TWI715208B - Weighted hybrid type anc system and controller - Google Patents

Abstract

The present invention provides a weighted hybrid type ANC system comprising a feedforward microphone, a feedback microphone, a speaker, and a controller coupled to the feedforward microphone, the feedback microphone, and the speaker. The controller includes a feedforward noise reduction filter module, a feedback noise reduction filter module, a mixer, a noise detection module, and a weight distribution module. The weight distribution module includes one or more weight modulation units disposed between the feedback microphone and the feedback noise reduction filter module, wherein the noise detection module is detected by the feedforward microphone or the feedback microphone The control signal is transmitted to the weight distribution module when the high frequency noise is transmitted, and the weight distribution module reduces the weight value of the weight modulation unit to reduce the proportion of the feedback signal of the feedback noise reduction filter module.

Description

Weighted mixed type active anti-noise system and controller

The present invention relates to an active anti-noise system and controller, and particularly relates to a weighted hybrid active anti-noise system and controller.

The active noise reduction (ANC) technology currently used in headphones has two modes, namely feed-forward noise reduction and feedback noise reduction, which are composed of both feed-forward noise reduction and feedback noise reduction. The common combination is called Hybrid (Hybrid) noise reduction. Different active noise reduction technologies have their own limitations in noise reduction depth and bandwidth, which are mainly determined by the acoustic structure of the headset, signal processing, and system signal delay.

The working principle of the feedforward noise reduction system is mainly to output a signal with the same frequency response as the environmental noise but opposite phase to achieve noise reduction. The reference microphone detects the noise and generates an inverted signal through the filter circuit. The inverted signal and the noise signal are cancelled at the ear drum, thereby reducing the noise level heard by the human ear. The filter circuit here is mainly used to compensate for the difference between the noise detected at the ear drum and the microphone, and it also compensates for the response ability of the speaker itself in the noise reduction signal.

The working principle of the feedback noise reduction system is mainly to detect the noise in the ear drum area, and then form a basic feedback loop to minimize the noise level in this area. The entire loop is composed of the response of the speaker and the microphone and the filter. As the filter gain (and its loop gain) increases, the noise residue becomes smaller, and the noise reduction performance is improved. But if the phase of the loop is close to ±180°, the "loop" signal will be reversed, and the ‘+’ on the denominator will become ‘-’. In this case, the loop gain size adjustment is limited, because when it increases from 0.0 to 1.0, the result is amplification, and when it is equal to 1.0, the result is "Divide by zero" easily causes system instability and often causes howling as the frequency response increases.

The Hybrid ANC (Hybrid ANC) combines the feedforward noise reduction system and the feedback noise reduction system to effectively improve the individual deficiencies of the two. A hybrid active noise reduction system usually includes a pair of microphones. The feedforward noise reduction system uses an external microphone to measure environmental noise before entering the ear, processes the signal to ensure accurate reverse signals, and the system’s speakers effectively eliminate the environment noise. The filter of the feedback anti-noise system is used to collect the acoustic signal near the error microphone and feedback it for error correction. However, in the traditional hybrid active anti-noise architecture, when the noise is at a higher frequency and regularity, the feedback noise cancellation system is prone to unstable conditions, resulting in poor noise cancellation effects.

The object of the present invention is to provide a weighted hybrid active noise reduction system, which includes a reference microphone, an error microphone, a speaker, and a controller connected to the reference microphone, the error microphone, and the speaker. The controller includes a feedforward noise reduction filter module, a feedback noise reduction filter module, a mixer, a noise detection module, and a weight distribution module. The feedforward noise reduction filter module The reference signal received by the reference microphone obtains a feedforward noise reduction signal after feedforward noise reduction, and the feedback noise reduction filter module obtains a feedback noise reduction signal after the error audio source signal received by the error microphone is feedback noise reduction , And send the feedforward noise reduction signal and the feedback noise reduction signal to the mixer for mixing, and output the mixed noise reduction signal to the speaker. The weight distribution module includes one or more settings The weight modulation unit between the error microphone and the feedback noise reduction filter module transmits a control signal to the weight distribution when the noise detection module detects high frequency noise by the reference microphone or the error microphone Module, the weight distribution module reduces the weight value of the weight modulation unit to reduce the proportion of the feedback signal of the feedback noise reduction filter module.

