JPH06164278A - Howling suppressing device - Google Patents

Abstract

PURPOSE:To provide a howling suppression device capable of improving a howling margin by setting such characteristic for a digital filter as to detect a howling frequency even in a state where a background noise exists and to control the frequency. CONSTITUTION:This device is comprised of a microphone 1 which collects sound at a reproducing place, an A/D converter 2 which converts an input signal to a digital value, a howling suppression means 6 which suppresses howling, a frequency analysis means 7 which performs the frequency analysis of a signal from the A/D converter 2, a howling detecting means 10 which detects the howling from the result of the frequency analysis means 7 based on difference between the level of the peak frequency of the howling frequency and the mean level of the whole and finds the frequency, an arithmetic means 8 which computes the coefficient of the digital filter which sets the howling frequency on the howling suppression means 6, and a control means 9 which sets the coefficient on the howling suppression means 6.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a howling suppressing device for preventing howling between a microphone and a speaker.

[0002]

2. Description of the Related Art FIG. 10 is a block diagram showing a conventional typical howling suppressing device.

In FIG. 10, 1 is a microphone for collecting performances and voices in the field, 16 is a microphone amplifier, 17 is a graphic equalizer for suppressing howling, 4 is an amplifier, and 5 is a speaker.

The operation of the conventional howling suppressing device configured as described above will be described.

During the lecture, howling occurs due to the movement of the lecturer and changes in the conditions of the lecture place. When such howling occurs, the mixer lowers the level of the frequency band considered by the graphic equalizer 17, or lowers the overall output level. Then, when the sound collection conditions have changed, such as howling has stopped and the position of the speaker has changed,
The characteristics of the graphic equalizer, or the overall level was restored.

As described above, conventionally, the mixer performs this work and suppresses howling each time howling occurs.

[0007]

However, in the above configuration, the mixer must always lower the frequency of the graphic equalizer when the howling occurs, and it takes time to suppress the howling and at the same time, the graphic There is a problem that the frequency band for lowering the equalizer 17 is not immediately known accurately and it takes time to suppress howling.

In view of the above problems, it is an object of the present invention to provide a howling suppressing device which can automatically detect howling and suppress howling even under a condition where the background noise level is relatively large.

[0009]

According to the present invention, there is provided a microphone for picking up a sound at a reproduction place, an A / D converter for converting an input signal into a digital signal, a howling suppressing means for suppressing howling, and an A / D converter. Frequency analysis means for performing frequency analysis in real time with a signal from the converter; howling detection means for detecting howling; computing means for computing coefficients of digital filters set in the howling suppressing means;
It is constituted by control means for setting the coefficient in the howling suppressing means.

[0010]

With this configuration, it is possible to accurately detect howling in a short time, calculate its frequency, and suppress howling even under a condition where the background noise level is high or there is sound.

[0011]

Embodiments of the howling suppressing apparatus of the present invention will be described below with reference to the drawings.

An object of the present invention is to accurately detect howling and suppress howling even when background noise is large.

FIG. 1 is a block diagram of a howling suppressing apparatus according to an embodiment of the present invention.

In FIG. 1, 1 is a microphone for picking up a sound at a reproduction place, 2 is an A / D converter for A / D converting the picked up sound into a digital signal, and 3 is a D / A conversion. A D / A converter for converting into an analog signal, 4 an amplifier, 5 a reproducing speaker, 6 a howling suppressing means for lowering the level of the howling frequency by notch filtering the signal from the A / D converter 2. Is frequency analysis means for frequency-converting the signal from the A / D converter 2, and 10 is frequency analysis means 7.
Howling detection means for detecting howling based on the result of 8), 8 means for calculating the characteristic set in the howling suppressing means 6, 9 for setting the characteristic of the digital filter obtained by the calculating means 8 in the howling suppressing means 6. However, it is a control means for suppressing howling. The same components as those in the conventional embodiment are designated by the same reference numerals.

The operation of the howling suppressor of this embodiment having the above-described structure will be described below.

