CN110720121A

CN110720121A - Compensation and automatic gain control in active noise reduction devices

Info

Publication number: CN110720121A
Application number: CN201880036193.2A
Authority: CN
Inventors: J·H·卡特尔; M·奥科内尔; R·F·卡雷拉斯; D·瓦肯丁
Original assignee: BOSS Co Ltd
Current assignee: BOSS Co Ltd
Priority date: 2017-03-30
Filing date: 2018-03-29
Publication date: 2020-01-21
Anticipated expiration: 2038-03-29
Also published as: WO2018183714A3; CN116741138A; EP3602538B1; EP3602538A2; EP3618058A1; EP3627493B1; CN110720121B; EP3627493A1; EP3618058B1; WO2018183714A2

Abstract

The technology described in this document may be embodied by a method comprising: an input signal representing audio captured by a microphone of an Active Noise Reduction (ANR) earpiece is received, and a first range of frequencies of the input signal is processed by a first compensator to generate a first signal for a sound transducer of the ANR earpiece. The method also includes processing, by a second compensator disposed in parallel with the first compensator, a second frequency range of the input signal to generate a second signal for the acoustic transducer. The first frequency range includes frequencies higher than frequencies in the second frequency range. The method also includes detecting, by the one or more processing devices, whether the second signal satisfies a threshold condition, and attenuating the second signal in response to determining that the second signal satisfies the threshold condition.

Description

Compensation and automatic gain control in active noise reduction devices

Technical Field

The present disclosure relates generally to techniques for controlling overload conditions in Active Noise Reduction (ANR) devices.

Background

ANR devices may utilize one or more Digital Signal Processors (DSPs) to implement various signal flow topologies. Examples of such DSPs are described in U.S. patents 8,073,150 and 8,073,151, which are incorporated herein by reference in their entirety.

Disclosure of Invention

In general, in one aspect, this document features a method that includes: an input signal representing audio captured by a microphone of an Active Noise Reduction (ANR) earpiece is received, and a first range of frequencies of the input signal is processed by a first compensator to generate a first signal for a sound transducer of the ANR earpiece. The method also includes processing, by a second compensator disposed in parallel with the first compensator, a second frequency range of the input signal to generate a second signal for the acoustic transducer. The first frequency range includes frequencies higher than frequencies in the second frequency range. The method also includes detecting, by the one or more processing devices, whether the second signal satisfies a threshold condition, and attenuating the second signal in response to determining that the second signal satisfies the threshold condition.

In another aspect, this document features an Active Noise Reduction (ANR) device that includes one or more sensors configured to generate an input signal indicative of an environment external to the ANR device. The apparatus further comprises: a first compensator configured to process a first range of frequencies of an input signal to generate a first signal for an acoustic transducer of an ANR earpiece; and a second compensator arranged in parallel with the first compensator. The second compensator filter is configured to process a second frequency range of the input signal to generate a second signal for the acoustic transducer, wherein the first frequency range includes frequencies higher than frequencies in the second frequency range. The device also includes one or more processing devices configured to detect whether the second signal satisfies a threshold condition and attenuate the second signal in response to determining that the second signal satisfies the threshold condition.

In another aspect, this document features one or more machine-readable storage devices having encoded thereon computer-readable instructions for causing one or more processing devices to perform various operations. These operations include: the method includes receiving an input signal representing audio captured by a microphone of an Active Noise Reduction (ANR) earpiece, causing a first compensator to process a first range of frequencies of the input signal to generate a first signal for an acoustic transducer of the ANR earpiece, and causing a second compensator to process a second range of frequencies of the input signal to generate a second signal for the acoustic transducer. The second compensator is arranged in parallel with the first compensator and the first frequency range comprises higher frequencies than in the second frequency range. The operations further include: detecting whether the second signal satisfies a threshold condition, and attenuating the second signal in response to determining that the second signal satisfies the threshold condition.

Implementations of the above aspects may include one or more of the following features.

The combined signal for the acoustic transducer may be generated by adding the second signal to the first signal or by adding the attenuated second signal to the first signal. The combined signal may be used to drive an acoustic transducer. The upper limit of the second frequency range may be substantially equal to 100 Hz. The first frequency range may include at least a portion of the second frequency range. Detecting whether the second signal satisfies the threshold condition may include determining whether a voltage level representative of the second signal meets or exceeds a threshold to indicate an overload condition. Detecting whether the second signal satisfies the threshold condition may include: the second signal is filtered using a digital filter and the filtered second signal is compared to a value associated with a threshold condition. A set of coefficients of the digital filter may be selected based on an operating mode of the ANR earpiece. Attenuating the second signal may include adjusting a Variable Gain Amplifier (VGA) that processes the second signal. Processing the first frequency range of the input signal to generate the first signal may include: the input signal is processed by a first filter to generate a first filtered signal, and the first filtered signal is processed by a first compensator to generate a first signal. The first filter may have a pass band including a first frequency range, and the first signal may represent an anti-noise signal configured to reduce a noise signal in the first filtered signal. Processing the second frequency range of the input signal to generate the second signal may include: the input signal is processed by a second filter to generate a second filtered signal, and the second filtered signal is processed by a second compensator to generate a second signal. The second filter may have a pass band that includes a second frequency range, and the second signal may represent an anti-noise signal configured to reduce a noise signal in the second filtered signal. The input signal may represent audio captured by a feedforward microphone of the ANR earpiece. Each of the first compensator and the second compensator may be a feedforward compensator disposed in a feedforward signal flow path of the ANR earpiece. Each of the first compensator and the second compensator may be a feedback compensator disposed in a feedback signal flow path of the ANR earpiece.

In another aspect, this document features a method that includes: an input signal representing audio captured by one or more sensors of an Active Noise Reduction (ANR) earpiece is received, and based on the input signal, a first signal is generated by a compensator disposed in an ANR signal flow path of the ANR earpiece. The method also includes determining one or more characteristics of the first signal and selecting a plurality of filter coefficients of a digital filter disposed in series with a compensator in the ANR signal flow path based on the one or more characteristics of the first signal. The filter coefficients are selected according to a target frequency response of the digital filter. The method also includes generating a feedback control signal for an electroacoustic transducer of the ANR earpiece by processing the input signal using a plurality of filter coefficients of the digital filter.

In another aspect, the invention features an Active Noise Reduction (ANR) device that includes: one or more sensors configured to generate an input signal indicative of an environment external to the ANR device; and a compensator disposed in an ANR signal flow path of the ANR device. The compensator is configured to generate a first signal based on an input signal. The ANR device also includes a tunable digital filter disposed in series with the compensator in the ANR signal flow path, wherein the tunable digital filter is configured to generate a control signal for an electroacoustic transducer of the ANR device. The ANR device also includes one or more processing devices configured to determine one or more characteristics of the first signal and select a plurality of filter coefficients for the tunable digital filter according to a target frequency response of the tunable digital filter based on the one or more characteristics of the first signal.

In another aspect, this document features one or more machine-readable storage devices having encoded thereon computer-readable instructions for causing one or more processing devices to perform various operations. These operations include: an input signal representing audio captured by one or more sensors of an Active Noise Reduction (ANR) earpiece is received, and a compensator disposed in an ANR signal flow path of the ANR earpiece is caused to generate a first signal based on the input signal. The operations further include determining one or more characteristics of the first signal, and selecting a plurality of filter coefficients of a digital filter disposed in series with a compensator in the ANR signal flow path based on the one or more characteristics of the first signal. The filter coefficients are selected according to a target frequency response of the digital filter. The operations further include generating a feedback control signal for an electroacoustic transducer of an ANR earpiece by causing a digital filter to process an input signal using a plurality of filter coefficients.

The digital filter may be placed before or after the compensator in the signal flow path. Based on the one or more characteristics, it may be determined that a portion of the input signal within the particular frequency range is causing the first signal to trigger an overload condition in the acoustic transducer, and the plurality of filter coefficients may be selected such that the selected filter coefficients configure the digital filter to attenuate the portion of the input signal within the particular frequency range. The one or more characteristics may include a voltage level. The electro-acoustic transducer may be driven using a feedback control signal. The digital filter may be a high pass filter or a notch filter. The digital filter may be an Infinite Impulse Response (IIR) filter. The ANR signal flow path may include a feedforward path disposed between a feedforward microphone and an electroacoustic transducer of the ANR earpiece. The ANR signal flow path may include a feedback path disposed between a feedback microphone and an electroacoustic transducer of the ANR earpiece.

In another aspect, this document features a method that includes: an input signal captured by one or more sensors associated with an ANR earpiece is received, and one or more characteristics of a first portion of the input signal are determined by one or more processing devices in an ANR signal flow path. The method further comprises the following steps: automatically adjusting, by the one or more processing devices, a gain of a Variable Gain Amplifier (VGA) disposed in the ANR signal flow path based on one or more characteristics of the first portion of the input signal, and selecting, by the one or more processing devices, a set of coefficients for a tunable digital filter disposed in the ANR signal flow path. The set of coefficients is selected according to the gain of the VGA. The method also includes processing a second portion of the input signal in the ANR signal flow path using the adjusted gain and the selected set of coefficients to generate a second output signal for an electroacoustic transducer of the ANR earpiece.

