JP2014116727A - Sound extraction device, sound extraction method, and sound extraction program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a sound extraction device, sound extraction method, and sound extraction program capable of accurately extracting the sound of a test object.SOLUTION: A sound extraction device includes: a detection unit for detecting the passage of a tested body in the neighborhood of a sound recording microphone; a storage unit for storing the output sound of a tested subject as sound recording data with the use of a sound recording microphone when the detection unit has detected the passage; a recognition unit for recognizing the program search signal of a frequency individually determined for a tested body from the sound recording data; and an extraction unit for extracting multiple pieces of period data on the basis of the program search signal recognized by the recognition unit and extracting the output sound of a tested body by acquiring the logical product of the period data.

Description

  The present invention relates to a sound extraction device, a sound extraction method, and a sound extraction program.

  There has been disclosed a technique for determining whether a sound reproduction function is normal by reproducing a test sound on a test object flowing on a line and analyzing the test sound (see, for example, Patent Document 1).

JP 2004-297368 A

  However, when the test is to be performed asynchronously without communicating with the device under test, the device under test reproduces the test sound continuously. In this case, since the other test object reproduces the test sound while inspecting the specific test object, the test sound of the other test object becomes noise.

  The present invention has been made in view of the above problems, and an object of the present invention is to provide a sound extraction device, a sound extraction method, and a sound extraction program that can accurately extract a test target sound.

  In order to solve the above-described problem, the sound extraction device disclosed in the specification includes a detection unit that detects that a device under test passes in the vicinity of a recording microphone, and the recording unit when the passage is detected by the detection unit. A storage unit for storing the output sound of the device under test as recording data using a microphone; a recognition unit for recognizing a cueing signal having a frequency individually determined for the device under test; and the recognition unit An extraction unit that extracts a plurality of period data based on the cue signal recognized by the step C, and extracts an output sound of the DUT by obtaining a logical product of the period data.

  According to the sound extraction device, the sound extraction method, and the sound extraction program disclosed in the specification, it is possible to accurately extract the sound to be tested.

It is a figure for demonstrating the outline of the sound test apparatus used for a sound test. (A) And (b) is a figure showing an output sound. It is a block diagram for demonstrating the structure of the sound extraction apparatus which concerns on Example 1, and a test object. It is a block diagram showing each function of a processing part. It is a schematic diagram for demonstrating the positional relationship of a passage detection sensor, a barcode reader, a recording microphone, and a device under test. It is a block diagram for demonstrating the hardware constitutions of a process part. (A) And (b) is a figure showing a sound source signal. (A) represents each pattern, (b) is a figure showing a table. (A) is a figure showing the superimposition of the output sound of patterns A-C, (b) is a figure showing the result of the frequency analysis with respect to the synthesized sound of (a). (A) is a figure showing cut-out of sound-recording data, (b) is a figure showing the cut-out period data. (A)-(d) represents each period data, (e)-(g) is the result of the minimum value calculation. (A)-(c) is a figure showing the example of the sound test with respect to the output sound of the extracted pattern B. FIG. (A) And (b) is a figure showing volume correction | amendment. (A)-(c) is a figure showing volume correction. (A) And (b) is a figure for demonstrating an example of the flowchart performed in the case of a sound test. It is a figure for demonstrating an example of the flowchart performed in the case of a sound test. It is a flowchart showing the detail of step S30. It is an example which extracts an output sound based on an amplitude.

  Prior to the description of the examples, the sound test will be described. The sound test is a test in which a sound output device provided in a personal computer, a mobile phone, an acoustic device, or the like is used as a test object and the sound output characteristics of the test object are inspected. For example, the presence or absence of sound output and sound quality abnormality (frequency characteristics, noise, chatter, sound cracking, etc.) are inspected. FIG. 1 is a diagram for explaining an outline of a sound test apparatus used for a sound test. With reference to FIG. 1, in the sound test, a sound signal output from a speaker 2 of a device under test moving on the conveyor line 1 is detected by a sound sensor (recording microphone) 3, and characteristics of the detected sound signal Is measured.

