Noise reduction function control device and noise reduction function control method

Abstract

The present disclosure relates to a noise elimination function control device capable of allowing a user to identify keywords detected with respect to an external sound. A noise elimination function control device according to the present disclosure comprises: an external sound recording unit that records, in an external sound memory, external sound data input from the outside and information related to a time of input of the external sound data; a keyword detection unit that performs comparison processing for detecting whether one or more keywords set in advance are included in the external sound data recorded in the external sound memory and, when a keyword is included in the external sound data, outputs keyword detection data including information related to a start time and an end time of the keyword; and an external sound reproduction unit that reproduces the external sound including the keyword from among the external sound data recorded in the external sound memory on the basis of the keyword detection data.

Description

[Technical Field]

[0001] The present disclosure relates to a noise reduction function control device and a noise reduction function control method for controlling a noise reduction function in a playback device. [Background technology]

[0002] Conventional noise cancellation function control devices have the ability to detect whether sounds coming from outside (ambient sounds) contain the voice of a pre-set keyword, and use this detection as a trigger to stop the noise cancellation function of noise-canceling headphones (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] U.S. Patent No. 9,398,367 Summary of the Invention [Problem to be solved by the invention]

[0004] Since keyword detection is not guaranteed to be 100% accurate, it is inevitable that the noise reduction function will stop at unexpected times due to false keyword detection. Since users do not know what external sound the keyword was detected in, they have no choice but to determine whether the keyword was detected correctly by continuing to listen to the ambient sounds entering the headphones with the noise reduction function stopped, or by removing the headphones and listening to the external sounds directly.

[0005] The present disclosure has been made to solve such problems, and relates to a noise reduction function control device and a noise reduction function control method that enable a user to understand what external sound a keyword was detected for. [Means for solving the problem]

[0006] In order to solve the above-mentioned problems, the noise elimination function control device according to the present disclosure includes an external sound recording unit that records external sound data input from the outside and information on the time when the external sound data was input in an external sound memory; a keyword detection unit that performs a matching process to detect whether or not one or more preset keywords are included in the external sound data recorded in the external sound memory, and outputs keyword detection data that includes information on the start time and end time of the keyword if the keyword is included in the external sound data; and an external sound playback unit that plays external sounds that include the keyword from the external sound data recorded in the external sound memory based on the keyword detection data. The audio playback device further includes a playback start / stop time generation unit that generates a playback start time and a playback stop time of the external sound based on the keyword detection data output by the keyword detection unit and outputs the generated time as an external sound playback command; an external sound superimposition unit that superimposes an external sound playback signal corresponding to the external sound played by the external sound playback unit on a playback sound source signal input from a playback sound source and outputs the superimposed signal to a playback device with a noise reduction function; and a cancel command input unit that receives a cancel command from a user of the playback device when the user determines that an erroneous keyword detection has occurred, the external sound playback unit plays the external sound based on the external sound playback command, and when the playback of the external sound ends, outputs a keyword detection stop signal that stops keyword detection to the keyword detection unit and the playback start / stop time generation unit, and outputs a noise reduction control signal that stops the noise reduction function to the playback device, and the cancel command includes a command to cancel playback of the external sound, a command to cancel stopping keyword detection, and a command to cancel stopping the noise reduction function. . [Effects of the Invention]

[0007] According to the present disclosure, it is possible to allow a user to understand what external sound caused a keyword to be detected.

[0008] The objects, features, aspects, and advantages of the present disclosure will become more apparent from the following detailed description and the accompanying drawings. [Brief explanation of the drawings]