In another aspect, this document features an Active Noise Reduction (ANR) device. The apparatus comprises: the system includes one or more sensors configured to generate an input signal indicative of an environment external to the ANR device, and a Variable Gain Amplifier (VGA) and a tunable digital filter disposed in an ANR signal flow path of the ANR device. The ANR signal flow path is connected to an electroacoustic transducer of the ANR device. The device also includes one or more processing devices configured to: the method includes determining one or more characteristics of a first portion of an input signal, automatically adjusting a gain of the VGA based on the one or more characteristics of the first portion of the input signal, and selecting a set of coefficients for the tunable digital filter, wherein the set of coefficients is selected according to the gain of the VGA.

In another aspect, this document features one or more machine-readable storage devices having encoded thereon computer-readable instructions for causing one or more processing devices to perform various operations. These operations include: an input signal captured by one or more sensors associated with the ANR earpiece is received, and one or more characteristics of a first portion of the input signal are determined. The operations further include: automatically adjusting a gain of a Variable Gain Amplifier (VGA) disposed in an ANR signal flow path based on one or more characteristics of a first portion of an input signal; and selecting a set of coefficients for a tunable digital filter disposed in the ANR signal flow path. The set of coefficients is selected according to the gain of the VGA. The operations further include processing a second portion of the input signal in the ANR signal flow path using the adjusted gain and the selected set of coefficients to generate a second output signal for an electroacoustic transducer of the ANR earpiece.

Implementations of the above aspects may include one or more of the following features. Determining one or more characteristics of the first portion of the input signal may include: processing a first portion of an input signal to generate a first output signal for an electroacoustic transducer of an ANR device; and determining one or more characteristics of the first portion of the input signal based on the one or more characteristics of the first output signal. The one or more sensors may include one of: a feedforward microphone of the ANR device, a feedback microphone of the ANR device, a pressure sensor, an accelerometer, and a displacement sensor configured to sense an offset of the electro-acoustic transducer. The ANR signal flow path may include a feedforward path disposed between a feedforward microphone and an electroacoustic transducer of the ANR device. The ANR signal flow path may include a feedback path disposed between a feedback microphone and an electroacoustic transducer of the ANR earpiece. The VGA gain adjustment may be deactivated in response to a first user input received by an input device of the ANR device. The VGA gain adjustment may also be reactivated in response to a second user input received by an input device of the ANR earpiece. The gain of the VGA may be periodically adjusted during operation of the ANR earpiece. The gain of the VGA may be adjusted in response to determining that one or more characteristics of the first portion of the input signal satisfy a threshold condition. One or more characteristics of the first portion of the input signal may be indicative of a noise floor of an environment external to the ANR earpiece. Determining one or more characteristics of the first portion of the input signal based on the one or more characteristics of the first output signal may include determining that the first output signal is clipped. The one or more characteristics of the first portion of the input signal may indicate a likelihood that an output signal resulting from processing the first portion of the input signal by the ANR signal flow path will be clipped.

Various embodiments described herein may provide one or more of the following advantages. By suppressing compensation in only a selected portion of the frequency range under overload conditions, performance of a low-profile ANR device (e.g., an in-ear headphone) may be improved. For example, selective suppression in the low frequency range may allow for mitigation of overload conditions while avoiding potentially objectionable noise modulation that may occur due to turning off the entire feed forward compensation. Since the sensitivity of the human ear to low frequencies (e.g., below 100Hz) is relatively low, suppressing compensation at such low frequencies may have little impact on the psychoacoustic experience of the user when an overload condition is detected, and thus may improve the overall user experience compared to a device that completely turns off compensation in the ANR signal flow path (e.g., feedforward or feedback path) when an overload is detected. In addition to or independent of processing in one ANR signal flow path (e.g., a feedforward path), tunable filters may be provided in the same or another ANR signal flow path (e.g., a feedback path) to mitigate overload conditions due to low frequency stimuli detected by a corresponding microphone (e.g., a feedback microphone in this particular example). In some cases, the tunable filter (which may be implemented as a high pass or notch filter) may improve user experience and driver lifetime by reducing low frequency displacement of the driver due to, for example, jaw movement or walking. In some implementations, noise reduction performance of an ANR device can be adaptively balanced with its overload characteristics by providing a Variable Gain Amplifier (VGA) disposed in series with a tunable filter in a signal flow path (e.g., a feedback path or a feedforward path). For example, in some cases, an increase in increasing VGA may result in better signal-to-noise ratio (SNR), but at the cost of reduced dynamic range and/or increased likelihood of being driven into an overload state. Thus, automatic and simultaneous adjustment of both the VGA and the tunable filter may be used to adapt the ANR device to various different environments, thereby improving the overall user experience.

Two or more features described in this disclosure, including those described in this summary, can be combined to form embodiments not specifically described herein. The details of one or more implementations are set forth in the accompanying drawings and the description below. Other features, objects, and advantages will be apparent from the description and drawings, and from the claims.

Drawings

Fig. 1 illustrates an in-ear Active Noise Reduction (ANR) earpiece.

FIG. 2 is a block diagram of an ANR device.

FIG. 3A is an example of a block diagram of an ANR device with feedforward compression.

FIG. 3B is an example of a block diagram of an ANR device with parallel feedforward compression.

Fig. 4 is a graph illustrating a pressure change in an ear canal due to a jaw movement of a user using a sealed in-ear headphone.

Fig. 5A is a block diagram of a feedback path filter including a tunable filter configured to mitigate overload conditions due to low frequency stimuli.

FIG. 5B is a block diagram of an example combination of a Variable Gain Amplifier (VGA) and a tunable filter disposed in a signal flow path of an ANR device.

Fig. 6A-6C are amplitude responses of different tunable high pass filters.

FIG. 7 is a block diagram of an example biquad notch filter that may be used as a tunable filter in an ANR signal flow path.

Fig. 8A shows the magnitude and phase response of the biquad notch filter of fig. 7 for different combinations of filter coefficients.

Fig. 8B shows the variation of the poles and zeros of the biquad notch filter with respect to the tuning parameter n.

Fig. 8C and 8D show the variation of the coefficient values of the biquad notch filter with respect to the tuning parameter n.

FIG. 9A illustrates the magnitude and phase response of the feedback path loop gain of an example ANR device without a tunable filter.

Fig. 9B shows the magnitude and phase response of the feedback path loop gain of fig. 9A but with a tunable filter.

FIG. 10A illustrates the sensitivity of the feedback path of an example ANR device without a tunable filter.

Fig. 10B shows the sensitivity of the feedback path of fig. 10A but with a tunable filter.

Fig. 11 shows the variation of the output voltage of the feedback compensator for various values of the tuning parameter of the series connected tunable filters.

Figure 12A illustrates an amplitude response of another example of a notch filter that may be used as a tunable filter in an ANR signal flow path.

Fig. 12B shows changes in the coefficient values of the notch filter shown in fig. 12A.

FIG. 13 is a flow chart of an exemplary process for implementing parallel feed forward compression in accordance with the techniques described herein.

Fig. 14 is a flow diagram of an example process for implementing a tunable filter in a feedback path of an ANR device in accordance with the techniques described herein.

FIG. 15 is a flow diagram of an exemplary process for implementing a combination of a Variable Gain Amplifier (VGA) and a tunable filter in a signal flow path of an ANR device.

Detailed Description

An Active Noise Reduction (ANR) device may include a configurable Digital Signal Processor (DSP) that may be used to implement various signal flow topologies and filter configurations. Examples of such DSPs are described in U.S. patents 8,073,150 and 8,073,151, which are incorporated herein by reference in their entirety. Us patent 9,082,388 (also incorporated herein by reference in its entirety) describes an acoustic implementation of an in-ear Active Noise Reduction (ANR) earpiece as shown in fig. 1. The headset 100 includes a feedforward microphone 102, a feedback microphone 104, an output transducer 106 (which may also be referred to as an electroacoustic transducer or an acoustic transducer), and a noise reduction circuit (not shown) coupled to the two microphones and the output transducer to provide an anti-noise signal to the output transducer based on signals detected at the two microphones. An additional input (not shown in fig. 1) of the circuit provides an additional audio signal, such as music or a communication signal, for playback on the output transducer 106 independent of the noise reduction signal.

The term "earpiece" as used interchangeably herein with the term "headset" includes various types of personal acoustic devices, such as in-ear, circum-ear or circum-ear headsets, earphones, and hearing aids. The headset or earpiece may include an ear plug or ear muff for each ear. The earplugs or earmuffs may be physically tied to each other, such as by a cord, a head bridge, or a headband, or a behind-the-head retention structure. In some implementations, the earplugs or earmuffs of the headphones may be connected to each other via a wireless link.

Various signal flow topologies may be implemented in ANR devices to implement functions such as audio equalization, feedback noise cancellation, feedforward noise cancellation, and so on. For example, as shown in the exemplary block diagram of the ANR device 200 in fig. 2, the signal flow topology may include a feedforward noise reduction path 110 that drives the output transducer 106 to generate an anti-noise signal (e.g., using the feedforward compensator 112) to reduce the effect of the noise signal picked up by the feedforward microphone 102. As another example, the signal flow topology may include a feedback noise reduction path 114 that drives the output transducer 106 to generate an anti-noise signal (using, for example, a feedback compensator 116) to reduce the effect of the noise signal picked up by the feedback microphone 104. The signal flow topology may also include an audio path 118 that includes circuitry (e.g., an equalizer 120) for processing the input audio signal 108, such as music or communication signals, for playback on the output transducer 106.

