Sound source separation device, sound source detection device, acoustic device, sound source direction detection device, and sound source separation method

Abstract

To provide a sound source separation device, a sound source detection device, an acoustic device, a sound source direction detection device, and a sound source separation method that accurately separate a periodic signal having an unknown frequency from an acoustic signal.SOLUTION: A sound source separation device 100 for separating a periodic signal having periodicity from an acoustic signal collected by a microphone and input from an input terminal 101 comprises: a first delay unit 111 for generating a first delay signal obtained by delaying the acoustic signal; a first adaptive filter 112 for outputting a periodic signal obtained by filtering the first delay signal using a first filter coefficient; a first subtraction unit 113 for generating a first error signal obtained by subtracting the periodic signal from the acoustic signal; and an update amount control unit 114 for controlling an update amount of the first filter coefficient. The update amount control unit 114 determines whether or not the acoustic signal includes the periodic signal and controls the update amount's size depending on whether or not the acoustic signal includes the periodic signal.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to a sound source separation device, a sound source detection device, an acoustic device, a sound source direction detection device, and a sound source separation method. [Background technology]

[0002] In recent years, the quietness of vehicle interiors has improved, so that drivers of vehicles may have difficulty hearing sirens of police vehicles, fire engines, ambulances, and the like.

[0003] For example, Patent Document 1 or Patent Document 2 describes a detection device that detects sounds such as the siren of an emergency vehicle buried in surrounding noise. The detection device described in Patent Document 1 or Patent Document 2 detects the siren sound by passing signals in a specific frequency band using a bandpass filter.

[0004] However, there are siren sounds and warning sounds with various frequency characteristics. For example, in Japan, the characteristics of the siren sounds of police vehicles and the siren sounds of ambulance vehicles are different. For example, in Japan, the siren sounds of police vehicles have a periodicity of about 4 to 5 seconds, and the siren sounds of ambulance vehicles have a periodicity of about 1 to 2 seconds. Also, in countries other than Japan, siren sounds may have a periodicity of less than 1 second. For this reason, when warning sounds with various frequency characteristics are extracted using a bandpass filter, the detection device has to be equipped with a wideband bandpass filter or with many types of bandpass filters. However, when the surrounding noise is large, such a detection device has a small S / N ratio and cannot accurately detect siren sounds and warning sounds. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Publication No. 3-6797 [Patent Document 2] JP 2020-161932 A Summary of the Invention [Problem to be solved by the invention]

[0006] The present disclosure provides a sound source separation device, a sound source detection device, an acoustic device, a sound source direction detection device, and a sound source separation method that accurately separate a periodic signal with an unknown frequency from an acoustic signal. [Means for solving the problem]

[0007] A first form of a sound source separation device according to the present disclosure is a sound source separation device that separates a periodic signal having periodicity from an acoustic signal picked up by a microphone, and includes a first delay unit that generates a first delayed signal by delaying the acoustic signal, a first adaptive filter that outputs the periodic signal by filtering the first delayed signal with a first filter coefficient, a first subtraction unit that generates a first error signal by subtracting the periodic signal from the acoustic signal, and an update amount control unit that controls an update amount of the first filter coefficient, wherein the update amount control unit determines whether or not the periodic signal is included in the acoustic signal, and controls a magnitude of the update amount depending on whether or not the periodic signal is included in the acoustic signal.

[0008] A sound source separation method of a second form according to the present disclosure is a sound source separation method in which an information processing device separates a periodic signal having periodicity from an acoustic signal picked up by a microphone, wherein the information processing device generates a first delayed signal by delaying the acoustic signal, the information processing device outputs the periodic signal by filtering the first delayed signal with a first filter coefficient using a first adaptive filter, the information processing device generates a first error signal by subtracting the periodic signal from the acoustic signal, the information processing device controls an update amount of the first filter coefficient, the information processing device determines whether the acoustic signal includes the periodic signal, and controls a magnitude of the update amount depending on whether the acoustic signal includes the periodic signal. Effect of the Invention

[0009] According to the sound source separation device, sound source detection device, acoustic device, sound source direction detection device, and sound source separation method according to the present disclosure, a periodic signal with an unknown frequency can be separated from an acoustic signal with high accuracy. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1 is a diagram showing the configuration of a sound source separation device according to the first embodiment. [Diagram 2] FIG. 2 is a diagram illustrating the configuration of the update amount control unit according to the first embodiment. [Diagram 3] FIG. 3 is a flowchart showing a flow of processing by the update amount control unit according to the first embodiment. [Figure 4] FIG. 4 is a diagram illustrating a configuration of the update amount control unit according to the second embodiment. [Diagram 5] FIG. 5 is a flowchart showing a flow of processing by the update amount control unit according to the second embodiment. [Figure 6] FIG. 6 is a diagram illustrating a configuration of the update amount control unit according to the third embodiment. [Figure 7] FIG. 7 is a flowchart showing a flow of processing by the update amount control unit according to the third embodiment. [Figure 8] FIG. 8 is a diagram showing the configuration of a sound source separation device according to the fourth embodiment. [Figure 9] FIG. 9 is a diagram showing the configuration of a sound source separation device according to the fifth embodiment. [Figure 10] FIG. 10 is a diagram illustrating a configuration of the update amount control unit according to the fifth embodiment. [Figure 11] FIG. 11 is a diagram showing the configuration of a sound source separation device according to the sixth embodiment. [Figure 12] FIG. 12 is a diagram showing the configuration of a sound source detection device according to the seventh embodiment. [Figure 13] FIG. 13 is a diagram showing the configuration of a sound source direction detecting device according to the eighth embodiment. [Figure 14] FIG. 14 is a diagram showing the configuration of an audio device according to the ninth embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, an embodiment will be described with reference to the drawings.

[0012] (First embodiment) FIG. 1 is a diagram showing a configuration of a sound source separation device 100 according to the first embodiment.

[0013] The sound source separation device 100 is provided in, for example, a vehicle. The sound source separation device 100 includes an input terminal 101, an A / D converter 102, a data processing unit 103, a D / A converter 104, and an output terminal 105.

[0014] An analog sound signal picked up by a microphone is input to an input terminal 101. For example, the microphone picks up sounds around the vehicle.

[0015] The A / D converter 102 performs analog / digital conversion on the analog audio signal input from the input terminal 101 to convert it into a digital audio signal.

[0016] The data processing unit 103 receives the digital acoustic signal from the A / D converter 102. The data processing unit 103 separates a periodic signal having periodicity from the acoustic signal. The data processing unit 103 outputs the digital periodic signal separated from the digital acoustic signal.