  Specific test object No. Only the output of the speaker 2 is turned on. 1, No. 1 In order to turn off the output of the speaker 2 of No. 3, a synchronization facility such as a wireless LAN is provided in the inspection apparatus and each pair to be tested. In this method, the test object No. No. 2 in the period during which 2 passes under the recording microphone 3. No. 2 is turned on, and other DUT No. 1, No. 1 3 output is turned off. Thereby, referring to FIG. 2 output sounds can be tested with high accuracy.

  However, when a communication abnormality or the like occurs, another device under test No. 1, No. 1 3 cannot be turned off. For this reason, referring to FIG. The sound output by 1 is detected as noise. As a result, the test object No. 2 may be determined to be poor. Therefore, in the following embodiments, a sound extraction device, a sound extraction method, and a sound extraction program that can accurately extract the output sound of a test object to be subjected to a sound test will be described.

  FIG. 3 is a block diagram for explaining the configuration of the sound extraction device 100 and the DUT 200 according to the first embodiment. Referring to FIG. 3, the sound extraction device 100 includes a passage detection sensor 10, a barcode reader 20, a communication control unit 30, a recording microphone 40, a processing unit 50, and the like. The DUT 200 includes a sound source data storage unit 201, a control unit 202, a speaker 203, and the like. FIG. 4 is a block diagram illustrating each function of the processing unit 50. Referring to FIG. 4, processing unit 50 functions as recording data storage unit 51, sound source information storage unit 52, Fourier transform unit 53, extraction unit 54, volume correction unit 55, determination unit 56, and output unit 57.

  FIG. 5 is a schematic diagram for explaining the positional relationship among the passage detection sensor 10, the barcode reader 20, the recording microphone 40, and the DUT 200. With reference to FIG. 5, the DUT 200 is placed on a belt conveyor 204. The passage detection sensor 10, the barcode reader 20, and the recording microphone 40 are disposed at positions separated from the DUT 200 on the belt conveyor 204.

  FIG. 6 is a block diagram for explaining the hardware configuration of the processing unit 50. Referring to FIG. 6, the processing unit 50 includes a CPU 101, a RAM 102, a storage device 103, an interface 104, and the like. Each of these devices is connected by a bus or the like. A CPU (Central Processing Unit) 101 is a central processing unit. The CPU 101 includes one or more cores. A RAM (Random Access Memory) 102 is a volatile memory that temporarily stores programs executed by the CPU 101, data processed by the CPU 101, and the like. The storage device 103 is a nonvolatile storage device. As the storage device 103, for example, a ROM (Read Only Memory), a solid state drive (SSD) such as a flash memory, a hard disk driven by a hard disk drive, or the like can be used. When the CPU 101 executes the sound extraction program, the processing unit 50 includes a recording data storage unit 51, a sound source information storage unit 52, a Fourier transform unit 53, an extraction unit 54, a volume correction unit 55, a determination unit 56, and an output unit 57. Function as.

  Next, the operation of each part of the DUT 200 and the sound extraction device 100 will be described. The sound source data storage unit 201 is a nonvolatile memory such as a ROM (read only memory), a hard disk, or a flash memory, and stores sound source signal data for sound tests. The control unit 202 includes a CPU (Central Processing Unit), a RAM (Random Access Memory), and the like, and generates a sound source signal for sound test from the sound source signal data stored in the sound source data storage unit 201. The speaker 203 outputs sound according to the sound source signal.

  The passage detection sensor 10 is a sensor that detects the passage of an object. In the present embodiment, the passage detection sensor 10 detects whether the DUT 200 has started passing near the recording microphone 40 in the moving direction of the belt conveyor 204. The barcode reader 20 is a sensor that reads a barcode associated with the device under test 200. The barcode is an ID for identifying each device under test 200 or the like. The location of the barcode is not particularly limited. For example, the barcode may be attached to the device under test 200 or may be attached to a tray or the like on which the device under test 200 is placed. The barcode reader 20 reads the barcode when the passage detection sensor 10 detects the passage of the device under test 200, and the read barcode is sent to the processing unit 50 via the communication control unit 30. Thereby, the device under test in which the passage is detected is specified. The recording microphone 40 is a sound sensor and detects sound output from the speaker 203.