Under most operating conditions, the acoustic noise energy that ANR devices attempt to reduce is small enough to keep the system hardware within capacity. However, in some cases, discrete acoustic signals or low frequency pressure disturbances (e.g., loud pops, explosions, door pops, etc.) picked up by the feed-forward or feedback microphones may cause the noise reduction circuit to exceed the capacity of the electronics or output transducer 106 in an attempt to reduce the generated noise, thereby creating audible artifacts that may be objectionable to some users. These conditions, referred to herein as overload conditions, may be manifested by, for example, clipping of the amplifier, hard offset limiting of the acoustic driver or transducer, or an offset level that causes sufficient change in the acoustic response to cause oscillation. The problem of overload conditions can be particularly acute in small-form-factor ANR devices, such as in-ear headphones. For example, to compensate for low frequency pressure disturbances (e.g., sounds of buses passing through potholes, door pops, or aircraft takeoff), the feedforward compensator may generate a signal that requires the acoustic transducer to exceed a corresponding physical excursion limit. Due to acoustic leakage, the offset or drive displacement that produces a given pressure generally increases as the frequency decreases. For example, a particular acoustic transducer may require a displacement of 1mm to generate an anti-noise signal for 100Hz noise, a displacement of 2mm to generate an anti-noise signal for 50Hz noise, and so on. Many acoustic transducers, especially the small transducers used in small-form-factor ANR devices, are physically incapable of producing such large displacements. In such cases, the need for high displacements by the compensator may cause the transducer to generate sounds that cause auditory artifacts, which may result in an objectionable user experience. Auditory artifacts may include oscillations, potentially objectionable transient sounds (e.g., "pounding," "cracking," "popping," or "clicking") or cracking/beeping.

In some cases, such artifacts may be reduced by temporarily reducing the gain (also referred to herein as "suppression") along selected portions of the signal processing path, such that the transients in noise caused by the gain reduction are compared to the artifacts to be processedThe increase may not cause a sense of incongruity to the user. For example, as shown in the block diagram of the example ANR device 300 in fig. 3A, the feedforward path 110 may include a Variable Gain Amplifier (VGA)125 whose gain may be reduced (dampened) upon detection of a signal that may potentially overload the output transducer 106. This may be accomplished, for example, using a side-chain filter 128, which is a filter applied to a signal sampled from the main signal stream to generate a test signal used to determine whether to suppress the gain of VGA 125. For example, the output of VGA125 may be passed through side chain filter 128 (whose transfer function is denoted as K) connected to feedforward path 110_{sc_ff}) And compares the output of the side-chain filter to a predetermined threshold T (e.g., using comparator 130)_ff132 to determine if an overload condition exists. The output of the comparator 130 is provided to a Variable Gain Amplifier (VGA) 125. If the comparator 130 detects that the filtered output signal is greater than the threshold value T_ffThe gain of the VGA125 is adjusted to suppress the signal in the feed-forward path 110 to mitigate the overload condition. Although fig. 3A shows a side-chain filter for the feed-forward path only, a similar side-chain filter may be implemented in the feedback path. In addition, the side-chain filter may be placed before or after the corresponding

main compensator

112 or 116, respectively, in the feed-forward or feedback path.

In some cases, reducing the gain of the entire feed-forward or feedback path may also generate some undesirable audible artifacts and/or noise modulation. For example, if the noise causing the overload condition has significant energy at low and high frequencies, turning off or significantly reducing the gain of the feedforward compensator may allow the noise to pass through without attenuation and may create an uncomfortable or objectionable experience for some users. The techniques described herein may improve the user experience in such situations by allowing gain adjustment only within a selected frequency range when an overload condition is detected, while allowing compensation signals to be generated at frequencies outside of the selected range. For example, since noise reduction compensation at the lower end of the spectrum (e.g., below 100Hz) is often the primary cause of overload conditions, feed forward compensation may be suppressed only in the low frequency portion of the spectrum, while allowing feed forward compensation to continue for other frequencies. This may provide an improved psychoacoustic experience for the user, since the feed forward compensation is only temporarily suspended in a selected part of the frequency range when the low frequency disturbance is detected. If the selected frequency range is in the region below 100Hz, the user experience is not significantly degraded for most users, since the human ear is generally less sensitive to noise in this frequency range.

FIG. 3B is a block diagram of an example of an ANR device 350 with parallel feedforward compression, where feedforward compensation is suppressed only in a portion of a range of frequencies of operation when an overload condition is detected within the range of frequencies. The apparatus 350 includes at least two parallel paths in the feed-forward path 110, each path processing a different portion of the operating frequency range. For example, the feedforward path 110 may include a main feedforward compensator 133 (represented by transfer function K)_ffMIndicative) that process a frequency range that substantially excludes frequencies at which overload conditions are expected to occur. The feedforward path 110 also includes an auxiliary feedforward compensator 134 (represented by transfer function K) connected in parallel to the main feedforward compensator 133_ffPRepresentation). The auxiliary feedforward compensator 134 processes the frequency range in which an overload condition is expected to occur. The VGA125, the side chain filter 128 and the comparator 130 are connected to the output of the auxiliary compensator 134 to suppress compensation in the corresponding frequency range when an overload condition is detected. However, even when the VGA125 suppresses the feed-forward compensation of the auxiliary compensator 134, the main feed-forward compensator 133 continues to provide compensation in the corresponding frequency range. For example, the auxiliary compensator may be configured to process signals only in the range below 100Hz in order to suppress feed forward compensation due to low frequency pressure disturbances using VGA 125. Even when the feedforward compensation by the auxiliary compensator 134 is suppressed, the main feedforward compensator 133 (which may be configured to process signals above 100Hz) continues to provide feedforward compensation to reduce noise in the corresponding frequency range. In some implementations, this improves the overall noise reduction performance of the corresponding ANR device while limiting audible artifacts that may be caused by low-frequency pressure disturbances.

The threshold 132 associated with the comparator 130 may be determined in various ways. In some implementations, the threshold 132 may be determined based on characteristics of the output transducer. For example, the threshold 132 may be set as a voltage reference point to prevent the drive voltage output by the VGA125 from causing the output transducer 106 to reach a mechanical limit, or to reach a drive level where sound distortion due to mechanical, magnetic, or electrical characteristics is considered undesirable. In some cases, these limitations may be related to the equivalent pressure level in the ear canal. For example, as the size of the output transducer becomes smaller, these limitations may occur at lower equivalent pressure levels in the ear canal.

In some implementations, the main feedforward compensator 133 and the auxiliary feedforward compensator 134 may include filters for isolating the corresponding operating frequency ranges. For example, the auxiliary compensator 134 may comprise a low pass filter having a passband cutoff frequency substantially equal to 100 Hz. The main feedforward compensator 133 may comprise, for example, a high-pass filter having a stop-band cutoff frequency substantially equal to 100 Hz. Other configurations may also be used depending on, for example, the corresponding application. For example, the main feedforward compensator 133 may include a band pass filter to isolate a range of frequencies that exclude frequencies where an overload condition is expected to occur. In some implementations, the passbands of the main and auxiliary feedforward compensators may partially overlap. Although fig. 3B only depicts the filter topology in the feed-forward path 110, a similar parallel topology may also be used in the feedback path.

In some implementations, multiple side-chain filters may be used in conjunction with the auxiliary compensator 134. For example, the side-chain filters may be implemented as a filter bank, with the particular side-chain filter selected based on the mode of operation of the ANR device. For example, if the ANR device is used in a mode that allows some environmental sounds (e.g., human sounds) to pass through, the selected side-chain filter may be different than the side-chain filter selected in the mode in which feedforward compensation is performed for the entire operating frequency range. In some implementations, the filter bank may be implemented using a DSP, where a different set of filter coefficients and/or thresholds are selected for the side-chain filter based on the identified mode of operation. In some implementations, the main feedforward compensator 133 may be configured to provide noise attenuation in the corresponding frequency range as long as the signal of the feedforward microphone 102 is not clipped. In some implementations, the side-chain filter can operate based on input from one or more additional sensors. For example, an accelerometer may be used to identify a user's motion (e.g., running, jogging, etc.) that may cause an overload condition. In some implementations, historical information about user behavior can be used to predict events that may cause overload conditions. For example, if it is known that a user enters a car at 7:30 am and re-enters the car at 5 pm each day, and each door slap causes an overload condition, this information may be used to enable active suppression of the low frequency portion of the feed forward signal path.

In some implementations, compensation in an ANR signal flow path (e.g., a feedforward path or a feedback path) corresponding to an acoustic transducer for one ear may be coordinated with compensation in a signal flow path corresponding to an acoustic transducer for the other ear. For example, if a user wears two earplugs of an earpiece at the same time, such coordination between corresponding signal flow paths may ensure that ANR performance in both ears is substantially similar. In some implementations, the side-chain filters of the signal flow path may be adjusted based on determining whether the user is wearing both earpieces of the headset. Sensors that may be used for this purpose include, for example, capacitive sensors or infrared sensors disposed on the ear buds or ear muffs to determine whether the ear buds or ear muffs are worn by the user.