[0009] [Figure 1] 1 is a hardware configuration diagram showing an example of the configuration of a noise removal function control device according to a first embodiment. [Figure 2] 1 is a hardware configuration diagram showing an example of the configuration of noise-canceling headphones according to a first embodiment. [Figure 3] 1 is a hardware configuration diagram showing an example of the configuration of a noise removal function control device according to a first embodiment. [Figure 4] 1 is a hardware configuration diagram showing an example of the configuration of noise-canceling headphones according to a first embodiment. [Figure 5] 1 is a functional block diagram showing an example of the configuration of a noise removal function control device according to a first embodiment. [Figure 6]4 is a flowchart showing an example of the operation of the noise removal function control device according to the first embodiment. [Figure 7] 4 is a diagram for explaining a data arrangement in an external sound memory according to the first embodiment. FIG. [Figure 8] 4 is a flowchart showing an example of the operation of the noise removal function control device according to the first embodiment. [Figure 9] 4 is a flowchart showing an example of the operation of the noise removal function control device according to the first embodiment. [Figure 10] FIG. 10 is a functional block diagram showing an example of the configuration of a noise removal function control device according to a second embodiment. [Figure 11] 10 is a flowchart showing an example of the operation of the noise removal function control device according to the second embodiment. [Figure 12] FIG. 11 is a hardware configuration diagram showing an example of the configuration of a noise removal function control device according to a third embodiment. [Figure 13] FIG. 11 is a hardware configuration diagram showing an example of the configuration of a noise removal function control device according to a third embodiment. [Figure 14] FIG. 11 is a functional block diagram showing an example of the configuration of a noise removal function control device according to a third embodiment. [Figure 15] 11 is a flowchart showing an example of the operation of the noise removal function control device according to the third embodiment. [Figure 16] 11 is a flowchart showing an example of the operation of the noise removal function control device according to the third embodiment. [Figure 17] FIG. 10 is a functional block diagram showing an example of the configuration of a noise removal function control device according to a fourth embodiment. [Figure 18] 10 is a flowchart showing an example of the operation of the noise removal function control device according to the fourth embodiment. [Figure 19] FIG. 11 is a functional block diagram showing an example of the configuration of a noise removal function control device according to a fifth embodiment. [Figure 20] 13 is a flowchart showing an example of the operation of the noise removal function control device according to the fifth embodiment. [Figure 21]FIG. 13 is a diagram illustrating detection of a subsequent utterance according to the fifth embodiment. [Figure 22] FIG. 13 is a hardware configuration diagram showing an example of the configuration of a noise removal function control device according to a sixth embodiment. [Figure 23] FIG. 13 is a hardware configuration diagram showing an example of the configuration of a noise removal function control device according to a sixth embodiment. [Figure 24] FIG. 13 is a functional block diagram showing an example of the configuration of a noise removal function control device according to a sixth embodiment. [Figure 25] 13 is a flowchart showing an example of the operation of the noise removal function control device according to the sixth embodiment. [Figure 26] 13 is a flowchart showing an example of the operation of the noise removal function control device according to the sixth embodiment. [Figure 27] FIG. 13 is a functional block diagram showing an example of the configuration of a noise removal function control device according to a seventh embodiment. [Figure 28] 13 is a flowchart showing an example of the operation of the noise removal function control device according to the seventh embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0010] <First Embodiment> <Configuration> The noise removal function control device according to the first embodiment will be described with reference to FIGS.

[0011] FIG. 1 is a hardware configuration diagram showing an example of the configuration of a noise removal function control device according to the first embodiment.

[0012] The external sound input microphone 1 is provided to capture external sound. The external sound captured by the external sound input microphone 1 is input to an A / D converter 3 via a microphone amplifier 2. The number of external sound input microphones 1 is not limited to one, and multiple microphone amplifiers 2 may be provided. When multiple external sound input microphones 1 are provided, it is sufficient to provide the same number of microphone amplifiers 2 as the number of external sound input microphones 1.

[0013] The A / D converter 3 performs A / D conversion on the external sound at, for example, 16 kHz / 16 bits, and outputs the converted digital data (external sound data) to a CPU (Central Processing Unit) 4. If multiple external sound input microphones 1 are provided, external sound data of multiple channels is output to the CPU 4.

[0014] The CPU 4 stores the external sound data input from the A / D converter 3 and information about the time the external sound was input to the external sound input microphone 1 in the external sound memory 5. The CPU 4 also performs a comparison process to detect whether the input external sound data contains a preset keyword. When the CPU 4 detects a keyword, it reads out the external sound data (external sound containing the keyword) stored in the external sound memory 5 based on the information about the start time and end time of the keyword, and outputs the read external sound data to the D / A converter 6.

[0015] The D / A converter 6 converts the external sound data from digital to analog at, for example, 16 kHz / 16 bits, and outputs the converted external sound to the audio amplifier 7. The audio amplifier 7 amplifies the external sound and outputs the amplified external sound reproduction signal 108 (FIG. 5) to the audio mixer 8.

