JP2006270368A - Howling canceler - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a howling canceler capable of corresponding to occurrence of howling due to change of a sound field environment, and also, capable of facilitating setting during initial installation. <P>SOLUTION: A signal generation part 9 outputs a measuring signal to an adder 5 and an automatic setting part 10. The measuring signal is fed back to a microphone 1 from a speaker 6 through a sound feedback path. The automatic setting part 10 measures direct-sound arrival time and attenuation of an impulse response on the basis of the impulse response of the measuring signal. The direct-sound arrival time and attenuation are measured by using the impulse response estimated with an adaptive filter 8. A delay amount of a delay circuit 7 is set from the measured direct-sound arrival time. The tap number of the adaptive filter 8 is set from the attenuation of the impulse response. By this, parameters optimum for the sound field environment where the howling canceler is installed can be set. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a howling canceller that is used in a loudspeaker system installed in a lecture hall, a hall, or the like and suppresses howling.

  In general, when a loudspeaker is used in a lecture hall, a hall, or the like, sound output from a speaker is input to a microphone again through an acoustic path having a certain transfer function. That is, a closed loop is formed by the path of microphone-amplifier-speaker-acoustic path-microphone. When the gain of the closed loop exceeds 1, the sound returned from the speaker to the microphone increases, and howling occurs.

  As a method for preventing this howling, there is a method in which an equalizer, a compressor (limiter), or the like is inserted into an amplification path from a microphone to a speaker and these are manually adjusted.

  In such a system, it is necessary to actually generate howling, manually detect the frequency at which howling is generated using a graphic equalizer or a parametric equalizer, and manually lower the volume level of the frequency. In such a method, not only adjustment time is required, but also skill is required for adjustment to suppress howling.

In order to solve the above problem, a method has been proposed in which a peak frequency at which howling occurs is automatically detected, an equalizer is adjusted to automatically reduce the volume of the peak frequency when howling occurs, and a compressor is applied. . (For example, see Patent Document 1)
On the other hand, in order to efficiently prevent howling, a howling canceller that uses an adaptive filter (adaptive digital filter) to prevent howling has been proposed (for example, see Non-Patent Document 1).

  FIG. 8 shows the above-described howling canceller. The microphone 101 and the speaker 104 are installed in the same acoustic space such as a lecture hall or a hall. Here, the audio signal input from the microphone 101 is amplified by a front-end microphone amplifier and then converted into a digital signal y (k) by an A / D converter.

  The signal y (k) is supplied to the amplifier 103 via the adder 102 and amplified. G (z) is a transfer function of the amplifier 103. The signal x (k) output from the amplifier 103 is sounded from the speaker 104 after being converted into an audio signal by the D / A converter.

  The sound generated from the speaker 104 returns to the microphone 101 via the acoustic feedback path 105. The acoustic return path 105 is an acoustic path from the speaker 104 to the microphone 101. H (z) is a transfer function of the acoustic feedback path 105. A feedback signal d (k) fed back through the acoustic feedback path 105 is input to the microphone 101 together with a sound source signal s (k) generated by a sound source such as a speaker. The microphone 101 converts the input voice into a digital signal and outputs it as y (k).

  In such a loudspeaker, a closed loop is formed by the path of the microphone 101 -the amplifier 103 -the speaker 104 -the acoustic feedback path 105 -the microphone 101. When the gain of the closed loop exceeds 1, the feedback signal d (k) is increased and howling occurs. The loudspeaker shown in FIG. 1 has a howling canceller including a delay circuit 106, an adaptive filter 107, and an adder 102 in order to prevent the occurrence of such howling.

  The delay circuit 106 gives a delay time τ corresponding to the time delay of the acoustic feedback path 105 to the output signal x (k) of the amplifier 103 and outputs it to the adaptive filter 107 as a signal x (k−τ). The adaptive filter 107 has a filter unit 107a and a filter coefficient estimation unit 107b as shown in FIG. 9, and the signal x (k−τ) is input to both the filter unit 107a and the filter coefficient estimation unit 107b. .

