System and method for canceling road noise in a microphone signal

Abstract

The present disclosure relates to an audio system including: an accelerometer positioned to generate an accelerometer signal representative of road noise within a vehicle cabin; a microphone disposed within the vehicle cabin such that the microphone receives the road noise and generates a microphone signal having a road noise component; and a road noise canceller including a road noise cancellation filter configured to receive the accelerometer signal and the microphone signal and, based on the accelerometer signal, minimize the road noise component of the microphone signal to generate an estimated microphone signal.

Description

Background Art

[0001] The present disclosure relates generally to systems and methods for road noise cancellation in microphone signals, and more particularly to systems and methods for road noise cancellation in microphone signals based on accelerometer signals representing road noise in a vehicle cabin. Summary of the Invention

[0002] All examples and features mentioned below can be combined in any technically possible way.

[0003] According to one aspect, an audio system includes an accelerometer positioned to generate an accelerometer signal representative of road noise within a vehicle cabin; a microphone disposed within the vehicle cabin such that the microphone receives the road noise and generates a microphone signal having a road noise component; and a road noise canceller including a road noise cancellation filter configured to receive the accelerometer signal and the microphone signal and minimize the road noise component of the microphone signal based on the accelerometer signal to generate an estimated microphone signal.

[0004] In one example, the road noise cancellation filter is configured to provide an estimated road noise signal based on the accelerometer signal, wherein the road noise canceller is configured to subtract the estimated road noise signal from the microphone signal such that a road noise component of the microphone signal is minimized.

[0005] In one example, the road noise cancellation filter is a fixed filter.

[0006] In one example, the road noise cancellation filter is an adaptive filter configured to minimize error signals.

[0007] In one example, the audio system further includes an echo cancellation filter configured to minimize an echo component of the estimated microphone signal to generate a residual signal, the echo component being acoustically generated by at least one sound transducer disposed within the vehicle cabin.

[0008] In one example, the adaptive filter is included in a multi-channel adaptive filter, which also includes an echo cancellation filter, which is configured to minimize an echo component of the microphone signal, the echo component being formed by acoustic generation of at least one sound transducer disposed within a vehicle cabin.

[0009] In one example, the road noise cancellation filter is configured to receive the microphone signal and the accelerometer signal, the road noise cancellation filter being optimized to minimize a road noise component of the microphone signal based on the microphone signal and the accelerometer signal.

[0010] According to one aspect, a method for canceling road noise in a microphone signal includes receiving an accelerometer signal representing road noise within a vehicle cabin from an accelerometer; receiving a microphone signal having a road noise component from a microphone operatively positioned within the vehicle cabin; and minimizing the road noise component of the microphone signal based on the accelerometer signal using a road noise cancellation filter to produce an estimated microphone signal.

[0011] In one example, the minimizing step includes generating an estimated road noise signal using the road noise cancellation filter based on the accelerometer signal, and subtracting the estimated road noise signal from the microphone signal such that a road noise component of the microphone signal is minimized.

[0012] In one example, the road noise cancellation filter is a fixed filter.

[0013] In one example, the road noise cancellation filter is an adaptive filter, wherein coefficients of the adaptive filter are adapted according to the error signal.

[0014] In one example, the method further includes minimizing an echo component of the estimated microphone signal using an echo cancellation filter to produce a residual signal, the echo component being acoustically generated by at least one acoustic transducer disposed within the vehicle cabin.

[0015] In one example, the method further includes minimizing an echo component of the microphone signal generated by acoustic generation of at least one sound transducer disposed within the vehicle using an echo cancellation filter included in a multi-channel adaptive filter together with the adaptive filter.

[0016] In one example, the step of minimizing the road noise component of the microphone signal is performed based on both the accelerometer signal and the microphone signal.

[0017] According to another aspect, a non-transitory storage medium storing program code that, when executed by a processor, includes the steps of: receiving an accelerometer signal representing road noise within a vehicle cabin from an accelerometer; receiving a microphone signal having a road noise component from a microphone operatively positioned within the vehicle; and minimizing the road noise component of the microphone signal based on the accelerometer signal using a road noise cancellation filter to produce an estimated microphone signal.

[0018] In one example, the minimizing step includes generating an estimated road noise signal using the road noise cancellation filter based on the accelerometer signal, and subtracting the estimated road noise signal from the microphone signal such that a road noise component of the microphone signal is minimized.

[0019] In one example, the road noise cancellation filter is a fixed filter.

[0020] In one example, the road noise cancellation filter is an adaptive filter, wherein coefficients of the adaptive filter are adapted according to the error signal.

[0021] In one example, the program code further includes the step of minimizing an echo component of the estimated microphone signal using an echo cancellation filter to generate a residual signal, the echo component being acoustically generated by at least one sound transducer disposed within a vehicle cabin, wherein the error signal is the residual signal.

[0022] In one example, the program code further includes the step of minimizing an echo component of the microphone signal using an echo cancellation filter included in a multi-channel adaptive filter together with the adaptive filter, the echo component being acoustically generated by at least one sound transducer disposed in a vehicle cabin.

[0023] The details of one or more implementations are discussed in the accompanying drawings and the description below. Other features, objects, and advantages will be apparent from the description, drawings, and claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1A Depicted is a schematic diagram of an audio system including a road noise canceller for canceling a road noise component of a microphone signal, according to an example.

[0025] Figure 1B Depicted is a partial schematic diagram of an audio system including a road noise canceller for canceling a road noise component of a microphone signal, according to an example.

