JP4774512B2 - Output control device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an output controller capable of effectively suppressing howlings which occur simultaneously at different two or more frequencies. <P>SOLUTION: The output controller 10 is provided with a difference calculator 11, a criterion information storage 12, a difference comparator 13, an abnormality determination portion 14, and an output control unit 15. The calculator 11 calculates a level difference between signals detected by two microphones 102a, 102b, and the comparator 13 compares the level difference with criterion information stored beforehand in the storage 12. Based on the comparison by the comparator 13, the determination portion 14 determines whether or not an abnormal signal enters. The output control unit 15 suppresses frequency components specified by the determination portion 14. <P>COPYRIGHT: (C)2007,JPO&amp;INPIT

Description

  The present invention relates to an output control device for suppressing generation of abnormal signals by being applied to a system in which a signal detection device and a signal output device such as a binaural hearing aid system are close to each other.

  When a signal detection device that handles various signals such as sound waves, mechanical vibrations, and electromagnetic waves and a signal output device are used in close proximity, signal wraparound may occur between the devices, resulting in so-called howling, etc. An abnormal energy divergence like this may occur.

  Such abnormal energy divergence may have a detrimental effect on human bodies and objects. Taking howling as a typical example, this howling is basically an abnormal divergence of sonic energy in the audible frequency band, which is meaningless and is likely to occur in a relatively high frequency range. It can even be other factors such as hearing impairment. Moreover, it is known that frequent occurrence of howling causes a great deal of stress on the listener.

In order to prevent the occurrence of such howling, a proposal has been made in the past to preliminarily estimate a signal and a signal transmission path, and to suppress only a specific frequency component that may be generated by the wraparound based on the estimation. Patent Document 1).
JP 2004-297304 A

  However, the abnormal signal such as the above-mentioned howling is not generated only in the vicinity of a single frequency but often generated at a plurality of different frequencies at the same time. For this reason, it has been difficult or impossible to sufficiently suppress howling with the conventional techniques as described above. In addition, the frequency band that becomes an abnormal signal is not always constant, and is often different from time to time. Therefore, only a single frequency is assumed as in the conventional technique described above. Then, there was a problem that an accurate suppression effect could not be obtained.

  The present invention has been made in view of such problems, and an object thereof is to provide an output control device capable of sufficiently effectively suppressing abnormal energy divergence such as howling that occurs simultaneously at a plurality of different frequencies. There is to do.

  An output control apparatus according to the present invention is an output control in a signal output apparatus in a system including at least two signal detection apparatuses that detect signals and a signal output apparatus that outputs signals detected by these signal detection apparatuses. The generation of abnormal signals is effectively suppressed by providing the following elements.

(1) Difference calculation means for calculating a level difference for each frequency component for signals detected by two signal detection devices (2) 2 detected in a state where no abnormal signal is detected in the two signal detection devices Level calculated by determination criterion information storage means (3) difference calculation means for storing, as determination reference information, the maximum absolute value of the level difference of sound waves detected by a plurality of subjects for each frequency component for one normal signal Difference comparison means for comparing the difference with the determination reference information (4) Abnormality determination means for determining whether an abnormal signal is taken in the signal detection device based on the comparison result of the difference comparison means (5) Abnormality determination means Output control means for controlling the signal output device based on the determination result

  In this output control device, a level difference between signals such as sound waves detected by the two signal detection devices is calculated. Then, the difference value is compared with the determination criterion information obtained in advance, and based on the comparison result, it is determined whether or not an abnormal signal has been taken in, and output control in the signal output device is performed based on the result. .

  More specifically, the abnormality determination unit has a function of specifying the frequency component when an abnormal signal is captured, and the output control unit determines the frequency component specified by the abnormality determination unit. Output control is performed to suppress it. In addition, the abnormality determination unit has a function of specifying the frequency components of abnormal signals at once for a plurality of frequency components.

  Further, in the output control by the output control means, it is desirable to perform weighting by multiplying the specified frequency component by a coefficient obtained by a predetermined weighting function (for example, cosine function).

  The output control device of the present invention can be suitably used for a hearing aid or the like provided as a single headset so that, for example, a pair of earphones and microphones are respectively applied to the left and right ears. That is, a microphone that detects a sound wave signal is used as the signal detection device, and a speaker or an earphone is used as the signal output device. In that case, an experiment is performed in a state where abnormal signals are not captured by a plurality of subjects wearing headsets, and the maximum absolute value of the level difference between the two signals captured at this time is calculated. , Can be used as determination criterion information.

  According to the output control device of the present invention, the level difference between the signals detected by the two signal detection devices is calculated, the difference value is compared with the determination criterion information obtained in advance, and based on the comparison result, Since it is determined whether or not an abnormal signal has been taken in and output control is performed by the signal output device based on the result, abnormal energy divergence such as howling in an acoustic system including an earphone and a microphone is used. Can be suppressed sufficiently effectively.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

  FIG. 1 shows a block configuration of a binaural hearing aid system to which an output control apparatus according to an embodiment of the present invention is applied, and FIG. 2 shows a state in use of this system. The binaural hearing aid system has a so-called ear-hook type headset 100, and microphones 102a and 102b (signal detection devices) are respectively mounted in the vicinity of the left earphone 101a and the right earphone 101b (signal output device). is there. In the present embodiment, the output control device 10 is provided between the earphones 101a and 101b and the microphones 102a and 102b.