Another object of the present invention is to provide a controller including a feedforward noise reduction filter module, a feedback noise reduction filter module, a mixer, a noise detection module, and a weight distribution module. The feedforward noise reduction filter module obtains the reference signal received by a reference microphone through feedforward noise reduction A feedforward noise reduction signal. The feedback noise reduction filter module obtains a feedback noise reduction signal after the error audio source signal received by an error microphone undergoes feedback noise reduction. The mixer mixes the feedforward noise reduction signal and the feedback noise reduction signal and outputs a noise reduction signal. The noise detection module detects high frequency noise from the input of the feedforward noise reduction filter module or the feedback noise reduction filter module, and outputs a control signal when high frequency noise is detected. The weight distribution module includes one or more weight modulation units disposed between the error microphone and the feedback noise reduction filter module. When the control signal is received, the weight value of the weight modulation unit is reduced to Reduce the proportion of the feedback signal of the feedback noise reduction filter module.

Therefore, the present invention has the following advantages compared with the conventional technology:

1. The present invention adds a weight adjustment function to the feedback path, and when the performance deteriorates, the weight can be appropriately reduced on the feedback path, and the feedforward path is led to achieve a true adaptive effect.

2. The present invention can strengthen the processing effect corresponding to irregular noise and regular noise, so as to improve the traditional hybrid active anti-noise architecture that is prone to instability when processing high-frequency regular noise.

100‧‧‧Weighted Hybrid Active Noise Control System

10‧‧‧Reference microphone

20‧‧‧Error microphone

30‧‧‧Speaker

31‧‧‧Speaker unit

32‧‧‧Drive unit

40‧‧‧Controller

F1‧‧‧Feedforward noise reduction filter module

F11‧‧‧Secondary Path Filter

F12‧‧‧Adaptive arithmetic unit

F13‧‧‧Digital Filter

F2‧‧‧Feedback noise reduction filter module

F21‧‧‧Secondary Path Filter

F22‧‧‧Adaptive arithmetic unit

F23‧‧‧Digital Filter

F3‧‧‧Mixer

F4‧‧‧Noise detection module

F5‧‧‧Weight Distribution Module

F5A‧‧‧Weight Modulation Unit

F51‧‧‧The first complementary weight modulator

F52‧‧‧Second Complementary Weight Modulator

F53‧‧‧First weight modulator

F54‧‧‧Second weight modulator

F6‧‧‧Mixer

F7‧‧‧Secondary Path Filter

S01-S06‧‧‧Step

Figure 1 is a block diagram (1) of the weighted hybrid active anti-noise system of the present invention.

Figure 2 is a block diagram (2) of the weighted hybrid active anti-noise system of the present invention.

Fig. 3 is a schematic diagram (1) of the control flow of the weighted hybrid active anti-noise system of the present invention.

Figure 4 is a schematic diagram (2) of the control flow of the weighted hybrid active anti-noise system of the present invention.

The detailed description and technical content of the present invention will now be described with the drawings as follows. Furthermore, for the convenience of description, the figures in the present invention are not necessarily drawn according to actual proportions. These figures and their proportions are not intended to limit the scope of the present invention, and are described here first.

The embodiments of the present invention can be implemented in a noise reduction device or a noise reduction controller in a personal listening system including wired headsets, smart phones, wireless earphones or other head-mounted audio devices. Be restricted. The controller described in the present invention may be composed of a single chip or a plurality of chips. In another embodiment, the controller may be a chip provided by an audio device (such as a mobile device), or be integrated or separated from a wireless Earphones and headsets are composed of audio chips, and these implementations are not within the scope of the present invention. Specifically, the controller may be, for example, a microprocessor (Microprocessor), a digital signal processor (DSP), or other similar devices or a combination of these devices, which is not limited in the present invention.

The following describes a preferred embodiment of the present invention, please refer to "Figure 1" and "Figure 2" together, which are the block diagram (1) and block diagram of the weighted hybrid active noise reduction system of the present invention (2) As shown in the figure: This embodiment mode discloses a weighted hybrid active noise reduction system 100, which mainly includes a reference microphone 10, an error microphone 20, a speaker 30, and a reference microphone connected to the reference microphone. 10. The error microphone 20 and the controller 40 of the speaker 30.

