JPH08193876A - Apparatus and method for detection of howling frequency - Google Patents

Apparatus and method for detection of howling frequency

Info

Publication number
JPH08193876A
JPH08193876A JP624795A JP624795A JPH08193876A JP H08193876 A JPH08193876 A JP H08193876A JP 624795 A JP624795 A JP 624795A JP 624795 A JP624795 A JP 624795A JP H08193876 A JPH08193876 A JP H08193876A
Authority
JP
Japan
Prior art keywords
frequency
howling
power spectrum
detecting
peak
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
JP624795A
Other languages
Japanese (ja)
Inventor
Michiaki Yoneda
道昭 米田
Original Assignee
Sony Corp
ソニー株式会社
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Sony Corp, ソニー株式会社 filed Critical Sony Corp
Priority to JP624795A priority Critical patent/JPH08193876A/en
Publication of JPH08193876A publication Critical patent/JPH08193876A/en
Withdrawn legal-status Critical Current

Links

Abstract

PURPOSE: To simply detect a howling operation and a howling frequency by increasing the point number of an FFT (fast Fourier transform) processing operation. CONSTITUTION: A sound which is collected by a microphone 1 is converted into a digital signal by an A/D converter 3 through an amplifier 2, it is then FPT-processed by a howling-frequency detection circuit 4, the frequency point of the peak of a power spectrum is found, and whether the sound is generated by a howling operation or not is detected. When the sound is generated by the howling operation, a frequency obtained by counting the cycle of the digital signal from the A/D converter 3 is compared with the frequency point of the peak, and a howling frequency is decided.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a case where a microphone is used between a microphone and a speaker when a musical instrument performance or voice is picked up by a microphone and a speaker is used to reproduce sound. The present invention relates to a howling frequency detection device and a howling frequency detection method for detecting a frequency of howling that occurs.

[0002]

2. Description of the Related Art Conventionally, howling occurs between a microphone and a speaker when a musical instrument performance or voice is picked up by using a microphone and sound reproduction is performed by using a speaker in, for example, a concert. When detecting the frequency of the howling, a howling frequency detecting device is used.

Here, a schematic structure of the howling frequency detecting device is shown in FIG.

In the howling detection apparatus of FIG. 6, the microphone 11 picks up a sound at a reproduction place such as a musical instrument performance or voice. The signal from the microphone 11 is input to the A / D (analog / digital) converter 12 via the microphone amplifier 26 and converted into a digital signal, and then input to the first howling detection means 17.

In the first howling detection means 17,
Coarse precision howling detection is performed using the transmitted digital signal. In addition, the second howling detection means 20
Then, based on the output signal from the first howling detecting means 17, the howling and the howling frequency are obtained with fine accuracy.

Further, the second howling detecting means 2
The howling and howling frequency obtained at 0 are sent to the calculating means 18, and the coefficient which constitutes the digital filter for reducing the gain of only the detected howling frequency component is calculated. This coefficient is controlled by the control means 19 and sent to the howling suppression means 16.

The howling suppressing means 16 has the A
The digital signal from the / D converter 12 is being sent,
The howling suppressing means 16 reduces the gain of only the howling frequency component of the sent digital signal. The digital signal in which the howling is suppressed is converted into an analog signal by the D / A (digital / analog) converter 13, and then output from the speaker 15 via the amplifier 14.

The first howling detecting means 17 is also provided.
FIG. 7 shows a specific configuration of the above.

The first howling detecting means 17 is provided with a plurality of band division filters and measures the level ratio between the fundamental frequency of howling and other frequency components. In the first howling detection means 17, the above A /
The output from the D converter 12 is fed to the bandpass filter 23 1
.About.23 n are passed to be divided into frequencies for each band and input to the howling determination means 21.

The howling judging means 21 makes a judgment for each frequency band by providing a threshold level of howling level with respect to a frequency for each band.
If the threshold level is exceeded in a certain band, look at the levels in other bands, and if the overall level is high, it is judged that the possibility of a music signal is high and the possibility of howling is low. Continue detection. on the other hand,
When the overall level is low and the level of only a specific frequency band is high, it is regarded as howling and rough howling frequency information is transferred to the second howling detection means 20.

