JPS6387027A - Howling suppressing device - Google Patents

Abstract

PURPOSE:To always ensure the stable howling suppressing effect by estimating the impulse response between a speaker and a microphone every time an input signal is cut off to the microphone. CONSTITUTION:A timing signal is applied to an impulse response estimating means 2 from a power calculating means 3 every time an input signal is cut off to a microphone 6. Then the means 2 estimates the impulse response between a speaker 5 and the microphone 6 based on the output of a white color noise generating means 1 and the output of the microphone 6 and by means of a learning type identification method. These actions are repeated several times to obtain estimated impulse responses. these responses are averaged to decide an impulse response of that time point. Then a convolutional operation is carried out between said impulse response and the output of the means 1 for artificial production of a feedback signal 9. This signal is subtracted from the output of the microphone 6 and outputted.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a howling suppression device capable of suppressing howling occurring in a public address system including a microphone, an amplifier, and a speaker.

(Prior Art) FIG. 2 shows the configuration of a conventional howling suppression device. In FIG. 2, 6 is a microphone, 4 is an amplifier, 5 is a speaker, 7 is a voice from a speaker, 9 is a feedback signal from the speaker to the microphone, 2 is an impulse response estimation circuit, 8 is a white noise generation circuit, 11 is a changeover switch.

Next, the operation of the above conventional example will be explained. In Figure 2, before starting to use this loudspeaker, first
The position of the micro is fixed, the selector switch 11 is turned to the white noise generating circuit 8 side, white noise is outputted from the speaker 5 through the amplifier 4, and the signal is received by the micro. At this time, the micro must prevent other sounds from entering.

Next, using the output of the white noise generation circuit 8 and the output of the microphone 7, the impulse response estimation circuit 2 estimates the impulse response between the speaker 5 and the microphone 7 using a learning identification method. When the loudspeaker is used as a public address system, the selector switch 11 is set to the output side of the impulse response estimation circuit 2, and the impulse response estimated in the preparation is convoluted with the amplifier input, and the output from the loudspeaker is Howling was suppressed by generating a pseudo-return signal to the microphone and subtracting it from the microphone output.

(Problem to be Solved by the Invention) However, in the conventional howling suppression device described above, in order to estimate the impulse response between the speaker and the microphone, white noise is manually output from the speaker before starting to use the loudspeaker. In addition, preliminary preparation is required, and since the impulse response is estimated only by -degrees, there is a problem in that it is not possible to follow fluctuations in the impulse response due to physical changes around the loudspeaker.

The present invention solves such conventional problems, and aims to provide an excellent howling suppression device that can obtain a highly accurate impulse response by averaging several impulse responses. That is.

SUMMARY OF THE INVENTION An object of the present invention is to provide an excellent howling suppression device that can estimate the impulse response at any time when the audio signal is interrupted and can follow fluctuations in the impulse response due to physical changes in the loudspeaker system. It is something to do.

Another object of the present invention is to provide an excellent howling suppression device that does not require preparation for impulse response estimation because impulse response estimation is performed at any time.

Furthermore, it is an object of the present invention to provide an excellent howling suppression device that eliminates the need for mechanical switches used in preparation for impulse response estimation, thereby eliminating failures due to contact failure. It is.

(Means for Solving the Problems) In order to solve the above-mentioned problems, the present invention provides a white noise noise generator comprising a circuit that sends out an output at a level lower than the average power of the microphone output using the output signal from the power calculation means. Generation means; an arithmetic circuit for estimating the impulse response between the speaker and the microphone using a learning identification method from the output of the white noise generation means and the audio output according to the output signal from the power calculation means; a feedback signal pseudo generation circuit that obtains a feedback signal by convolving the average and the output of the white noise generation means; and a circuit that sends out an output signal obtained by subtracting the pseudo feedback signal generated by the circuit from the microphone output. impulse response estimation means; power calculation means comprising a circuit for detecting the point in time when the power of the microphone output becomes almost zero; and a circuit for sending a start signal to the white noise generation means and the impulse response estimation means at each time point; ;

(Function) Therefore, according to the present invention, the power calculation circuit detects the point in time when the power of the microphone output becomes almost zero.