[0016] An audio mixer (external sound superimposing unit) 8 mixes a signal input from a playback sound source input terminal 9 with an external sound playback signal 108 input from an audio amplifier 7, and outputs the mixed signal to an audio signal input terminal 11 of noise-canceling headphones 10. The signal input from the playback sound source input terminal 9 refers to a signal that the user intends to listen to, such as an audio signal output from a smartphone, for example.

[0017] After the output of the external sound data is completed, the CPU 4 outputs a noise cancellation control signal 110 (FIG. 5) to the noise cancellation control signal input terminal 12 of the noise cancellation headphones 10 to cancel the noise cancellation function of the noise cancellation headphones 10.

[0018] The cancel command input switch 13 is a switch for inputting a cancel command 114 (FIG. 5) that cancels the processing executed as a result of keyword detection when the user determines that a keyword has been erroneously detected as a result of listening to external sounds through the noise-canceling headphones 10. The cancel command 114 output from the cancel command input switch 13 is input to the CPU 4.

[0019] When the cancel command 114 is input, the CPU 4 stops the playback of the external sound if the external sound is being played back, and also cancels the output of the noise cancellation control signal 110 that cancels the noise cancellation function of the noise cancellation headphones 10, which was scheduled to be output to the noise cancellation headphones 10 after the playback of the external sound has finished. Furthermore, if the playback of the external sound has already finished and the noise cancellation control signal 110 that cancels the noise cancellation function of the noise cancellation headphones 10 has already been output when the cancel command 114 is input, the CPU 4 outputs the noise cancellation control signal 110 that executes the noise cancellation function of the noise cancellation headphones 10.

[0020] The program executed by the CPU 4, the data to be read, the data to be written, and the keywords are stored in the memory 14.

[0021] Noise-canceling headphones 10 (a playback device with a noise cancellation function) reproduces audio signals input from audio mixer 8 using driver unit 15. Furthermore, noise-canceling headphones 10 switch the noise cancellation function ON or OFF in accordance with a noise cancellation control signal 110 input from CPU 4.

[0022] 2 is a hardware configuration diagram showing an example of the configuration of the noise-canceling headphones 10. The configuration of the noise-canceling headphones 10 is the same as the configuration of commercially available noise-canceling headphones.

[0023] The external sound signal input microphone 17 is attached to the outside of the noise canceling headphones 10 and captures external sounds.

[0024] The error signal input microphone 18 is provided in the vicinity of the driver unit 15 inside the noise cancellation headphones 10.

[0025] The noise cancellation circuit 16 uses one or both of the signals from the external sound signal input microphone 17 and the error signal input microphone 18 to generate a signal that cancels external noise that enters through the main body of the noise cancellation headphones 10 and the earmuffs, and superimposes the generated signal on the signal input to the audio signal input terminal 11 to output it as a noise-canceled playback signal 19.

[0026] The driver unit 15 reproduces the noise-reduced playback signal 19 in the noise-canceling headphones 10.

[0027] A noise elimination control signal that controls ON / OFF of the noise elimination function is input to the noise elimination control signal input terminal 12 .

[0028] Fig. 3 is a hardware configuration diagram showing an example of another configuration of the noise cancellation function control device according to embodiment 1. Fig. 3 shows a configuration in which an external sound signal input microphone 17 provided in a noise cancellation headphone 10 is used in place of the external sound input microphone 1 shown in Fig. 1. The other configurations are the same as those shown in Fig. 1.

[0029] Fig. 4 is a hardware configuration diagram showing another example of the configuration of the noise-canceling headphones 10 according to embodiment 1. The noise-canceling headphones 10 shown in Fig. 4 are provided with an external sound microphone signal output terminal 20 that extracts the signal output by the external sound signal input microphone 17 to the outside. The other configurations are the same as those shown in Fig. 2.

[0030] FIG. 5 is a functional block diagram illustrating an example of the configuration of the noise removal function control device according to the first embodiment.

[0031] The external sound data input unit 101 receives the external sound data output from the A / D converter 3 .

[0032] The external sound recording unit 102 writes the external sound data and information about the time when the external sound data was input to the external sound data input unit 101 into the external sound memory 5 without delay.