[0017] The periodic signal may be any signal that is included in an acoustic signal picked up by a microphone and has a periodic acoustic component. An example of the periodic signal is a signal component of a siren sound from a police vehicle, a fire engine, an ambulance, or the like. The periodic signal may be a signal component of an alarm sound emitted at a crossing gate at a railway crossing or the like. The periodic signal may be a signal component of an alarm sound such as a horn emitted from a vehicle, or an alarm sound such as a whistle emitted from a railway vehicle. The periodic signal may also be a signal component of the sound of vehicle braking, etc.

[0018] The data processing unit 103 is realized, for example, by an ECU (Electronic Control Unit) provided in the vehicle. The data processing unit 103 may be a dedicated semiconductor device that executes data processing. The data processing unit 103 may also be realized by a microcontroller or an information processing device that includes a processor other than the ECU.

[0019] The D / A converter 104 acquires the digital periodic signal output from the data processing unit 103. The D / A converter 104 performs digital-to-analog conversion of the digital periodic signal to convert it into an analog periodic signal. The D / A converter 104 provides the analog periodic signal to an output terminal 105.

[0020] The output terminal 105 provides the analog periodic signal output from the D / A converter 104 to an audio output device including, for example, an amplifier and a speaker, to generate sound corresponding to the periodic signal. For example, the audio output device generates sound corresponding to the periodic signal in the passenger compartment of the vehicle. In this way, the sound source separation device 100 can make the driver or the like in the passenger compartment of the vehicle hear the siren and alarm sounds of a police vehicle, a fire engine, an ambulance, or the like. The output terminal 105 may provide the periodic signal to a communication device, a storage device, or the like. The output terminal 105 may also obtain a digital periodic signal from the data processing unit 103 and provide the digital periodic signal to an audio output device, a communication device, a storage device, or the like.

[0021] The data processing unit 103 has a first delay unit 111, a first adaptive filter 112, a first subtraction unit 113, and an update amount control unit 114. The first delay unit 111, the first adaptive filter 112, the first subtraction unit 113, and the update amount control unit 114 are realized by digital processing using a processor or the like.

[0022] The first delay unit 111 receives the acoustic signal output from the A / D converter 102. The first delay unit 111 generates a first delayed signal by delaying the acoustic signal by a preset first delay time. The first delay time is set to a value that prevents correlation from occurring between the acoustic signal and the first delayed signal when the acoustic signal is, for example, only noise.

[0023] The first adaptive filter 112 receives a first delay signal. The first adaptive filter 112 outputs a periodic signal obtained by filtering the first delay signal using a first filter coefficient. For example, the first adaptive filter 112 performs a convolution operation on the input signal. The first adaptive filter 112 is, for example, a Finite Impulse Response (FIR) filter in which the number of taps and the connection relationship of the delay units are predetermined. For example, the first filter coefficient includes a plurality of tap coefficients set in the FIR filter. Note that the first adaptive filter 112 may be an Infinite Impulse Response (IIR) filter.

[0024] The first subtraction unit 113 receives the acoustic signal output from the A / D converter 102 and the periodic signal output from the first adaptive filter 112. The first subtraction unit 113 generates a first error signal by subtracting the periodic signal from the acoustic signal.

[0025] Here, the first adaptive filter 112 receives a first error signal. Then, the first adaptive filter 112 updates the values ​​of the multiple filter coefficients included in the first filter coefficients at each predetermined sampling time using a predetermined adaptive algorithm so as to minimize the power of the first error signal. For example, the first adaptive filter 112 updates the first filter coefficients so as to minimize the mean square of the first error signal.

[0026] The power of the first error signal is minimized when the non-periodic components are completely removed from the first delayed signal by the first adaptive filter 112. The non-periodic components contained in the first delayed signal are signal components that are not correlated with the acoustic signal. Therefore, the first adaptive filter 112 updates the first filter coefficients so as to output a signal that is correlated between the acoustic signal and the first delayed signal, i.e., a signal component having periodicity contained in the acoustic signal. In this way, the first adaptive filter 112 can output, as a periodic signal, a signal obtained by removing the non-periodic components from the acoustic signal, i.e., a signal obtained by extracting the components having periodicity from the acoustic signal.

[0027] The first adaptive filter 112 updates the first filter coefficients by, for example, a normalized least mean square (NLMS) algorithm. The first adaptive filter 112 may update the first filter coefficients by other known adaptive algorithms instead of the NLMS algorithm.

[0028] The update amount control unit 114 receives the acoustic signal output from the A / D converter 102. The update amount control unit 114 controls the update amount of the first filter coefficient in the first adaptive filter 112 based on the acoustic signal. More specifically, the update amount control unit 114 determines whether or not the acoustic signal contains a periodic signal by adaptive filtering separate from the first adaptive filter 112. Then, the update amount control unit 114 controls the magnitude of the update amount depending on whether or not the acoustic signal contains a periodic signal.

[0029] For example, the first adaptive filter 112 updates the first filter coefficient at every predetermined sampling time. The larger the update amount, the more the value of the first filter coefficient changes from before the update. Therefore, when a relatively large first update amount is set, the first adaptive filter 112 changes the filter coefficient at high speed, and can quickly track characteristic changes of periodic components contained in the audio signal. Furthermore, when a second update amount smaller than the first update amount is set, the first adaptive filter 112 changes the filter coefficient at a slower speed than when the first update amount is set, and can accurately extract periodic components contained in the audio signal.

[0030] In this way, the larger the update amount, the faster the system can follow characteristic changes in periodic components contained in the audio signal, but the worse the extraction accuracy. On the other hand, the smaller the update amount, the better the extraction accuracy of periodic components contained in the audio signal, but the less quickly the system can follow characteristic changes. In other words, the size of the update amount has a trade-off relationship between responsiveness and accuracy.

[0031] Therefore, the update amount control unit 114 controls the magnitude of the update amount in the first adaptive filter 112 depending on whether or not the acoustic signal contains a periodic signal, thereby enabling the first adaptive filter 112 to perform adaptive filtering with appropriate responsiveness and accuracy according to the situation.

[0032] The NLMS algorithm employs a variable step size, and the amount of update in the NLMS algorithm corresponds to a step size parameter that indicates the magnitude of the step size.

[0033] FIG. 2 is a diagram showing the configuration of the update amount control unit 114 according to the first embodiment.

[0034] The update amount control unit 114 includes a second delay unit 201 , a second adaptive filter 202 , a second subtraction unit 203 , a first level calculation unit 204 , a second level calculation unit 205 , a ratio calculation unit 206 , and a determination unit 207 .