The player's performance is picked up by the microphone 1, passed through the microphone amplifier 16, the A / D converter 2, the howling suppressing means 6, and the D / A converter 3, amplified by the amplifier 4, and reproduced from the speaker 5. To be done. Normally, the reproduction level of the speaker 5 is adjusted so that howling does not occur in this state. However, if the player moves or the direction of the microphone 1 changes, the loop gain of the reproduction system increases, and if the gain exceeds 1, howling occurs.

The signal from the microphone 1 is converted into a digital signal by the A / D converter 2 and enters the howling suppressing means 6. The signal from the A / D converter 2 is also input to the frequency analysis means 7, converted into a frequency by the means, and the howling detection means 10 constantly detects howling. In the present embodiment, the frequency analysis means 7 analyzes by FFT processing.

Next, a howling detection method by the howling detection means 10 will be described. FIG. 2 is a flowchart of howling detection means. First, the highest level among frequency components is searched. Next, the average value is obtained by the method of (Equation 1).

[0019]

[Equation 1]

As a method of obtaining the average value, all the levels except the N-th power level are added in descending order of power level, and the average is obtained by dividing by the added number. Next, the difference between the peak level and the average value is calculated based on (Equation 2), and if the difference exceeds a certain value, the power level of the peak frequency is higher than the average value. , Howling can be considered.

[0021]

[Equation 2]

Although howling occurs at a single frequency, the power level of frequencies before and after the howling frequency becomes higher than other frequencies by the method of frequency analysis. Therefore, as the howling increases, the average power level of the whole also rises, and the howling has a great influence on the average level. Therefore, as shown in (Equation 1), by not adding the power levels up to the Nth in descending order of power level to the total added value, the difference between the power levels other than howling and howling becomes larger, and howling I try to emphasize it.

FIG. 3 shows a comparison of the difference between the average power level obtained by the conventional method and the peak level based on the power level obtained in this embodiment when the howling has occurred. It is a figure. (A) is a howling waveform,
(B) shows the comparison by the method of obtaining. It can be seen that as the howling grows, the peak level of the howling frequency increases with respect to the overall level, and when the peak frequency level of the howling is added, the average power level of the entire also tends to increase considerably.

FIG. 4 shows howling waveforms and the results of analysis of the waveforms using the method of this embodiment. (A) shows a howling waveform, (b) shows a difference power level change by this method, and (c) shows a peak frequency. Thus, it can be seen that as the howling signal increases, the difference between the average value of the overall levels and the peak level increases. Here, with respect to the difference value, an appropriate threshold level is set as shown in (FIG. 4) (b). Howling detection means 1
At 0, if the difference exceeds the threshold level, it is regarded as howling, and at the same time, howling frequency is calculated.

In the case of footsteps and steady background noise, the frequency band is wide, so the difference between the peak and the average level becomes small,
Not considered howling.

When the howling frequency is calculated and the howling frequency is calculated, the calculating means 8 calculates a coefficient that constitutes a digital filter for reducing the gain of only the howling frequency component of the howling suppressing means 6.
The calculated coefficient of the digital filter is set in the howling suppressing means 6 by the control means 9. In this embodiment, a notch filter is used as the digital filter.

By the above operation, when howling occurs, the frequency component is removed by a digital filter to suppress the howling.

As described above, according to the present embodiment, frequency analysis is performed using the frequency analysis means 7 and the howling detection means 10, and howling is judged based on the difference between the peak level and its average value to determine the frequency. By obtaining and removing the howling frequency component with a notch filter, howling can be removed even when the background noise is large.

Although the howling between one microphone and the speaker has been described in the present embodiment, the same effect can be obtained even when a plurality of microphones and speakers are used.

A howling suppressing apparatus according to a second embodiment of the present invention will be described below with reference to the drawings.

An object of the present invention is to suppress howling. The operation up to the frequency analysis means 7 is the same as that of the first embodiment. FIG. 5 shows a howling waveform (a) when a plurality of howlings occur simultaneously and a frequency characteristic (b) analyzed by the frequency analyzing means 7.