The reference microphone 10 (Feedfoward Microphone) is mainly used to receive a reference audio signal. The reference audio source signal is mainly environmental noise, which is the noise mainly filtered by the anti-noise system. In one embodiment, the reference microphone 10 can be a microphone, a pickup, or other devices that can receive ambient sound waves and further convert them into analog and digital audio.

The error microphone 20 (Feedback Microphone) is mainly used to receive error audio signals. The error microphone 20 is generally set at a reference position within the anti-noise area. The audio signal received by the error microphone 20 is equivalent to the difference between the reference audio signal and the reverse signal output by the speaker 30. The difference is defined here It is the error audio signal. Similar to the reference microphone 10, the error microphone 20 in an embodiment can be, for example, a microphone, a pickup, or other types that can be used to receive the environment A device that converts sound waves into analog and digital audio.

The speaker 30 is mainly used to output a reverse signal to cancel the noise in the environment. In an embodiment, the speaker 30 may be, for example, a headset, a speaker, or other similar devices for outputting reverse signals to cancel noise. The speaker 30 mainly includes a speaker unit 31 and a driving unit 32 connected to the speaker unit 31. The driving unit 32 is used to convert the received digital signal into an analog signal for the speaker unit 31 to output as sound.

The controller 40 is connected or coupled to the reference microphone 10, the error microphone 20, and the speaker 30 via specific pins, so as to coordinate the work of the devices and process the transmitted signals. Specifically, the controller 40 mainly includes a feedforward noise reduction filter module F1, a feedback noise reduction filter module F2, a mixer F3, a noise detection module F4, and a weight distribution module F5. It is important to note that these functional modules can be integrated into a single controller for execution, or can be executed by multiple independent controllers (for example, a single or multiple controllers can execute individual function blocks), and these equal changes It is not within the scope of the present invention.

In terms of circuit configuration, the feedforward noise reduction filter module F1 receives the reference signal of the reference microphone 10, and transmits the reference signal as the first input signal (feedforward noise reduction signal) after the feedforward noise reduction process The mixer F3 and the feedback noise reduction filter module F2 receive the error audio source signal of the error microphone 20, and send the error audio source signal through the feedback noise reduction process as a second input signal (feedback noise reduction signal) to the mixer The mixer F3 mixes the first input signal and the second input signal through the mixer F3 to obtain a noise reduction signal, and outputs the mixed noise reduction signal to the speaker 30.

The noise detection module F4 receives audio via the reference microphone 10 or the error microphone 20, and confirms the noise state from the audio. The noise detection module F4 can be, for example, a spectral noise detector or a time domain noise detector. Detector and adaptive noise filter detector, etc. The weight distribution module F5 includes one or more weight adjustments arranged between the error microphone 20 and the feedback noise reduction filter module F2. The variable unit F5A, when the noise detection module F4 detects high frequency noise by the reference microphone 10 or the error microphone 20, transmits a control signal to the weight distribution module F5, and the weight distribution module F5 receives the control After the signal, the weight value of the weight modulation unit F5A is reduced to reduce the output weight of the feedback signal of the feedback noise reduction filter module F2.

Regarding the feedforward noise reduction filter module F1 and the feedback noise reduction filter module F2, a specific implementation aspect will be described below. Please also refer to "Figure 3", which is a block diagram (2) of the weighted hybrid active noise reduction system of the present invention, as shown in the figure: the feedforward noise reduction filter module F1 mainly includes a primary path Filter F11 (Secondary Path Filter), an adaptive algorithm F12 (Adaptation Algorithm Module), and a digital filter F13 (Digital Filter). The secondary path filter F11 is mainly defined by the designer according to the path that the noise reduction signal y(n) transforms into the error audio signal e(n), which is used to correct the input digital signal to compensate the signal, thereby obtaining the Reference signal waveform. The adaptive operator F12 updates the filter coefficients of the digital filter F13 according to the reference audio source signal and the error audio source signal. The adaptive operator F12 may be, for example, a least mean square filter (LMS Filter). No restrictions on it. The digital filter F13 filters the reference signal of the reference microphone 10 according to the updated filter coefficients to obtain a feedforward noise reduction signal and then transmits it to the mixer F3. The digital filter F13 is in a feasible implementation state. The sample may be a finite impulse response filter (FIR filter), or a biquad filter (Biquad Filter), etc., which are not limited in the present invention.