[0035] The second delay unit 201 receives the acoustic signal from the A / D converter 102. The second delay unit 201 generates a second delayed signal by delaying the acoustic signal by a preset second delay time. The second delay time is set to a value that does not cause correlation with the second delayed signal when the acoustic signal is, for example, only noise. Note that the second delay time may be the same as the first delay time, or may be a different value.

[0036] The second adaptive filter 202 receives a second delayed signal. The second adaptive filter 202 outputs an intermediate signal obtained by filtering the second delayed signal using a second filter coefficient. For example, the second adaptive filter 202 performs a convolution operation on the input signal. The second adaptive filter 202 is, for example, an FIR filter in which the number of taps and the connection relationship of the delay units are predetermined. For example, the second filter coefficient includes a plurality of tap coefficients in an FIR filter.

[0037] The filter configuration and the number of taps in the second adaptive filter 202 may be the same as or different from those of the first adaptive filter 112. For example, the number of taps in the second adaptive filter 202 may be smaller than that of the first adaptive filter 112. This allows the second adaptive filter 202 to have a simpler configuration than the first adaptive filter 112.

[0038] The second subtraction unit 203 receives the acoustic signal and the intermediate signal output from the second adaptive filter 202. The second subtraction unit 203 generates a second error signal by subtracting the intermediate signal from the acoustic signal.

[0039] Here, the second adaptive filter 202 receives the second error signal. Then, the second adaptive filter 202 updates the values ​​of each of the multiple filter coefficients included in the second filter coefficients at each predetermined sampling time using a predetermined adaptive algorithm so as to minimize the power of the second error signal. For example, the second adaptive filter 202 updates the second filter coefficients so as to minimize the mean square of the second error signal. Thereby, the second adaptive filter 202, like the first adaptive filter 112, can output a signal obtained by extracting a component having periodicity from the acoustic signal as an intermediate signal.

[0040] The second adaptive filter 202 updates the second filter coefficients by, for example, the NLMS algorithm or other known adaptive algorithms. The second adaptive filter 202 may update the second filter coefficients by the same algorithm as the first adaptive filter 112, or may update the second filter coefficients by a different algorithm.

[0041] The first level calculation unit 204 receives the intermediate signal output from the second adaptive filter 202. The first level calculation unit 204 calculates a first level that is the signal level of the intermediate signal. For example, the first level calculation unit 204 may calculate a root mean square value or a peak value as the signal level of the intermediate signal.

[0042] The second level calculation unit 205 receives the second error signal output from the second subtraction unit 203. The second level calculation unit 205 calculates a second level that is the signal level of the second error signal. For example, the second level calculation unit 205 may calculate a root mean square value or a peak value as the signal level of the second error signal.

[0043] The ratio calculation unit 206 acquires the first level and the second level. The ratio calculation unit 206 calculates a level ratio that is a ratio between the first level and the second level. In this embodiment, the ratio calculation unit 206 calculates the ratio of the first level to the second level as the level ratio.

[0044] The determination unit 207 receives the level ratio output from the ratio calculation unit 206. The determination unit 207 determines whether or not the audio signal includes a periodic signal based on a comparison result between the level ratio and a preset first threshold value. For example, when the level ratio is the ratio of the first level to the second level, the determination unit 207 determines that the audio signal includes a periodic signal if the level ratio is greater than the first threshold value.

[0045] When the sound signal contains a signal component of a siren sound of a police vehicle, a fire engine, an ambulance, or the like, the signal component of the periodic signal becomes relatively large and the non-periodic component becomes relatively small. Therefore, the determination unit 207 can determine whether or not the sound signal contains a periodic signal based on the result of comparing the level ratio with the first threshold value.

[0046] Furthermore, the determining unit 207 controls the amount of update of the first filter coefficient in the first adaptive filter 112 according to the determination result.

[0047] For example, when the acoustic signal contains a periodic signal, the determination unit 207 increases the update amount compared to when the acoustic signal does not contain a periodic signal. For example, when the determination unit 207 determines that the acoustic signal contains a periodic signal, it sets the magnitude of the update amount to a first update amount, and when it determines that the acoustic signal does not contain a periodic signal, it sets the magnitude of the update amount to a second update amount smaller than the first update amount. In this way, when it is determined that the acoustic signal contains a periodic signal, the update amount control unit 114 can cause the first filter coefficient set in the first adaptive filter 112 to follow the characteristics of the periodic signal more quickly than when it is determined that the acoustic signal contains a periodic signal.

[0048] Also, for example, the determination unit 207 may set the update amount to be smaller when the acoustic signal contains a periodic signal than when the acoustic signal does not contain a periodic signal. For example, when the determination unit 207 determines that the acoustic signal contains a periodic signal, it sets the magnitude of the update amount to a second update amount, and when it determines that the acoustic signal does not contain a periodic signal, it sets the magnitude of the update amount to a first update amount larger than the second update amount. In this way, when it is determined that the acoustic signal contains a periodic signal, the update amount control unit 114 can update the first filter coefficient set in the first adaptive filter 112 so as to extract the periodic signal with higher accuracy than when it is determined that the acoustic signal contains a periodic signal.

[0049] Further, the sound source separation device 100 may detect a periodic signal whose signal level changes rapidly over time. In such a case, the determination unit 207 further calculates a time rate of change of the signal level of the periodic signal based on the first level. For example, the determination unit 207 may determine whether or not the signal level of the periodic signal is repeatedly changed in intensity at intervals of several hundreds of milliseconds. In such a case, when the calculated time rate of change is greater than a preset first rate of change, the determination unit 207 sets the update amount to be greater than when the acoustic signal does not include a periodic signal or when the time rate of change is equal to or less than the first rate of change. For example, when the calculated time rate of change is equal to or less than the first rate of change, the determination unit 207 sets the magnitude of the update amount to the second update amount, and when the calculated time rate of change is greater than the first rate of change, the determination unit 207 sets the magnitude of the update amount to the first update amount, which is greater than the second update amount. In this way, when the acoustic signal includes a periodic signal whose signal level changes over time, the update amount control unit 114 can cause the first adaptive filter 112 to quickly follow the characteristics of the periodic signal included in the acoustic signal.

[0050] Furthermore, when the acoustic signal contains a periodic signal for a continuous period of time, the determination unit 207 may increase the update amount stepwise over time up to a preset upper limit value. Furthermore, when the acoustic signal does not contain a periodic signal for a continuous period of time, the determination unit 207 may decrease the update amount stepwise over time up to a preset lower limit value.