The second howling detecting means 20 obtains a howling frequency by performing a fast Fourier transform (FFT) process on an input signal, and always performs an FFT process to obtain a maximum amplitude value. Monitoring the frequency component. Then, the change in the maximum frequency component is monitored, and if the change is large, howling is determined, and if the change is small and the high frequency component is small, howling is determined.

When the howling is detected, the howling frequency is calculated at the same time.

The howling frequency detecting device is provided with an autocorrelation function calculating means for obtaining an autocorrelation function using the output from the A / D converter 12, and the output from the autocorrelation function calculating means is used as a second value. It may be supplied to the howling detection means 20 and the howling and the howling frequency may be obtained with fine accuracy.

As described above, the howling frequency detecting device removes howling using the detected howling frequency and outputs a music signal.

[0015]

By the way, in practice,
The howling kurtosis so-called Q is very sharp. Therefore, unless the center frequency of the howling is accurately measured, howling is performed using a parametric equalizer or a notch filter by a digital filter based on the result measured by the above-mentioned howling frequency detector. When removing the, it often happens that the surrounding frequencies other than howling are also removed.

However, in order to accurately measure the howling frequency in the howling frequency detecting apparatus using the FFT processing as described above, it is necessary to increase the number of points in the FFT processing. However, if the number of points in the FFT processing is increased. The processing time increases and the hardware configuration increases.

Therefore, in view of the above situation, the present invention provides an FF.
Provided are a howling frequency detecting device and a howling frequency detecting method capable of accurately measuring a howling frequency without increasing the number of T processing points.

[0018]

A howling frequency detecting apparatus according to the present invention includes a digital signal converting means for converting a picked-up signal into a digital signal, and a fast Fourier transform process for an output from the digital signal converting means. Fast Fourier transforming means for performing, power spectrum calculating means for calculating a power spectrum using the spectrum from the fast Fourier transforming means, and frequency point detecting means for obtaining the frequency point of the peak of the power spectrum from the power spectrum calculating means , It is determined whether or not the peak of the power spectrum is howling according to the result of comparison between the level of the frequency point of the peak of the power spectrum from the frequency point detection means and the level of another frequency point or the presence or absence of harmonics. Howling determination means When the ring discrimination means determines that the peak of the power spectrum is howling, frequency detection means for detecting a frequency based on the output from the digital signal conversion means, output from the frequency detection means and frequency point detection Comparing means for comparing the frequency of the peak of the power spectrum from the means, and howling frequency determining means for making the output from the frequency detecting means the howling frequency in accordance with the comparison result from the comparing means. Solves the above-mentioned problem.

Here, the frequency detecting means counts the outputs from the digital converting means a plurality of times to obtain an average, and the comparing means calculates the average output from the frequency detecting means and the power spectrum from the frequency point detecting means. And the howling frequency determining means determines whether or not the comparison result from the comparing means is within the interval of the frequency points in the fast Fourier transforming means.

Further, the center frequency of the bandpass filter is moved to the frequency point of the peak of the power spectrum discriminated as howling by the howling discriminating means, and the frequency detecting means outputs the output from the digital signal converting means. It is characterized in that the frequency is detected after passing through a bandpass filter.

The howling frequency detecting method according to the present invention comprises a digital signal converting step for converting a picked-up signal into a digital signal, and a fast Fourier transform for applying a fast Fourier transform process to the output from the digital signal converting step. Step, a power spectrum calculation step of calculating a power spectrum using the spectrum from the fast Fourier transform step, a frequency point detection step of obtaining a frequency point of a peak of the power spectrum from the power spectrum calculation step, and the frequency point A howling determination step of determining whether or not the peak of the power spectrum is howling according to the result of comparison between the level of the frequency point of the peak of the power spectrum from the detection step and the level of another frequency point or the presence or absence of harmonics When,
When the howling determination step determines that the peak of the power spectrum is howling, a frequency detection step of detecting a frequency based on the output from the digital signal conversion step, an output from the frequency detection step, and the frequency point It comprises a comparison step of comparing the frequency of the peak of the power spectrum from the detection step, and a howling frequency determination step of making the output from the frequency detection step a howling frequency in accordance with the comparison result from the comparison step. It is characterized by

Here, in the frequency detecting step, the outputs from the digital converting step are counted a plurality of times to obtain an average, and in the comparing step, the average output from the frequency detecting step and the power spectrum from the frequency point detecting step. Is compared with the frequency point of the peak, and it is determined in the howling frequency determining step whether or not the comparison result from the comparing step is within the frequency point interval in the fast Fourier transform step.