[0026] Figure 1C Depicted is a partial schematic diagram of an audio system including a road noise canceller for canceling a road noise component of a microphone signal, according to an example.

[0027] Figure 2 Depicted is a schematic diagram of an audio system including an adaptive road noise canceller for canceling a road noise component of a microphone signal, according to an example.

[0028] Figure 3 Depicted is a schematic diagram of an audio system including an adaptive road noise canceller for canceling a road noise component of a microphone signal combined with an echo canceller for canceling an echo component of the microphone signal, according to an example.

[0029] Figure 4Depicted is a schematic diagram of an audio system including an adaptive road noise canceller for canceling a road noise component of a microphone signal combined with an echo canceller for canceling an echo component of the microphone signal, according to an example. DETAILED DESCRIPTION

[0030] A hands-free phone system implemented in a vehicle typically includes a microphone positioned within the vehicle to receive the user's voice. The signal from the microphone is then typically routed to a mobile device. Because the microphone is located within the vehicle cabin, road noise generated by vibrations from the vehicle structure can be present in the microphone signal and detected. This road noise in the microphone signal can be heard by the user receiving the call and often degrades the quality of the call. Therefore, there is a need in the art for a method to minimize the presence of road noise in the microphone signal sent to a hands-free phone system.

[0031] Various examples described herein relate to systems and methods for minimizing road noise present in microphone signals by utilizing accelerometer signals representing road noise in the vehicle cabin. FIG1 shows an example of an audio system 100 typically implemented in a vehicle, comprising one or more acoustic transducers 102, one or more microphones 104, and an audio processing subsystem, such as a road noise canceller 106, an echo canceller 108, and a post-filter subsystem 110. The audio system 100 receives one or more content signals u(n) on one or more channels 112. The program content signal u(n) can be a single type of program content signal, such as a speech signal, presented as, for example, a left and right pair on multiple channels 112 (e.g., channels 112a and 112b). Alternatively or in combination, multiple types of program content signals u(n), such as speech, navigation, or music, can be presented separately on one or more channels 112. The program content signal u(n) may be an analog or digital signal and may be provided as a compressed and / or packetized stream, and additional information may be received as part of such a stream, such as instructions, commands, or parameters from another system for controlling and / or configuring additional processing such as the soundstage rendering 114, the road noise canceller 106, or other components.

[0032] The content signal is converted into an acoustic signal by one or more acoustic transducers 102. The acoustic transducers 102 may have additional processing components, such as a sound field rendering 114, which provides various processing, such as equalization and speaker routing, to drive the acoustic transducers 102 to generate an acoustic sound field based on the various content signals and sound field parameters. In one example, one or more acoustic transducers 102 may be positioned within a vehicle cabin, with each of the acoustic transducers 102 being located within a respective door of the vehicle and configured to project sound into the cabin. Alternatively or in addition, the acoustic transducers 102 may be positioned within headrests or elsewhere within the vehicle cabin.

[0033] The block diagrams shown in the figures such as Figures 1 to Figure 4 The example audio system 100 is a schematic diagram and does not necessarily illustrate individual hardware components. For example, in some examples, each of the road noise canceller 106, the echo canceller 108, the post-filter subsystem 110, the soundstage rendering 114, and other components, and / or any portion or combination of these, may be implemented in a set of circuits, such as a digital signal processor, a controller, or other logic circuitry, and may include instructions stored on a non-transitory storage medium for the circuitry to perform the functions described herein. In alternative examples, various portions or combinations of these may be distributed across various sets of circuits.

[0034] A microphone, such as microphone 104, may receive each of the following: an acoustic speech signal s(n) from a user, a noise signal v(n), an acoustic echo signal d(n), and other acoustic signals, such as background noise within a vehicle. Microphone 104 converts the acoustic signals into, for example, electrical signals and provides them to road noise canceller 106. Specifically, microphone 104 provides the speech signal s(n) when the user is speaking, provides the noise signal v(n) at least when the vehicle is moving, and provides the echo signal d(n) (i.e., the component of the combined signal formed by the acoustic generation of the acoustic transducer 102) when the acoustic transducer 102 is active, as a combined signal y mic A portion of (n) is provided to the road noise canceller 106. The acoustic road noise signal v(n) will include at least a component v related to the road noise. a (n) (i.e., the acoustic signal inside the vehicle cabin caused by the vibration of the vehicle structure or the vibration of the engine when the vehicle travels on the road or other surface) and wind noise v r (n) (i.e., the acoustic signal within the cabin formed by the air passing through the vehicle as the vehicle travels). (The parameter n in this disclosure represents a discrete-time signal.)

[0035] The road noise canceller 106 is used to try to extract the mic (n) Remove or make the road noise component v a (n) is minimized to provide a road noise cancellation signal y(n). In one example, the road noise canceller 106 generates an estimated road noise signal y(n) by processing the accelerometer signal a(n) received from, for example, one or more accelerometers 116 using a road noise cancellation filter 118. To remove the road noise component v a (n). In at least one example, the estimated road noise signal is based on road noise measured at one or more accelerometers 116 operatively positioned around the vehicle to measure road noise. is an estimate of the road noise present at microphone 104 .

[0036] As used herein, "accelerometer" should be understood to encompass any sensor suitable for detecting vibrations in the vehicle structure, resulting from the vehicle's travel across a road or other surface or from engine vibrations, which are transduced into sound within the vehicle cabin.