[0051] Furthermore, when the acoustic signal contains a periodic signal for a continuous period of time, the determination unit 207 may gradually decrease the update amount over time to a preset lower limit value. Furthermore, when the acoustic signal does not contain a periodic signal for a continuous period of time, the determination unit 207 may gradually increase the update amount over time to a preset upper limit value. By gradually increasing or decreasing the update amount in this manner, the determination unit 207 can smoothly change the first filter coefficient of the first adaptive filter 112.

[0052] In this embodiment, the second level calculation unit 205 calculates the signal level of the second error signal as the second level. Alternatively, the second level calculation unit 205 may acquire the acoustic signal output from the A / D converter 102 and calculate the signal level of the acoustic signal as the second level. When a signal component of a periodic signal is included in the acoustic signal, the signal level of the periodic signal becomes relatively large and the signal level of the acoustic signal becomes relatively small. Therefore, even in this manner, the determination unit 207 can determine whether or not the acoustic signal includes a periodic signal.

[0053] 3 is a flowchart showing the flow of processing by the update amount control unit 114 according to the first embodiment. The update amount control unit 114 according to the first embodiment executes processing in the flow shown in FIG.

[0054] First, in S301, the update amount control unit 114 calculates a first level which is the signal level of the intermediate signal output from the second adaptive filter 202 and a second level which is the signal level of the second error signal output from the second subtraction unit 203.

[0055] Next, in S302, the update amount control unit 114 calculates a level ratio, which is a ratio between the first level and the second level. In this embodiment, the update amount control unit 114 calculates the ratio of the first level to the second level as the level ratio.

[0056] Next, in S303, the update amount control unit 114 compares the level ratio with a preset first threshold value. If the level ratio is greater than the first threshold value (Yes in S303), the update amount control unit 114 advances the process to S304. In S304, the update amount control unit 114 determines that the sound signal includes a periodic signal.

[0057] On the other hand, if the level ratio is equal to or less than the first threshold value (No in S303), the update amount control unit 114 advances the process to S305. In S305, the update amount control unit 114 determines that the sound signal does not include a periodic signal.

[0058] Then, the update amount control unit 114 controls the magnitude of the update amount of the first filter coefficient in the first adaptive filter 112 according to the result of the determination as to whether or not the acoustic signal contains a periodic signal. For example, when the update amount control unit 114 determines that the acoustic signal contains a periodic signal, it sets the magnitude of the update amount to a first update amount, and when it determines that the acoustic signal does not contain a periodic signal, it sets the magnitude of the update amount to a second update amount smaller than the first update amount. In this way, when the update amount control unit 114 determines that the acoustic signal contains a periodic signal, it is possible to improve the responsiveness of the first adaptive filter 112 in detecting the periodic signal.

[0059] For example, the update amount control unit 114 may set the magnitude of the update amount to the second update amount when it is determined that the acoustic signal contains a periodic signal, and may set the magnitude of the update amount to the first update amount when it is determined that the acoustic signal does not contain a periodic signal. In this way, when it is determined that the acoustic signal contains a periodic signal, the update amount control unit 114 can improve the accuracy of extraction of the periodic signal in the first adaptive filter 112.

[0060] As described above, the sound source separation device 100 according to the first embodiment can separate a periodic signal from an acoustic signal by the first adaptive filter 112. Furthermore, the sound source separation device 100 according to the first embodiment determines whether or not a periodic signal is included in the acoustic signal by using the second adaptive filter 202, and controls the magnitude of the update amount of the first filter coefficient in the first adaptive filter 112 depending on whether or not a periodic signal is included in the acoustic signal. In this way, the sound source separation device 100 according to the first embodiment can cause the first adaptive filter 112 to extract a periodic signal from the acoustic signal with appropriate responsiveness or appropriate accuracy, and can therefore accurately separate a periodic signal with an unknown frequency buried in noise from the acoustic signal.

[0061] Second embodiment Next, the sound source separation device 100 according to the second embodiment will be described.

[0062] The sound source separating device 100 according to the second embodiment has substantially the same configuration and function as the first embodiment described with reference to Fig. 1 to Fig. 3. Therefore, in the description of the sound source separating device 100 according to the second embodiment, components having substantially the same configuration and function are denoted by the same reference numerals, and detailed description is omitted except for the differences. The same applies to the third and subsequent embodiments.

[0063] FIG. 4 is a diagram showing the configuration of the update amount control unit 114 according to the second embodiment.

[0064] The update amount control unit 114 according to the second embodiment includes a second delay unit 201, a second adaptive filter 202, a second subtraction unit 203, a first level calculation unit 204, and a determination unit 207.

[0065] In the second embodiment, the determination unit 207 receives the first level from the first level calculation unit 204. The determination unit 207 determines whether or not the acoustic signal includes a periodic signal based on a comparison result between the first level and a preset second threshold value. For example, when the first level is greater than the second threshold value, the determination unit 207 determines that the acoustic signal includes a periodic signal.

[0066] When the signal component of a siren of a police vehicle, a fire engine, an ambulance, or the like is included in the sound signal, the signal component of the periodic signal becomes large. Therefore, the determination unit 207 can determine whether or not the sound signal includes a periodic signal based on the comparison result between the first level and the second threshold. Then, the determination unit 207 controls the magnitude of the update amount of the first filter coefficient in the first adaptive filter 112 according to the determination result, as in the first embodiment.

[0067] The update amount control unit 114 according to the second embodiment may include a second level calculation unit 205 instead of the first level calculation unit 204. In this case, the determination unit 207 receives a second level, which is a signal level of a non-periodic signal included in the sound signal, from the second level calculation unit 205. Then, the determination unit 207 determines whether or not a periodic signal is included in the sound signal based on a comparison result between the second level and a second threshold set in advance. For example, the determination unit 207 determines that the sound signal includes a periodic signal when the second level is equal to or less than the second threshold. When a signal component of a siren sound of a police vehicle, a fire engine, an ambulance, or the like is included in the sound signal, the signal component of the periodic signal becomes large, while the non-periodic component included in the sound signal becomes relatively small. Therefore, the determination unit 207 can also determine whether or not a periodic signal is included in the sound signal based on whether or not the second level is equal to or less than the second threshold.

[0068] 5 is a flowchart showing the flow of processing by the update amount control unit 114 according to the second embodiment. The update amount control unit 114 according to the second embodiment executes processing in the flow shown in FIG.

[0069] First, in S501, the update amount control unit 114 calculates a first level which is the signal level of the intermediate signal output from the second adaptive filter 202.