  The processing unit 50 records the sound detected by the recording microphone 40. Specifically, the recording data storage unit 51 stores the sound detected by the recording microphone 40 as recording data. The sound source information storage unit 52 stores information related to the sound source signal for the sound test. The processing unit 50 inspects the sound output characteristics of the device under test 200 based on the recording data stored in the recording data storage unit 51 and the information on the sound source signal stored in the sound source information storage unit 52. Next, details of the operation of the processing unit 50 will be described.

  First, an example of a sound source signal reproduced by the device under test 200 will be described. FIG. 7A and FIG. 7B show examples of sound source signals. Referring to FIGS. 7A and 7B, the sound source signal includes a cue signal having a constant frequency and a signal under test after the cue signal. The cue signal is a sound source signal for identifying each device under test 200. The frequency of the cueing signal is different for each device under test 200. In the example of FIG. 7A, the frequency of the cue signal is constant at 1 KHz. In the example of FIG. 7B, the frequency of the cueing signal is constant at 2 KHz.

  The signal under test is a sound source signal that can be used for a sound output characteristic test, and is, for example, a chirp signal. The chirp signal is a swept sine wave signal having an audible range of 20 Hz to 20 kHz or the like, and is also called a swept sine wave signal. Specifically, the chirp signal is a signal whose frequency is continuously changed (swept) from a low frequency to a high frequency, or a signal whose frequency is continuously changed (swept) from a high frequency to a low frequency. The reproduction time of the signal under test is different for each device under test 200. In the example of FIG. 7A, the reproduction time tn of the signal under test is 5 seconds as an example. In the example of FIG. 7B, the reproduction time tn of the signal under test is 3 seconds as an example.

  Note that the reproduction time t1 of the cue signal may be the same for each device under test 200 or may be different. Further, the interval t2 from the end time of the cueing signal to the start time of the signal under test may be the same or different for each device under test 200. The reproduction cycle of each sound source signal is (reproduction time t1 + interval t2 + reproduction time tn). The reproduction cycle of each sound source signal is different for each device under test 200. This reproduction cycle is preferably set so that each of the DUTs 200 is not an integral multiple of each other.

  Referring to FIG. 8A, each device under test 200 repeatedly reproduces the sound source signal described above. That is, each device under test 200 alternately reproduces the cue signal and the signal under test. The frequency of the cueing signal of each device under test 200 and the reproduction cycle of the signal under test are determined in advance. As an example, the combination of the frequency of the cue signal and the reproduction period of the signal under test is No. No. 1 to be tested 200 is pattern A. No. 2 to be tested 200 is pattern B. In the test object 200 of No. 3, it is the pattern C.

  Referring to FIG. 8B, the sound source information storage unit 52 holds the pattern of each device under test 200 as a table. The sound extraction device 100 acquires the pattern of the device under test 200 in which passage is detected as a result of reading by the barcode reader 20. The sound extraction device 100 extracts the output sound of the device under test 200 that is closest to the recording microphone 40 (No. 2 in the example of FIG. 8A) according to the acquired pattern. Details of the output sound extraction will be described below.

  FIG. 9A is a diagram illustrating the superposition (synthetic sound) of the output sounds of patterns A to C. In the example of FIG. 1-No. Since the third device under test 200 reproduces the sound source signal, the recording data stored in the recording data storage unit 51 is a synthesized sound as shown in FIG. The Fourier transform unit 53 performs frequency analysis (such as short-time Fourier transform) on the synthesized sound. Short-time Fourier transform (StFt: Short-time Fourier Transform) is a method of analyzing a time change of a frequency spectrum by multiplying a window function while shifting it by a minute time Δt and sequentially performing Fourier transform. The short-time Fourier transform is a method generally used for a time-varying signal such as sound. A wavelet transform may be used for the same purpose.