  In the filter unit 107a, the filter coefficient is set so that the signal input from the microphone 101 is attenuated by the transfer function F (z) simulating the transfer function H (z) of the acoustic feedback path 105. Therefore, the signal do (k) output from the adaptive filter 107 is a signal obtained by filtering the signal x (k−τ) with the transfer function F (z) simulating the transfer function H (z) of the acoustic feedback path 105. Therefore, the feedback signal d (k) transmitted from the speaker 104 through the acoustic feedback path 105 and re-input to the microphone 101 is simulated.

  The adder 102 is a signal do (k) simulating the feedback signal d (k) from the signal y (k) input from the microphone 101 (where y (k) is a signal obtained by adding the sound source signal and the feedback signal). ) Is subtracted to generate a signal e (k). Thereby, the feedback signal is removed from the input signal, and howling can be canceled.

The filter coefficient estimating unit 107b uses an adaptive algorithm so that the transfer function F (z) matches or approximates the transfer function H (z) based on the signals x (k−τ) and e (k). Are sequentially updated. As a result, a signal do (k) simulating the signal d (k) is obtained, and howling can be prevented.
JP-A-6-46490 Inazumi, Imai, Konishi: "Preventing howling in loudspeakers using the LMS algorithm", Proc. 417-418 (1991, 3)

  The howling prevention processor described in Patent Document 1 is to operate after measuring how often the howling is generated by manually increasing the gain at the time of installation or the like. Therefore, there is a problem that it is impossible to cope with howling that occurs in a band different from the frequency at the time of measurement due to a change in the positional relationship between the speaker and the microphone or a change in the installation environment (for example, a change in temperature and humidity).

  On the other hand, the howling canceller described in Non-Patent Document 1 adapts to changes in the positional relationship between speakers and microphones and changes in the installation environment. Etc. were set manually. Therefore, skill is required to accurately match the return time from the speaker to the microphone in the installation environment, and it has not been possible to cope with a large environmental change after setting.

  In view of the above circumstances, an object of the present invention is to provide a howling canceller that can cope with the occurrence of howling due to a change in sound field environment and facilitates the setting at the time of initial installation.

  The invention according to claim 1 is a howling canceller including an adaptive filter that estimates a transfer function of an acoustic feedback path from a speaker to a microphone and cancels a feedback audio signal from the input signal of the microphone. A delay circuit that delays the residual signal, which is a signal after canceling the feedback audio signal, for a predetermined time, a sound source unit that inputs the measurement signal to the speaker and the delay circuit, and a feedback audio signal of the measurement signal estimated by the adaptive filter And setting means for setting at least the delay amount of the delay circuit.

  In the present invention, a measurement signal is input to the delay circuit and the speaker. The measurement signal is, for example, white noise or pink noise. When a user or the like instructs the start of measurement, a measurement signal is generated from the speaker and is fed back to the microphone. The setting means measures the direct sound arrival time from the speaker to the microphone based on the impulse response of the measurement signal, and sets the delay amount of the delay circuit. An impulse response estimated by an adaptive filter is used. For example, the direct sound arrival time is determined by measuring the timing at which the power of the impulse response exceeds a predetermined threshold.

  The invention according to claim 2 is characterized in that, in the above invention, the setting means sets parameters of the adaptive filter based on a feedback audio signal of a measurement signal estimated by the adaptive filter.

  In the present invention, the setting means further measures the reverberation time of the acoustic feedback path based on the impulse response of the measurement signal, and sets the number of taps of the adaptive filter according to the reverberation time. As a result, the number of useless taps can be reduced and the amount of calculation can be suppressed. Further, the convergence parameter of the adaptive filter is set by measuring the ratio of the direct sound from the speaker to the microphone and the indirect sound (noise, etc.). When many indirect sound components are included, the convergence parameter is decreased so that the convergence of the adaptive filter becomes slow. This can prevent an increase in error.

  The invention described in claim 3 is characterized in that, in the above-mentioned invention, switch means for cutting off the residual signal is further provided.

  In the present invention, the signal is cut off at the time of measurement so that the measurement signal fed back to the microphone is not sounded again from the speaker. Thereby, howling can be prevented during measurement.

  As described above, according to the present invention, the measurement signal is input to the delay circuit and the speaker, and the delay circuit and the like are set based on the feedback signal to the microphone estimated by the adaptive filter. It is possible to set the most suitable parameter for the adaptive filter. Thereby, it is possible to cope with the occurrence of howling due to a change in the sound field environment with high accuracy and to facilitate the setting at the initial installation.