[0070] Next, in S502, the update amount control unit 114 compares the first level with a preset second threshold. If the first level is greater than the second threshold (Yes in S502), the update amount control unit 114 advances the process to S503. In S503, the update amount control unit 114 determines that the sound signal includes a periodic signal.

[0071] On the other hand, if the first level is equal to or lower than the second threshold (No in S502), the update amount control unit 114 advances the process to S504. In S504, the update amount control unit 114 determines that the sound signal does not include a periodic signal.

[0072] Then, the update amount control unit 114 controls the magnitude of the update amount of the first filter coefficient in the first adaptive filter 112 according to the result of the determination as to whether or not the acoustic signal includes a periodic signal, in the same way as in the first embodiment.

[0073] The sound source separation device 100 according to the second embodiment described above can accurately separate a periodic signal with an unknown frequency buried in noise from an acoustic signal, similar to the first embodiment.

[0074] Third embodiment Next, a sound source separation device 100 according to a third embodiment will be described.

[0075] FIG. 6 is a diagram showing the configuration of the update amount control unit 114 according to the third embodiment.

[0076] The update amount control section 114 according to the third embodiment includes a second delay section 201, a second adaptive filter 202, a second subtraction section 203, a band division section 601, a band level calculation section 602, and a determination section 207.

[0077] The band dividing unit 601 receives the intermediate signal output from the second adaptive filter 202. The band dividing unit 601 generates a band signal representing a signal component included in the intermediate signal for each of a plurality of frequency bands. The band dividing unit 601 may generate each signal component of a plurality of frequency bands included in the intermediate signal using, for example, a fast Fourier transform. The band dividing unit 601 may also generate each signal component of a plurality of frequency bands included in the intermediate signal using a filter bank including a plurality of FIR filters or a plurality of IIR filters, each of which extracts a component of a specific frequency band. The band dividing unit 601 may also generate each signal component of a plurality of frequency bands included in the intermediate signal using other known frequency division methods.

[0078] The band level calculation unit 602 receives signal components of each of the multiple frequency bands included in the intermediate signal from the band division unit 601. The band level calculation unit 602 calculates the signal level of each of the multiple frequency bands in the intermediate signal. For example, the band level calculation unit 602 may calculate the root mean square value or the peak value of the signal components for each of the multiple frequency bands.

[0079] In the third embodiment, the determination unit 207 acquires the signal levels of the multiple frequency bands in the intermediate signal from the band level calculation unit 602. The determination unit 207 determines whether or not the audio signal includes a periodic signal based on a comparison result between the signal level of at least one of the multiple frequency bands and a preset third threshold value. For example, the determination unit 207 determines that the audio signal includes a periodic signal when the signal level of at least one of the multiple frequency bands is greater than the third threshold value.

[0080] When the signal component of a siren sound of a police vehicle, a fire engine, an ambulance, or the like is included in the sound signal, the signal component of the periodic signal in a specific frequency band becomes large, and a peak occurs at a predetermined frequency on the frequency axis, for example. Therefore, when the signal level of at least one of the multiple frequency bands is greater than the third threshold value, the determination unit 207 can determine whether or not the sound signal includes a periodic signal. Then, the determination unit 207 controls the magnitude of the update amount of the first filter coefficient in the first adaptive filter 112 according to the determination result, as in the first embodiment.

[0081] Alternatively, the determination unit 207 may calculate an average value of the signal levels of a plurality of frequency bands, and calculate, for each of the plurality of frequency bands, a component ratio of the signal level of that frequency band to the calculated average value. In this case, the determination unit 207 determines that the acoustic signal contains a periodic signal when the component ratio of at least one of the plurality of frequency bands is greater than a third threshold, and determines that the acoustic signal does not contain a periodic signal when the component ratio of any of the plurality of frequency bands is equal to or less than the third threshold.

[0082] Furthermore, the determination unit 207 may calculate a difference or ratio of signal components for each pair of two adjacent frequency bands in the multiple frequency bands. In this case, the determination unit 207 determines that the acoustic signal contains a periodic signal when the difference or ratio of signal components of at least one pair of two adjacent frequency bands is greater than a third threshold, and determines that the acoustic signal does not contain a periodic signal when the difference or ratio of any signal component is equal to or less than the third threshold. That is, the determination unit 207 detects whether or not the intermediate signal has a peak on the frequency axis based on the signal levels of the multiple frequency bands. The determination unit 207 determines that the acoustic signal contains a periodic signal when the intermediate signal has a peak on the frequency axis, and determines that the acoustic signal does not contain a periodic signal when the intermediate signal does not have a peak.

[0083] Note that the determining unit 207 may set the third threshold to a different value for each of a plurality of frequency bands.

[0084] 7 is a flowchart showing the flow of processing by the update amount control unit 114 according to the third embodiment. The update amount control unit 114 according to the third embodiment executes processing in the flow shown in FIG.

[0085] First, in S701, the update amount control unit 114 divides the intermediate signal into bands, and generates band signals representing signal components included in the intermediate signal for each of a plurality of frequency bands.

[0086] Next, in S702, the update amount control unit 114 calculates the signal level of each of a plurality of frequency bands in the intermediate signal.

[0087] Next, in S703, the update amount control unit 114 compares the signal level of the signal component in the intermediate signal with a preset third threshold for each of the multiple frequency bands. If there is a band among the multiple frequency bands whose signal level is greater than the third threshold (Yes in S703), the update amount control unit 114 advances the process to S704. In S704, the update amount control unit 114 determines that the sound signal includes a periodic signal.

[0088] Furthermore, if the signal level is equal to or lower than the third threshold for any of the frequency bands (No in S703), the update amount control unit 114 advances the process to S705. In S705, the update amount control unit 114 determines that the sound signal does not include a periodic signal.

[0089] Then, the update amount control unit 114 controls the magnitude of the update amount of the first filter coefficient in the first adaptive filter 112 according to the result of the determination as to whether or not the acoustic signal includes a periodic signal, in the same way as in the first embodiment.

[0090] The sound source separation device 100 according to the third embodiment described above can accurately separate a periodic signal with an unknown frequency buried in noise from an acoustic signal, similar to the first embodiment.

[0091] (Fourth embodiment) Next, the sound source separation device 100 according to the fourth embodiment will be described.

[0092] FIG. 8 is a diagram showing a configuration of a sound source separation device 100 according to the fourth embodiment.

[0093] The data processing unit 103 according to the fourth embodiment has a band limiting filter 801 , a first delay unit 111 , a first adaptive filter 112 , a first subtraction unit 113 , and an update amount control unit 114 .