The above discussion describes the overload problem of the acoustic transducer primarily with respect to the feed-forward path 110. The electro-acoustic transducer 106 may also be driven to an overload state due to the stimulus picked up by the feedback microphone 104. For example, in the case of an in-ear ANR earpiece that seals tightly in the ear canal, low frequency stimulation of lower jaw movement can produce large pressure changes that can be picked up by the feedback microphone 104. Fig. 4 is a graph showing the pressure change in the ear canal, detected by the feedback microphone, over a two second time period. In some cases, for a tightly sealed in-ear headphone, low frequency pressure variations (about 15Hz) may be produced when the user walks on a hard surface. Such high amplitude, low frequency pressure variations, when detected by the feedback microphone 104, may cause the feedback compensator 116 to generate a feedback compensation signal that drives the acoustic transducer to an overload state. This, in turn, may cause the acoustic transducer 106 to generate auditory artifacts and reduce the performance of the ANR device. Although the user is exemplified herein as walking on a hard surface (which produces low frequency variations of about 15Hz), other events may cause low frequency variations of different frequencies. The techniques described herein are generally applicable to events that cause low frequency variations that may result in overload conditions regardless of the particular frequency.

In some implementations, auditory artifacts generated due to low frequency pressure variations in the ear canal can be mitigated by using a tunable filter in the feedback compensator. Fig. 5A is a block diagram of an example of such a feedback compensator 500, which includes a tunable filter 502 configured to mitigate overload conditions due to low frequency stimuli picked up by the feedback microphone 104. Feedback compensator 500 also includes a fixed filter 504 configured to generate a feedback compensation signal for transducer 106. In some implementations, upon detection of high amplitude low frequency stimuli (e.g., due to mandibular motion or walking), tunable filter 502 may be configured to filter out such stimuli from the input signal provided to fixed filter 504 to generate the feedback compensation signal. According to such a signal flow scheme, the feedback compensator may continue to generate the feedback compensation signal for noise reduction without being driven to an overload state even in the presence of high amplitude low frequency stimuli.

The parameters of the tunable filter may be selected, for example, by a parameter selector module 508 that determines an appropriate set of parameters based on the output of the estimator 506. In some implementations, the estimator 506 determines from the feedback compensation signal generated by the fixed filter 504 whether the feedback compensation signal is likely to potentially drive the acoustic transducer 106 to an overload state. Based on the output of the estimator 506, the parameter selector module 508 may be configured to select one or more parameters (or a set of filter coefficients) for the tunable filter 502 such that the tunable filter 502 filters out stimuli that result in the generation of a larger feedback compensation signal. The parameter selector module 508 may be configured to access a look-up table based on the degree of driver displacement reported by the estimator 506 to select one or more parameters (or a set of filter coefficients) for the tunable filter 502. In some implementations, the estimator may be configured to monitor the output of the fixed filter 504 to reduce the likelihood that the output voltage exceeds a threshold condition related to, for example, driving the output transducer 106 to an unacceptably high displacement or clipping the electrical output.

In some implementations, the parameter selector module 508 may be configured to select one or more parameters or coefficients of the tunable filter 502 such that the tunable filter 502 acts as a high pass filter. Fig. 6A to 6C show the magnitude response of different tunable high-pass filters parameterized by parameter a. The transfer function of the filter corresponding to fig. 6A is given by:

the transfer function of the filter corresponding to fig. 6B is given by:

the transfer function of the filter corresponding to fig. 6C is given by:

for each transfer function described above, choosing a value for a equal to 1 results in an all-pass filter. However, upon detection of a low frequency stimulus that drives the acoustic transducer to an overload state, the value of a (or the resulting set of filter coefficients) may be selected according to the desired magnitude response that will filter out the low frequency stimulus.

In some implementations, the parameter selector module 508 may be configured to select one or more parameters or filter coefficients of the tunable filter 502 such that the tunable filter 502 acts as a notch filter. This may be useful, for example, when the pressure variations causing the overload condition are within a narrow frequency range. For example, when a user walks on a hard surface wearing a tightly sealed in-ear headphone, high amplitude pressure variations of about 15Hz may occur. In this case, a notch filter may be used to prevent such pressure variations from generating a feedback signal that may drive the acoustic transducer to overload. Because the use of notch filters only suppresses a narrow range of frequencies, such filters may only significantly degrade the feedback compensation performance of the ANR device.

While the description so far used examples using parallel compression in the feed-forward signal flow path (fig. 3B), and tunable filters in the feedback signal path (fig. 5A), each of these techniques may be used in other signal flow paths. For example, feedforward compression techniques may be used in feedback ANR signal flow paths, while tunable filters may be used in feedforward ANR signal flow paths. FIG. 5B illustrates a block diagram of another system that may be used to either feedforward or feedback ANR signal flow paths. In particular, fig. 5B is a block diagram of an example system 550 that uses a combination of a Variable Gain Amplifier (VGA)552 and a tunable filter 554 disposed in a signal flow path of an ANR device. The signal flow path including a sensor (e.g., microphone 557 and/or non-microphone sensor 555) at one end and acoustic transducer 106 at the other end may include a feedback path or a feedforward path of an ANR device, for example. Tunable filter 554 may be referred to as a feed-forward compensator if the signal flow path in which system 550 is disposed is a feed-forward path. Tunable filter 554 may be referred to as a feedback compensator if the signal flow path in which system 550 is disposed is a feedback path.

In some implementations, the noise reduction performance of an ANR device may be balanced against its overload performance by adaptively adjusting the VGA552 and the tunable filter 554 based on the environment of the ANR device. In some implementations, noise reduction performance can be improved by increasing the gain of VGA 552. For example, the ANR device may introduce system-generated noise (e.g., noise produced by electronics disposed in the signal flow path), which may appear as a substantially constant audible "hissing" generated by the acoustic transducer 106. In such cases, increasing the gain of VGA552 may improve the signal-to-noise ratio (SNR) in some cases and reduce the undesirable hissing sound generated by acoustic transducer 106. This may also be referred to as reducing the "noise floor" and improving the user experience, especially in low noise environments. However, the gain of pre-amplifying VGA552 may enhance any signal captured using microphone 557 (e.g., a feedback microphone and/or a feed-forward microphone), which may in some cases result in clipping of the input signal. For example, if the gain of the VGA552 is increased to reduce the noise floor, the dynamic range of the system may also be reduced, making the system (e.g., the electronics of the signal flow path and/or the acoustic transducer 106) more susceptible to overloading. In some cases, such overload conditions may cause the acoustic transducer 106 to produce audible pop and click sounds, which in turn may affect the improved user experience due to the reduced noise floor.

The signal flow paths shown in fig. 5B are examples of systems that may be used to balance noise reduction performance with overload performance of an ANR device. For example, in a quiet environment where the likelihood of low frequency interference (e.g., interference caused by low frequency pressure variations in the environment) is low, the gain of the VGA may be adjusted to a relatively high value to reduce the audible noise floor. This in turn may make it more likely that the ANR device is driven to an overload state. Thus, once the estimator 556 detects an overload condition, or when the ANR device moves into an environment where the likelihood of low frequency pressure changes is higher, the parameter selector 558 may be configured to adjust the gain of the VGA to a lower value to reduce the likelihood that the system is driven to an overload state. In some cases, if the VGA gain is reduced in a noisy environment, the psychoacoustic effects of increasing the noise floor may not have a significant impact on the user. However, the adaptive reduction in VGA552 gain may result in audible pop and click mitigation that may otherwise degrade the user experience due to the occurrence of overload conditions.

In some implementations, when the gain of VGA552 is adjusted to a particular level, the filter coefficients of tunable filter 554 are also adjusted accordingly to compensate for the gain variation of VGA 552. For example, if parameter selector 558 increases the gain of VGA552 by 6dB, parameter selector 558 may be further configured to select an appropriate set of filter coefficients for tunable filter 554 such that the magnitude response of the tunable filter is reduced by approximately 6dB to compensate for the increased gain of VGA 552. In some cases, such simultaneous adjustment of the VGA and the tunable filter ensures that the overall gain of the signal flow path is substantially constant and the user experience is substantially consistent.

VGA552 is configured to process signals captured by one or more sensors, such as microphone 557 and/or non-microphone sensor 555. The microphone 557 may be of various types, possibly depending on, for example, the signal flow path in which the system 550 is disposed. For example, if the system 550 is disposed in a feedforward ANR path, the microphone 557 may comprise a feedforward microphone of an ANR device, such as the microphone 102 described above. As another example, if the system 550 is disposed in a feedback ANR path, the microphone 557 may include a feedback microphone, such as the microphone 104 described above. The sensor 555 may also be of various types. In some implementations, the non-microphone sensor 555 may include, for example, a pressure sensor, an accelerometer, or a gyroscope. Such non-microphone sensors 555 may be used, for example, to detect pressure changes or activity that may prompt the VGA552 and/or the tunable filter 554 for a change in settings. For example, based on the output of an accelerometer disposed in the ANR earpiece, it may be determined that the user is running or jogging, which in turn may produce low frequency pressure changes at a particular frequency. Based on such a determination, the parameter selector 558 may be configured to adjust the gain associated with the VGA and the filter coefficients of the tunable filter 554. Although the system 550 shown in fig. 5B includes both a non-microphone sensor 555 and a microphone 557, systems that include only the microphone 557 or only the non-microphone sensor 555 are also possible.