Further, the center frequency of the bandpass filter is moved to the frequency point of the peak of the power spectrum which is discriminated to be howling in the howling discrimination step, and the output from the digital signal converting step is converted to the above in the frequency detecting step. It is characterized in that the frequency is detected after passing through a bandpass filter.

[0024]

In the present invention, the picked-up signal is converted into a digital signal, the fast Fourier transform process is performed to calculate the power spectrum, and the level of the frequency point at the peak of this power spectrum and the levels of other frequency points are calculated. Whether or not the peak of the power spectrum is howling is determined according to the result of comparison with the above or the presence or absence of harmonics.
Further, when it is determined that the peak of the power spectrum is howling, the frequency is detected from the digital signal output, and the howling frequency is determined according to the result of comparison between the detected frequency and the frequency of the peak of the power spectrum. By determining, the howling and the howling frequency can be easily detected by using different properties of the transient music signal and the steady howling.

Here, the digital signal output is counted a plurality of times to obtain an average, and the result of comparing the average output with the frequency point of the peak of the power spectrum is within the interval of the frequency points of the fast Fourier transform processing. The howling frequency can be accurately measured by determining whether or not there is any.

Further, the center frequency of the bandpass filter is moved to the frequency point of the peak of the power spectrum discriminated as the howling, and the frequency is detected after the digital signal output is passed through the bandpass filter. The frequency detection accuracy can be improved.

[0027]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT A preferred embodiment of the present invention will be described below with reference to the drawings. FIG. 1 shows a schematic configuration of a howling frequency detecting device according to the present invention.

The microphone 1 shown in FIG. 1 picks up a sound at a reproduction place such as a musical instrument performance or voice. The signal from the microphone 1 is A / D (analog / digital) via the amplifier 2.
The signal is input to the converter 3, converted into a digital signal, and then supplied to the howling frequency detection circuit 4.

The howling frequency detection circuit 4 performs FFT processing on the input signal to obtain a power spectrum, and detects the maximum value of the power spectrum, that is, the peak frequency point. Further, the value of the power spectrum at a frequency point that is an integral multiple of this frequency point is also detected.

Next, howling discrimination processing is performed by using the maximum value of the detected power spectrum and the frequency points which are integral multiples thereof. In this howling discrimination process, the maximum value of the power spectrum has a certain difference or more as compared with the average value of the values of other power spectra except the maximum value of the power spectrum, and the integer multiple. Has a certain difference or more with respect to the maximum value of the power spectrum at the frequency point of, and at the frequency point of the integral multiple of the value of the other power spectrum except the value of the integral power spectrum of the maximum value. When the difference does not exceed a certain value with respect to the average value, the frequency of the maximum value of the power spectrum is set as the howling frequency point.

Now, howling discrimination processing will be specifically described using the power spectrum shown in FIG.

As shown in FIG. 2, when the maximum value of the power spectrum exists at the frequency f 0 , in order to determine whether or not this frequency f 0 is the frequency point of howling, first, this power It is determined whether or not the difference a between the average q of the values of the other power spectra except the maximum value of the spectrum and the maximum value of the power spectrum is a certain value or more.

Since howling occurs at a single frequency, there are no harmonics that are integral multiples. When an integral multiple of harmonics is present, its power spectrum is nothing but an instrument or a human voice. Therefore, at frequency points that are integer multiples of the frequency point of the maximum value of the power spectrum, the differences b 1 and b from the maximum value of the power spectrum
It is determined whether 2 , b 3 , ... Are above a certain level.

Further, at the frequency points of the integral multiples, the differences c 1 , c 2 , with the average of the values of the power spectra other than the value of the power spectrum of the integral multiples of the maximum value.
It is determined whether or not c 3 , ... Is not more than a certain value.

For example, the difference a is 20 dB or more,
Moreover, the differences b 1 , b 2 , b 3 , ... Are each 20
dB or more, and the differences c 1 , c 2 , c 3 , ...
When is not more than 20 dB, it is determined that the frequency having the maximum value of the power spectrum is the howling frequency point.

Therefore, the number of points of the FFT processing for detecting the frequency points of the howling can be determined by determining whether or not there is a second peak at a frequency point that is an integral multiple of the frequency of the maximum value of the power spectrum. Good, you don't need a lot of points.