  FIG. 9B is a diagram showing the result of frequency analysis for the synthesized sound of FIG. 9A. In FIG. 9B, the inclination differs depending on the pattern because the reproduction time of the signal under test described above is different. Since the reproduction time of the pattern A is the longest, the inclination is the smallest. Since the reproduction time of the pattern C is the shortest, the inclination is the largest. Referring to FIG. 9A and FIG. 9B, even if the first cue signal is output at the same time, the output timing of the next cue signal is different.

  Next, referring to FIG. 10A, the extraction unit 54 cuts out the recording data at the reproduction cycle of the device under test 200 that is the subject of the sound test. Specifically, the extraction unit 54 recognizes a part where the frequency of the cueing signal of the pattern B is continuous for a predetermined time (reproduction time t1), and cuts out the recording data at the period of the cueing signal. FIG. 10B is a diagram showing recorded data (periodic data) cut out a plurality of times.

  Next, the extraction part 54 extracts the output sound of the pattern B by calculating | requiring the logical product of each period data of FIG.10 (b). For example, the extraction unit 54 performs the minimum value calculation of the amplitude value for each minute section (Δt × Δf) of the time resolution (Δt) and the frequency resolution (Δf). In this case, since the output sound of pattern B is included in all periodic data, it remains as the minimum value. On the other hand, output sound and intermittent external noise components that are not subjected to the sound test are removed if they are not included even once in the plurality of periodic data. Therefore, the output sound of pattern B is extracted as a result of the minimum value calculation. An average value calculation may be performed.

  Fig.11 (a)-FIG.11 (d) represent each period data. FIG. 11E shows the result of the minimum value calculation between the period data in FIG. 11A and the period data in FIG. FIG. 11F shows the result of the minimum value calculation between the result of FIG. 11E and the period data of FIG. 11C. FIG. 11G shows the result of the minimum value calculation between the result of FIG. 11F and the period data of FIG. 11D. Referring to FIGS. 11 (e) to 11 (g), the output sound that is not the subject of the sound test is removed. Although the accuracy increases as the number of period data increases, the output sound that is the object of the sound test can be extracted by obtaining the logical product of at least two period data.

  12A to 12C are diagrams illustrating an example of a sound test for the output sound of the extracted pattern B. FIG. FIG. 12A shows an example when the output sound of pattern B is normal. Referring to FIG. 12A, if the output sound is normal, a cueing signal of pattern B and a signal under test are obtained as a result of a logical product operation on a plurality of period data.

  FIG. 12B shows an example in which the sound source signal of pattern B is not reproduced. Referring to FIG. 12B, the pattern B is not included in each period data. In this case, the output sound is not extracted as a result of the AND operation. Therefore, it can be determined that the sound source signal of pattern B is not reproduced. Even if the logical product operation is not performed, it may be determined that the sound source signal of pattern B is not reproduced when the cueing signal of pattern B is not detected.

  FIG. 12C shows an example of the case where the output sound of the pattern B includes harmonics such as resonance due to poor housing attachment. Referring to FIG. 12C, a specific frequency component is detected as a result of the logical product operation of each period data. In this way, it is possible to determine whether the output sound contains chatter, cracks, noise, or the like according to components detected in addition to the cue signal and the signal under test.

  By the way, since the device under test 200 is moving on the belt conveyor 204, it is conceivable that the volume detected by the recording microphone 40 changes. Therefore, the volume of the output sound that is the subject of the sound test may be corrected. Referring to FIG. 13A, the volume correction unit 55 reads the specific frequency (Fp) stored in the sound source information storage unit 52, and acquires the peak amplitude value Ap of the specific frequency (Fp) from each chirp signal. To do. The volume correction unit 55 obtains peak amplitude values Ap1, Ap2,..., ApN at a specific frequency (Fp) for each chirp signal. Since the DUT 200 moves relative to the recording microphone 40, each peak amplitude value becomes a different value with reference to FIG. 13B.

