JPH07154467A - Sound equipment - Google Patents

Abstract

PURPOSE:To prevent howling by providing an analog/digital conversion circuit and a digital signal processor which removes a resonance frequency component from a digital signal. CONSTITUTION:A signal in which a sound fed back from a speaker 8 and a resonance frequency 13 are mixed is inputted to a microphone 9. In such a case, the signal including the resonance frequency 13 is converted into a digital signal by a second analog/digital conversion circuit 3, and the resonance frequency is detected by checking the character proper to the resonance by the resonance frequency detection circuit 11 of the digital signal processor 6, and frequency information is sent to a resonance frequency elimination circuit 12. The resonance frequency elimination circuit 12 eliminates a level of resonance frequency included in the digital signal. In such a way, a signal from which the resonance frequency is eliminated is converted into an analog signal (signal C) by a second digital/analog conversion circuit 5. The signal is converted into the one of two-wire system by a hybrid circuit 1, then, it is sent out to a line.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an acoustic device having a howling prevention function.

[0002]

2. Description of the Related Art With the increase in functionality of telephone devices, a function (hereinafter referred to as "speakerphone") for making a call from the vicinity of a handset without a handset is being developed. The speakerphone has the advantage that both hands can realize free calls and conference calls. However, the speakerphone has a larger output of the speaker and amplification of the microphone than the case of using a handset, so one loop can be created due to acoustic coupling between the speaker and the microphone and wraparound in the side noise prevention circuit to cause howling It may happen.

Conventionally, there has been known an echo suppressor which also prevents the howling from preventing the echo. Echo suppressors are basically voice operated switches,
When one is busy, an attenuation amount of several tens of dB or more is added to the path where the echo of the 4-wire line returns. However, when both speak at the same time, the echo suppressor sometimes switches the head so that the speaker with the stronger signal hears the voice of the stronger echo suppressor, and sometimes the echo suppressor cuts off the voice or, in the case of simultaneous speakers, the voice There is a problem of separating one.

Such a problem can be overcome by an echo canceller. The echo canceller calculates an estimated amount of echo by using a signal correlation technique, generates a pseudo signal, and removes the pseudo signal from the signal on the return path, and is disclosed in Japanese Patent Application No. 4293411 of the same applicant as the present application. Has been done.

[0005]

The occurrence of howling is not limited to the sound directly returning from the speaker to the microphone. When a speaker emits sound, members such as a structure and a cabinet holding the sound vibrate and resonate when the frequency condition is met. If this resonance frequency is high, it may be transmitted to the microphone, and may loop once through the sound-proof sound circuit, causing howling. It is difficult to estimate and remove such howling by the echo canceller.

[0006]

SUMMARY OF THE INVENTION The present invention has been made in view of the above points, and is characterized in that in an audio device having a speaker and a microphone, an analog / digital converter for converting a signal coming from the microphone into a digital signal. A digital conversion circuit and a digital signal processor for detecting a resonance frequency of resonance generated by the reproduced sound of the speaker from the digital signal and removing the resonance frequency component from the digital signal.

[0007]

In general, the resonance frequency is a single tone, and the frequency component of this tone is higher than other frequency components.
Further on. Focusing on this point, only the frequency component that matches this condition is detected, and since this is removed, howling can be prevented without affecting the sound.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention is applied to a telephone and will be described with reference to FIG. The present invention has an echo canceller,
The echo canceller is omitted from FIG. 1 in order to explain the essential points of the present invention. Hybrid circuit 1, first analog / digital conversion circuit 2, second analog / digital conversion circuit 3, first digital / analog circuit 4, second digital / analog circuit 5, digital signal processor 6, reception amplification circuit The digital signal processor 6 has at least a resonance frequency detecting circuit 11 and a resonance frequency removing circuit 12. The hybrid circuit 1 performs 2-line to 4-line conversion, the analog / digital conversion circuits 2 and 3 convert analog signals into digital signals, and the digital / analog conversion circuits 4 and 5 convert digital signals into analog signals. Is to be converted to. The digital signal processor 6 controls a transmission signal and a reception signal by a digital signal, a reception amplification circuit 7 amplifies the reception signal in order to emit a voice from the speaker 8, and the transmission amplification circuit 10 is a microphone. The signal from 9 is amplified.