[0094] The band-limiting filter 801 receives the acoustic signal output from the A / D converter 102. The band-limiting filter 801 limits the acoustic signal to a predetermined frequency band. The band-limiting filter 801 may be, for example, a low-pass filter, a high-pass filter, or a band-pass filter. Such a band-limiting filter 801 can pass signal components in a band near the periodic signal to be separated.

[0095] In the fourth embodiment, the first delay unit 111 receives an acoustic signal that has been band-limited by the band-limiting filter 801. Then, in the fourth embodiment, the first delay unit 111 generates a first delayed signal by delaying the band-limited acoustic signal by a first delay time.

[0096] In the fourth embodiment, the update amount control unit 114 receives an audio signal that has been band-limited by the band limiting filter 801. Then, in the fourth embodiment, the update amount control unit 114 controls the update amount of the first filter coefficient based on the band-limited audio signal.

[0097] The sound source separation device 100 according to the fourth embodiment can improve the separation accuracy of a periodic signal buried in noise when the frequency band of the periodic signal to be separated is included in a predetermined frequency range.

[0098] The sound source separating device 100 according to the fourth embodiment may have the band-limiting filter 801 in various positions. For example, the sound source separating device 100 may be configured to include the band-limiting filter 801 in a stage preceding the first delay unit 111, and to provide the update amount control unit 114 with a non-band-limited acoustic signal directly from the A / D converter 102. The sound source separating device 100 according to the fourth embodiment may be configured to include the band-limiting filter 801 in a stage preceding the update amount control unit 114, and to provide the first delay unit 111 with a non-band-limited acoustic signal directly from the A / D converter 102.

[0099] Fifth embodiment Next, the sound source separation device 100 according to the fifth embodiment will be described.

[0100] FIG. 9 is a diagram showing a configuration of a sound source separation device 100 according to the fifth embodiment.

[0101] The data processing unit 103 according to the fifth embodiment has a first delay unit 111, a first adaptive filter 112, a first subtraction unit 113, and an update amount control unit 114.

[0102] In the fifth embodiment, the update amount control unit 114 receives the acoustic signal output from the A / D converter 102 and the first delay signal output from the first delay unit 111. Then, in the fifth embodiment, the update amount control unit 114 controls the update amount of the first filter coefficient in the first adaptive filter 112 based on the acoustic signal and the first delay signal.

[0103] FIG. 10 is a diagram showing the configuration of the update amount control unit 114 according to the fifth embodiment.

[0104] In the fifth embodiment, the update amount control unit 114 includes a second adaptive filter 202 , a second subtraction unit 203 , a first level calculation unit 204 , a second level calculation unit 205 , a ratio calculation unit 206 , and a determination unit 207 .

[0105] The second adaptive filter 202 receives the first delayed signal output from the first delay unit 111. The second adaptive filter 202 outputs an intermediate signal obtained by filtering the first delayed signal using a second filter coefficient.

[0106] The second subtraction unit 203 receives the acoustic signal output from the A / D converter 102 and the intermediate signal output from the second adaptive filter 202. The second subtraction unit 203 generates a second error signal by subtracting the intermediate signal from the acoustic signal.

[0107] The sound source separating device 100 according to the fifth embodiment can accurately separate a periodic signal buried in noise and having an unknown frequency from an acoustic signal, similarly to the first embodiment, even if the update amount control unit 114 does not include the second delay unit 201. As a result, the sound source separating device 100 according to the fifth embodiment can accurately separate a periodic signal from an acoustic signal with a simple configuration.

[0108] Note that the sound source separation device 100 according to the fifth embodiment may be configured such that the data processing unit 103 does not include the first delay unit 111, and the update amount control unit 114 includes the second delay unit 201. In this case, the first adaptive filter 112 included in the data processing unit 103 receives a second delay signal from the second delay unit 201, instead of receiving the first delay signal from the first delay unit 111. Even with such a configuration, the sound source separation device 100 can accurately separate a periodic signal with an unknown frequency buried in noise from an acoustic signal, similar to the first embodiment.

[0109] Sixth embodiment Next, the sound source separation device 100 according to a sixth embodiment will be described.

[0110] FIG. 11 is a diagram showing a configuration of a sound source separation device 100 according to the sixth embodiment.

[0111] The sound source separation device 100 includes an input terminal 101, an A / D converter 102, a data processing unit 103, a D / A converter 104, and an output terminal 105, and further includes a notification unit 1101.

[0112] The notification unit 1101 receives a determination result as to whether or not a periodic signal is included in the audio signal from the update amount control unit 114. When the update amount control unit 114 determines that the audio signal includes a periodic signal, the notification unit 1101 notifies an external device that the audio signal includes a periodic signal.

[0113] For example, when it is determined that the audio signal contains a periodic signal, the notification unit 1101 may display information indicating that the audio signal contains a periodic signal on the display device. When it is determined that the audio signal contains a periodic signal, the notification unit 1101 may output a predetermined voice guidance from a speaker. Furthermore, when it is determined that the audio signal contains a periodic signal, the notification unit 1101 may cause a light-emitting diode of a predetermined color among a plurality of colored light-emitting diodes to emit light or change the intervals at which the light-emitting diode lights up or blinks.

[0114] When a periodic signal such as a siren of a police vehicle, a fire engine, or an ambulance is included in the surrounding acoustic signal, the sound source separation device 100 according to the sixth embodiment can visually or audibly notify a user, such as a driver in a vehicle, of the periodic signal. This allows the user to recognize that a periodic signal such as a siren of a police vehicle, a fire engine, or an ambulance is occurring in the surroundings, even in an environment where it is difficult to hear surrounding sounds.

[0115] Seventh embodiment Next, a sound source detection device 1200 according to the seventh embodiment will be described.

[0116] FIG. 12 is a diagram showing the configuration of a sound source detection device 1200 according to the seventh embodiment.

[0117] The sound source detection device 1200 includes a microphone 1201 , the sound source separation device 100 , and a sound source detection unit 1202 .

[0118] The microphone 1201 picks up surrounding sounds and outputs an analog audio signal. For example, the microphone 1201 is provided in a vehicle and picks up sounds around the vehicle.

[0119] The sound source separation device 100 is any one of the sound source separation devices 100 according to the first to sixth embodiments. The sound source separation device 100 receives an analog acoustic signal from a microphone 1201. The sound source separation device 100 separates a periodic signal from the acoustic signal and outputs the separated periodic signal. Note that the sound source separation device 100 outputs a digital periodic signal instead of the analog periodic signal.