[0037] The combined signal y provided by the microphone 104 can then be mic (n) minus the estimated road noise signal So that the combined signal y mic (n) The road noise component v a (n) is minimized. Therefore, if the road noise cancellation filter 119 provides an estimated road noise signal , the road noise canceller 106 will generate a signal from the combined signal y provided by the microphone 104. mic (n) Remove the road noise component v a (n) performs well.

[0038] like Figure 1A As shown, the road noise cancellation filter 118 may be configured to apply a set of fixed coefficients to the accelerometer signal a(n) to generate an estimated road noise signal The road noise cancellation filter 118 can be thought of as applying a transfer function It is an estimate of the transfer function g(n) between the accelerometer 116 and the microphone 104, such that the accelerometer signal a(n) received at the road noise cancellation filter 118 is converted by the road noise cancellation filter 118 to an estimate of the road noise present at the microphone In such Figure 1A In the case of using multiple accelerometers 116, the estimated transfer function is It may represent an estimate of the sum of the transfer functions between each accelerometer 116 and the microphone 104. For example, the transfer function The transfer function between the accelerometer 116a and the microphone 104 may be The transfer function between the accelerometer 116L and the microphone 104 is The estimated road noise signal is subtracted from the combined signal output of microphone 104 Thus, a road noise cancellation signal microphone signal y(n) is generated. (It should be understood that the road noise cancellation signal y(n) may still include a road noise component; however, if operating properly, the road noise component v of the road noise cancellation signal y(n) is a (n) should be at least relative to the combined signal y mic(n) Minimized.)

[0039] Likewise, if the microphone 104 is a microphone array, such as e.g. Figure 1B As shown, the road noise cancellation filter 118 may estimate the sum of the transfer functions from each accelerometer 116 to each corresponding microphone 104. Thus, for example, the road noise cancellation filter 118 may roughly estimate the transfer function from accelerometer 116a to microphone 104a to microphone 104j. The process is repeated for each accelerometer, up to accelerometer 116L. In practice, because accelerometers 116 and microphones 104 are spatially distributed at different locations around the vehicle, the transfer function from each accelerometer to each corresponding microphone may vary, and thus may be envisioned as the transfer function between each accelerometer and each microphone. Alternatively, the estimated transfer function It can be thought of as a transfer function between each accelerometer 116 and an equivalent microphone (ie, a combined microphone including microphone 104 ), the properties of the equivalent microphone being determined by the spatial relationship of microphones 104 .

[0040] In implementation, the coefficients of the road noise cancellation filter 118 (and thus the estimated transfer function) may be determined empirically according to a suitable method (eg, combined signal processing). ) in order to minimize the road noise component of the road noise cancellation signal y(n). For example, a vehicle including both microphone 104 and accelerometer 116 can be driven on various road surfaces and the signals from both recorded. From this data, an estimated road noise signal can be generated. A set of optimized coefficients, when the combined signal y mic (n) is subtracted from the estimated road noise signal to make the combined signal y mic (n) The road noise component v a (n) Minimization.

[0041] like Figure 1B As shown, the microphone signal y output from the microphone 104 is mic1…micJ (n) can also be input to implement the transfer function The fixed microphone filter 120 can be configured to convert the microphone signal y mic1…micN (n) Combined into a single microphone signal y mic (n), and applying any other necessary or useful signal processing, such as projecting the microphone 104 to a position near the user's mouth. Insofar as such signal processing is applied by the microphone filter 120, the estimated transfer function may represent an estimated transfer function between each accelerometer 116 and the projected position of each microphone 104. Alternatively or additionally, the microphone filter 120 may steer the beam toward the source of the desired acoustic signal and / or away from noise sources, and may additionally or alternatively steer the null toward the noise source.

[0042] In practice, when the microphone filter 120 is used, the coefficients of the road noise cancellation filter 118 may be determined empirically in the same manner as described above, such that y mic The road noise component of (n) is minimized to produce the road noise cancellation signal y(n). Figure 1B A microphone filter 120 is shown, but it should be understood that similar microphone filters may be implemented with any example including the microphones described herein.

[0043] like Figure 1C As shown, in an alternative example, the road noise canceller 106 may be implemented as a filter configured to receive the combined signal y from the microphone 104. mic (n) and receives the road noise signal a(n) from the accelerometer 116 and implements the estimated transfer function The estimated transfer function is optimized based on the relationship between the accelerometer 116 and the microphone 104 to minimize the road noise component of the road noise cancellation signal y(n). In this example, the road noise canceller 106 does not extract the road noise component from y(n). mic Instead of subtracting the estimated road noise signal from y(n), y(n) is generated directly using a road noise cancellation filter 118 using inputs from the microphone 104 and the accelerometer 116. The road noise cancellation filter 118 can be optimized empirically to achieve a combined signal y mic (n) The road noise component v a (n) as described in the example above. For example, a vehicle including both microphone 104 and accelerometer 116 can be driven on various road surfaces and the signals from both recorded. From this data, a set of optimized coefficients can be determined according to any suitable array processing method, which optimizes the combined signal y to be mic (n) The road noise component v a (n) Minimization.