  The binaural hearing aid system has spatial resolution, has an operation characteristic that estimates the direction of a sound source and emphasizes only the sound source in a specific direction, but there is a gap between the earphone 101a and the microphone 102a and between the earphone 101b and the microphone 102b. Since they are close to each other, howling, which is an abnormal divergence of sonic energy, may occur. The output control apparatus 1 suppresses such howling sufficiently effectively by performing howling cancellation using the interaural level difference.

  The output control device 10 includes a difference calculation unit 11, a determination reference information storage unit 12, a difference comparison unit 13, an abnormality determination unit 14, and an output control unit 15.

Among these, the difference calculation unit 11 calculates a phase difference and a level difference of the sound wave signals captured between the left and right microphones 102a and 102b. Further, in the determination criterion information storage unit 12, the phase difference and the level difference of two signals obtained in a state where normal sound wave signals are captured are stored in advance as determination criterion information. Here, the level difference in the criterion information is used as information for output control in the present embodiment, and the phase difference is used for estimation and separation of the sound source direction together with the level difference.
Specifically, the headset 100 is worn by a large number of subjects (for example, 96 people), and the phase of two sound wave signals captured from the left and right microphones 102a and 102b in a state where no howling has occurred. Find the difference and level difference. The reason why data from a large number of subjects such as 96 is obtained is that it is possible to provide accurate judgment criteria for different users, and here, the absolute value of the level difference for each frequency. Is the criterion information.

  In the difference comparison unit 13, the level difference calculated by the difference calculation unit 11 is compared with the determination reference information stored in the determination reference information storage unit 12. The abnormality determination unit 14 determines whether or not an abnormal signal has been taken in from the microphones 102a and 102b based on the comparison result in the difference comparison unit 13, and specifies the frequency of the abnormal signal. That is, when the level difference calculated by the difference calculation unit 11 is larger than the determination reference information, it is determined that an abnormal signal has been captured. In addition, when there are a plurality of frequency components of abnormal sound wave signals, the abnormality determination unit 14 has a function of specifying them at a time.

  The output control unit 15 suppresses abnormal output by weighting the frequency component identified as abnormal by the abnormality determination unit 14.

  Next, a specific operation of the output control device 10 in this binaural hearing aid system will be described.

  First, an interaural phase difference (IPD) θmap (k, φ, ψ) is calculated for each frequency component of the binaural observation signal by performing an experiment with the headset 100 in advance and a calculation using fast Fourier transform. And interaural level difference (ILD) ζmap (k, φ, ψ). Here, k represents a frequency index, φ represents an azimuth angle, and ψ represents an elevation angle.

  The database (determination reference information) stored in the determination information storage unit 12 is the maximum absolute value of the interaural level difference ζmax (for each of the frequency components obtained as described above) for a large number of subjects (for example, 96 people). k) is an accumulation of data obtained. This can be regarded as information representing the maximum level difference that can occur between both ears. FIG. 3 is a graph showing an example of the information content of such a database.

  The difference calculation unit 11 calculates the interaural phase difference and the interaural level difference for each frequency component from the sound wave signals (observation signals) detected by the left and right microphones 102a and 102. Then, the difference comparison unit 13 compares the calculated interaural phase difference and interaural level difference with each determination criterion information stored in the determination information storage unit 12, and uses the difference. Sound source direction estimation information for each frequency component is obtained.

  The abnormality determination unit 14 determines whether or not an abnormality has occurred based on the comparison result in the difference comparison unit 13. That is, when the level difference ζin (k) of each frequency component of the captured signal exceeds ζmax (k) of the determination reference information, a level difference that cannot be generated between both ears at normal time (ie, It is determined that (howling) has occurred.

When the abnormality determination unit 14 determines that howling has occurred, the output control unit 15 causes the howling prevention filter H (k) 15 </ b> A to function and performs howling suppression. This howling prevention filter H (k) 15A is generated by software by the output control unit 15 corresponding to the frequency component of howling occurring at any given time. More specifically, for example, as schematically shown in FIG. 4, the spectrum near the frequency component in which howling is generated is multiplied by a coefficient obtained by a weighting function (for example, a cosine function of a half wavelength). It attenuates by this. Here, if the width of the frequency index to be attenuated is kwid and the attenuation coefficient is γ, the howling prevention filter 15A is defined as shown in the following equation. Here, m is an integer in a range of −kwid / 2 ≦ m ≦ kwid / 2.
This weighting is applied to the frequency component determined to be abnormal.