[0033] Through initialization processing, the keyword detection unit 103 performs a process of comparing the external sound data written to the external sound memory 5 by the external sound recording unit 102 with a preset keyword. When a keyword is detected in the external sound data, the keyword detection unit 103 outputs information on the start time and end time of the keyword and the ID of the keyword as keyword detection data 104. The keyword detection unit 103 outputs keyword detection data 104 every time it detects a keyword until a keyword detection stop signal 111, which will be described later, is input. Here, the start time of a keyword refers to the start time when the keyword is uttered, and the end time of a keyword refers to the end time when the keyword is uttered.

[0034] The playback start / stop time generating unit 105 generates the playback start time and playback stop time for the external sound data stored in the external sound memory 5 according to the input keyword detection data 104. The generation rules for generating the playback start time and playback stop time are as follows: Playback start time = Keyword start time - 0.5 seconds Playback stop time = Keyword end time + 0.5 seconds

[0035] By using such a generation rule, even if the matching algorithm has a tendency to determine the start and end times of the keyword more in-depth than the actual utterance, the start and end parts of the uttered keyword are reproduced without any loss.

[0036] The playback start / stop time generation unit 105 outputs the generated playback start time and playback stop time to the external sound playback unit 107 as an external sound playback command 106. The playback start / stop time generation unit 105 repeats the above operation every time keyword detection data 104 is input, until a keyword detection stop signal 111 (described later) is input.

[0037] The external sound playback unit 107 reads the external sound data corresponding to the time between the playback start time and playback stop time included in the external sound playback command 106 from the external sound memory 5, and outputs the read external sound data to the D / A converter 6.

[0038] An external sound reproduction signal 108 obtained via the D / A converter 6 and the audio amplifier 7 is mixed in an audio mixer 8 with a signal input from a reproduction sound source input terminal 9 (a signal intended for listening by the user, such as an audio output signal from a smartphone). The audio mixer 8 outputs the mixed signal to an audio signal input terminal 11 of noise-canceling headphones 10.

[0039] The noise cancellation unit 109 in the noise cancellation headphones 10 is currently performing a noise cancellation function through initialization processing, and uses the output signals from the external sound signal input microphone 17 and the error signal input microphone 18 to generate a signal that cancels external noise that has passed through the main body and earmuffs of the noise cancellation headphones 10 and entered the noise cancellation headphones 10, and outputs the resulting noise-canceled playback signal 19 to the driver unit 15 by superimposing this signal on the signal input from the audio signal input terminal 11.

[0040] The driver unit 15 reproduces the noise-reduced playback signal 19 in the noise-canceling headphones 10.

[0041] The external sound reproducing unit 107 performs the following three processes when the reproduction of the external sound corresponding to the reproduction start time and the reproduction stop time of the external sound reproducing command 106 ends. A noise cancellation control signal 110 for stopping the noise cancellation function is output to the noise cancellation unit 109 of the noise cancellation headphones 10 via the noise cancellation control signal input terminal 12. A keyword detection stop signal 111 is sent to the keyword detection unit 103 and the playback start / stop time generation unit 105. The output of the external sound data stored in the external sound memory 5 to the D / A converter 6 is stopped (the playback of the external sound is stopped).

[0042] The cancel command input unit 113 accepts a cancel command 114 input from the cancel command input switch 13 .

[0043] When the external sound reproducing unit 107 receives the cancel command 114, it performs processing according to the following condition 1 or condition 2. Condition 1 "External sound is being played": The output of external sound data to the D / A converter 6 is stopped, and a keyword detection stop signal 111 is temporarily output. Next, a keyword detection resume signal 112 containing information on the final end time of external sound playback is output, and input of the next external sound playback command 106 is awaited. Condition 2 "External sound playback has already been completed": A noise removal control signal 110 is output to the noise removal unit 109 to turn on the noise removal function, and then a keyword detection resume signal 112 (but does not include time information) is output to the keyword detection unit 103 to resume keyword detection, and input of the next external sound playback command 106 is awaited.

[0044] In addition, when a new external sound playback command 106 is input before the playback of the external sound corresponding to the playback start time and playback stop time of the external sound playback command 106 ends, the external sound playback unit 107 continues playing the external sound until the playback stop time included in the new external sound playback command 106.

[0045] The noise cancellation function control device according to the first embodiment may be configured integrally with the noise cancellation headphones 10, or may be configured integrally with a playback sound source (for example, a smartphone) connected to the playback sound source input terminal 9.