[0044] Go to Figure 2 , shows an alternative example audio system 200 in which the road noise canceller 106 includes one or more adaptive road noise cancellation filters 118 that focus on satisfactory parameters that produce a sufficiently accurate estimate of the road noise signal according to an adaptive algorithm. Figure 1A and Figure 1BAs in the example of , the road noise cancellation filter 118 may apply a set of filter coefficients to the accelerometer signal a(n) to produce an estimated road noise signal The coefficients of the adaptive road noise cancellation filter 118 may be updated according to an adaptive algorithm to minimize the error signal (here shown as the road noise cancellation signal y(n)). Examples of adaptive algorithms that may be employed include, for example, a least mean square (LMS) algorithm, a normalized least mean square (NLMS) algorithm, a recursive least square (RLS) algorithm, or any combination or variation of these or other algorithms. The adaptive road noise cancellation filter 118, adjusted by the adaptive algorithm, converges to apply the estimated transfer function As described above, the estimated transfer function represents the transfer function g(n) between the accelerometer 116 and the microphone 104, such that the accelerometer signal a(n) received at the road noise cancellation filter 118 is converted by the road noise cancellation filter 118 to an estimate of the road noise present at the microphone.

[0045] like Figure 2 As shown, multiple adaptive road noise cancellation filters 118 may together form a multi-channel adaptive filter. Each constituent road noise cancellation filter 118 of the multi-channel adaptive filter is associated with (i.e., receives a signal from) a corresponding accelerometer 116. For example, adaptive road noise cancellation filter 118a is associated with accelerometer 116a and receives signal a1(n) therefrom and may apply a corresponding transfer function representing a transfer function between accelerometer 116a and microphone 104. Likewise, the remaining adaptive filter 118L may be associated with the accelerometer 116L and receive a signal a therefrom. L (n) and applying the corresponding transfer function between the corresponding accelerometer 116L and the microphone 104 The respective transfer function of each adaptive road noise cancellation filter 118 is adjusted to minimize the error signal, shown here as the road noise cancellation signal y(n). Accordingly, the output of each adaptive road noise cancellation filter 118 will be based on the signal received from the associated accelerometer 116 and the estimated transfer function of the adaptive road noise cancellation filter 118. to represent an estimate of the road noise at microphone 104. The outputs of the adaptive road noise cancellation filters 118 may be summed to produce an estimated road noise signal

[0046] In alternative embodiments, the residual signal e(n) (at the output of the echo canceller 108) or the estimated speech signal 18a than the accelerometer signal a1(n) received at adaptive filter 118L. L As the power of y(n) increases, the coefficients of adaptive road noise cancellation filter 118a will receive larger updates relative to the updates of the coefficients of adaptive road noise cancellation filter 118L. Thus, the channels that are most responsible for the errors observed in the road noise cancellation signal y(n) will receive the largest updates.

[0047] Generally, the adaptive algorithm updates the road noise cancellation filter 118 during times when the user is not speaking, but in some examples, the adaptive algorithm can be updated at any time. To this end, the double-ended detector 204 can detect when the user is speaking and instruct or otherwise cause the adaptive road noise cancellation filter 118 to stop updating.

[0048] Figure 1 and Figure 2 As shown, the road noise canceller 106 is implemented in conjunction with an echo canceller 108 and a post-filter subsystem 110 (the functionality and operation of which will be briefly described below). However, it should be understood that in various examples, the road noise canceller 106 may be implemented without one or both of the echo canceller or the post-filter (in the sense that these subsystems function independently of the road noise canceller 106), and FIG. Figure 2 The audio systems 100 , 200 are provided merely as examples of audio systems in which a road noise canceller may be implemented.

[0049] The echo canceller 108 is configured to attempt to remove the echo signal from the road noise cancellation signal y(n) to provide a residual signal e(n). The echo canceller 108 removes the echo signal by processing the program content signal u(n) on the channel 112 using one or more echo cancellation filters 124 to produce an estimated echo signal d(n) that is subtracted from the signal provided by the microphone 104. In various alternative embodiments, the output of the sound stage rendering 114, b(n), rather than the program content signal u(n), may be used as the reference signal for the echo canceller 108. In fact, any signal associated with at least one program content signal u(n) and suitable for minimizing the echo signal d(n) present in the road noise cancellation signal y(n) may be used as the reference signal for the echo canceller 108.

[0050] The echo canceller 108 may include an adaptive algorithm to periodically update the adaptive echo cancellation filter 124 to improve the estimated echo signal. Over time, the adaptive algorithm causes the adaptive echo cancellation filter 124 to converge on producing a sufficiently accurate estimate of the echo signal. , to minimize the error in the residual signal e(n). Generally, the adaptive algorithm updates the adaptive echo cancellation filter 124 during the time when the double-ended detector 204 detects that the user is not speaking, but in some examples, the adaptive algorithm can be updated at any time. When the user speaks, this is considered "double-ended" and the microphone 104 picks up both the acoustic echo signal d(n) and the speech signal s(n).

[0051] The adaptive echo cancellation filter 124 may apply a set of filter coefficients to the program content signal u(n) to produce an estimated echo signal The adaptive algorithm may use any of a variety of techniques to determine the filter coefficients and update or change the filter coefficients to improve the performance of the adaptive echo cancellation filter 124. Such adaptive algorithms, whether operating on the active filter or the background filter, may include, for example, a least mean square (LMS) algorithm, a normalized least mean square (NLMS) algorithm, a recursive least square (RLS) algorithm, or any combination or variation of these or other algorithms. As adjusted by the adaptive algorithm, the echo cancellation filter 124 converges to apply the estimated transfer function It represents the response of the echo path between the acoustic transducer 102 and the microphone 104 .