When the number of howling frequencies is large, the average level of the whole rises, the difference from the peak becomes small, and the howling detection accuracy deteriorates. Therefore, the change in the peak power level and the continuity of the peak frequency are used as parameters for howling detection. FIG. 6 shows a howling waveform (a), a peak level change (b), and a peak frequency change with time (c). In (Fig. 6), the peak frequency is stable even under the condition of multiple howling,
It can be seen that the power level at that peak is increasing. In this embodiment, howling is determined by using the continuity of peak frequencies, increase / decrease in peak power level, and increase / decrease in overall power level as howling conditions. If the peak frequency continues M times, the peak power level increases from the analysis result one time before, and the overall power level also increases A times or more, it is considered as howling. The number of continuous peak frequencies is determined by the time required for one frequency analysis and the characteristics of background noise.

As described above, when the howling grows, the peak level also increases and the frequency becomes constant, so that the howling can be detected by the above determination.
According to this method, general noise has large peak frequency fluctuations and the peak level rarely monotonically increases, so that it is less likely to be erroneously detected as howling. For pulsed noise, it is possible to prevent erroneous detection by appropriately determining the number of times the peak frequency continues.
The subsequent processing is the same as in the first embodiment.

As described above, according to the present embodiment, howling can be detected relatively accurately even under the condition of noise or a plurality of howlings by judging the increase / decrease of the peak power level and the continuity of the peak frequency. You can

When the level of the low frequency is high due to the background noise, the detection processing may be performed on the signal from which the background noise in the low range is removed by using the low pass filter to further reduce the erroneous detection. Is possible.

A howling suppressing apparatus according to a third embodiment of the present invention will be described below with reference to the drawings.

An object of the present invention is to suppress howling. (FIG. 7) is a diagram showing a structure of a howling suppressing apparatus according to a third embodiment of the present invention. 11 is A /
It is a voice discriminating means for detecting a voice section based on a signal from the D converter 2, and the other parts have the same configuration as that of the first embodiment.

The operation of the howling suppressing device configured as described above will be described below.

The operation up to the frequency analysis means 7 is the same as that of the first embodiment. The voice discriminating means 11 discriminates voice / non-voice based on the signal from the A / D converter 2. When it is determined that the sound is a sound, the threshold level for howling detection of the howling detection means 10 is changed. Second
In the embodiment, when the difference between the maximum power level and the average power level is larger than a certain threshold level a, howling is determined, whereas in the present embodiment, the voice section has the threshold level a. Decrease the value and perform detection.

As described above, according to the present embodiment, howling can be more accurately performed by changing the threshold level for howling detection of the howling detecting means 10 in the audio portion judged by the sound judging means 11. Can be detected.

In this embodiment, the howling is detected by the value of the difference component between the peak value and the average level. However, the method of the second embodiment or simply the threshold level having a power level is used. It can be used in various cases, such as howling is regarded as howling when the value exceeds.

A howling suppressing apparatus according to the fourth embodiment of the present invention will be described below with reference to the drawings.

An object of the present invention is to suppress howling. (FIG. 8) is a diagram showing a structure of a howling suppression device in a fourth embodiment. Reference numeral 12 is a threshold level control means for changing the threshold level of the howling detection means 10. The other structure is the same as that of the third embodiment.

The operation of the howling suppressing device configured as described above will be described below.

The operation up to the voice discriminating means 11 is the same as that of the third embodiment. When the voice discriminating means 11 discriminates the voice, the threshold level control means 12 changes the threshold level for howling detection by the howling detecting means 6. The threshold level is
It is set for several frequency bands.

As a howling detection method, if howling is performed when the level of the frequency band is equal to or higher than the threshold level, in the case of voice, a pitch frequency component is generally erroneously detected as howling, so that a frequency band near the voice pitch is detected. False detection is reduced by increasing the threshold level for howling detection and making the detection sensitivity dull.

FIG. 9 shows an example of the frequency characteristic (a) of the frequency analysis means 7 in the presence of voice and the threshold changing method (b). The sound has a high power level at a frequency near 200 Hz, and at a normal reference level for howling detection, the frequency is determined to be howling. Therefore, as an example of the case of voice, the threshold level of the pitch frequency band (see FIG.
By setting like (b)), the 200Hz
Even if the level of the nearby band becomes higher due to the voice pitch, it becomes possible to prevent erroneous detection of howling.