  DESCRIPTION OF EMBODIMENTS Hereinafter, a loudspeaker system with a built-in adaptive howling canceller according to an embodiment of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a block diagram of a loudspeaker built-in howling canceller according to an embodiment of the present invention. As shown in the figure, this loudspeaker system includes a microphone 1, an adder 2, a switch circuit 3, an amplifier 4, an adder 5, a speaker 6, a delay circuit 7, an adaptive filter 8, a signal generation unit 9, and an automatic setting unit. The microphone 1 and the speaker 6 are arranged in a lecture hall or a hall.

  The audio signal input to the microphone 1 is input to the adder 2 as a digital signal by A / D conversion processing. The adder 2 subtracts the output signal of the adaptive filter 8 from the input signal of the microphone 1 and outputs the result. The output signal of the adder 2 is input to the amplifier 4 through the switch circuit 3. The amplifier 4 amplifies the signal input from the adder 2. The signal amplified by the amplifier 4 is output to the speaker 6 through the adder 5. The signal transmitted to the speaker 6 is converted into an analog audio signal and emitted. Here, the sound generated from the speaker 6 is re-input to the microphone 1 as a feedback signal.

  In the howling canceller in this example, the audio signal input from the microphone 1 is propagated through the acoustic space in which the amplifier 4, the speaker 6, and the microphone 1 are installed by the delay circuit 7 and the adaptive filter 8 and is input to the microphone 1 again. This simulates the transmission characteristics of a series of voice transmission paths up to the point.

  The delay circuit 7 provides a time delay that estimates the time delay of the feedback signal that is fed back from the speaker 6 to the microphone 1. Specifically, the time delay of the delay circuit 7 is set in advance so that the signal output from the amplifier 4 matches the amount of time until it returns to the microphone 1 via the speaker 6. The accuracy of howling suppression increases as the set delay amount is closer to the actual time delay of the feedback signal. The signal output after being given a time delay by the delay circuit 7 is input to the adaptive filter 8.

  The adaptive filter 8 is a filter that simulates the transfer function of the acoustic feedback path, and is a finite length filter (FIR filter) that filters the signal delayed by the delay circuit 7. The filtered output signal is output to the adder 2 as a simulation signal. The adaptive filter 8 has the number of taps assuming the reverberation time of the acoustic space. Specifically, the number of taps of the adaptive filter 8 is set in advance so that the signal output from the speaker 6 matches the time for reverberation in the acoustic space. The accuracy of howling suppression increases as the set number of taps (tap length) is closer to the reverberation time in the actual acoustic space.

  As shown in FIG. 2, the adaptive filter 8 includes a filter unit 8a and a filter coefficient estimation unit 8b, and the delayed signals from the delay circuit 7 are input to the filter unit 8a and the filter coefficient estimation unit 8b, respectively. The filter unit 8 a filters the input signal and outputs it to the adder 2. The adder 2 subtracts the output signal of the filter unit 8a from the input signal of the microphone.

  The filter coefficient estimator 8b detects an erasure error of the feedback signal based on the past signal delayed by the delay circuit 7 and the current signal that is the output signal of the adder 2, and matches the simulated signal with the feedback signal. The transfer function of the filter unit 8a is automatically updated to approximate it.

  An adaptive algorithm is used to update the transfer function of the filter coefficient estimation unit 8b. As the adaptive algorithm, for example, a LMS (Least Mean Square) algorithm is used.

  The howling canceller in this embodiment performs a setting operation that is a mode in which the delay amount of the delay circuit 7 and the number of taps of the adaptive filter 8 can be changed.

  The signal generator 9 inputs the measurement signal to the adder 5 in the setting operation. The measurement signal is, for example, white noise or pink noise. The adder 5 inputs the measurement signal to the speaker 6 and the delay circuit 7.

  In the setting operation, the automatic setting unit 10 instructs the switch circuit 3 to be turned off, and the direct sound arrival time from the speaker 6 to the microphone 1 based on the transfer function estimated by the adaptive filter 8 or the reverberation in the acoustic space. Measure time etc. Specifically, the filter coefficient estimated by the adaptive filter 8 is measured, and the time calculated from the tap number where the power of the filter coefficient exceeds a predetermined threshold is regarded as the direct sound arrival time. Also, the reverberation time of the acoustic space is measured by examining the attenuation of the impulse response. Further, the delay amount of the delay circuit 7 and the number of taps of the adaptive filter 8 are set based on the measured direct sound arrival time, reverberation time, and the like. Also, the convergence parameter of the adaptive filter 8 is set.