The average q of the values of the other power spectra except the maximum value of the power spectrum and the average of the values of the other power spectra except the values of the power spectrum which are integral multiples of the maximum value are substantially equal. Therefore, the average q is used as the average of the values of the other power spectra except the value of the power spectrum that is an integral multiple of the maximum value.

If the frequency of the maximum value of the power spectrum is determined to be the howling frequency point by the howling determination processing, the A / D converter 3 is used.
The output signal from is taken in a certain amount in real time, and frequency detection processing or frequency counting processing is performed to detect the frequency by counting the period of the taken signal that crosses the zero crossing.

Specifically, as shown in FIG. 3, in consideration of the DC offset, the period p for zero crossing on the plus side and the period m for zero crossing on the minus side are respectively counted.

Cycles p [0], p [1], p [2], ...
The average value of [n] is represented by the following equation (1), and the zero crossing period m [0], m on the minus side is counted.
The average value of [1], m [2], ..., M [n] is expressed by the following equation (2).

[0041]

[Equation 1]

[0042]

[Equation 2]

Further, these average values are taken and output as a measurement result. When this measurement result is R, the measurement result R is expressed by equation (3).

[0044]

(Equation 3)

After that, if the difference between the frequency obtained by the frequency counting process and the detected howling frequency point is within the interval of the frequency points in the FFT process, the frequency obtained by the frequency counting process. Is regarded as the howling frequency.

By repeating the above frequency counting process a plurality of times and averaging, the average value approaches the value of the howling frequency which is a stationary signal.

Further, the larger the number of times of averaging, the closer the value to the howling frequency. Compared with the average value of the power spectrum value of the point, it limits the detection of the howling frequency point only when there is a certain difference or more, so it approaches the value of the howling frequency with a relatively small number of times. .

As described above, in this embodiment, an accurate howling frequency can be obtained only by repeating the zero-cross frequency count a plurality of times, and the processing for obtaining the howling frequency can be simplified.

It is also conceivable that the output signal from the A / D converter 3 is passed through a bandpass filter (BPF) before the frequency counting process and the frequency counting process is performed using this output. In this case, after detecting the frequency point of the howling, the center frequency of the bandpass filter is moved to the frequency point of the howling to pass the output signal from the A / D converter 3, and this output is used. By performing the frequency counting process, the accuracy of the counted frequency can be improved.

In the frequency counting process,
Instead of using the zero cross, a certain threshold level may be set and the period crossing this threshold level may be counted.

The signal from the amplifier 2 is sent to the equalizer 5 to have its phase adjusted and output from the speaker 7 via the power amplifier 6.

Here, the howling frequency detection circuit 4 is used.
Then, the coefficient of a digital filter forming a parametric equalizer or a notch filter whose center frequency is the detected howling frequency is calculated, and a digital filter using this coefficient is inserted between the microphone 1 and the speaker 7 to perform howling. May be removed.

Next, a flowchart of the first specific procedure of the howling frequency detecting method is shown in FIG. 4, and the howling frequency detecting method will be described below.

First, in step S1, FFT processing is performed on the digital signal digitally converted by the A / D converter 3 in FIG. 1, and the power spectrum is calculated in step S2. In step S3, the maximum value of the power spectrum calculated above and its frequency point are detected. Further, in step S4, a power spectrum of a frequency point that is an integral multiple of the detected frequency point is detected.

Thereafter, in step S5, it is determined whether or not the frequency of the maximum value of the detected power spectrum is the howling frequency point by performing the above-described howling determination processing. Accordingly, if it is determined that the frequency of the maximum value of the detected power spectrum is not the howling frequency point, step S6.
And howling is not detected, and the step S
Returning to 1, the FFT processing is performed by the next frequency point.

If it is determined in step S5 that the frequency of the detected maximum value of the power spectrum is the frequency point of howling, the process proceeds to step S7 and the frequency of the output signal from the A / D converter 3 is reached. The counting process is performed a plurality of times, and in step S8, the values obtained by the frequency counting process are averaged to obtain the frequency.

Then, in step S9, the detected howling frequency point is compared with the frequency obtained in the process of step S8. By this comparison, the detected howling frequency points and the above-described step S8
If the difference from the frequency obtained by the process of step S1 is within the interval of the frequency points in the FFT process, step S
It is assumed that the howling frequency is detected in step 10.
Further, the difference between the detected howling frequency point and the frequency obtained in the process of step S8 is the FF.
If it is not within the interval of the frequency points in the T processing, the procedure goes to step S6 and it is considered that the howling frequency is not detected.