  The sound volume correction unit 55 generates an approximate curve f (t) most applicable to the point sequence data of each peak amplitude value. The approximate curve (regression curve) can be generated by a numerical calculation method such as high-order polynomial curve fitting (Curve Fitting) such as the least square method. The approximate curve is at least a quadratic expression, and is generated from point sequence data of three (degree of polynomial + 1) or more. FIG. 14A is a diagram for explaining an example of an approximate curve calculated from each peak amplitude value.

  The maximum amplitude value Apc, which is the maximum value of the approximate curve f (t), is estimated to correspond to the sound volume when the recording microphone 40 passes through the point of the shortest distance from the speaker 203. Further, the time Tpc corresponding to the maximum amplitude value Apc is estimated to correspond to the time when the recording microphone 40 passes through the point of the shortest distance from the speaker 203. Thus, even when the specific frequency (Fp) does not appear at the shortest distance between the speaker 203 and the recording microphone 40, the sound volume and time when the recording microphone 40 passes through the shortest distance from the speaker 203 are estimated. be able to.

  The volume correction unit 55 calculates a volume correction curve [Apc / f (t)] by multiplying the reciprocal of the approximate curve f (t) by the maximum amplitude value Apc. Next, referring to FIG. 14B, the volume correction unit 55 corrects the recording signal by matching the time and multiplying the recording signal by a volume correction curve [Apc / f (t)]. To do. In particular, since the maximum amplitude value Apc is used, each recording signal can be corrected to a state where the recording microphone 40 and the speaker 203 are located at the shortest distance point. FIG. 14C shows the corrected recording signal. Referring to FIG. 14C, the influence of the relative movement of the speaker 203 is avoided in the corrected recording signal.

  FIGS. 15A, 15 </ b> B, and 16 are diagrams for explaining an example of a flowchart executed in the sound test by the sound extraction device 100. The flowchart in FIG. 15A is a flowchart executed by the device under test 200. Referring to FIG. 15A, each device under test 200 outputs sound from the speaker 203 according to the sound source signal data stored in the sound source data storage unit 201 (step S1).

  In parallel with the flowchart of FIG. 15A, the flowchart of FIG. 15B is executed by the sound extraction device 100. Referring to FIG. 15B, the passage detection sensor 10 monitors the passage of the device under test 200 (step S11). Next, the passage detection sensor 10 determines whether or not the passage of any device under test 200 has been started (step S12). Specifically, it is determined whether or not the passage detection sensor 10 is turned on. If it is determined “No” in step S12, step S11 is executed again.

  When it is determined as “Yes” in Step S12, the barcode reader 20 reads a barcode associated with the DUT 200 in which passage is detected (Step S13). Next, the extraction unit 54 acquires an audio pattern corresponding to the read barcode (step S14). Next, the recording data storage unit 51 starts recording the output sound (step S15). Next, the passage detection sensor 10 monitors the passage of the device under test 200 (step S16). Next, the passage detection sensor 10 determines whether or not the passage of the DUT 200 that has passed has been completed (step S17). Specifically, it is determined whether or not the passage detection sensor 10 is turned off. If it is determined “No” in step S17, step S16 is executed again.

  If it is determined as “Yes” in step S17, the recording data storage unit 51 ends the recording of the output sound (step S18). Next, the recording data storage unit 51 stores the recording data (step S19). Next, the extraction unit 54 performs an extraction process of the output sound that is the subject of the sound test (step S20). FIG. 16 is a flowchart showing details of step S20.

  Referring to FIG. 16, volume correction unit 55 reads the recording data stored in recording data storage unit 51, and corrects the volume according to the procedure described in FIGS. 13 (a) to 14 (b). (Step S21). Next, the Fourier transform unit 53 performs short-time Fourier transform on the recording data corrected in Step 21 (Step S22). Next, the extraction unit 54 detects the frequency component of the cueing signal of the voice pattern read in step S14, and when the cueing signal is detected, sets a flag for each minute section (Δt × Δf) (step S23). ).