The digital filter F13 of the feedforward noise reduction filter module F1 is mainly used to estimate the error caused by unknown environmental factors (such as earmuffs). The unknown environmental factor and the digital filter F13 both receive the same input signal x(n ). By introducing the filtering FXLMS algorithm, the secondary path problem can be solved, the transformation set of the analog filter signal and the error audio signal in the path from the electrical signal to the audio signal. During the electro-acoustic conversion process, the effect of errors generated therebetween is minimized by delaying or changing the signal.

The feedback noise reduction filter module F2 mainly includes a primary path filter F21, An adaptive operator F22 (Secondary Path Filter) and a digital filter F23 (Digital Filter). Same as the feedforward noise reduction filter module F1, the secondary path filter F21 is defined by the designer according to the path that the noise reduction signal y(n) transforms into the error audio signal e(n) to convert the input The digital signal is used to correct the compensation signal. Among them, the adaptive filter F22 of the feedback noise reduction filter module F2 consists of the error signal e(n) and the result signal y(n-1) output from the previous noise reduction signal from the secondary path and the error audio signal e(n) The filter coefficients of the digital filter F23 are updated by mixing the noise reduction signal-error audio source signal mixer F6 as the input signal. The adaptive operator F22 may be, for example, a least mean square filter (LMS Filter). In the present invention No restrictions. The digital filter F23 filters the resultant signal y(n-1) and the signal after the error audio source signal e(n) according to the updated filter coefficients according to the updated coefficients, and then outputs the feedback noise reduction signal to the mixed The second input terminal of the filter F3, where the digital filter F23 can be a finite impulse response filter (FIR filter), or a biquad filter (Biquad Filter) in a feasible implementation mode, which is not in the present invention. Be restricted. At the front end of the noise reduction signal-error audio signal mixer F6, a secondary path filter F7 is provided to correct the input noise reduction signal to compensate the signal.

In this embodiment, the weight distribution module F5 includes a first complementary weight modulator F51 arranged between the mixer F3 and the feedforward noise reduction filter module F1, and a first complementary weight modulator F51 arranged on the A second complementary weight modulator F52 between the mixer F3 and the feedback noise reduction filter module F2, and a second complementary weight modulator F52 disposed between the error microphone 20 and the adaptive filter F22 of the feedback noise reduction filter module F2 A first weight modulator F53, and a second weight modulator F54 arranged between the error microphone 20 and the noise reduction signal-error audio signal mixer F6. The total weight of the first complementary weight modulator F51 and the second complementary weight modulator F52 is a constant value, for example, 1. The weight modulator can be, for example, split modulation, weighted modulation, motion detection modulation, etc., and each of the weight modulators is modulated by software or firmware stored in the controller (or storage unit) To determine the signal amplification ratio of the weight modulator.

In another feasible implementation aspect, the first weight modulator F53 can also be A weight modulator (not shown in the figure) is set before the common node of the second weight modulator F54, and a single weight modulator controls the weight of the input signal of the feedback noise reduction filter module F2, but the efficiency is compared with the previous one. Low implementation status.

Regarding the control logic of the weight distribution module F5, please refer to "Figure 3" and "Figure 4" below. The control flow diagram (1) and control flow diagram (2) of the weighted hybrid active anti-noise system of the present invention , As shown in the figure: When the system is started, the noise detection module F4 continuously detects the reference audio signal of the reference microphone 10 (and/or the error audio signal of the error microphone 20), and detects whether noise is received ( Step S01).

When noise is detected, the noise detection module F4 confirms the type of the noise, and confirms that the type of noise is a regular noise signal or an irregular noise signal (step S02). The noise pattern can be determined through the spectrum noise detection Detectors, time-domain noise detectors, and adaptive noise filter detectors are confirmed to distinguish between regular noise and irregular noise.

When the detected noise is high frequency and regular, unstable conditions are likely to occur on the path of the feedback noise reduction filter module F2, which reduces the effect of noise reduction. At this time, the noise detection module F4 transmits a first control signal to the weight distribution module F5 to switch the weight distribution module F5 to the first working mode (step S03). In the state of the first working mode, the weight distribution module F5 gradually reduces the weight of the second complementary weight modulator F52 according to timing priority, and secondarily reduces the weight of the first weight modulator F53 and the second The weight of the weight modulator F54.