[0052] Generally speaking, as shown in Figures 1 and Figure 2 As shown, multiple echo cancellation filters 124 may together form a multi-channel adaptive echo cancellation filter, with each constituent echo cancellation filter 124 receiving an associated reference signal (e.g., program content signal u(n)). For example, adaptive echo cancellation filter 124a is associated with program content channel 112a and receives signal u1(n) therefrom and may apply a corresponding transfer function representing echo path h1(n) (and any additional processed responses, as will be described below). Similarly, the remaining adaptive echo cancellation filters 124M may each be associated with a program content channel 112M and receive a signal u therefrom. M (n), and apply the corresponding transfer function The respective transfer function of each adaptive echo cancellation filter 124 is adjusted to minimize the error signals, shown here as road noise and the echo cancellation residual signal e(n).

[0053] It will be appreciated that the number of adaptive echo cancellation filters 124 will generally depend on the number of reference signals received. Thus, if program content signals u(n) are used as reference signals, then M echo cancellation filters 124 may be implemented, each associated with one of the M program content signals u(n), while if soundstage rendering outputs b(n) are used, then N echo cancellation filters 124 may be implemented, each associated with one of the N soundstage rendering outputs b(n). It will also be appreciated that in some examples, fewer adaptive echo cancellation filters 124 may be used than, for example, program content signals u(n) or soundstage rendering outputs b(n). For example, fewer echo cancellation filters 124 may be used if certain program content signals u(n), such as a set of left bass, left grace notes, and left bird song program content signals u(n), are summed together and provided as reference signals to a single echo cancellation filter 124, or if only a subset of the reference signals are required to achieve effective echo cancellation.

[0054] In addition to estimating the echo path h(n), the estimated transfer function may represent an estimate of any processing provided between the location where the reference signal (eg, program content signal u(n)) is obtained and the echo canceller 108. Thus, if Figure 1A As shown, in the case where the reference signal is the program content signal u(n), in addition to the response of the echo path h(n), the estimated transfer function = would represent the responses of the soundstage rendering 114, the acoustic transducer 102, the microphone 104, and any processing associated with the microphone 104 (e.g., array processing). Thus, in combination with the responses and any processing performed at the microphone 104, the estimated transfer function is a representation of how the program content signal u(n) is converted from the form in which it is received to the echo signal d(n). However, if a reference signal is obtained at the output of the soundstage rendering 114, b(n), then the estimated transfer function , collectively represent the responses of the acoustic transducer 102, the echo path h(n), the microphone 104, and any processing associated with the microphone 104. Thus, although FIG. Figure 2 Depicts M estimated echo signals Instead of N estimated echo signals d(n), since the response of the sound field rendering 114 is included in the estimated transfer function So each estimated echo signal This would include processing the associated program content signal u(n) by the sound field rendering 114. Thus, the M estimated echo signals The sum of d(n) will estimate the sum of N echo signals d(n).

[0055] Although the echo canceller 108 typically cancels the linear aspects of the microphone signal y(n) associated with the program content channel, rapid variations and / or nonlinearities in the echo path prevent the echo canceller 108 from providing an accurate estimate of the echo signal, and residual echo thus remains in the residual signal e(n). Therefore, the post-filter subsystem 110 is used to suppress the residual echo component with spectral filtering to produce an improved estimated speech signal. Such post filters are generally known in the art, however a brief description of one example will be provided below.

[0056] As shown, the post-filter subsystem 110 may include a coefficient calculator 126 and a post-filter 128. In some examples, the post-filter 128 suppresses residual echo (from the echo canceller 108) in the residual signal e(n) by reducing the spectral content of the residual signal e(n) by an amount that is related to the possible difference in residual echo signal power relative to the total signal power (e.g., speech and residual echo). In one example, the post-filter 128 may compare each frequency bin (indicated by index "k") of the residual signal e(n) with the filter coefficient H pf(k) Multiplied together, the filter coefficients are calculated by coefficient calculator 126 according to the following example formula:

[0057]

[0058] where ΔH i (k) is the spectrum mismatch, S ee (k) is the power spectral density of the residual signal e(n), and is the power spectral density of the program content signal u(n) on the i-th content channel. The minimum multiplier H min is applied to each frequency bin, thereby ensuring that no frequency bin is multiplied by a value less than the minimum value. It will be appreciated that multiplying by a lower value is equivalent to greater attenuation. It should also be noted that in the example of formula (1), each frequency bin is at most multiplied by the unit element, but other examples may use different methods to calculate the filter coefficients. The β factor is a scaling or overestimation factor that can be used to adjust how strongly the post-filter subsystem 110 suppresses signal content, or in some examples can be effectively removed by being equal to the unit element. The p factor is a regularization factor used to avoid division by zero.

[0059] The spectral mismatch ΔH_i(k) represents the spectral mismatch between the echo path h(n) and the acoustic echo canceller 108. The spectral mismatch ΔH_i(k) can be calculated as the cross power spectral density of the residual error signal e(n) and the program content signal on the i-th content channel ui(n) The power spectral density of the program content signal u(n) on the i-th content channel This rate

[0060]

[0061] In some examples, the power spectral density used may be time-averaged or otherwise smoothed or low-pass filtered to prevent sudden changes (eg, rapid or significant changes) in the calculated spectral mismatch.