  Next, the extraction unit 54 determines whether one or more flags are detected (step S24). When it is determined as “Yes” in Step S24, the extraction unit 54 sets the detection counter N to “0” (Step S25). Next, the extraction unit 54 determines whether or not the flag ON period is equal to the reproduction time t1 of the cue signal (step S26). If it is determined as “Yes” in step S26, the extraction unit 54 sets the start time of the detected cueing signal as a start time variable, and adds “1” to the detection counter N (step S27).

  After the execution of step S27 or when it is determined “No” in step S26, the extraction unit 54 determines whether or not all the flags have been scanned (step S28). If it is determined “No” in step S28, step S26 is executed again. When it is determined as “Yes” in Step S <b> 28, the extraction unit 54 determines whether or not the start time variable is 4 or more (Step S <b> 29). That is, the extraction unit 54 determines whether four pieces of periodic data have been acquired. When it is determined as “Yes” in step S29, the extraction unit 54 extracts the output sound that is the subject of the sound test by performing the minimum value calculation using the four period data (step S30). Next, the determination unit 56 performs a test such as the frequency characteristics of the extracted output sound, and the output unit 57 outputs information for notifying the test result (step S31). Thereafter, the execution of the flowchart ends.

  If “No” is determined in step S24 or “No” is determined in step S29, the output unit 57 has an abnormality in the sound output of the DUT 200 that is the subject of the sound test. Information for notifying is output (step S32). Thereafter, the execution of the flowchart ends.

  FIG. 17 is a flowchart showing details of step S30. Referring to FIG. 17, the extraction unit 54 clears various variables UF [m] [n], m, n (step S41). The variable UF [m] “n” is a variable of the output sound that is the subject of the sound test. The variable m is a variable with a time resolution Δt. The variable n is a variable of the frequency resolution Δf. Next, the extraction unit 54 stores the minimum amplitude value in the variable UF [m] “n” in the minute sections of the four periodic data (step S42). Specifically, the extraction unit 54 converts the minimum amplitude component F [T] [m] [n] from the period T = 0, 1, 2, 3 of the four period data to the variable UF [m] “n”. (T: period of minimum value).

  Next, the extraction unit 54 determines whether or not the variable n is the maximum value (step S43). When it is determined as “No” in Step S43, the extraction unit 54 adds “1” to the variable n (Step S44). After step S44 is executed, step S42 is executed again. When it is determined as “Yes” in Step S44, the extraction unit 54 determines whether or not the variable m is the maximum value (Step S45). When it is determined as “Yes” in Step S45, the extraction unit 54 adds “1” to the variable m (Step S46). After execution of step S46, step S42 is executed again. When it is determined “Yes” in step S44, the flowchart of FIG. 17 ends.

  By repeating steps S42 to S44, the minimum value calculation is repeated until the variable n of the frequency resolution Δf reaches the maximum value. Further, by repeating steps S42, S45, and S46, the minimum value calculation is repeated until the variable m of the time resolution Δt reaches the maximum value.

  According to the present embodiment, since a cueing signal having a frequency individually determined for the device under test 200 is detected, the reproduction period of the output sound to be subjected to the sound test from the output sound of the plurality of devices under test 200. Can be specified. Further, by performing a logical product operation on a plurality of period data, it is possible to extract an output sound to be subjected to a sound test. As a result, even if a plurality of devices under test 200 output sound simultaneously, it is possible to extract the output sound to be subjected to the sound test. Further, since the output sound to be extracted is switched when the DUT 200 whose passage is detected by the passage detection sensor 10 is switched, the test can be sequentially performed on a plurality of DUTs 200 on the same line. . Further, since it is not necessary to turn on / off the output of each device under test 200, the sound extraction device 100 and the device under test 200 need not be synchronized. Thereby, a synchronous installation etc. become unnecessary and it can suppress cost.

  In the present embodiment, the passage detection sensor 10 functions as a detection unit that detects that the DUT 200 passes near the recording microphone 40. The recording data storage unit 51 functions as a storage unit that stores the output sound of the DUT 200 as recording data. The extracting unit 54 functions as a recognizing unit that recognizes a cue signal from the recorded data, and functions as an extracting unit that extracts the output sound of the DUT 200. The sound source information storage unit 52 functions as a holding unit that holds the relationship between the DUT 200 and the frequency of the cue signal determined individually for each DUT 200.