As described above, by determining the frequency point of the peak of the power spectrum of the picked-up signal and comparing the level of this frequency point with the level of another frequency point, whether the peak of the power spectrum is howling or not. The FFT processing is performed by determining whether or not, counting the digital signal output obtained by digitally converting the collected signal, and comparing the count output with the peak frequency of the power spectrum to determine the howling frequency. The howling frequency can be accurately measured without increasing the number of points.

FIG. 5 shows a flowchart of the second specific procedure of the howling frequency detecting method.

Here, steps S11 to S16
Up to step S in the first concrete procedure described above
The same operation as the operation from 1 to step S6 is performed.

If it is determined in step S15 that the frequency of the maximum value of the detected power spectrum is the howling frequency point, the process proceeds to step S17, in which the bandpass filter of the detected howling frequency point is set. The center frequency is moved, and the output signal from the A / D converter is passed through this bandpass filter. After that, in step S18, frequency count processing is performed a plurality of times using the output from the band pass filter,
In step S19, the values obtained by the frequency counting process are averaged to obtain the frequency.

Then, in step S20, the detected howling frequency point is compared with the frequency obtained in the process of step S19. By this comparison, if the difference between the detected howling frequency point and the frequency obtained in the process of step S19 is within the frequency point interval in the FFT process, the process proceeds to step S21, and the howling frequency is detected. Regard as If the difference between the detected howling frequency point and the frequency obtained in the process of step S19 is not within the frequency point interval in the FFT process, the process proceeds to step S16 and the howling frequency is not detected. I reckon.

As described above, the center frequency of the bandpass filter is moved to the detected howling frequency point, the digital signal output is passed through the bandpass filter, and then the frequency count processing is performed to obtain the howling frequency. Can be measured more accurately.

[0064]

As is apparent from the above description, the howling frequency detecting device according to the present invention includes a digital signal converting means for converting a picked-up signal into a digital signal, and an output from the digital signal converting means. A fast Fourier transform means for performing a fast Fourier transform process, a power spectrum calculation means for calculating a power spectrum using the spectrum from the fast Fourier transform means, and a frequency point of a peak of the power spectrum from the power spectrum calculation means. Obtaining frequency point detection means, the power spectrum peak by howling depending on the result of comparison of the frequency point level of the peak of the power spectrum from the frequency point detection means and the level of another frequency point or the presence or absence of harmonics. Howling to determine whether there is When the howling determining means determines that the peak of the power spectrum is howling, the frequency detecting means detects a frequency based on the output from the digital signal converting means, and the output from the frequency detecting means. And a comparing means for comparing the frequency of the peak of the power spectrum from the frequency point detecting means, and a howling frequency determining means for making the output from the frequency detecting means a howling frequency in accordance with the comparison result from the comparing means. By including howling by using different properties of transient music signals and stationary howling, the FFT
It is possible to easily detect howling and howling frequency without increasing the number of processing points.

Here, the frequency detecting means counts the outputs from the digital converting means a plurality of times to obtain an average, and the comparing means calculates the average output from the frequency detecting means and the power spectrum from the frequency point detecting means. Of the howling frequency is determined by comparing the frequency points of the peaks of the above, and determining whether the comparison result from the comparing means is within the frequency point interval in the fast Fourier transform means in the howling frequency determining means. It can be measured well.

Further, the center frequency of the bandpass filter is moved to the frequency point of the peak of the power spectrum discriminated as howling by the howling discriminating means, and the frequency detecting means outputs the output from the digital signal converting means as described above. By detecting the frequency after passing through the bandpass filter, the frequency count accuracy can be further increased, and thus a more accurate howling frequency can be obtained.