  Note that turning off the switch circuit 3 during the setting operation ensures that no measurement signal is input to the speaker 6 again, so that no howling is generated, but this is the case without the switch circuit 3. However, the setting operation of this embodiment can be performed.

  If the delay amount of the delay circuit 7 is too large compared to the time delay of feedback from the speaker 6 to the microphone 1, a signal delayed more than the time delay of the actual feedback audio signal is input to the adaptive filter 8, and the feedback audio signal Therefore, it is difficult to suppress howling.

  The number of taps of the adaptive filter 8 corresponds to the reverberation time, that is, the attenuation of the impulse response. If the number of taps is larger than the actual reverberation time, an antiphase signal is added without contributing to the cancellation of the feedback speech signal. Sometimes, another signal is added. In addition, if the number of taps is large, the amount of calculation increases, and the processing of the adaptive filter 8 is burdened.

  The filter coefficient of the adaptive filter 8 is updated with a convergence parameter called a forgetting coefficient λ or a step size α. The forgetting factor λ is a coefficient to be multiplied by the filter coefficient so far and is set in the range of 0-1. If the forgetting factor λ is reduced, the filter factor up to that point is deleted and the update is promoted. The step size α is a coefficient representing the magnitude of the correction. When the step size α is increased, the corrected filter coefficient is used more and the update is promoted. Increasing the step size α in a state where many errors are included in the impulse response increases this error.

  The loudspeaker system with a built-in adaptive howling canceller according to the present invention is capable of stably preventing howling by the automatic setting unit 10 setting the above parameters to optimum values.

  Next, the operation of the sounding system with a howling canceller will be described in detail.

  FIG. 3 is a diagram showing the transmission characteristics of the sounding system with a built-in howling canceller according to the embodiment of the present invention. FIG. 2A is a diagram showing the transfer characteristics of a sounding system with a howling canceller during normal operation. As shown in the figure, the audio signal input via the microphone 1 is converted into a digital signal y (k) by A / D conversion processing and input to the amplifier 4 via the adder 2. The amplifier 4 is for amplifying the input signal y (k). G (z) is a transfer function of the amplifier 4.

  The signal x (k) output from the amplifier 4 is converted into an analog audio signal by D / A conversion processing, and a sound is produced from the speaker 6. The sound generated from the speaker 6 returns to the microphone 1 through the acoustic return path 11. The acoustic return path 11 is an acoustic path from the speaker 6 to the microphone 1. H (z) is a transfer function of the acoustic feedback path 11. A feedback signal d (k) fed back via the acoustic feedback path 11 is input to the microphone 1 together with a sound source signal s (k) generated by a sound source such as a speaker. The microphone 1 converts the input voice signal into a digital signal and outputs it as y (k).

  The signal x (k) output from the amplifier 4 is also input to the delay circuit 7. The delay circuit 7 gives a time delay to the input signal x (k) and outputs it. Here, the delay circuit 7 calculates a time delay τ that estimates the time delay of the feedback audio signal fed back from the speaker 6 to the microphone 1. It is given. The signal x (k−τ) output with the time delay τ added by the delay circuit 7 is input to the adaptive filter 8.

As shown in FIG. 2, the adaptive filter 8 includes a filter unit 8a and a filter coefficient estimation unit 8b. The filter unit 8a and the filter coefficient estimation unit 8b each receive a signal x (k− τ) is input. The filter unit 8a outputs a signal do (k) simulating the feedback signal d (k) from the speaker 6 to the microphone 1, and the feedback signal d from the signal y (k) re-input from the microphone 1 by the adder 2. A signal do (k) simulating (k) is subtracted. A signal do (k) simulating the feedback signal d (k) is determined based on the signal x (k−τ) output from the delay circuit 7 according to the transfer function F (z). The filter coefficient estimation unit 8b uses the signal x (k−τ) output from the delay circuit 7 and the signal do (k) simulating the feedback signal d (k) among the signals transmitted from the microphone 1 to the amplifier 4. Based on the subtracted signal e (k), an adaptive algorithm is used to filter the filter coefficient of the filter unit 8a so that the signal do (k) simulating the feedback signal matches or approximates the actual feedback signal d (k). Is to be updated. As the adaptive algorithm, for example, an LMS algorithm is used. If the root mean square value of the signal e (k) is J = E [e (k) ² ] (where E [•] is an expected value), a filter coefficient that minimizes J is estimated by calculation. The filter coefficient of the filter unit 8a is updated using the estimated filter coefficient.