[0062] It should be understood that equations (1) and (2) as a whole relate to the case where the reference signals are uncorrelated. If the reference signals are not necessarily uncorrelated (e.g., the left and right channel pairs share some common content), the coefficient calculator 126 can calculate the filter coefficients H according to the following equations: pf(k) :

[0063]

[0064] where ΔH H denotes the Hermitian of ΔH, which is the complex conjugate transpose of ΔH, and where ΔH is given by:

[0065]

[0066] S uu is the matrix of the power spectral density and cross power spectral density of the program content channels. ΔH is the vector containing the spectral mismatch of all channels, and S ue is a vector containing the cross power spectral density of each reference channel with the error signal.

[0067] Although the above formulas have been provided for a post-filter subsystem 110 configured to suppress residual echo from multiple content channels, in an alternative example, the post-filter subsystem 110 may be configured to suppress residual echo from only one content channel.

[0068] In various examples, post-filter subsystem 110 can be configured to operate in frequency domain or time domain. Therefore, the use of term "filter coefficient" is not intended to limit post-filter subsystem 110 to operate in time domain. Term "filter coefficient" or other similar terms can refer to any value set applied to or incorporated into the filter to cause a desired response or desired transfer function. In certain examples, post-filter subsystem 110 can be a digital frequency domain filter that operates on the digital version of the estimated speech signal to multiply the signal content in a plurality of individual frequency intervals with different values ​​that are usually less than or equal to the unit element. A group of different values ​​can be considered as filter coefficients.

[0069] It should be understood that in various alternative examples, the road noise canceller 106 can be positioned to receive the estimated residual error signal e(n) rather than the combined signal from the microphone 104. That is, the road noise canceller 106 can be placed after the echo canceller 108 in the processing chain. This can improve the performance of the road noise canceller 106 because the echo signal will not be present or will be minimally present in the error signal used by the adaptive road noise cancellation filter 118 to adjust the filter coefficients.

[0070] In one example, the road noise canceller 106 and the echo canceller 108 may be sub-banded. That is, the road noise canceller 106 and the echo canceller 108 may be duplicated, with each duplicate associated with a specific frequency band. For each sub-band, the order of the road noise canceller 106 and the echo canceller 108 in the processing chain may be from the echo signal d(n) to the road noise component v a (n) is determined by the signal-to-noise ratio (SNR). For example, the combined signal y mic (n) can be filtered, for example, using a low-pass filter, to create a low-frequency sub-band, for example <400 Hz. In this frequency range, the road noise signal v a The power of the echo signal d(n) is usually higher than the power of the echo signal d(n) (i.e., the combined signal y mic (n) will typically have an SNR < 0 dB), accordingly, the road noise canceller 106 may be positioned before the echo canceller 108 in the processing chain (ie, in FIG. 1 and FIG. Figure 2 ). If the echo canceller is placed before the road noise canceller 106 in this frequency band, the road noise component v a (n) may dominate the erroneous signal received at the echo canceller 108, thereby preventing the echo canceller 108 from being properly adjusted.

[0071] Similarly, the combined signal y mic (n) can be filtered, for example, using a bandpass filter, to an intermediate range of, for example, 400 Hz to 1 kHz, where the echo signal d(n) will dominate the combined signal ymic (n) (i.e., the combined signal y mic (n) will typically have an SNR > 0 dB). In this frequency band, the echo canceller 108 may be positioned before the road noise canceller 106 in the processing chain. Otherwise, the combined signal y mic The power of the echo signal d(n) in (n) will prevent the road noise canceller 106 from being adjusted properly.

[0072] Finally, the combined signal y mic (n) can be filtered, for example, with a high-pass filter to a high frequency band, for example >1 kHz, where the echo signal d(n) will largely dominate the combined signal y mic (n) (i.e., the combined signal y mic (n) will typically have an SNR of >>0 dB). In this example, the road noise canceller 106 may be omitted entirely to avoid unnecessary processing.

[0073] It should be understood that the above frequency bands are provided merely as examples to illustrate the concept that the order of the road noise canceller 106 and the echo canceller 108 in the processing chain can be determined by the SNR of a particular frequency band. More specifically, for frequency bands where the SNR is typically <0 dB, the road noise canceller 106 can be positioned before the echo canceller 108. For frequency bands where the SNR is typically >0 dB, the road noise canceller 106 can be positioned after the echo canceller 108. And for frequency bands where the SNR is typically >>0 dB, the road noise canceller 106 can be omitted entirely.

[0074] As mentioned above, Figure 2 The adaptive filters 124, 118 of the road noise cancellation system 108 may update coefficients based on the power of the reference signals relative to the sum of the powers of each of the reference signals. However, because the road noise cancellation adaptive filter 118 is implemented separately from the adaptive filter of the echo canceller 108, the adaptive road noise cancellation filter 118 only compares the relative powers of the signals received from the accelerometer 116. However, errors present in the road noise cancellation signal y(n) may be partially attributed to the echo received at the microphone 104, which is still present in the road noise cancellation signal y(n). Therefore, it is advantageous to combine the adaptive echo cancellation filter 124 of the echo canceller with the adaptive road noise cancellation filter 118 of the road noise canceller into a combined multi-channel adaptive filter 302, as shown in FIG. Figure 3 The combined multi-channel adaptive filter 302 updates the coefficients of each signal relative to the total power of the reference signal (including the content program signal u(n) and the accelerometer signal a(n)). For example, Figure 3As shown, the program content signals u1(n), u2(n) through u2(n) may be considered when calculating the coefficients for the adaptive echo cancellation filters 124a, 124b through 124M and the road noise cancellation filters 118a through 118L. M (n) and accelerometer signal a1(n) until a L As described above, in one example, the size of the update to each adaptive filter can be proportional to the ratio of the power of the adaptive filter reference signal to the sum of the powers of all reference signals. Thus, for example, the size of the update to the adaptive road noise cancellation filter 118a can be proportional to the ratio of the power of the accelerometer signal a1(n) to the power of the program content signals u1(n), u2(n), and so on. M (n) and accelerometer signal a1(n) until a L Therefore, the summed output of the adaptive filter 302 will represent the estimated echo signal and the estimated road noise component The relative power of the accelerometer signal and the content channel is also considered during the update, resulting in more accurate attribution of errors to each program content channel 112 and accelerometer channel.