In some implementations, the gain of VGA552 and the filter coefficients of tunable filter 554 may be adjusted based on predicting the beginning of a particular event. In some implementations, the user's environment may be determined based on an output of a Global Positioning System (GPS) (e.g., a global positioning system disposed in an ANR device or in a mobile phone connected to an ANR device), and settings of VGA552 and tunable filter 554 may be adjusted according to the determination. For example, if it is determined that the user of the ANR device is at a library or office, the parameter selector 558 may be configured to adjust settings of the ANR device according to circumstances typically used in quiet environments. Conversely, if it is determined that the user is on a train during the commute time, the parameter selector 558 may be configured to adjust settings of the ANR device according to circumstances typically used in noisy environments. In some implementations, the environment of the user may be detected based on one or more applications executing on the ANR device and/or a mobile device connected to the ANR device. For example, upon determining that the user has just started an application that tracks the number of steps the user runs, it may be inferred that the user is about to start running. Accordingly, the parameter selector 558 may be configured to adjust the VGA552 and the tunable filter 554 to account for corresponding expected low voltage variations in the ANR device. In some implementations, information regarding both the environment and activity of the user may be used to determine operating parameters of VGA552 and filter coefficients of the tunable filter.

Parameter selector 558 may be configured to select operating parameters of VGA552 and tunable filter 554 in various ways. In some implementations, the parameter selector can be configured to access a computer-readable storage device that stores a representation of a lookup table that stores different sets of filter coefficients for tunable filter 554 linked to different gain values of VGA 552. In some implementations, the parameter selector may be configured to calculate filter coefficients for tunable filter 554 based on a predefined relationship with the selected gain value. The gain values to be used in different environments may be determined or calculated empirically based on the output of one or more sensors, such as pressure sensors or microphones. In some implementations, the gain level of VGA552 may also be changed based on user input received via a user interface. The user interface may be a control such as a switch, knob, or dial provided on the ANR device, or a software-based graphical user interface displayed on a display device, such as a graphical user interface displayed on a connected mobile device.

Estimator 556 may be configured to determine whether any adjustments need to be made to VGA552 and/or tunable filter 554. Accordingly, estimator 556 may be configured to signal parameter selector 558 to adjust one or both of VGA552 and tunable filter 554. In some implementations, the estimator 556 is substantially similar to the estimator 506 described above with reference to fig. 5A. In some implementations, the estimator 556 can be a displacement estimator configured to estimate whether the drive signal for the acoustic transducer can potentially cause the transducer to exceed its excursion limit. In some implementations, the estimator 556 may comprise a pressure estimator configured to detect pressure disturbances in the environment.

The system 550 may operate in various modes. In some implementations, the system 550 may be configured in some implementations to run substantially continuously at initialization. For example, if the system 550 is disposed in a feedforward or feedback path of an ANR earpiece, the system may be initialized when the ANR function of the earpiece is activated and then allowed to run during operation of the earpiece. However, in some cases, such modes of operation may result in multiple pops and clicks, which may degrade the user experience to some extent. In some implementations, the system 550 can include controls (e.g., buttons) that deactivate/activate the system 550 based on user input. In some implementations, rather than performing continuous adjustments when active, system 550 can be configured such that parameter selector 558 adjusts VGA552 and tunable filter 554 according to the current environment and then turns off or enters a standby mode. The system may be reactivated based on user input indicating that the environment has changed or otherwise needs to be readjusted. In some implementations, the system 550 or the ANR device in which it is deployed may include one or more controls (e.g., hardware buttons and/or software controls presented on a user interface) for selecting an operating mode of the system 550.

In the mode of operation where the system 550 automatically adjusts the gain of the VGA552 and the filter coefficients of the tunable filter 554, the adjustment may be performed in various ways. In some implementations, the adjustment is performed substantially periodically. For example, the adjustment may be performed for a time period of about 100ms or longer. For example, the frequency of adjustment may be selected empirically to allow the system 550 to adjust appropriately to accommodate changing environments. In some implementations, the adjustment may be performed upon detecting a change in the environment. For example, if the estimator 556 detects a signal indicative of an environmental change (e.g., occurrence of a low frequency pressure event), the estimator may signal the parameter selector 558 to adjust the VGA552 and the tunable filter 554 accordingly.

In some implementations, to prevent the system 550 from adjusting too frequently, a decision threshold may be associated with the adjustment. In some implementations, the adjustment may be made only when the required amount of change in the gain of the VGA552 exceeds a threshold amount. For example, the adjustment can be made only when the gain adjustment is 2.25dB or more. The threshold amount may be determined empirically, for example, to prevent adjustment too frequently.

The adjustment of the gain level of the VGA552 may be performed in various ways. In some implementations, the adjustment can be made in a single step. In some implementations, the adjustment can be performed as a series of multiple steps. For example, if a 6dB adjustment is required, the adjustment may be made as a single step change of 6dB, or as a series of six steps, each implementing a 1dB change, or another combination of steps. The step size may be determined empirically, for example, based on tolerance to any associated auditory artifact generated by the step change. In some implementations, the time gap between steps may also be adjusted, for example, to reduce the likelihood of multiple auditory artifacts merging into a single larger auditory artifact. However, increasing the gap between these steps also increases the total adjustment time. Thus, the interval between steps may be selected empirically based on a target tradeoff between settling time and tolerable audible artifacts.

FIG. 7 is a block diagram of an example biquad notch filter that may be used as a tunable filter in an ANR signal flow path. The transfer function of the filter is given by:

in implementation, multiple biquad notch filters may be cascaded to achieve a desired level of rejection. Fig. 8A shows the magnitude and phase response of a single biquad notch filter (as shown in fig. 7) for different combinations of filter coefficients. The filter coefficients are parameterized by a parameter n. Specifically, curves 802, 804, 806, and 808 represent magnitude responses for parameter values n-1, n-2, n-3, and n-4, respectively.

Curves

810, 812, 814 and 816 represent the phase response for parameter values n-1, n-2, n-3 and n-4, respectively. Fig. 8B shows the variation of the poles and zeros of the biquad notch filter with respect to the tuning parameter n in frequency and biquad singularities Q. Q is also referred to as a quality factor. In fig. 8B, the curves depicting the pole and zero frequencies and Qs as a function of tuning change show that the notch clarity and notch center frequency are coupled. To maintain the desired feedback ANR performance, the notch depth increases with increasing frequency.

Fig. 8C and 8D show the variation of the filter coefficient values of the biquad notch filter with respect to the tuning parameter n. In the particular examples shown in fig. 8C and 8D, the coefficients b1 and a1 have values around 2, and the coefficients b2 and a2 have values around one. This is relevant for a particular implementation that uses a sampling rate of 384000 samples per second that is greater than the desired 15Hz notch frequency. In some implementations, the filter coefficient values (e.g., B1, B2, a1, and a2 in this example) may be stored in a lookup table or derived from a mapping rule such as the frequency/Q mapping shown in fig. 8B.

Fig. 9A to 9B and fig. 10A to 10B show the performance of the tunable filter in the feedback path. Specifically, fig. 9A and 9B show the loop gain of the feedback path (provided as the driver voltage versus feedback microphone voltage transfer function and the feedback compensator transfer function K) without and with a tunable filter, respectively_fbThe product of) the amplitude and phase response. The particular tunable filter used in this example includes twelve cascaded biquad notch filters, each substantially similar to the biquad notch filter shown in fig. 7.As shown in fig. 9B, the tunable filter remains stable and exhibits consistent loop gain behavior for various values of the tuning parameter n. Furthermore, as shown in fig. 10A and 10B, which illustrate the sensitivity of the feedback path of the example ANR device without a tunable filter (fig. 10A) and with a tunable filter (fig. 10B), the sensitivity of these filters also remains consistent for various values of tuning parameters.

Fig. 11 shows the variation of the output voltage of the feedback compensator 116 for various values of the tuning parameter n. Specifically, curves 1102, 1104, 1106, and 1108 represent the change in feedback compensator output for parameter values n-1, n-2, n-3, and n-4, respectively. As these curves show, the parameter values can be adjusted to achieve different levels of rejection around the desired 15Hz frequency without significantly affecting the feedback compensator output at other frequencies.

Fig. 12A shows the magnitude response of another example of a notch filter that may be used as a tunable filter in the feedback path. The notch filter is another biquad notch filter, such as the notch filter shown in fig. 7, but where the coefficients a1 and a2 remain constant. In this example, the frequencies of the complex pole and the zero are equal, and the Q of the zero is varied to change the notch depth, which results in only the coefficients b1 and b2 being varied. Such a change in the coefficient is shown in fig. 12B.