[0120] The sound source detection unit 1202 receives a periodic signal from the sound source separation device 100. The sound source detection unit 1202 analyzes the periodic signal output from the sound source separation device 100, detects whether or not a predetermined type of signal is included, and outputs the detection result.

[0121] For example, the sound source detection unit 1202 detects whether or not a siren of a police vehicle is being generated by analyzing the received periodic signal. Furthermore, the sound source detection unit 1202 may detect whether a siren of a fire engine or an ambulance is being generated. For example, the sound source detection unit 1202 recognizes the type of periodic signal by a known method such as pattern matching between the periodic signal and a database prepared in advance.

[0122] Then, the sound source detection unit 1202 displays the detection result on a display device or outputs it as voice guidance, for example, so that the driver of the vehicle can recognize what type of periodic signal is being generated.

[0123] The sound source detection device 1200 according to the seventh embodiment can accurately recognize the type of periodic signal from the periodic signal accurately separated from the acoustic signal.

[0124] Eighth embodiment Next, a sound source direction detecting device 1300 according to an eighth embodiment will be described.

[0125] FIG. 13 is a diagram showing the configuration of a sound source direction detecting device 1300 according to the eighth embodiment.

[0126] The sound source direction detecting device 1300 includes a first microphone 1301, a second microphone 1302, a first sound source separating device 1303 (100), a second sound source separating device 1304 (100), and a sound source direction detecting unit 1305.

[0127] The first microphone 1301 picks up sound from a first direction with higher sensitivity than a second direction different from the first direction. That is, the first microphone 1301 picks up sound with high directivity in the first direction. The first microphone 1301 outputs an analog sound signal. For example, the first microphone 1301 is provided in a vehicle. For example, the first microphone 1301 is provided to have directivity toward the front of the vehicle.

[0128] The second microphone 1302 picks up sound from the second direction with higher sensitivity than the first direction. That is, the second microphone 1302 picks up sound with high directivity in the second direction. The second microphone 1302 outputs an analog sound signal. For example, the second microphone 1302 is provided in a vehicle. For example, the second microphone 1302 is provided to have directivity toward the rear of the vehicle.

[0129] The first sound source separation device 1303 is any one of the sound source separation devices 100 according to the first to sixth embodiments. The first sound source separation device 1303 receives an analog acoustic signal from the first microphone 1301. The first sound source separation device 1303 separates a periodic signal from the acoustic signal picked up by the first microphone 1301, and outputs the separated periodic signal. Note that the first sound source separation device 1303 outputs a digital periodic signal instead of an analog periodic signal.

[0130] The second sound source separation device 1304 is any one of the sound source separation devices 100 according to the first to sixth embodiments. The second sound source separation device 1304 receives an analog acoustic signal from the second microphone 1302. The second sound source separation device 1304 separates a periodic signal from the acoustic signal picked up by the second microphone 1302, and outputs the separated periodic signal. Note that the second sound source separation device 1304 outputs a digital periodic signal instead of an analog periodic signal.

[0131] The sound source direction detection unit 1305 detects the direction of a sound source from which a sound that is the source of the components of the periodic signals is generated, based on the periodic signals separated by the first sound source separation device 1303 and the periodic signals separated by the second sound source separation device 1304. The sound source direction detection unit 1305 detects the direction of the sound source by, for example, analyzing the amplitude difference or time difference between the two periodic signals using a known method. Then, the sound source direction detection unit 1305 displays the detected direction on a display device or outputs it as voice guidance, for example, so that the driver of the vehicle can recognize from which direction the periodic signal is generated.

[0132] The sound source direction detection device 1300 according to the eighth embodiment can accurately detect the direction from which the sound that is the source of the components of the periodic signals originates, from two periodic signals that are accurately separated from an acoustic signal picked up with high directivity in a first direction and an acoustic signal picked up with high directivity in a second direction.

[0133] The sound source direction detection device 1300 may include N microphones (N is an integer equal to or greater than 3) with different sensitive directions, and N sound source separation devices 100 in one-to-one correspondence with the N microphones. Each of the N sound source separation devices 100 separates a periodic signal from an acoustic signal picked up by a corresponding one of the N microphones. In this case, the sound source direction detection unit 1305 acquires the periodic signal from each of the N sound source separation devices 100 and detects the sound source direction.

[0134] Ninth embodiment Next, an audio device 1400 according to a ninth embodiment will be described.

[0135] FIG. 14 is a diagram showing the configuration of an audio device 1400 according to the ninth embodiment.

[0136] The acoustic device 1400 includes a microphone 1401 , the sound source separation device 100 , an amplifier device 1402 , and a speaker 1403 .

[0137] The microphone 1401 picks up surrounding sounds and outputs an analog sound signal. For example, the microphone 1401 is provided in a vehicle and picks up sounds around the vehicle. The microphone 1401 is provided, for example, outside the vehicle. Alternatively, the microphone 1401 may be provided inside the vehicle.

[0138] The sound source separation device 100 is any one of the sound source separation devices 100 according to the first to sixth embodiments. The sound source separation device 100 receives an analog acoustic signal from a microphone 1401. The sound source separation device 100 separates a periodic signal from the acoustic signal, and outputs the separated periodic signal.

[0139] The amplifier 1402 receives the periodic signal from the sound source separation device 100. The amplifier 1402 amplifies the received periodic signal and provides it to the speaker 1403.

[0140] The speaker 1403 receives the periodic signal from the amplifier 1402 and generates sound corresponding to the received periodic signal. The speaker 1403 is provided in the cabin of the vehicle. This allows, for example, the driver of the vehicle to hear sound that is an amplified version of a periodic signal generated from a sound source outside the vehicle, from the speaker 1403 provided in the cabin of the vehicle.

[0141] The audio device 1400 according to the ninth embodiment can reproduce, as audio, a periodic signal that has been accurately separated from an audio signal.

[0142] The vehicle may include a plurality of acoustic devices 1400. In this case, the plurality of microphones 1401 included in the plurality of acoustic devices 1400 are arranged so that the directions in which the microphones 1401 have high sensitivity for picking up sound are different from each other. Also, the plurality of speakers 1403 included in the plurality of acoustic devices 1400 are provided at positions in the vehicle interior corresponding to the microphones 1401 included in the same acoustic device 1400.

[0143] For example, a first acoustic device 1400 of the multiple acoustic devices 1400 includes a microphone 1401 having directivity toward the front of the vehicle, and a speaker 1403 provided at the front of the vehicle interior. A second acoustic device 1400 of the multiple acoustic devices 1400 includes a microphone 1401 having directivity toward the rear of the vehicle, and a speaker 1403 provided at the rear of the vehicle interior. This allows the multiple acoustic devices 1400 to reproduce the periodic signal, which has been accurately separated from the acoustic signal, as sound with a sense of direction.