[0075] In addition to combining the multi-channel adaptive filter 302, Figure 3 The structure and components are largely combined with Figure 2 The functionality described is identical, so no additional explanation is needed.

[0076] like Figure 4 As shown, the post-filter subsystem can be further configured to receive the accelerometer signal a(n) as a reference signal to suppress residual road noise present in the residual signal e(n) in addition to the residual echo in the residual signal e(n). In conjunction with this disclosure, one of ordinary skill in the art will understand how to modify the post-filter subsystem 110 and the above formula to suppress residual road noise in the residual signal e(n).

[0077] In addition to the combined post-filter subsystem 110 being modified to suppress road noise in the residual signal e(n), Figure 4 The structure and components are largely combined with Figure 3 The functionality described is the same and therefore no additional explanation is required. However, it should be understood that in various alternative examples, the modified post-filter subsystem 110 can be included in an audio system that does not have the combined multi-channel adaptive filter 302. For example, the modified post-filter subsystem 110 can be included with both audio systems 100 and 200.

[0078] The road noise canceller 106, the echo canceller 108, and the post-filter subsystem 110 can be configured to calculate the adaptive filter coefficients and the post-filter subsystem 110 coefficients, respectively, only during periods when a double-ended condition is not detected (e.g., by the double-ended detector 204). As described above, when a user speaks within the acoustic environment of the audio system 100, 200, 300, 400, the combined microphone signal y mic (n) includes a component that is the user's voice. In this case, since the user is speaking, the combined signal y mic (n) not only represents the echo from the acoustic transducer 102, but the residual signal e(n) does not represent the residual echo, for example, a mismatch of the echo canceller 108 with respect to the actual echo path. Thus, the double-ended detector 204 is used to indicate that when a double-ended is detected, no new coefficients may be calculated during that time period, and the coefficients that were valid at the beginning of the user's speech or just before the user spoke may be used when the user speaks. The double-ended detector 204 may be any suitable system, component, algorithm, or combination thereof.

[0079] The output of the audio system 100, 200, 300, 400 or any variation thereof (e.g., the estimated speech signal ) is provided to another subsystem or device for use in various applications and / or processing. In fact, the output of the audio system 100, 200, 300, 400 can be provided for any application in which a noise-reduced speech signal is useful, including, for example, telephone communications (e.g., providing output to a far-end recipient via a cellular connection), virtual personal assistants, speech-to-text applications, speech recognition (e.g., identity recognition), or audio recording.

[0080] It should be understood that in this disclosure, capital letters used as identifiers or subscripts represent any number of structures or signals using the subscript or identifier. Thus, channel 112M represents the concept that any number of channels 112 can be implemented in various examples. In fact, in some examples, only one channel 112 can be implemented for a program content signal. Similarly, program content signal u M (n) represents the concept that any number of program content signals may be used. Insofar as different letters are used as subscripts, it is generally understood that the number of those signals and structures may be different from other structures with different letters. Thus, there may be different numbers of sound field rendering outputs b N (n) and program content signal u M (n). However, it will be appreciated that in some examples, the same number of soundstage rendering outputs b may be used. N (n) and program content channel u M (n). Finally, it should be understood that the different signals or structures (e.g., program content signals uM (n) and the estimated echo signal ) represent the general case where the same number of specific signals or structures are present. Thus, in the general case, when the program content signal u(n) is used as the reference signal for the echo canceller, the same number of estimated echo signals will be present. and program content signal u M (n). However, this general case should not be considered limiting. In conjunction with the review of this disclosure, one of ordinary skill in the art will understand that in some examples, a different number of such signals or structures may be used. Thus, in some examples (e.g., where certain program content signals u(n) are summed together to form a single reference for a single echo cancellation filter 124), there may be a different number of estimated echo signals and program content signal u M (n).

[0081] The functions or parts thereof described herein, as well as various modifications thereof (hereinafter referred to as "functions") may be implemented at least in part via a computer program product, for example, a computer program tangibly embodied in an information carrier, such as one or more non-transitory machine-readable media or storage devices, for execution, or controlling the operation of one or more data processing apparatuses, such as a programmable processor, a computer, multiple computers and / or programmable logic components.

[0082] 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 on one computer or executed on multiple computers distributed at one site or multiple sites and interconnected by a network.

[0083] The actions 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 functionality of the calibration process. All or part of the functionality may be implemented as special-purpose logic circuitry, such as an FPGA and / or an ASIC (application-specific integrated circuit).

[0084] Processors suitable for executing a computer program include, for example, both general-purpose and special-purpose microprocessors, and any one or more processors of any type of digital computer. Generally, a processor receives 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.