  The sound source separation is performed based on the sound source direction estimation information obtained by the difference comparison unit 13. A filter for extracting a spectral component estimated to have arrived from the target sound source direction and a howling prevention filter function, and an inverse fast Fourier transform process is performed on the filter to obtain an enhanced signal. Note that IPD and ILD, which are inter-separated signal interaural information, are retained, and localization information is still included in the processed voice.

  In order to confirm the effect of the above-described howling prevention, a feedback model as shown in FIG. 5 was created and a computer simulation was performed.

  Hl (n) and hr (n) in the feedback model are impulse responses between the left and right earphones 101a and 101b and the microphones 102a and 102b of the binaural hearing aid system. Those hl (n) and hr (n) were measured by attaching a binaural hearing aid system to a dummy head in an anechoic chamber. The input was a signal in which HRTF from the front direction of the dummy head was convoluted with the voice <a:>. The sampling frequency was 16 [kHz]. The inverse fast Fourier transform (inverse FFT) was performed every frame period (32 [ms]).

  Here, in order to operate the function of the howling prevention filter 15A, two parameters of the attenuation coefficient γ and the frequency width kwid to be attenuated must be determined. In this experiment, those parameters were determined by conducting a listening experiment with three normal hearing persons in advance.

  As a result, the transition of power for each frame of the output signal in the feedback model is as shown in FIG. When the howling prevention filter 15A is not functioning, the power of the signal increases with time due to the occurrence of howling. In contrast, when the howling prevention filter 15A is operated, even if howling occurs, the power of the signal decreases with time and converges to a constant power.

  Further, the power spectrum of the output signal after 1500 [ms] from the start of the simulation is as shown in FIG. When the howling prevention filter 15A did not function, a characteristic peak lobe waveform of howling was observed in the vicinity of 4000 [Hz]. In contrast, when the howling prevention filter 15A is operated, such a peak lobe is not observed, and it has been confirmed that howling is effectively suppressed. In addition, the suppression effect of this howling could be confirmed clearly in the experiment.

  Thus, in the binaural hearing aid system using the output control device 10 of the present embodiment, howling that occurs simultaneously at a plurality of different frequencies can be sufficiently effectively suppressed.

  Although the present invention has been described with reference to the embodiment, the present invention is not limited to the above embodiment and can be variously modified. For example, in the above-described embodiment, the example in which the output control apparatus 10 of the present invention is applied to a binaural hearing aid system has been described. However, the present invention is not limited to such a system, and howling that occurs between a microphone and a speaker in the acoustic field is suppressed. Therefore, the present invention can be applied to all of the devices. In addition, not only in the acoustic field, but in a field where abnormal energy divergence such as vibration and noise due to mechanical rotational motion is likely to occur, as a device for suppressing the generated abnormal signal by multipoint control, Furthermore, since the electromagnetic wave is a kind of signal, the output control device of the present invention can be applied to multipoint control of the electromagnetic wave.

It is a block diagram showing the structure of the binaural hearing aid system using the output control apparatus which concerns on one embodiment of this invention. It is a figure showing the use condition of a binaural hearing aid system. It is the figure which showed an example of the information content of a database. It is the figure which showed the howling prevention filter H (k) typically. It is the block diagram which showed the structure of the principal part of the feedback model for confirming the effect of the operation of the howling prevention by filter H (k). It is a figure showing transition of the power for every frame of the output signal in the feedback model shown in FIG. FIG. 6 is a diagram showing a power spectrum of an output signal after 1500 [ms] from the start of simulation by the feedback model shown in FIG. 5.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 ... Output control apparatus, 11 ... Difference calculation part, 12 ... Determination reference | standard information storage part, 13 ... Difference comparison part, 14 ... Abnormality determination part, 15 ... Output control part, 15A ... Howling prevention filter, 100 ... Headset, 101a , 101b ... earphone, 102a, 102b ... microphone

Claims (4)

An output control device that performs output control in the signal output device of a system that includes at least two signal detection devices that detect signals and a signal output device that outputs signals detected by these signal detection devices. And
A difference calculating means for calculating a level difference for each frequency component of the signals detected by the two signal detection devices;
Determination of the maximum absolute value of the level difference of sound waves detected by a plurality of subjects for each frequency component for two normal signals detected in a state where no abnormal signal is detected in the two signal detection devices Determination criterion information storage means for storing as reference information;
Difference comparison means for comparing the level difference calculated in the difference calculation means with the determination criterion information;
An abnormality determination means for determining whether or not an abnormal signal is taken in the signal detection device based on a comparison result in the difference comparison means;
An output control device comprising: output control means for performing output control on the signal output device based on a determination result of the abnormality determination means. The abnormality determination means has a function of specifying a frequency component when an abnormal signal is captured,
The output control device according to claim 1, wherein the output control unit performs output control so as to suppress the frequency component specified by the abnormality determination unit. The output control device according to claim 2 , wherein the output control unit performs output control by performing weighting by multiplying the identified frequency component by a coefficient obtained by a predetermined weighting function. The output control device according to any one of claims 1 to 3 , wherein the signal detection device is a microphone that detects sound waves, and the signal output device is a speaker or an earphone.