[0144] Although the embodiments and modifications of the present invention have been described, these embodiments and modifications are presented as examples and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and modifications are included in the scope of the invention and its equivalents described in the claims, as well as in the scope and spirit of the invention. [Explanation of symbols]

[0145] 100 Sound source separation device, 101 Input terminal, 102 A / D converter, 103 Data processing unit, 104 D / A converter, 105 Output terminal, 111 First delay unit, 112 First adaptive filter, 113 First subtraction unit, 114 Update amount control unit, 201 Second delay unit, 202 Second adaptive filter, 203 Second subtraction unit, 204 First level calculation unit, 205 Second level calculation unit, 206 Ratio calculation unit, 207 Judgment unit, 601 Band division unit, 602 Band level calculation unit, 1101 Notification unit, 1200 Sound source detection device, 1201 Microphone, 1300 Sound source direction detection device, 1301 First microphone, 1302 Second microphone, 1303 First sound source separation device, 1304 Second sound source separation device, 1305 Sound source direction detection unit, 1400 Acoustic equipment, 1401 microphone, 1402 amplifier, 1403 speaker

Claims

1. A sound source separation device that separates a periodic signal having periodicity from an acoustic signal collected by a microphone, a first delay unit that generates a first delayed signal by delaying the acoustic signal; a first adaptive filter that outputs the periodic signal obtained by filtering the first delayed signal using a first filter coefficient; a first subtraction unit that generates a first error signal by subtracting the periodic signal from the acoustic signal; an update amount control unit for controlling an update amount of the first filter coefficient; Equipped with The update amount control unit determines whether the acoustic signal includes the periodic signal, and controls the magnitude of the update amount depending on whether the acoustic signal includes the periodic signal. Sound source separation device.

2. The update amount control unit increases the update amount when the acoustic signal includes the periodic signal compared to when the acoustic signal does not include the periodic signal. The sound source separation device according to claim 1 .

3. When the acoustic signal includes the periodic signal, the update amount control unit further detects a time rate of change of a signal level of the periodic signal; When the time rate of change of the signal level of the periodic signal is greater than a preset first rate of change, the update amount is made larger than when the acoustic signal does not include the periodic signal or when the time rate of change of the signal level of the periodic signal is equal to or smaller than the first rate of change. The sound source separation device according to claim 1 .

4. The update amount control unit reduces the update amount when the acoustic signal includes the periodic signal compared to when the acoustic signal does not include the periodic signal. The sound source separation device according to claim 1 .

5. The update amount control unit a second delay unit that generates a second delayed signal by delaying the acoustic signal; a second adaptive filter that outputs an intermediate signal obtained by filtering the second delayed signal using a second filter coefficient; a second subtraction unit that generates a second error signal by subtracting the intermediate signal from the acoustic signal; A determination unit; Including, The determination unit determines whether the acoustic signal includes the periodic signal based on at least one of a signal level of the intermediate signal and a signal level of the second error signal. The sound source separation device according to claim 1 .

6. The update amount control unit a first level calculation unit that calculates a first level that is a signal level of the intermediate signal; a second level calculation unit that calculates a second level that is a signal level of the second error signal; a ratio calculation unit that calculates a level ratio that is a ratio between the first level and the second level; Further comprising: The determination unit determines whether or not the acoustic signal includes the periodic signal based on a comparison result between the level ratio and a preset first threshold value. The sound source separation device according to claim 5 .

7. The update amount control unit a first level calculation unit that calculates a first level that is a signal level of the intermediate signal, The determination unit determines whether or not the acoustic signal includes the periodic signal based on a comparison result between the first level and a preset second threshold value. The sound source separation device according to claim 5 .

8. The update amount control unit a band level calculation unit that calculates a signal level of each of a plurality of frequency bands in the intermediate signal, The determination unit determines whether the acoustic signal includes the periodic signal based on a comparison result between a signal level of at least one of the plurality of frequency bands and a preset third threshold value. The sound source separation device according to claim 5 .

9. a band limiting filter for limiting the acoustic signal to a predetermined frequency band; In addition, The first delay unit delays the acoustic signal limited to the predetermined frequency band. The sound source separation device according to claim 1 .

10. The update amount control unit a second adaptive filter that outputs an intermediate signal obtained by filtering the first delayed signal through a filter to which the second filter coefficient is set, the second filter coefficient being updated; a second subtraction unit that generates a second error signal by subtracting the intermediate signal from the acoustic signal; A determination unit; Including, The determination unit determines whether the acoustic signal includes the periodic signal based on at least one of a signal level of the intermediate signal and a signal level of the second error signal. The sound source separation device according to claim 1 .

11. The audio signal further includes a notification unit that notifies an external device that the audio signal includes the periodic signal when the update amount control unit determines that the audio signal includes the periodic signal. The sound source separation device according to claim 1 .

12. The sound source separation device according to claim 1 ; a sound source detection unit that analyzes the periodic signal output from the sound source separation device to detect whether or not a predetermined type of signal is included, and outputs a detection result; A sound source detection device comprising:

13. a first microphone that picks up sound from a first direction with higher sensitivity than a second direction different from the first direction; a second microphone that picks up sound from the second direction with higher sensitivity than the first direction; A first sound source separation device according to claim 1 , which separates the periodic signal from the acoustic signal picked up by the first microphone; A second sound source separation device according to claim 1 , which separates the periodic signal from the acoustic signal picked up by the second microphone; a sound source direction detection unit that detects a sound source direction based on the periodic signal separated by the first sound source separation device and the periodic signal separated by the second sound source separation device; A sound source direction detection device comprising:

14. The sound source separation device according to claim 1 ; a speaker that outputs the periodic signal output from the sound source separation device as sound; An audio device comprising:

15. A sound source separation method for separating a periodic signal having periodicity from an acoustic signal collected by a microphone by an information processing device, comprising: the information processing device generates a first delayed signal by delaying the acoustic signal; the information processing device outputs the periodic signal obtained by filtering the first delayed signal with a first filter coefficient using a first adaptive filter; the information processing device generates a first error signal by subtracting the periodic signal from the acoustic signal; the information processing device controls an update amount of the first filter coefficient; The information processing device determines whether the acoustic signal includes the periodic signal, and controls the magnitude of the update amount depending on whether the acoustic signal includes the periodic signal. Sound source separation method.

Images