Further, the howling frequency detecting method according to the present invention comprises a digital signal converting step for converting a picked-up signal into a digital signal, and a fast Fourier transform for subjecting the output from the digital signal converting step to the fast Fourier transform processing. Step, a power spectrum calculation step of calculating a power spectrum using the spectrum from the fast Fourier transform step, a frequency point detection step of obtaining a frequency point of a peak of the power spectrum from the power spectrum calculation step, and the frequency point A howling determination step of determining whether or not the peak of the power spectrum is howling according to the result of comparison between the level of the frequency point of the peak of the power spectrum from the detection step and the level of another frequency point or the presence or absence of harmonics When,
When the howling determination step determines that the peak of the power spectrum is howling, a frequency detection step of detecting a frequency based on the output from the digital signal conversion step, an output from the frequency detection step, and the frequency point It comprises a comparison step of comparing the frequency of the peak of the power spectrum from the detection step, and a howling frequency determination step of making the output from the frequency detection step a howling frequency in accordance with the comparison result from the comparison step. As a result, howling is detected by using different properties between the transient music signal and the steady howling, and thus howling and howling frequency can be easily detected without increasing the number of points in the FFT processing.

Here, in the frequency detecting step, the outputs from the digital converting step are counted a plurality of times to obtain an average, and in the comparing step, the average output from the frequency detecting step and the power spectrum from the frequency point detecting step. The frequency point of the peak, and in the howling frequency determination step, by determining whether the comparison result from the comparison step is within the frequency point interval in the fast Fourier transform step, It can be measured well.

Further, the center frequency of the bandpass filter is moved to the frequency point of the peak of the power spectrum discriminated as howling in the howling discrimination step, and the output from the digital signal conversion step is converted into the frequency in the frequency detection step. By detecting the frequency after passing through the bandpass filter, the frequency count accuracy can be further increased, and thus a more accurate howling frequency can be obtained.

[Brief description of drawings]

FIG. 1 is a diagram showing a schematic configuration of a howling frequency detection device according to the present invention.

FIG. 2 is a diagram for explaining howling frequency point detection.

FIG. 3 is a diagram for explaining a frequency counting method.

FIG. 4 is a flowchart of a first specific procedure of the howling frequency detection method according to the present invention.

FIG. 5 is a flowchart of a second specific procedure of the howling frequency detection method according to the present invention.

FIG. 6 is a diagram showing a schematic configuration of a conventional howling frequency detection device.

FIG. 7 is a diagram showing a schematic configuration of a first howling detection means.

[Explanation of symbols]

 1 Microphone 2 Amplifier 3 A / D converter 4 Howling frequency detection circuit 5 Equalizer 6 Power amplifier 7 Speaker

Claims (6)

[Claims]
1. A howling frequency detecting device for detecting a howling frequency of a signal picked up by a measuring microphone, comprising: a digital signal converting means for converting the picked up signal into a digital signal; and the digital signal converting means. Fast Fourier transforming means for subjecting the output of the fast Fourier transforming means, a power spectrum calculating means for calculating a power spectrum using the spectrum from the fast Fourier transforming means, and a peak frequency of the power spectrum from the power spectrum calculating means. Frequency point detecting means for obtaining points, the peak of the power spectrum according to the result of comparison between the frequency point level of the peak of the power spectrum from the frequency point detecting means and the level of another frequency point or the presence or absence of harmonics Howling Howling determining means for determining whether or not, when the howling determining means determines that the peak of the power spectrum is howling, frequency detecting means for detecting a frequency based on the output from the digital signal converting means, Comparing means for comparing the output from the frequency detecting means and the frequency of the peak of the power spectrum from the frequency point detecting means, and the output from the frequency detecting means in accordance with the comparison result from the comparing means is a howling frequency. And a howling frequency determining means.
2. The frequency detecting means counts the outputs from the digital converting means a plurality of times to obtain an average,
The comparison means compares the average output from the frequency detection means with the frequency point of the peak of the power spectrum from the frequency point detection means, and the howling frequency determination means compares the comparison result from the comparison means with the fast Fourier transform. The howling frequency detection device according to claim 1, wherein it is determined whether or not the frequency point is within the interval of the frequency points in the means.
3. The center frequency of the bandpass filter is moved to the frequency point of the peak of the power spectrum determined to be howling by the howling determination means,
2. The howling frequency detecting device according to claim 1, wherein the frequency detecting means detects the frequency after passing the output from the digital signal converting means through the band pass filter.
4. A howling frequency detecting method for detecting a howling frequency of a signal picked up by a measuring microphone, comprising: a digital signal converting step for converting the picked up signal into a digital signal; and the digital signal converting step. Fast Fourier transform process for applying a fast Fourier transform process to the output of, a power spectrum calculation process for calculating a power spectrum using the spectrum from the fast Fourier transform process, and a frequency of the peak of the power spectrum from the power spectrum calculation process. The frequency point detection step to obtain points, the peak of the power spectrum depending on the result of comparison of the frequency point level of the peak of the power spectrum from the frequency point detection step and the level of other frequency points or the presence or absence of harmonics Howling A howling determining step of determining whether or not, when the howling determining step determines that the peak of the power spectrum is howling, a frequency detecting step of detecting a frequency based on the output from the digital signal converting step, A comparison step of comparing the output from the frequency detection step and the frequency of the peak of the power spectrum from the frequency point detection step, and the output from the frequency detection step according to the comparison result from the comparison step and a howling frequency. And a howling frequency determining step.
5. The frequency detecting step counts the outputs from the digital converting step a plurality of times to obtain an average,
In the comparison step, the average output from the frequency detection step and the frequency point of the peak of the power spectrum from the frequency point detection step are compared, and in the howling frequency determination step, the comparison result from the comparison step is the fast Fourier transform. The howling frequency detection method according to claim 4, wherein it is determined whether or not it is within an interval of frequency points in the process.
6. The center frequency of the bandpass filter is moved to the frequency point of the peak of the power spectrum determined to be howling in the howling determination step,
5. The howling frequency detecting method according to claim 4, wherein in the frequency detecting step, the frequency is detected after the output from the digital signal converting step is passed through the band pass filter.
JP624795A 1995-01-19 1995-01-19 Apparatus and method for detection of howling frequency Withdrawn JPH08193876A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP624795A JPH08193876A (en) 1995-01-19 1995-01-19 Apparatus and method for detection of howling frequency