FIG. 13 is a flow diagram of an exemplary process 1300 for implementing parallel feed forward compression in accordance with the techniques described above. At least a portion of process 1300 may be implemented using one or more processing devices, such as the DSPs described in U.S. patents 8,073,150 and 8,073,151, which are incorporated by reference herein in their entirety. The operations of process 1300 include receiving an input signal representing audio captured by an ANR device, such as a feedforward microphone of an ANR earpiece (1302). In some implementations, the ANR device may be an in-ear earpiece, such as the earpiece described with reference to fig. 1. In some implementations, the ANR device may include, for example, a circumaural earpiece, an open earpiece, a hearing aid, or other personal acoustic device. In some implementations, the feed-forward microphone may be part of a microphone array.

The operations of process 1300 also include processing the first range of frequencies of the input signal to generate a first feedforward signal for an acoustic transducer of an ANR earpiece (1304). This may be accomplished using a first feedforward compensator disposed in the ANR device to generate the anti-noise signal to reduce or eliminate the noise signal picked up by the feedforward microphone. In some implementations, generating the first feed-forward signal includes: the input signal is processed by a first filter to generate a first filtered signal, and the first filtered signal is processed by a first feedforward compensator to generate a first feedforward signal. The first filter may be a high pass or band pass filter having a pass band including a first frequency range. The first feedforward signal represents an anti-noise signal configured to reduce a noise signal in the first filtered signal.

Process 1300 also includes processing a second frequency range of the input signal to generate a second feed-forward signal for the acoustic transducer (1306). This may be done, for example, by a second feedforward compensator arranged in parallel with the first feedforward compensator. In some implementations, the first frequency range includes frequencies higher than frequencies in the second frequency range. For example, the upper limit of the second frequency range may be substantially equal to 100Hz, while the lower limit of the first frequency range may be greater than or substantially equal to 100 Hz. In some implementations, the first frequency range may include at least a portion of the second frequency range. In some implementations, generating the second feed-forward signal includes: the input signal is processed by a second filter to generate a second filtered signal, and the second filtered signal is processed by a second feedforward compensator to generate a second feedforward signal. The second filter may have a passband that includes a second range of frequencies, and the second feedforward signal may represent the anti-noise signal configured to reduce the noise signal in the second filtered signal.

The operations of process 1300 also include detecting if the second feed-forward signal satisfies a threshold condition (1308). This may include, for example, determining whether a voltage level representative of the second feed-forward signal meets or exceeds a threshold value to indicate an overdrive condition of the electro-acoustic transducer. This may also include filtering the second feed-forward signal, for example, using a digital filter, and comparing the filtered second feed-forward signal to a value associated with a threshold condition. A set of coefficients of the digital filter may be selected based on an operating mode of the ANR earpiece.

The operations of process 1300 further include attenuating the second feed-forward signal in response to determining that the second feed-forward signal satisfies a threshold condition (1310). For example, if the second feedforward signal satisfies the threshold condition, it is determined that the second feedforward signal will drive the acoustic transducer or other portion of the associated electronic device to overload and the variable gain amplifier in the signal path of the second feedforward signal is adjusted accordingly to attenuate the second feedforward signal.

The operations of the process may also include generating a combined feedforward signal for the acoustic transducer, for example, by adding the second feedforward signal to the first feedforward signal or by adding the attenuated second feedforward signal to the first feedforward signal. The combined feed forward signal may then be used in part to drive the acoustic transducer.

All of the various signal topologies and filter designs described above can be implemented in the configurable digital signal processor described in the referenced patent. These topology and filter designs may also be implemented in analog circuits or in a combination of analog and digital circuits using conventional circuit design techniques, but the resulting products may be larger or less flexible than products implemented using integrated configurable digital signal processors.

Fig. 14 is a flow diagram of an example process 1400 for implementing a tunable filter in a feedback path of an ANR device in accordance with the techniques described above. At least a portion of process 1400 may be implemented using one or more processing devices, such as the DSPs described in U.S. patents 8,073,150 and 8,073,151, which are incorporated by reference herein in their entirety. The operations of process 1400 include receiving an input signal representing audio captured by an ANR device, such as a feedback microphone of an ANR earpiece (1402). In some implementations, the ANR device may be an in-ear earpiece, such as the earpiece described with reference to fig. 1. In some implementations, the ANR device may include, for example, a circumaural earpiece, an open earpiece, a hearing aid, or other personal acoustic device. In some implementations, the feedback microphone may be part of a microphone array.

The operations of process 1400 also include generating, by the feedback compensator and based on the input signal, a first signal (1404). In some implementations, the feedback compensator may be substantially similar to the fixed filter 504 described above with reference to fig. 5A. For example, the first signal may include an anti-noise signal generated in response to noise detected by the feedback microphone, where the anti-noise signal is configured to cancel or at least reduce the effects of the noise.

The operations of process 1400 include determining one or more characteristics of the first signal (1406). This may be accomplished, for example, using modules substantially similar to estimator 506 described above with reference to fig. 5A. The one or more characteristics may include a voltage level of the first signal indicating an amount of driver displacement or offset of the corresponding acoustic transducer 106. In some implementations, the one or more characteristics may be indicative of a frequency or range of frequencies of the input signal in which the underlying noise is detected.

The operations of process 1400 also include selecting a plurality of filter coefficients of a digital filter disposed in series with the feedback compensator based on one or more characteristics of the first signal (1408). The plurality of filter coefficients may be selected, for example, based on a target frequency response of the digital filter. For example, if one or more characteristics of the first signal indicate that the voltage level of the first signal can potentially drive the corresponding acoustic transducer to an overload state, and the underlying noise in the input signal is near 15Hz, then the coefficients may be selected to configure the digital filter as a high pass or notch filter to suppress or attenuate components of the input signal around 15 Hz. In some implementations, the plurality of coefficients may be selected by accessing a pre-stored look-up table that includes parameters or coefficient values for the digital filter for various combinations of one or more characteristics determined for the first signal. In some implementations, the digital filter is substantially similar to the tunable filter described above with reference to fig. 5A and 5B.

The digital filter may be disposed in the feedback path of the ANR device in series with the feedback compensator and before or after the feedback compensator. In some implementations, the digital filter may be integrated with the feedback compensator in the form of a combined set of coefficients. For example, referring to fig. 5A, tunable filter 502 and fixed filter 504 may be combined in the form of a unified filter that provides feedback compensation. The digital filter may be implemented in various forms, including, for example, as an Infinite Impulse Response (IIR) filter or a Finite Impulse Response (FIR) filter.

The operations of process 1400 also include generating a feedback compensation signal for an acoustic transducer of an ANR earpiece by processing the input signal using a plurality of filter coefficients of the digital filter (1410). In some implementations, once the input signal is processed by the digital filter having the selected coefficients, a portion of the input signal that results in the out-of-range feedback compensation signal may be attenuated, thereby preventing any potential overload condition in the acoustic transducer. In some implementations, this may improve the user experience by avoiding audible artifacts otherwise generated by such overload conditions.

FIG. 15 is a flow diagram of an example process 1500 for implementing a combination of a Variable Gain Amplifier (VGA) and a tunable filter in a signal flow path of an ANR device in accordance with the techniques described above. At least a portion of process 1500 may be implemented using one or more processing devices, such as the DSPs described in U.S. patents 8,073,150 and 8,073,151, which are incorporated by reference herein in their entirety. The operations of the process 1500 include receiving input signals captured by one or more sensors associated with an ANR device, such as an ANR earpiece (1502). In some implementations, the one or more sensors may include one or more of a feedforward microphone and a feedback microphone of the ANR earpiece. In some implementations, the one or more sensors include one or more of a pressure sensor, an accelerometer, and a displacement sensor configured to sense an offset of the electroacoustic transducer.

The operations of process 1500 also include determining one or more characteristics of the first portion of the input signal (1506). In some implementations, the one or more characteristics of the first portion of the input signal are indicative of a noise floor associated with an environment external to the ANR earpiece. In some implementations, determining one or more characteristics of the first portion of the input signal includes: the method includes processing a first portion of an input signal to generate a first output signal for an electroacoustic transducer of an ANR earpiece, and determining one or more characteristics of the first portion of the input signal based on one or more characteristics of the first output signal. For example, determining one or more characteristics of a first portion of the input signal based on one or more characteristics of the first output signal may include determining that the first output signal is clipped. In some implementations, the one or more characteristics of the first portion of the input signal indicate a likelihood that an output signal resulting from processing the first portion of the input signal by the ANR signal flow path will be clipped. In some implementations, determining the one or more characteristics of the first portion of the input signal includes determining a non-linear relationship between the first portion of the input signal and the first output signal. For example, the non-linear relationship may be embodied by an output signal that causes the acoustic transducer to generate an audible artifact, such as a pop or click sound.

The operations of the process 1500 also include automatically adjusting a gain of a Variable Gain Amplifier (VGA) disposed in the ANR signal flow path based on one or more characteristics of the first portion of the input signal (1508). In some implementations, the gain of the VGA is periodically adjusted during operation of the ANR earpiece. The period of time for adjustment may be empirically determined and may be, for example, at least about 100 ms. In some implementations, the gain of the VGA is adjusted in response to determining that one or more characteristics of the first portion of the input signal or the first output signal satisfy a threshold condition. The threshold condition may include, for example, a desired gain adjustment amount. For example, the gain of the VGA can be adjusted only when the required adjustment is at least 2.25 dB.