  As described above, the adaptive filter 8 simulates the feedback signal d (k) among the signal x (k−τ) output from the delay circuit 7 and the signal y (k) transmitted from the microphone 1 to the amplifier 4. Since the filter coefficient is updated based on the signal e (k) obtained by subtracting the signal do (k), it is possible to cancel the signal d (k) among the signals input from the microphone 1. When the gain of the closed loop formed by the microphone 1 -amplifier 4 -speaker 6 -acoustic feedback path 11 -microphone 1 path exceeds 1, the sound pressure value of a certain frequency in the feedback signal d (k) passes over time. It increases along with this to form a peak on the frequency spectrum, and howling occurs. In particular, the adaptive filter 8 updates the filter coefficient so as to attenuate this peak, and uses the generated simulated signal do (k) as the input signal y. Subtract from (k). Accordingly, when the signal y (k) obtained by adding the sound source signal s (k) input from the microphone and the feedback signal d (k) is a continuous sound having a constant frequency, the filter coefficient is sequentially updated to obtain the continuous sound. Can be attenuated.

  FIG. 5B is a diagram showing the transmission characteristics of the sounding system with a howling canceller during the setting operation. The setting operation is performed when a user or the like of the sounding system with a built-in howling canceller instructs setting.

  The signal generator 9 outputs the measurement signal n (k) to the adder 5. The adder 5 outputs the measurement signal n (k) to the speaker 6 and the delay circuit 7. The measurement signal n (k) is converted into an analog audio signal by D / A conversion processing, and a sound is produced from the speaker 6. The sound generated from the speaker 6 returns to the microphone 1 through the acoustic return path 11. A feedback signal d1 (k) fed back via the acoustic feedback path 11 is input to the microphone 1 together with a signal s1 (k) that is an indirect sound (noise or the like) from the outside. The microphone 1 converts the input voice signal into a digital signal and outputs it as y1 (k).

  The signal n (k) output from the signal generator 9 is also input to the delay circuit 7. The delay circuit 7 gives a time delay τ to the input signal n (k). The signal n (k−τ) output with the time delay τ added by the delay circuit 7 is input to the adaptive filter 8.

  The adaptive filter 8 outputs the signal d2 (k) simulating the feedback signal d1 (k) from the speaker 6 to the microphone 1 as described above, and the signal y1 (k) input from the microphone 1 by the adder 2 Is subtracted from the signal d2 (k) simulating the feedback signal d1 (k). A signal d2 (k) simulating the feedback signal d1 (k) is determined based on the signal n (k−τ) output from the delay circuit 7 according to the transfer function F1 (z). The filter coefficient estimating unit 8b performs an adaptive algorithm based on the signal n (k−τ) output from the delay circuit 7 and the signal e1 (k) obtained by subtracting the signal d2 (k) from the signal y1 (k). Used to update the filter coefficient of the filter unit 8a so that the signal d2 (k) simulating the feedback signal matches or approximates the actual feedback signal d1 (k).

  The automatic setting unit 10 refers to the filter coefficient estimation unit 8b of the adaptive filter 8, and measures the filter coefficient of the adaptive filter 8, here, the impulse response estimated by the filter coefficient estimation unit 8b. Various parameters are set based on the estimated impulse response.

  Next, the operation of the automatic setting unit 10 will be described in detail.

  FIG. 4 is a flowchart showing switching between the setting operation and the normal operation. This operation starts when the power is turned on. As shown in the figure, it is first determined whether or not to set (s1). When the user of the loudspeaker system with built-in adaptive howling canceller instructs the setting when the power is turned on, the operation proceeds to the setting operation (s2). If the user does not instruct anything, the past settings are read out. If no past setting has been made, the default setting of the howling canceller is read (s3). These settings are stored in a howling canceller memory (not shown) or the like. After finishing the setting mode or reading the setting, normal operation is started (s4).