  By the way, paying attention to the fact that the distance between the DUT 200 in which passage has been detected and the recording microphone 40 is reduced and the volume is increased, a method of filtering the amplitude of each output sound included in the recording data is conceivable. In this case, with reference to FIG. 18, it is considered that the output sound to be subjected to the sound test can be extracted by extracting only the signal whose amplitude exceeds the threshold value. However, if the volume of the device under test 200 that is not subject to the sound test is increased due to an abnormality or the like, the amplitude of the output sound that is not subject to the sound test exceeds the threshold value. From the above, it is difficult to extract only the output sound to be subjected to the sound test by the method of filtering the amplitude of each output sound.

  Although the embodiments of the present invention have been described in detail above, the present invention is not limited to such specific embodiments, and various modifications and changes can be made within the scope of the gist of the present invention described in the claims. It can be changed.

DESCRIPTION OF SYMBOLS 10 Pass detection sensor 20 Barcode reader 30 Communication control part 40 Recording microphone 50 Processing part 51 Recording data storage part 52 Sound source information storage part 53 Fourier transform part 54 Extraction part 55 Volume correction part 56 Determination part 57 Output part 100 Sound extraction apparatus 101 CPU
102 RAM
DESCRIPTION OF SYMBOLS 103 Storage apparatus 200 Test object 201 Sound source data storage part 202 Control part 203 Speaker 204 Belt conveyor

Claims (11)

A detector for detecting that the DUT passes through the vicinity of the recording microphone;
When the passage is detected by the detection unit, a storage unit that stores output sound of the DUT as recording data using the recording microphone;
A recognizing unit for recognizing a cueing signal having a frequency individually determined for the DUT from the recorded data;
An extraction unit that extracts a plurality of period data based on the cue signal recognized by the recognition unit and extracts an output sound of the DUT by obtaining a logical product of the period data. A featured sound extraction device.   2. The sound extraction device according to claim 1, wherein the periodic data is periodic data from a cue signal of the DUT in which the passage is detected to a next cue signal. A holding unit that holds the relationship between each DUT and the frequency of the cue signal determined individually for each DUT,
The sound extraction apparatus according to claim 1, wherein the recognition unit recognizes a cueing signal having a frequency corresponding to the DUT in which the passage is detected from the recorded data.   The sound extraction apparatus according to any one of claims 1 to 3, wherein a period of sound repeatedly reproduced by the DUT is different for each DUT.   The sound extraction apparatus according to claim 4, wherein the sound to be repeatedly reproduced includes the cue signal and the chirp signal. Detect that the DUT passes near the recording microphone,
When the passage is detected, the output sound of the device under test is stored as recording data using the recording microphone,
Recognizing the cue signal of the frequency individually determined for the test object from the recorded data,
A sound extraction method comprising: extracting a plurality of period data based on a recognized cueing signal; and extracting an output sound of the DUT by obtaining a logical product of the period data.   The sound extraction method according to claim 6, wherein the periodic data is periodic data from a cue signal of the DUT in which the passage is detected to a next cue signal. Maintains the relationship between each device under test and the frequency of the cue signal determined individually for each device under test,
8. The sound extraction method according to claim 6 or 7, wherein a cueing signal having a frequency corresponding to the test object in which the passage is detected is recognized from the recorded data.   The sound extraction method according to any one of claims 6 to 8, wherein the period of the sound repeatedly reproduced by the device under test is different for each device under test.   The sound extraction method according to claim 9, wherein the sound to be repeatedly reproduced includes the cue signal and the chirp signal. On the computer,
Detecting that the DUT passes near the recording microphone;
Storing the output sound of the device under test as recording data using the recording microphone when the passage is detected;
Recognizing a cueing signal having a frequency individually determined for the DUT from the recorded data;
Extracting a plurality of periodic data based on the recognized cue signal and extracting the output sound of the device under test by obtaining a logical product of the periodic data. Extraction program.

Images