Specifically, the weight distribution module F5 sets the weight of the first weight modulator F53 to 1 at the initial value, sets the weight of the second weight modulator F54 to 1, and the second complementary formula The weight of the weight modulator F52 is set to 0.5.

When irregular noise is detected, the weight of the second complementary weight modulator F52 is reduced according to the time sequence (the minimum can be adjusted to 0), so that the noise reduction effect of the entire system is more forward-feed noise reduction The filter module F1 handles irregular noise.

Next, the weight distribution module F5 then gradually reduces the weight of the first weight modulator F53 (the smallest can be adjusted to 0), so that the update mechanism of the feedback noise reduction filter module F2 gradually tends to a state of little contribution .

Finally, the weight of the second weight modulator F54 is adjusted to gradually decrease (the minimum can be adjusted to 0), so that the signal reconstructed by the final feedback noise reduction filter module F2 has the smallest influence on the error.

In the above process, the time sequence is adjusted in advance by the design engineer, and each of the weight modulators is triggered sequentially through time. In another feasible implementation aspect, the weight allocation module F5 can determine whether to continue to adjust the weight according to the error rate of the noise detection module F4.

When the noise detection module F4 detects that the irregular noise disappears, the weight distribution module F5 directly or gradually returns the weight of each weight modulator to the initial value (the first weight modulator F53 The weight of is set to 1, the weight of the second weight modulator F54 is set to 1, and the weight of the second complementary weight modulator F52 is set to 0.5) (step S04).

Following step S02, when the detected noise is regular, the noise detection module F4 transmits a second control signal to the weight distribution module F5 to switch the weight distribution module F5 to the second working mode (Step S05). In the state of the second working mode, the weight distribution module F5 gradually increases the weight of the second complementary weight modulator F52 (up to 1) according to timing priority.

Specifically, when the weight distribution module F5 is in the second working mode, the weight distribution module F5 increases the weight of the second complementary weight modulator F52 (the maximum can be adjusted to 1), so that the entire system The denoising effect is more biased towards the feedback denoising filter module F2 to deal with regular noise. Because the regular noise is basically processed by the feedback noise reduction filter module F2 to get the best effect, the first weight modulator F53 and the second weight modulator F54 both maintain the initial value of 1 to achieve The best feedback noise reduction state.

When the noise detection module F4 detects that the regular noise disappears, the weight is assigned The module F5 directly or gradually resets the weight of each weight modulator to the initial value (the weight of the first weight modulator F53 is set to 1, and the weight of the second weight modulator F54 is set to 1, The weight of the second complementary weight modulator F52 is set to 0.5) (step S06).

In summary, the present invention adds a weight adjustment function to the feedback path, and when the performance deteriorates, the weight can be appropriately reduced on the feedback path, and the feedforward path is led to achieve a true adaptive effect. In addition, the present invention can strengthen the processing effect corresponding to irregular noise and regular noise, so as to improve the traditional hybrid active anti-noise architecture that is prone to instability when processing high-frequency regular noise.

The present invention has been described in detail above, but what has been described above is only a preferred embodiment of the present invention. It should not be used to limit the scope of implementation of the present invention, that is, everything made in accordance with the scope of the patent application of the present invention is equal Changes and modifications should still fall within the scope of the patent of the present invention.

100‧‧‧Weighted Hybrid Active Noise Control System

10‧‧‧Reference microphone

20‧‧‧Error microphone

31‧‧‧Speaker unit

32‧‧‧Drive unit

F1‧‧‧Feedforward noise reduction filter module

F2‧‧‧Feedback noise reduction filter module

F3‧‧‧Mixer

F5‧‧‧Weight Distribution Module

F5A‧‧‧Weight Modulation Unit

Claims (8)