  FIG. 5 is a flowchart showing the setting operation. As shown in the figure, when the setting operation is started, first, all parameters of the delay circuit 7 and the adaptive filter 8 are reset (s5). The parameters to be reset are the delay amount of the delay circuit 7, the number of taps of the adaptive filter 8, and the convergence parameter. When reset, the delay amount of the delay circuit 7 is set to zero. The number of taps and the convergence parameter are set to the maximum values that the adaptive filter 8 can set. Of the convergence parameters, the forgetting factor λ may be set to 1 which is the maximum value. The step size α is set to a maximum value according to the adaptive algorithm used.

  After resetting the parameters, the switch circuit 3 is turned off and the measurement signal is output from the signal generator 9 (s6). The measurement signal is white noise or pink noise. If the switch circuit 3 is not provided, the process of turning off the switch circuit is not performed.

  After that, as described above, the adaptive filter 8 estimates the impulse response based on the residual signal after subtracting the simulation signal from the output signal of the delay circuit 7 and the input signal of the microphone 1, and reflects it in the filter coefficient. This is measured (s7). The power of the measured filter coefficient is calculated, and it is determined whether or not the peak power that maximizes the power exceeds a predetermined threshold value (threshold value a1) (s8). That is, it is determined whether or not the measurement signal generated from the speaker 6 is fed back to the microphone 1 based on whether or not the power of the filter coefficient exceeds a predetermined threshold value. For example, if the delay amount of the delay circuit 7 is 0, immediately after the measurement signal is input to the delay circuit 7 and the speaker 6, the measurement signal is input to the filter coefficient estimation unit 8 b of the adaptive filter 8, and the feedback from the microphone 1. Since no signal is input, the power of the filter coefficient is reduced. Note that the absolute value of each tap coefficient may be calculated, and the determination may be made based on whether or not the maximum value of the tap coefficient absolute values exceeds a predetermined threshold value (threshold value a2).

  When it is determined that the peak power is equal to or less than a predetermined threshold, it is determined whether or not the elapsed time after the start of measurement is equal to or greater than a predetermined time (s9), and the power of the filter coefficient is calculated until the time is equal to or greater than the predetermined time (s9 → s7). The predetermined time is set to a constant multiple (0.5 times, etc.) of the number of taps of the adaptive filter 8, for example. When the elapsed time exceeds the predetermined time, the delay amount of the delay circuit 7 is changed (s10), and the process is repeated from the measurement of the filter coefficient (s10 → s7).

  When the power of the measured filter coefficient exceeds the threshold value a1, the time for the voice signal to arrive at the microphone 1 is calculated from the tap indicating the power and the sampling frequency, and the time is regarded as the direct sound arrival time (s11). Actually, even if a reflected sound in the acoustic space is coming instead of a true direct sound coming from the speaker 6, this time is set as the direct sound arrival time in the present embodiment.

  If the power of the filter coefficient does not exceed the threshold value a1 even if the delay amount of the delay circuit 7 is increased because the positions of the microphone 1 and the speaker 6 are too far, the setting operation is stopped and the maximum delay of the delay circuit 7 is reached. Set the amount and go to normal operation.

  After measuring the direct sound arrival time, the delay amount of the delay circuit 7 is set based on the direct sound arrival time (s12). At this time, a time obtained by subtracting a predetermined correction time from the direct sound arrival time is set as the delay amount. For example, when the direct sound arrival time is 100 msec, the correction time is set to about 40 msec, and the delay amount is set to 60 msec. This is to provide a margin for the delay amount so that it can cope with a change in the positional relationship between the microphone 1 and the speaker 6 due to movement of a speaker or the like during normal operation.

Thereafter, the filter coefficient is measured again to check the reverberation time, that is, the impulse response (s13). The attenuation of the impulse response is expressed as follows using the square integral value S _f (m) of the impulse response.

The number of taps is set based on the square integral value S _f (m) of the impulse response (s14). The settings of the square integral value S _f (m) of the impulse response and the number of taps are determined by the conditions of normal operation. For example, when it is desired that the reverberation component can be reliably erased, the number of taps until the impulse response is sufficiently attenuated is set. That is, it is assumed that the square integral value S _f (m) is sufficiently small, for example, is attenuated by 20 dB. For example, when a high-speed response is required for the adaptive filter, the number of taps is set so as to include the periphery where the impulse response is detected. Attenuation should be 10 dB. These can be freely set by the user or the like. For example, a setting button or the like may be installed so that the user or the like can select one of the conditions.