As described above, according to the present embodiment, howling is more accurately performed in the voice portion judged by the voice judging means 11 by changing the threshold level for howling detection of the howling detecting means 10. Can be detected.

Although the threshold level change of howling detection in the case of voice is described in the present embodiment, the low range where the background noise level is large, the band where the noise of a specific frequency is large, and the acoustic of the room where howling detection is performed. By changing the threshold level for detection according to the conditions, it is possible to prevent erroneous howling detection under various conditions.

Further, even when the voice section is not detected by the voice discriminating means 11, the same effect can be obtained by setting the threshold level for each band in advance.

[0051]

As is apparent from the above description, according to the present invention, the voice discriminating means discriminates between voice and non-voice, and the howling detecting means obtains the level difference between the peak value and the average value in real time, thereby detecting the voice / non-voice. The characteristics of a digital filter that lowers the level of the howling frequency are determined by making a decision as to howling in accordance with the sound, and by removing the howling frequency component, the sound quality can be reliably and reliably maintained even under the conditions of large background noise and loud sound. It is possible to increase the howling margin in a state where there is little change in.

[Brief description of drawings]

FIG. 1 is a block diagram of a howling suppressing device according to a first embodiment of the present invention.

FIG. 2 is a diagram showing a howling detection method of howling detection means according to the first embodiment of the present invention.

FIG. 3 is a comparison diagram of the method according to the first embodiment of the present invention and the conventional method.

FIG. 4 is an explanatory diagram of howling detection in the first embodiment of the present invention.

FIG. 5 is an input signal waveform in the second embodiment of the present invention and a frequency characteristic diagram obtained by FFT of the same.

FIG. 6 is an explanatory diagram of howling detection in the second embodiment of the present invention.

FIG. 7 is a block diagram of a howling suppressing device according to a third embodiment of the present invention.

FIG. 8 is a block diagram of a howling suppressing device according to a fourth embodiment of the present invention.

FIG. 9 is a diagram showing a threshold level changing method in the fourth embodiment of the present invention.

FIG. 10 is a block diagram of a conventional howling suppression device.

[Explanation of symbols]

 1 Microphone 6 Howling Suppressing Means 7 Frequency Analyzing Means 8 Computing Means 9 Controlling Means 10 Howling Detecting Means 11 Voice Discriminating Means 12 Threshold Level Controlling Means

Continued Front Page (72) Inventor Hiroko Numazu 1006 Kadoma, Kadoma City, Osaka Prefecture Matsushita Electric Industrial Co., Ltd.

Claims (5)

[Claims] 1. A microphone for picking up a sound at a reproduction place, an A / D converter for converting an input signal from the microphone into a digital value, and a frequency characteristic of a signal input from the A / D converter. Controlling howling to suppress howling, frequency analyzing means for performing frequency analysis of the signal from the A / D converting means, howling detecting means for detecting howling based on the output of the frequency analyzing means, Calculation means for obtaining the coefficient of the digital filter to be set in the howling suppressing means so that an intended frequency characteristic can be obtained based on the signal from the howling detecting means, and the obtained coefficient is set in the howling suppressing means. A howling suppression device including a control means. 2. The howling detecting means obtains the power level of the peak frequency and the total frequency power level addition value from the value obtained by subtracting the Nth power level from the highest power level based on the result of the frequency analyzing means. The howling suppressing apparatus according to claim 1, wherein the detection is performed by judging the difference from the average value. 3. The howling detecting means detects howling by judging continuity of peak frequencies and increase / decrease in overall power level based on the result of the frequency analyzing means. Howling suppression device. 4. A voice discriminating means for discriminating between voice and non-voice based on a signal from the A / D converting means is provided. When the voice discriminating means discriminates a voice section, howling is detected. 2. The howling suppressing apparatus according to claim 1, wherein the threshold level of is changed. 5. A threshold level control means for setting a threshold level for howling detection in the howling detection means, wherein the threshold level is set for each of a plurality of frequency bands. The howling suppression device described.