGS. 2, 3 and 4 show the case where the phase difference expansion coefficient K _p = 1 and the channel multiplication coefficient K _c = 1, and the omnidirectional element, the unidirectional element and the secondary sound pressure gradient element are respectively shown. Used. In contrast, in FIGS. 5, 6 and 7, K _p =
5, K _c = 5 to reduce the directivity width. In FIGS. 8, 9 and 10, with respect to the improvement of the directivity width, shading is added to reduce the side lobe. However, since this increases the directivity width slightly, in FIGS. 11, 12 and 13, K _p =
8, K _c = 8 to reduce the directivity width again. Here, array length al = 15 cm, number of microphone channels m
= 2, frequency f = 1000 Hz, wavelength λ = 34 cm, 2
The distance a between the sensors of the next sound pressure gradient element is a = 5 cm. Also, the shading coefficient W _k = 0.54 + 0.4
6 cos (π (2k−n−1) / (n−1)).
"FIGS. 2, 3, and 4" and "FIGS. 5, 6, and 7"
And FIGS. 8, 9 and 10 and FIGS.
And Comparing Figure 13 ", it can be seen that the directivity width by increasing the K _p and K _c is reduced to a large congestion. When the microphones can be arranged in the vertical direction as in the case of a headphone-type wearing device, the directional characteristics in the vertical direction can be further improved in the same manner.

[Brief description of the drawings]

FIG. 1 is a system diagram of an embodiment of the present invention.

FIG.

FIG. 3 and

FIG. 4 shows a phase multiplication factor K _p = 1 and a channel multiplication factor K _c
Directivity characteristics when = 1.

FIG.

FIG. 6 and

FIG. 7 shows directivity characteristics when the directivity width is reduced with K _p = 5 and K _c = 5.

FIG.

FIG. 9 and

FIG. 10 shows directivity characteristics when the directivity width is reduced with K _p = 5 and K _c = 5, and the output of each channel is shaded to attenuate the side lobe.

FIG.

FIG. 12 and

FIG. 13 shows directivity characteristics when the directivity width slightly increased by shading is reduced again by setting K _p = 8 and K _c = 8.

[Explanation of symbols]

1, 2, 3, 4 ... microphone 5 ... microphone array 6 ... wavelet transform unit 7, 8, 9, 10 ... amplitude and phase spectrum 11 ... phase difference enlargement unit 12, 13, 14, 15 ... spectrum after phase difference enlargement Reference Signs List 16: Channel multiplying unit 17, 18, 19, 20 ... Spectrum after channel multiplication 21 ... Shading unit 22, 23, 24, 25 ... Spectrum after shading 26 ... Phasing and adding unit 27 ... Wavelet inverse transform unit

Claims (1)

[Claims] 1. A function of enlarging a phase difference between outputs of a plurality of microphones constituting a receiving array, calculating an interpolated output between the outputs, and obtaining an added output of these outputs. Directional receiving system.