The digital signal processor 6 will be described in detail. Digital signal processor 6 is disclosed in JP-A-6
As disclosed in Japanese Patent Laid-Open No. 3-5607, an audio signal can be divided into levels for each frequency band, and can be controlled for each frequency band by a program of the digital signal processor 6. For the purpose of explanation, the present invention refers to a resonance frequency detecting circuit 11 that determines a frequency having a monotonous character, which is a characteristic of the resonance frequency, and a level higher than a predetermined level and continuity higher than other frequency components from the control functions. A circuit for removing a resonance frequency component by generating a pseudo frequency having a level opposite to the resonance frequency received from the detection circuit 11 and adding the pseudo frequency to the input digital signal is called a resonance frequency removal circuit 12.

The operation when the caller's voice is received from the line will be described. The hybrid circuit 1 converts the reception signal sent from the line into four lines, and the first analog /
It is converted to digital by the digital conversion circuit 2. It is converted to the original analog by the first digital / analog circuit 4 via the digital signal processor 6 (signal A). The reception signal amplified by the reception amplification circuit 7 is converted into voice by the speaker 8.

Resonance occurs when the natural frequency of a member (not shown) such as a cabinet holding the speaker 8 and the frequency of the sound from the speaker 8 meet a specific condition.
The microphone 9 receives a signal in which the sound returned from the speaker 8 and the resonance frequency are mixed. A signal accompanied by this frequency signal is amplified by the transmission amplifier circuit 10 (signal B).
This resonance frequency (13) must be removed because it may loop through the hybrid circuit 1 to cause howling. A signal including the resonance frequency (13) is converted into a digital signal by the second analog / digital conversion circuit 3. The resonance frequency detection circuit 11 detects the resonance frequency by checking the characteristic peculiar to the resonance, and sends the frequency information to the resonance frequency removal circuit 12.
The resonance frequency removing circuit 12 removes the level of the resonance frequency included in the digital signal. The signal from which the resonance frequency has been removed is converted into analog by the second digital / analog circuit 5 (signal C). Then, it is converted into two lines by the hybrid circuit 1 and sent to the line.

In many cases, the resonance frequency does not usually extend to one point but several points, and it is possible to set in advance the frequency components to be removed in order of increasing level.

Even in a speech circuit, howling similar to the resonance frequency of the speaker and the microphone may occur due to the relationship between the line impedance and the sidetone circuit. The present invention can also be applied to prevent this howling.

The present invention can be applied not only to the telephone device but also to an acoustic device having a speaker and a microphone.

[0015]

It is difficult for the echo canceller to sufficiently prevent the howling caused by the resonance of the members such as the cabinet due to the sound emitted from the speaker. This resonance frequency is a single tone, has a higher level than other frequency components, and continues. Howling is prevented by detecting a frequency component that meets this condition and removing the resonance frequency thereof. Also, in order to detect only the resonance frequency,
Howling can be prevented without affecting the sound.

[Brief description of drawings]

FIG. 1 is a block diagram in which an embodiment of the present invention is applied to a telephone.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Hybrid circuit 2 1st analog / digital conversion circuit 3 2nd analog / digital conversion circuit 4 1st digital / analog circuit 5 2nd digital / analog circuit 6 Digital signal processor 7 Receiving / amplifying circuit 8 Speaker 9 Microphone 10 Transmission amplifier circuit 11 Resonance frequency detection circuit 12 Resonance frequency removal circuit

Claims (1)

[Claims] 1. In an acoustic device having a speaker and a microphone, an analog / digital conversion circuit for converting a signal coming from the microphone into a digital signal, and a resonance resonance caused by a reproduced sound of the speaker from the digital signal. A digital signal processor that detects a frequency and removes the resonance frequency component from the digital signal.