A weighted mixed-type active noise immunity system includes: a reference microphone; an error microphone; a speaker; and a controller connected to the reference microphone, the error microphone, and the speaker. The controller includes a feedforward Noise reduction filter module, a feedback noise reduction filter module, a mixer, a noise detection module, and a weight distribution module, the feedforward noise reduction filter module is the reference signal received by the reference microphone A feedforward noise reduction signal is obtained after the feedforward noise reduction. The feedback noise reduction filter module obtains a feedback noise reduction signal after the error audio source signal received by the error microphone is feedback noise reduction, and reduces the feedforward noise The noise signal and the feedback noise reduction signal are sent to the mixer for mixing, and the mixed noise reduction signal is output to the speaker. The weight distribution module includes one or more error microphones and the feedback The weight modulation unit between the noise reduction filter modules sends a control signal to the weight distribution module when the noise detection module detects high frequency noise from the reference microphone or the error microphone, and the weight distribution module The group reduces the weight value of the weight modulation unit to reduce the proportion of the feedback signal of the feedback noise reduction filter module, wherein the weight distribution module includes a set of the mixer and the feedforward noise reduction filter module Between the first complementary weight modulator, and a second complementary weight modulator disposed between the mixer and the feedback noise reduction filter module, the first complementary weight modulator and the second complementary weight modulator The total weight of the two complementary weight modulators is a constant value. According to the weighted hybrid active noise reduction system described in item 1 of the scope of patent application, the weight distribution module includes a set between the error microphone and the adaptive filter of the feedback noise reduction filter module A first weight modulator, and a second weight modulator arranged between the error microphone and the noise reduction signal-error audio signal mixer. The weighted hybrid type active noise immunity system described in item 2 of the scope of patent application, wherein, when the noise detection module detects irregular high-frequency noise by the reference microphone or the error microphone, a The first control signal is sent to the weight distribution module. When the weight distribution module receives the first control signal, the weight distribution module gradually reduces the first control signal according to the time sequence or the error rate collected by the error microphone. Two complementary weight modulator weights, and secondarily reduce the weights of the first weight modulator and the second weight modulator. The weighted hybrid active noise immunity system according to any one of the 2 or 3 of the scope of patent application, wherein the reference microphone or the error microphone detects the regular high frequency in the noise detection module When there is noise, a second control signal is transmitted to the weight distribution module. When the weight distribution module receives the second control signal, the weight distribution module gradually increases the second complementary weight modulation according to timing priority The weight of the device. A controller includes: a feedforward noise reduction filter module, the feedforward noise reduction filter module obtains a feedforward noise reduction signal after a reference signal received by a reference microphone undergoes feedforward noise reduction; A feedback noise reduction filter module, the feedback noise reduction filter module is to obtain a feedback noise reduction signal after the error audio source signal received by an error microphone is feedback noise reduction; a mixer, the feedforward noise reduction signal and The feedback noise reduction signal is mixed and outputs a noise reduction signal; a noise detection module, which detects high-frequency noise from the input of the feedforward noise reduction filter module or the feedback noise reduction filter module, and detects When high frequency noise is detected, a control signal is output; and a weight distribution module, the weight distribution module includes one or more weight modulation units arranged between the error microphone and the feedback noise reduction filter module, When the control signal is received, the weight value of the weight modulation unit is reduced to reduce the ratio of the feedback signal of the feedback noise reduction filter module; wherein, the weight distribution module includes a set of the mixer and the front A first complementary weight modulator between the feed noise reduction filter modules, and a second complementary weight modulator disposed between the mixer and the feedback noise reduction filter module, the first complementary weight modulator The total weight of the weight modulator and the second complementary weight modulator is a constant value. The controller according to claim 5, wherein the weight distribution module includes a first weight modulator arranged at the front end of the adaptive filter of the feedback noise reduction filter module, and a first weight modulator arranged on the The second weight modulator at the front end of the noise reduction signal-error audio signal mixer. The controller described in item 6 of the scope of patent application, wherein, when the noise detection module detects irregular high-frequency noise by the reference microphone or the error microphone, it transmits Send a first control signal to the weight distribution module, and when the weight distribution module receives the first control signal, the weight distribution module is based on the timing or the error rate collected by the noise detection module The size gradually reduces the weight of the second complementary weight modulator, and secondarily reduces the weights of the first weight modulator and the second weight modulator. The controller according to any one of items 6 or 7 of the scope of patent application, wherein, when the noise detection module detects regular high-frequency noise by the reference microphone or the error microphone, it transmits a second The control signal is sent to the weight distribution module. When the weight distribution module receives the second control signal, the weight distribution module gradually increases the weight of the second complementary weight modulator according to timing priority.

Images