The operations of the process 1500 also include selecting a set of coefficients for a tunable digital filter disposed in the ANR signal flow path according to the gain of the VGA (1510). For example, if the gain of the VGA is adjusted by a certain amount (e.g., 5dB), a set of coefficients for the tunable digital filter may be selected such that the magnitude response of the filter caused by the selected coefficients compensates for the gain adjustment of the VGA. This may be done, for example, to keep the overall gain of the signal flow path substantially constant.

The operations of the process 1500 also include processing a second portion of the input signal in the ANR signal flow path using the adjusted gain and the selected set of coefficients to generate a second output signal for an electroacoustic transducer of the ANR earpiece (1512). In some implementations, the second output signal reduces the likelihood that the system 550 is driven to an overload state compared to the first output signal. Thus, the process 1500 may be used to mitigate overload conditions in the feedforward or feedback path of an ANR earpiece.

The functions described herein, or portions thereof, and various modifications thereof (hereinafter "functions"), may be implemented at least in part via a computer program product, e.g., a computer program tangibly embodied in an information carrier, e.g., in one or more non-transitory machine-readable media or storage devices, for execution by, or to control the operation of, one or more data processing apparatus, e.g., a programmable processor, a computer, multiple computers, and/or programmable logic components.

A computer program can be written in any form of programming language, including compiled or interpreted languages, and it can be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment. A computer program can be deployed to be executed on one computer or distributed across one site or multiple sites and on multiple computers that are interconnected by a network.

The acts associated with implementing all or part of the functionality may be performed by one or more programmable processors executing one or more computer programs to perform the functions of the calibration process. All or part of the functionality can be implemented as, special purpose logic circuitry, e.g., an FPGA and/or an ASIC (application-specific integrated circuit). In some implementations, at least a portion of the functions may also be performed on a floating point or fixed point Digital Signal Processor (DSP), such as a super harvard architecture single chip microcomputer (SHARC) developed by Analog devices inc.

Processors suitable for the execution of a computer program include, by way of example, both general and special purpose microprocessors, and any one or more processors of any kind of digital computer. Generally, a processor will receive instructions and data from a read-only memory or a random access memory or both. Components of a computer include a processor for executing instructions and one or more memory devices for storing instructions and data.

Other embodiments and applications not specifically described herein are also within the scope of the following claims. For example, parallel feedforward compensation may be combined with a tunable digital filter in the feedback path. As another example, a tunable digital filter in the feed-forward path may be combined with a parallel compensation scheme in the feedback path. In some implementations, various combinations of parallel compensation techniques, tunable filter techniques, and VGA techniques may be used in an ANR signal flow path (e.g., a feedback path or a feedforward path) of an ANR device. In some implementations, the ANR signal flow path may include a tunable digital filter and a shunt compensation scheme to attenuate control signals generated in particular portions of a range of frequencies.

Elements of different embodiments described herein may be combined to form other embodiments not specifically set forth above. Some elements may be removed from the structures described herein without adversely affecting their operation. In addition, various separate elements may be combined into one or more separate elements to perform the functions described herein.

Claims

1. A method, comprising:

receiving an input signal representing audio captured by a microphone of an Active Noise Reduction (ANR) earpiece;

processing, by a first compensator, a first range of frequencies of the input signal to generate a first signal for an acoustic transducer of the ANR earpiece;

processing a second frequency range of the input signal by a second compensator arranged in parallel with the first compensator to generate a second signal for the acoustic transducer, wherein the first frequency range comprises frequencies higher than frequencies in the second frequency range;

detecting, by one or more processing devices, that the second signal satisfies a threshold condition; and

in response to determining that the second signal satisfies the threshold condition, attenuate the second signal.

2. The method of claim 1, further comprising: generating a combined signal for the acoustic transducer by adding the second signal to the first signal or by adding the attenuated second signal to the first signal.

3. The method of claim 2, further comprising driving the acoustic transducer using the combined signal.

4. The method of claim 1, wherein an upper limit of the second frequency range is substantially equal to 100 Hz.

5. The method of claim 1, wherein the first frequency range comprises at least a portion of the second frequency range.

6. The method of claim 1, wherein detecting that the second signal satisfies the threshold condition comprises: determining that a voltage level representative of the second signal meets or exceeds a threshold to indicate an overload condition.

7. The method of claim 1, wherein detecting that the second signal satisfies the threshold condition comprises: filtering the second signal using a digital filter and comparing the filtered second signal to a value associated with the threshold condition.

8. The method of claim 7, wherein a set of coefficients of the digital filter is selectable based on an operating mode of the ANR earpiece.

9. The method of claim 1, wherein attenuating the second signal comprises adjusting a variable gain amplifier that processes the second signal.

10. The method of claim 1, wherein processing the first frequency range of the input signal to generate the first signal comprises:

processing the input signal by a first filter to generate a first filtered signal, the first filter having a passband that includes the first frequency range; and

processing, by the first compensator, the first filtered signal to generate the first signal, wherein the first signal represents an anti-noise signal configured to reduce a noise signal in the first filtered signal.

11. The method of claim 1, wherein processing the second frequency range of the input signal to generate the second signal comprises:

processing the input signal by a second filter to generate a second filtered signal, the second filter having a passband that includes the second frequency range; and

processing, by the second compensator, the second filtered signal to generate the second signal, wherein the second signal represents an anti-noise signal configured to reduce a noise signal in the second filtered signal.

12. The method of claim 1, wherein the input signal represents audio captured by a feedforward microphone of the ANR earpiece.

13. The method of claim 1, wherein each of the first compensator and the second compensator is a feedforward compensator disposed in a feedforward signal flow path of the ANR earpiece.

14. The method of claim 1, wherein each of the first compensator and the second compensator is a feedback compensator disposed in a feedback signal flow path of the ANR earpiece.

15. An Active Noise Reduction (ANR) device, comprising:

one or more sensors configured to generate an input signal indicative of an environment external to the ANR device;

a first compensator configured to process a first range of frequencies of the input signal to generate a first signal for an acoustic transducer of the ANR earpiece;

a second compensator disposed in parallel with the first compensator, the second compensator filter configured to process a second frequency range of the input signal to generate a second signal for the acoustic transducer, wherein the first frequency range includes frequencies higher than frequencies in the second frequency range; and

one or more processing devices configured to:

detecting that the second signal satisfies a threshold condition, an

Attenuating the second signal in response to determining that the second signal satisfies the threshold condition.

16. The ANR device of claim 15, further comprising an adder configured to generate a combined signal for the acoustic transducer by adding the second signal to the first signal or by adding the attenuated second signal to the first signal.

17. The ANR device of claim 15, wherein an upper limit of the second range of frequencies is substantially equal to 100 Hz.

18. The ANR device of claim 15, wherein detecting that the second signal satisfies the threshold condition comprises: determining that a voltage level representative of the second signal meets or exceeds a threshold to indicate an overload condition.

19. The ANR device of claim 17, further comprising:

a digital filter for filtering the second signal; and

a comparator configured to compare the filtered second signal to a value associated with the threshold condition.

20. The ANR device of claim 18, wherein a set of coefficients of the digital filter is selectable based on a mode of operation of the ANR device.

21. The ANR device of claim 15, further comprising a Variable Gain Amplifier (VGA) that processes the second signal, and the one or more processing devices are configured to adjust a gain of the VGA to attenuate the second signal.

22. The ANR device of claim 15, further comprising:

a first filter configured to process the input signal to generate a first filtered signal, the first filter having a passband that includes the first frequency range,

wherein the first filtered signal is processed by the first compensator to generate the first signal, the first signal representing an anti-noise signal configured to reduce an effect of ambient noise on the first filtered signal compared to the effect of the ambient noise on the input signal.

23. The ANR device of claim 15, further comprising:

a second filter configured to process the input signal to generate a second filtered signal, the second filter having a passband that includes the second frequency range,

wherein the second filtered signal is processed by the second compensator to generate the second signal, the second signal representing an anti-noise signal configured to reduce an effect of ambient noise on the second filtered signal compared to the effect of the ambient noise on the input signal.

24. The ANR device of claim 15, wherein each of the first and second compensators is a feedforward compensator disposed in a feedforward signal flow path of the ANR earpiece.

25. The ANR device of claim 15, wherein each of the first and second compensators is a feedback compensator disposed in a feedback signal flow path of the ANR earpiece.

26. One or more machine-readable storage devices having encoded thereon computer-readable instructions for causing one or more processing devices to perform operations comprising:

causing a first compensator to process a first range of frequencies of the input signal to generate a first signal for an acoustic transducer of the ANR earpiece;

causing a second compensator disposed in parallel with the first compensator to process a second frequency range of the input signal to generate a second signal for the acoustic transducer, wherein the first frequency range includes frequencies higher than frequencies in the second frequency range;

detecting that the second signal satisfies a threshold condition; and

27. A method, comprising:

receiving an input signal representing audio captured by one or more sensors of an Active Noise Reduction (ANR) earpiece;

generating, by a compensator disposed in an ANR signal flow path of the ANR earpiece and based on the input signal, a first signal;

determining one or more characteristics of the first signal;

selecting a plurality of filter coefficients for a digital filter disposed in series with the compensator in the ANR signal flow path according to a target frequency response of the digital filter based on the one or more characteristics of the first signal;

generating a feedback control signal for an electroacoustic transducer of the ANR earpiece by processing the input signal using the plurality of filter coefficients of the digital filter.