Each time, it gives the start timing of output to the white noise generation circuit and the start timing of estimating the impulse response between the speaker and the microphone using the learning identification method to the impulse response estimation circuit.
The impulse response estimation circuit has a white noise generation circuit output,
Using the microphone output when the noise is output from the speaker through the amplifier and returned to the microphone, the impulse response between the speaker and the microphone is estimated, the average of several times is taken as the impulse response at that point, and it is Howling can be suppressed by generating a pseudo feedback signal from the speaker to the microphone and subtracting it from the microphone output.

(Example) FIG. 1 shows the configuration of an embodiment of the present invention.

In FIG. 1, 6 is the microphone, 4 is the amplifier, 5 is the speaker, 7 is the voice from the speaker, 9 is the feedback signal from the speaker to the microphone, 1 is the white noise generation stage, and the power calculation means 3
Based on the timing given by and the average power of the microphone output, the output of white noise whose level is lower than the average power is started. Reference numeral 2 denotes an impulse response estimating means, which estimates the impulse response between the speaker 5 and the micro using a learning identification method from the output of the white noise generating means 1 and the output of the micro when the timing is given from the power calculating means 3. The impulse response estimated by repeating 0 or more several times is averaged to obtain the impulse response at that point in time, and after that, a convolution operation is performed with the output of the white noise generating means 1 to generate the feedback signal 9. Generate a pseudo-generate, subtract it from the micro output, and use it as the output. Power calculation means 3
constantly calculates the power of the output of the micro, detects the point in time when the power becomes almost zero, and outputs a white noise output start signal and an impulse response estimation start signal 10 to the white noise generation means 1 and the impulse response estimation means 2.

According to the above embodiment, each time the input signal to the microphone is interrupted, the obtained impulse response between the speaker and the microphone is averaged several times, and this is used as the impulse response at that time. An impulse response can be obtained, and a stable howling suppression effect can always be obtained.

It goes without saying that the impulse response estimating means includes an A/D and D/A converter, and can be constructed from a known circuit including these converters. Further, the circuits constituting the white noise generation means and the power calculation means are not special circuits, but are made up of a combination of well-known circuits.

(Effects of the Invention) As is clear from the above examples, the present invention has the following effects.

(1) Impulse response estimation is automatically performed at any time by using the time when the speaker's voice is interrupted, so no special manual operation is required in advance.

(2) Since the impulse response is re-estimated using the time when the speaker's voice is interrupted, it has a stable howling suppression effect that is not affected by changes in the surrounding physical environment.

(3) Since the impulse response is averaged several times, a highly accurate impulse response can be obtained.

(4) A mechanical switch used for preparation for an impulse response is no longer necessary, and the invention can be constructed using only a combination of well-known analog or digital circuits, making it possible to provide an inexpensive howling suppression device that is free from failures.

[Brief explanation of the drawing]

FIG. 1 is a schematic block diagram of a howling suppression device according to an embodiment of the present invention, and FIG. 2 is a schematic block diagram of a conventional howling suppression device. DESCRIPTION OF SYMBOLS 1... White noise generation circuit (means), 2... Impulse response estimation circuit (means), 3... Power calculation means, 4... Amplifier, 5... Speaker, 6... Microphone, 7...Voice from the speaker, 9...Return signal from the speaker to the microphone. Patent applicant Matsushita Electric Industrial Co., Ltd. Figure 1 Figure 2, 8

Claims (1)

[Scope of Claims] White noise generation means comprising a circuit that transmits an output at a level lower than the average power of the microphone output according to the output signal from the power calculation means; and White noise generation means according to the output signal from the power calculation means. an arithmetic circuit that estimates the impulse response between the speaker and the microphone using a learning identification method from the output of Impulse response estimating means includes a feedback signal pseudo generation circuit that obtains a feedback signal, and a circuit that sends an output signal obtained by subtracting the pseudo feedback signal generated by the circuit from the microphone output, and the power of the microphone output becomes almost zero. A howling suppression device comprising: a circuit for detecting a time point; and a power calculation means comprising a circuit for sending a start signal to a white noise generating means and an impulse response estimating means at each time point.