JPH01149508A - Acoustic signal processing circuit - Google Patents

Abstract

PURPOSE:To satisfactorily suppress an unnecessary signal with a simple constitution by detecting the time change component of an input acoustic signal and controlling the input acoustic signal when such a detecting value goes to an arbitrary comparing level or above. CONSTITUTION:An acoustic signal from an input terminal 1 is extracted through a delaying circuit 2 and a multiplier 3 to an output terminal 4. The acoustic signal from the terminal 1 is supplied to an envelope detecting circuit 5, such a detecting signal is supplied to a difference circuit 6, and the time change component is obtained. The time change component is supplied to a comparator circuit 7, and it is compared with the arbitrary comparing level from a terminal 8. Thus, when the time change component is larger than the comparing level, an impulse is generated from the comparator circuit 7. The impulse is supplied to a time constant circuit 9, a waveform is smoothed, the smoothed signal is polarity-inverted and offset by a converting circuit 10, and it is supplied to the multiplier 3. Thus, the level of the input signal is lowered, and the external noise can be suppressed.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an acoustic signal processing circuit used in hearing aids and the like.

[Invention (already required)]

The present invention relates to an acoustic signal processing circuit that controls an input signal when the time-varying component of the input signal exceeds a given comparison level, thereby making it possible to effectively suppress unnecessary signals with a simple configuration. It is something.

[Conventional technology]

For example, for people who use hearing aids, external noise mixed into conversational sounds, especially transient, high-level signals such as the sound of doors opening and closing or the sound of dishes clinking, can be interpreted as conversational audio. Similarly, when amplified, it becomes extremely unpleasant and becomes an urgent problem for practical use.

Therefore, various studies are being conducted to reduce the sensory burden on the user by suppressing such extraneous noise mixed into conversational sounds.

That is, for example, by focusing on the fact that the energy of the above-mentioned external noise is generally concentrated in the low frequency range, this low frequency component is monitored, and when this component increases, for example, a bypass filter installed in the audio signal path is cut. It is conceivable to perform control so that the off frequency becomes high. However, this method requires a relatively long attack time for control in order to detect low-frequency signals, which causes a delay in operation, so it is not sufficient to deal with transient noise that includes high-frequency components. There was a possibility that it would not be possible.

[Problem that the invention seeks to solve]

As described above, the conventional techniques have problems such as the inability to satisfactorily suppress transient external noise mixed into conversational sounds.

[Means for solving problems]

The present invention detects a time-varying component of an input acoustic signal (input terminal (1)) (start/end detection circuit (5), difference circuit (6)), and when this detected value becomes larger than an arbitrary comparison level ( This is a sound mouth signal processing circuit configured to control the input acoustic signal (multiplier (3)) in the comparison circuit (7).

[Effect]

According to this, by detecting the time-varying component of the input signal, transient external noise can be determined extremely accurately, and unpleasant external noise can thereby be suppressed extremely well.

[Example] In Fig. 1, (1) is an input terminal to which a sampled and digitized audio signal is supplied, and the audio signal supplied to this input terminal (1) is transmitted through a delay port M (2).
, is taken out to the output terminal (4) through the multiplier (3).

In addition, the acoustic signal from the input terminal (1) is detected by the start/end detection circuit (5).
), and this detection signal is supplied to a difference circuit (6) to obtain a time-varying component.

Here, the start/end detection circuit (5) is configured as shown in FIG. 2, for example. In the figure, the signal from the input terminal 51) is a subtracter (52), a multiplier (53), an adder (
54) to the output terminal (55). The signal taken out to this output terminal (55) is supplied to the subtracter (52) through a delay circuit (56) of one sample period, and the signal from the delay circuit (56) is supplied to the multiplier (56).
57) to the adder (54). Furthermore, the signal from the subtracter (52) is supplied to a mode detection circuit (58), and the coefficients according to the detected mode are supplied to a multiplier (53). ``A predetermined coefficient from the terminal (59) is also supplied to the multiplier (57).

Accordingly, in the above circuit, a positive signal is output from the subtracter (52) in the so-called attack mode, and a negative signal is output in the recovery mode. Therefore, the detection circuit (58) determines whether the signal from the subtracter (52) is positive or negative, and when the signal is 0 or positive, the coefficient supplied to the multiplier (53) is, for example, 0.9998, and when it is negative, it is 0. shall be. Further, the coefficient supplied to the multiplier (57) is, for example, 0.002.

As a result, in the circuit described above, the attack time is approximately 1 when the zumbling frequency is 48kllz.
m5ec, the recovery time becomes about 5Qmsec, and the start and end signals of the signal supplied to the input terminal (51) are taken out to the output terminal (55).

By finding the difference between the beginning and end signals,
It is possible to extract time-varying components of the input signal.

This time-varying component is supplied to the comparator circuit (7), and the terminal (
8) is compared with any comparison level from 8). As a result, an impulse is generated from the comparator circuit (7) when the time-varying component is greater than the comparison level.

This impulse is supplied to the time constant v80) to smooth the waveform, and this smoothed signal is converted to the conversion port i? 8(1
0) and the polarity is reversed and offset by the multiplier (
3).

Therefore, in this circuit, the level of the above-mentioned time-varying component becomes large for transient external Th1E sounds, etc., so an impulse is generated at this time, and the polarity of this impulse is reversed and the °C multiplier (3 ), the level of the input signal is lowered and external noise can be suppressed.

In this way, according to this circuit, by detecting the time-varying component of the input signal, transient external noise can be determined extremely accurately, and unpleasant external noise can thereby be suppressed extremely well. .

Note that in the above-described circuit, good signal suppression can be achieved by smoothing the waveform in the time constant circuit (11). Furthermore, by inverting and offsetting the polarity of the °C signal at the conversion port 1 (12), it is possible to suppress the signal only when an impulse occurs. Further, a delay circuit (2) is provided for the purpose of correcting the delay of the control signal.

In addition, in the above circuit, the start/end detection circuit (5) described above requires a fairly complicated configuration, but this is done by accumulating the data every 1Q msec, for example, as shown in FIG. It is also possible to provide a circuit (31) for converting the blocks, supply this blocked signal to a maximum absolute value detection circuit (32), and use the calculated maximum absolute value difference between the blocks. According to this, substantially the same characteristics can be obtained with a simple configuration.

Furthermore, in this case, an average value between blocks or an effective value can be used instead of the maximum absolute value.

Furthermore, in the circuit described above, when the time increase rate of the start/end signal exceeds a certain value and then the time decrease rate exceeds a certain value, a control signal is obtained and the input signal is suppressed, but the same effect can be obtained. can be obtained.

Note that the above-described circuit can be applied not only to hearing aids but also to, for example, an input circuit for extracting only human speech in a speech recognition device.

〔Effect of the invention〕

According to this circuit, by detecting the time-varying component of the human input signal, transient external tIL sounds can be discriminated extremely accurately, and unpleasant external noise can thereby be suppressed extremely well. It became so.

[Brief explanation of the drawing]

FIG. 1 is a configuration diagram of an example of the present invention, and FIGS. 2 and 3 are diagrams for explaining the same. (1) is an input terminal, (2) is a delay circuit, (3) is a multiplier, (4) is an output terminal, (5) is a start/end detection circuit, (
6) is a differential circuit, (7) is a comparison circuit, (8) is a terminal, (
9) is a time constant circuit, and (10) is a conversion circuit.

Claims (1)

[Scope of Claims] An acoustic signal processing circuit that detects a time-varying component of an input acoustic signal, and controls the input acoustic signal when the detected value becomes larger than an arbitrary comparison level.