  Further, the convergence parameter of the adaptive filter 8 may be set based on the measured impulse response. The convergence parameter is set based on the ratio of the power accumulation around the peak power and the power accumulation of the entire filter coefficient. For example, the cumulative ratio is expressed as follows:

  This cumulative ratio is the ratio of the power accumulation of the tap number j before and after the tap number indicating the peak power and the power accumulation of the entire impulse response. j is an arbitrary number, but for example, is the number of taps corresponding to 50 msec. If this value is small, it can be determined that the collection of indirect sounds (noise etc.) other than the direct sound is large. That is, it can be determined that the error of the estimated impulse response is large, and if updating is promoted in this state, the error is increased. As described above, the convergence parameters are the step size α and the forgetting factor λ. When it is determined that the error of the estimated impulse response is large, the step size α is set small in order to delay the update. At the same time, the forgetting factor λ may be increased. Thereby, an increase in error can be prevented.

  Note that the automatic setting unit 10 may measure the loop gain of the acoustic transmission system. FIG. 6 is a diagram showing transfer characteristics of a loudspeaker built-in howling canceller when loop gain is measured. In addition, the same code | symbol is attached | subjected to the part similar to the transmission characteristic of the sound enhancement system with a howling canceller shown in FIG. 3, and detailed description is abbreviate | omitted. As shown in the figure, the signal generation unit 9 outputs the measurement signal n (k) to the speaker 6, the delay circuit 7, and the automatic setting unit 10. The automatic setting unit 10 also receives a signal y1 (k) from the microphone 1.

In this case, the automatic setting unit 10 can perform a fast Fourier transform or the like on the input signal. This makes it possible to measure the power spectrum of the input signal. The power spectrum _{PS y1} of the power spectrum PS _n and the input signal y1 of the measurement signal n (k) (k) is measured. PS _y1 / PS _n which is a ratio of the power spectrum PS _n and the power spectrum PS _y1 is _defined as a loop gain.

  It can be determined that the frequency at which the loop gain has a large value is likely to generate howling. If the loop gain is large, it can be determined that the power of the direct sound is large. Therefore, a predetermined threshold value is set for the loop gain, and the step size α of the adaptive filter 8 is increased for the frequency exceeding the threshold value and the surrounding frequency band in order to prevent the howling, thereby promoting the update. . Further, the forgetting factor λ may be set small.

  In this case, the adaptive filter 8 can perform adaptive processing in the frequency domain, and can set convergence parameters such as a step size α and a forgetting factor λ for each frequency band. .

  In addition, the following application examples are possible for the sound enhancement system with a howling canceller according to the embodiment of the present invention.

  FIG. 7 is a block diagram of a loudspeaker system incorporating a howling canceller according to an application example of the present invention. In addition, the same code | symbol is attached | subjected to the part similar to the above-mentioned sounding system with a howling canceller, and detailed description is abbreviate | omitted. As shown in the figure, the sounding system with a howling canceller further includes a delay circuit 12, an adaptive filter 13, and an adder 14. A signal output to the speaker 6 is input to the delay circuit 7 and the delay circuit 12, and an output signal of the delay circuit 12 is input to the adaptive filter 13. An adder 14 is connected to the output terminal of the adder 2. The adder 14 subtracts the output signal of the adaptive filter 13 from the output signal of the adder 2. The automatic setting unit 10 can set parameters of the delay circuit 7, the delay circuit 12, the adaptive filter 8, and the adaptive filter 13.

  The adaptive filter 13 has the same configuration as that of the adaptive filter 8 shown in FIG. Here, the adaptive filter 13 has a smaller number of taps and a shorter filter coefficient update interval than the adaptive filter 8.

  In this example, the adaptive filter 8 uses an adaptive algorithm that requires a small amount of calculation for updating the filter coefficient. For example, the STFT-CS (Short Time Fourier Transform and Cross Spectrum) algorithm is used. On the other hand, the adaptive filter 13 uses an LMS algorithm or the like with a fast filter coefficient update.