28. The method of claim 27, wherein the digital filter is disposed after the compensator in a signal flow path.

29. The method of claim 27, wherein the digital filter is disposed in a signal flow path before the compensator.

30. The method of claim 27, further comprising:

determining, based on the one or more characteristics, that a portion of the input signal within a particular frequency range is causing the first signal to trigger an overload condition in the electro-acoustic transducer; and

selecting the plurality of filter coefficients such that the selected filter coefficients configure the digital filter to attenuate the portion of the input signal within the particular frequency range.

31. The method of claim 27, wherein the one or more characteristics comprise a voltage level.

32. The method of claim 27, further comprising driving the electro-acoustic transducer using the feedback control signal.

33. The method of claim 27, wherein the digital filter is a high pass filter.

34. The method of claim 27, wherein the digital filter is a notch filter.

35. The method of claim 27, wherein the digital filter is an Infinite Impulse Response (IIR) filter.

36. The method of claim 27, wherein the ANR signal flow path comprises a feedforward path disposed between a feedforward microphone of the ANR earpiece and the electro-acoustic transducer.

37. The method of claim 27, wherein the ANR signal flow path comprises a feedback path disposed between a feedback microphone of the ANR earpiece and the electro-acoustic transducer.

38. An Active Noise Reduction (ANR) device, comprising:

a compensator disposed in an ANR signal flow path of the ANR device, the compensator configured to generate a first signal based on the input signal;

a tunable digital filter disposed in the ANR signal flow path in series with the compensator, wherein the tunable digital filter is configured to generate a control signal for an electroacoustic transducer of the ANR device; and

one or more processing devices configured to:

determining one or more characteristics of the first signal, an

Selecting a plurality of filter coefficients for the tunable digital filter according to a target frequency response of the tunable digital filter based on the one or more characteristics of the first signal.

39. The ANR device of claim 38, wherein the tunable digital filter is disposed after the compensator in a signal flow path.

40. The ANR device of claim 38, wherein the digital filter is disposed in a signal flow path before the compensator.

41. The ANR device of claim 38, wherein the one or more processing devices are further configured to:

selecting the plurality of filter coefficients such that the selected filter coefficients configure the tunable digital filter to attenuate the portion of the input signal within the particular frequency range.

42. The ANR device of claim 38, wherein the one or more characteristics comprise a voltage level.

43. The ANR device of claim 38, wherein the tunable digital filter is one of: a high pass filter or a notch filter.

44. The ANR device of claim 38, wherein the ANR signal flow path comprises a feedforward path disposed between a feedforward microphone and the electro-acoustic transducer of the ANR device.

45. The ANR device of claim 38, wherein the ANR signal flow path comprises a feedback path disposed between a feedback microphone of the ANR device and the electro-acoustic transducer.

46. One or more machine-readable storage devices having encoded thereon computer-readable instructions for causing one or more processing devices to perform operations comprising:

causing a compensator disposed in an ANR signal flow path of the ANR earpiece to generate a first signal based on the input signal;

determining one or more characteristics of the first signal;

selecting a plurality of filter coefficients for a digital filter disposed in series with the compensator in the ANR signal flow path according to a target frequency response of the digital filter based on the one or more characteristics of the first signal; and

generating a feedback control signal for an electroacoustic transducer of the ANR earpiece by causing the digital filter to process the input signal using the plurality of filter coefficients.

47. A method, comprising:

receiving input signals captured by one or more sensors associated with ANR headphones;

determining, by one or more processing devices in an ANR signal flow path, one or more characteristics of a first portion of the input signal;

automatically adjusting, by the one or more processing devices, a gain of a Variable Gain Amplifier (VGA) disposed in the ANR signal flow path based on the one or more characteristics of the first portion of the input signal;

selecting, by the one or more processing devices, a set of coefficients for a tunable digital filter disposed in the ANR signal flow path, wherein the set of coefficients is selected according to the gain of the VGA; and

processing a second portion of the input signal in the ANR signal flow path using the adjusted gain and the selected set of coefficients to generate a second output signal for an electroacoustic transducer of the ANR earpiece.

48. The method of claim 47, wherein determining the one or more characteristics of the first portion of the input signal comprises:

processing the first portion of the input signal to generate a first output signal for an electroacoustic transducer of the ANR earpiece; and

determining the one or more characteristics of the first portion of the input signal based on one or more characteristics of the first output signal.

49. The method of claim 47, wherein the one or more sensors comprise a feedforward microphone of the ANR earpiece.

50. The method of claim 47, wherein the one or more sensors comprise feedback microphones of the ANR headphones.

51. The method of claim 47, wherein the one or more sensors comprise one or more of a pressure sensor, an accelerometer, and a displacement sensor configured to sense an offset of the electroacoustic transducer.

52. The method of claim 47, wherein the ANR signal flow path comprises a feedforward path disposed between a feedforward microphone of the ANR earpiece and the electro-acoustic transducer.

53. The method of claim 47, wherein the ANR signal flow path comprises a feedback path disposed between a feedback microphone of the ANR earpiece and the electro-acoustic transducer.

54. The method of claim 47, further comprising disabling VGA gain adjustment in response to a first user input received by an input device of the ANR earpiece.

55. The method of claim 54, further comprising reactivating a VGA gain adjustment in response to a second user input received by the input device of the ANR earpiece.

56. The method of claim 47, wherein the gain of the VGA is adjusted periodically during operation of the ANR earpiece.

57. The method of claim 47, wherein the gain of the VGA is adjusted in response to determining that the one or more characteristics of the first portion of the input signal satisfy a threshold condition.

58. The method of claim 47, wherein the one or more characteristics of the first portion of the input signal are indicative of a noise floor of an environment external to the ANR headphones.

59. The method of claim 48, wherein determining the one or more characteristics of the first portion of the input signal based on one or more characteristics of the first output signal comprises determining that the first output signal is clipped.

60. The method of claim 47, wherein the one or more characteristics of the first portion of the input signal indicate a likelihood that an output signal resulting from processing the first portion of the input signal by the ANR signal flow path will be clipped.

61. An Active Noise Reduction (ANR) device, comprising:

a Variable Gain Amplifier (VGA) and a tunable digital filter disposed in an ANR signal flow path of the ANR device, wherein the ANR signal flow path is connected to an electroacoustic transducer of the ANR device; and

one or more processing devices configured to:

determining one or more characteristics of a first portion of the input signal,

automatically adjusting an gain of the VGA based on the one or more characteristics of the first portion of the input signal,

selecting a set of coefficients for the tunable digital filter, wherein the set of coefficients is selected according to the gain of the VGA.

62. The ANR device of claim 61, wherein determining the one or more characteristics of the first portion of the input signal comprises:

processing the first portion of the input signal to generate a first output signal for an electroacoustic transducer of the ANR device; and

63. The ANR device of claim 61, wherein the one or more sensors comprise one or more of: a feed-forward microphone, a feedback microphone, a pressure sensor, an accelerometer, and a displacement sensor configured to sense an offset of the electroacoustic transducer.

64. The ANR device of claim 61, wherein the ANR signal flow path comprises a feedforward path disposed between a feedforward microphone and the electro-acoustic transducer of the ANR device.

65. The ANR device of claim 61, wherein the ANR signal flow path comprises a feedback path disposed between a feedback microphone of the ANR device and the electro-acoustic transducer.

66. The ANR device of claim 61, further comprising an input device configured to receive a first user input, wherein the one or more processing devices are further configured to disable VGA gain adjustment in response to the first user input.

67. The ANR device of claim 66, wherein the input device is configured to receive a second user input, wherein the one or more processing devices are further configured to reactivate a VGA gain adjustment in response to the second user input.

68. The ANR device of claim 61, wherein the one or more processing devices are configured to periodically adjust the gain of the VGA during operation of the ANR device.

69. The ANR device of claim 61, wherein the one or more processing devices are configured to adjust the gain of the VGA in response to determining that the one or more characteristics of the first portion of the input signal satisfy a threshold condition.

70. The ANR device of claim 61, wherein the one or more characteristics of the first portion of the input signal are indicative of a noise floor of an environment external to the ANR device.

71. One or more machine-readable storage devices having encoded thereon computer-readable instructions for causing one or more processing devices to perform operations comprising:

receiving input signals captured by one or more sensors associated with an Active Noise Reduction (ANR) device;

determining one or more characteristics of a first portion of the input signal;

automatically adjusting a gain of a Variable Gain Amplifier (VGA) disposed in an ANR signal flow path of the ANR device based on the one or more characteristics of the first portion of the input signal;

selecting a set of coefficients for a tunable digital filter disposed in the ANR signal flow path, wherein the set of coefficients is selected according to the gain of the VGA; and