[0046] The functions of the external sound recording unit 102, keyword detection unit 103, playback start / stop time generation unit 105, and external sound playback unit 107 in the noise reduction function control device described above are realized by processing circuits. That is, the noise reduction function control device includes a processing circuit for storing external sound data and information on the time the external sound data was input in external sound memory 5 without delay, detecting keywords from the external sound data stored in external sound memory 5, generating playback start times and playback stop times for the external sound stored in external sound memory 5, and playing the external sound according to the playback start times and playback stop times. The processing circuit may be dedicated hardware, or may be a processor (also called a CPU, central processing unit, processing unit, arithmetic unit, microprocessor, microcomputer, or DSP (Digital Signal Processor)) that executes a program stored in memory.

[0047] When the processing circuit is dedicated hardware, the processing circuit may be, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or a combination thereof. The functions of the external sound recording unit 102, the keyword detection unit 103, the playback start / stop time generation unit 105, and the external sound playback unit 107 may be realized by separate processing circuits, or all of the functions may be realized by a single processing circuit.

[0048] When the processing circuit is a CPU 4, the functions of the external sound recording unit 102, keyword detection unit 103, playback start / stop time generation unit 105, and external sound playback unit 107 are realized by software, firmware, or a combination of software and firmware. The software or firmware is written as a program and stored in memory 14. The CPU 4 realizes each function by reading and executing the program recorded in memory 14. That is, the noise reduction function control device includes memory 14 for storing a program that ultimately executes the steps of: storing external sound data and information on the time the external sound data was input in external sound memory 5 without delay; detecting keywords from the external sound data stored in external sound memory 5; generating playback start times and playback stop times for the external sound stored in external sound memory 5; and playing the external sound according to the playback start times and playback stop times. These programs can also be said to cause a computer to execute the procedures or methods of the external sound recording unit 102, keyword detection unit 103, playback start / stop time generation unit 105, and external sound playback unit 107. Here, memory 14 may be, for example, a non-volatile or volatile semiconductor memory such as RAM (Random Access Memory), ROM (Read Only Memory), flash memory, EPROM (Erasable Programmable Read Only Memory), EEPROM (Electrically Erasable Programmable Read Only Memory), a magnetic disk, a flexible disk, an optical disk, a compact disk, a DVD (Digital Versatile Disc), or any storage medium that will be used in the future.

[0049] It is also possible to realize some of the functions of the external sound recording unit 102, keyword detection unit 103, playback start / stop time generation unit 105, and external sound playback unit 107 by using dedicated hardware, and other functions by using software or firmware.

[0050] Thus, the processing circuitry can implement each of the above-described functions through hardware, software, firmware, or a combination thereof.

[0051] Although the hardware configuration of the noise elimination function control device according to the first embodiment has been described above, the same applies to the hardware configurations of the noise elimination function control devices according to the second to seventh embodiments described below.

[0052] <Operation> Next, a detailed description will be given of the operation of the noise reduction function control device according to Embodiment 1. Fig. 6 is a flowchart showing the operation of the external sound recording unit 102 in the noise reduction function control device according to Embodiment 1. The operation shown in Fig. 6 is performed in parallel as a separate process from the operations shown in Figs. 8 and 9, which will be described later.

[0053] The operation shown in FIG. 6 starts when the noise cancellation function control device is powered on.

[0054] In step S1, the external sound recording unit 102 initializes the memory, data, and parameters related to recording of external sound data.

[0055] With regard to the external sound memory 5, which is one of the targets of the initialization process, an example of data and its arrangement in the external sound memory 5 is shown in Figure 7. The management information for the external sound data includes a first data index, a first data time, a last data index, and a last data time. In addition, an external sound data array DT(n) (n = 1 to N) that stores N samples of external sound data is arranged. The number N of external sound data arrays is assumed to be sufficient to store external sound data during one use of the noise reduction function control device according to embodiment 1 (one use is considered to be from power on to power off).