[0085] Although several inventive embodiments have been described and illustrated herein, a person of ordinary skill in the art will readily conceive of a variety of other devices and / or structures for performing the functions described herein and / or obtaining one or more of the results and / or advantages described herein, and each of such variations and / or modifications is considered to be within the scope of the inventive embodiments described herein. More generally, a person of ordinary skill in the art will readily understand that all parameters, dimensions, materials, and configurations described herein are intended to be exemplary, and that actual parameters, dimensions, materials, and / or configurations will depend on one or more specific applications in which the teachings of the present invention are used. A person of ordinary skill in the art will recognize or be able to ascertain, using only routine experimentation, many equivalents to the specific inventive embodiments described herein. Therefore, it should be understood that the above embodiments are presented by way of example only, and that within the scope of the appended claims and their equivalents, inventive embodiments may be practiced in a manner other than that specifically described and claimed. The inventive embodiments of the present disclosure relate to each individual feature, system, article, material, and / or method described herein. In addition, any combination of two or more such features, systems, articles, materials, and / or methods is included within the scope of the invention of the present disclosure if such features, systems, articles, materials, and / or methods are not mutually inconsistent.

Claims

1. An audio system comprising: an accelerometer positioned to generate an accelerometer signal representative of road noise within the vehicle cabin; a microphone disposed in the vehicle cabin so that the microphone receives the road noise and generates a microphone signal having a road noise component; as well as a road noise canceller comprising a road noise cancellation filter configured to receive the accelerometer signal and the microphone signal and minimize the road noise component of the microphone signal based on the accelerometer signal to generate an estimated microphone signal, wherein the road noise cancellation filter is included in a multi-channel adaptive filter, the multi-channel adaptive filter further comprising an echo cancellation filter configured to minimize an echo component of the microphone signal, the echo component being acoustically generated by at least one acoustic transducer disposed within the vehicle cabin.

2. The audio system of claim 1 , wherein the road noise cancellation filter is configured to provide an estimated road noise signal based on the accelerometer signal, wherein the road noise canceller is configured to subtract the estimated road noise signal from the microphone signal such that the road noise component of the microphone signal is minimized. 3 . The audio system of claim 2 , wherein the road noise cancellation filter is an adaptive filter configured to minimize an error signal.

4. The audio system of claim 1 , wherein the multi-channel adaptive filter is configured to update each of its coefficients relative to a total power of all reference signals, the reference signals comprising all content program content signals received by the audio system on one or more channels and all accelerometer signals received at the road noise cancellation filter. 5 . The audio system of claim 1 , wherein the road noise cancellation filter is configured to receive the microphone signal and the accelerometer signal, the road noise cancellation filter being optimized to minimize the road noise component of the microphone signal based on the microphone signal and the accelerometer signal.

6. A method for canceling road noise in a microphone signal, the method comprising: receiving an accelerometer signal from the accelerometer representing road noise within the vehicle cabin; receiving the microphone signal having a road noise component from a microphone operatively positioned within the vehicle cabin; as well as Minimizing the road noise component of the microphone signal based on the accelerometer signal using a road noise cancellation filter to generate an estimated microphone signal, wherein the road noise cancellation filter is included in a multi-channel adaptive filter that also includes an echo cancellation filter configured to minimize an echo component of the microphone signal that is acoustically generated by at least one acoustic transducer disposed within the vehicle cabin.

7. The method according to claim 6, wherein the step of minimizing comprises: generating an estimated road noise signal using the road noise cancellation filter based on the accelerometer signal, The estimated road noise signal is subtracted from the microphone signal such that the road noise component of the microphone signal is minimized. 8 . The method of claim 7 , wherein the road noise cancellation filter is an adaptive filter, wherein a plurality of coefficients of the adaptive filter are adapted according to an error signal.

9. The method of claim 6 , wherein the multi-channel adaptive filter is configured to update each of its coefficients relative to a total power of all reference signals, the reference signals comprising all content program content signals received on one or more channels and all accelerometer signals received at the road noise cancellation filter.

10. The method of claim 6, wherein the step of minimizing the road noise component of the microphone signal is performed based on both the accelerometer signal and the microphone signal.

11. A non-transitory storage medium storing program code, wherein the program code, when executed by a processor, comprises the following steps: receiving an accelerometer signal from the accelerometer representing road noise within the vehicle cabin; receiving a microphone signal having a road noise component from a microphone operatively positioned within the vehicle; as well as Minimizing the road noise component of the microphone signal based on the accelerometer signal using a road noise cancellation filter to generate an estimated microphone signal, wherein the road noise cancellation filter is included in a multi-channel adaptive filter that also includes an echo cancellation filter configured to minimize an echo component of the microphone signal that is acoustically generated by at least one acoustic transducer disposed within the vehicle cabin.

12. The non-transitory storage medium storing program code according to claim 11, wherein the step of minimizing comprises: generating an estimated road noise signal using the road noise cancellation filter based on the accelerometer signal, The estimated road noise signal is subtracted from the microphone signal such that the road noise component of the microphone signal is minimized.

13. The non-transitory storage medium storing program code of claim 12, wherein the road noise cancellation filter is an adaptive filter, wherein a plurality of coefficients of the adaptive filter are adapted according to an error signal.

14. The non-transitory storage medium storing program code of claim 11 , wherein the multi-channel adaptive filter is configured to update each of its coefficients relative to a total power of all reference signals, the reference signals comprising all content program content signals received on one or more channels and all accelerometer signals received at the road noise cancellation filter.

Images