  Since the adaptive filter 13 updates the filter coefficient at high speed, the amount of calculation increases remarkably as the number of taps increases. However, in this embodiment, the number of taps of the adaptive filter 8 is relatively large, so that the adaptive filter In 13, the number of taps can be reduced to reduce the amount of calculation. Thereby, even if the transfer function of the acoustic feedback path fluctuates rapidly and howling grows at a high speed, the adaptive filter 13 can suppress it. That is, the adaptive filter 13 can estimate an abrupt signal change that cannot be eliminated by the adaptive filter 8 and the adder 2, and mainly cancels the peak of the signal.

  In this example, the automatic setting unit 10 performs the same operation as the setting operation described above, but the delay circuit 7 and the delay circuit 12 set different delay amounts. As described above, in the delay circuit 7, a time obtained by subtracting a predetermined correction time from the measured direct sound arrival time is set as a delay amount. In the delay circuit 12, about half of the time corresponding to the number of taps of the adaptive filter 13 is set. Is subtracted as the correction time. The time corresponding to the number of taps of the adaptive filter 13 can be calculated by the sampling frequency. As a result, the adaptive filter 13 can cope with a change in the positional relationship between the microphone 1 and the speaker 6 due to movement of a speaker or the like.

  Although it is not possible to cope with a long reverberation time in the acoustic space by increasing the correction time, in the present embodiment, the number of taps of the adaptive filter 8 is relatively large, and the adaptive filter 13 is fast. Since the purpose is to suppress howling that occurs suddenly by updating the filter coefficient, the number of taps corresponding to the vicinity of the peak of the impulse response may be provided.

  As described above, the sound enhancement system with a howling canceller according to the embodiment of the present invention estimates the direct sound arrival time and reverberation time in the acoustic space by measuring the impulse response estimated by the adaptive filter using the measurement signal. It becomes possible to do. Based on this, by setting the delay amount of the delay circuit, the number of taps of the adaptive filter, etc., it is possible to set the most suitable parameters for the sound field environment to be installed, and to prevent howling efficiently and stably Is possible.

Block diagram of a loudspeaker system with a howling canceler according to the present invention Block diagram showing configuration of adaptive filter in detail The figure which shows the transmission characteristic of the loudspeaker system with a howling canceller of this invention Flow chart showing switching between setting operation and normal operation Flow chart showing setting operation The figure which shows the transfer characteristic of the loudspeaker system with a howling canceller when measuring loop gain Block diagram of a loudspeaker built-in howling canceller according to an application example of the present invention Block diagram showing the transmission characteristics of a conventional loudspeaker built-in adaptive howling canceller Block diagram showing the configuration of a conventional adaptive filter in detail

Explanation of symbols

1-microphone 2-adder 3-switch circuit 4-amplifier 5-adder 6-speaker 7-delay circuit 8-adaptive filter 9-signal generation unit 10-automatic setting unit 11-acoustic feedback path 12-delay circuit 13- Adaptive filter 14-adder 101-microphone 102 in conventional loudspeaker system-adder 103 in conventional loudspeaker system-amplifier 104 in conventional loudspeaker system-speaker 105 in conventional loudspeaker system-acoustic feedback path 106 in conventional loudspeaker system A delay circuit 107 in a conventional loudspeaker system, an adaptive filter 107a in a conventional loudspeaker system, a filter unit 107b in a conventional adaptive filter, and a filter coefficient estimator in a conventional adaptive filter.

Claims (3)

A howling canceller including an adaptive filter that estimates a transfer function of an acoustic feedback path from a speaker to a microphone and cancels a feedback audio signal from a microphone input signal,
A delay circuit that delays the residual signal, which is a signal after canceling the feedback audio signal from the input signal, for a predetermined time;
A sound source unit for inputting a measurement signal to a speaker and a delay circuit;
Setting means for setting at least the delay amount of the delay circuit based on the feedback audio signal of the measurement signal estimated by the adaptive filter;
A howling canceller characterized by comprising   The howling canceller according to claim 1, wherein the setting unit sets parameters of the adaptive filter based on a feedback audio signal of a measurement signal estimated by the adaptive filter.   3. The howling canceller according to claim 1, further comprising switch means for cutting off the residual signal.

Images