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP624795A JPH08193876A (en) 1995-01-19 1995-01-19 Apparatus and method for detection of howling frequency

Publications (1)

Publication Number Publication Date
JPH08193876A true JPH08193876A (en) 1996-07-30

Family

ID=11633174

Family Applications (1)

Application Number Title Priority Date Filing Date
JP624795A Withdrawn JPH08193876A (en) 1995-01-19 1995-01-19 Apparatus and method for detection of howling frequency

Country Status (1)

Country Link
JP (1) JPH08193876A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8428274B2 (en) 2008-07-01 2013-04-23 Sony Corporation Apparatus and method for detecting acoustic feedback

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8428274B2 (en) 2008-07-01 2013-04-23 Sony Corporation Apparatus and method for detecting acoustic feedback

Similar Documents

Publication Publication Date Title
CA2773294C (en) Sound detection and localization system
US7171357B2 (en) Voice-activity detection using energy ratios and periodicity
JP3963850B2 (en) Voice segment detection device
JP5507596B2 (en) Speech enhancement
JP4177755B2 (en) Utterance feature extraction system
JP3156975B2 (en) Audio / music discriminator for audio band signals
KR100873396B1 (en) Comparing audio using characterizations based on auditory events
Duxbury et al. A hybrid approach to musical note onset detection
ES2452557T3 (en) Spatial audio processor and method to provide spatial parameters based on an acoustic input signal
US5567162A (en) Karaoke system capable of scoring singing of a singer on accompaniment thereof
US6564181B2 (en) Method and system for measurement of speech distortion from samples of telephonic voice signals
DE60024963T2 (en) Method and device for band expansion of an audio signal
JP4845441B2 (en) Real-time power mask trigger generator
JP3235925B2 (en) Howling suppression device
JP4952698B2 (en) Audio processing apparatus, audio processing method and program
US5780759A (en) Method for pitch recognition, in particular for musical instruments which are excited by plucking or striking
US6950511B2 (en) Detection of both voice and tones using Goertzel filters
US4932063A (en) Noise suppression apparatus
US20120136655A1 (en) Speech processing apparatus and speech processing method
EP0459364B1 (en) Noise signal prediction system
US7190800B2 (en) Howling control device and howling control method
US8989403B2 (en) Noise suppression device
US20040267147A1 (en) Determining heart rate
JP6028502B2 (en) Audio signal processing apparatus, method and program
KR100580643B1 (en) Appratuses and methods for detecting and discriminating acoustical impact

Legal Events

Date Code Title Description
A300 Withdrawal of application because of no request for examination

Free format text: JAPANESE INTERMEDIATE CODE: A300

Effective date: 20020402