[0056] The first data index is the index value of the storage destination array of the oldest external sound data among the external sound data stored in the external sound memory 5 by the external sound recording unit 102. The first data time is the time when the oldest external sound data among the external sound data stored in the external sound memory 5 by the external sound recording unit 102 is taken into the external sound recording unit 102 from the external sound data input unit 101. The last data index is the index value of the storage destination array of the last external sound data among the external sound data stored in the external sound memory 5 by the external sound recording unit 102. The last data time is the time when the last external sound data among the external sound data stored in the external sound memory 5 by the external sound recording unit 102 is taken into the external sound recording unit 102 from the external sound data input unit 101.

[0057] By the initialization in step S1, the external sound recording unit 102 sets the first data index, the first data time, the last data index, and the last data time to zero. The external sound recording unit 102 also sets the external sound data array DT(n) (n=1 to N) to zero.

[0058] In step S2, the external sound recording unit 102 acquires current time information from the clock information in the CPU 4. The time information may be the actual time or a relative elapsed time starting from the time initialized in step S1.

[0059] In step S3, the external sound recording unit 102 reads the external sound data output by the A / D converter 3. The number of read samples is assumed to be K. Note that the number of read samples may vary each time the process of step S3 is executed.

[0060] In step S4, the external sound recording unit 102 writes the external sound data of the number of samples K read in step S3 to the external sound data array of the external sound memory 5. That is, the external sound data is written to DT (final data index+1) to DT (final data index+K) in the external sound memory 5.

[0061] In step S5, the external sound recording unit 102 updates the management information in the external sound memory 5. Specifically, the following two processes are performed. Update process 1: If the first data index is zero, the first data index is set to "1" and the first data time is set to the time information acquired in step S2. Update process 2: K is added to the final data index. Also, "(1 / Fs) x K" is added to the final data time. Fs is the sampling frequency of A / D conversion in the A / D converter 3. In the first embodiment, Fs=16000.

[0062] In step S6, the external sound recording unit 102 determines whether it is possible to write the next acquired external sound data (whether there is a destination to write the next external sound data). As described above, if the data array DT that stores external sound data has a number of arrays that is large enough to store all of the external sound data during one use of the noise reduction device according to embodiment 1, this determination always indicates that there is a destination to write to, and the process returns to step S2. On the other hand, if, for some reason, there is a situation in which external sound data is written that exceeds the number of arrays DT that are available, the determination indicates that there is no destination to write to, and the process proceeds to step S7.

[0063] In step S7, the noise reduction function control device according to the first embodiment executes a process for initializing the entire noise reduction function control device, and then proceeds to the initialization process in step S1.

[0064] 8 is a flowchart showing the operations of the keyword detection unit 103 and the playback start / stop time generation unit 105 in the noise removal function control device according to Embodiment 1. These operations are performed in parallel as separate processes from the operations of the external sound recording unit 102 shown in FIG. 6 and the operations shown in FIG. 9 described above.

[0065] The operation shown in FIG. 8 starts when the noise cancellation function control device is powered on.

[0066] In step S11, the noise reduction function control device initializes itself. Specifically, the noise reduction function control device performs the following three processes to initialize the keyword detection unit 103. · Obtain time information and set that time as the keyword detection start time Tkst. Load the data of keywords to be detected. -Initialize the matching process to detect keywords.

[0067] Furthermore, the noise removal function control device performs the following process to initialize the playback start / stop time generation unit 105. -Read the rules for generating the playback start time and playback stop time.

[0068] In step S12, the keyword detection unit 103 executes a matching process to detect a keyword. Specifically, the keyword detection unit 103 reads one or more samples of external sound data from the external sound memory 5, starting from the external sound data corresponding to the keyword detection start time Tkst up to the most recently written external sound data, and performs a matching process to detect a keyword. The keyword detection start time Tkst is updated by the time corresponding to the number of samples of the read external sound data. With this update, the keyword detection start time Tkst corresponds to the last (latest) external sound data input for the matching process.

[0069] In step S13, the keyword detection unit 103 determines whether or not a keyword has been detected. If a keyword has not been detected, the process proceeds to step S15. On the other hand, if a keyword has been detected, the process proceeds to step S14.

[0070] In step S14, the playback start / stop time generation unit 105 generates the playback start time and playback stop time for the external sound data in the external sound memory 5 using information on the start time and end time of the detected keyword (keyword detection data 104) and the playback start time and playback stop time generation rule read in step S11, and outputs this time information to the external sound playback unit 107 as an external sound playback command 106.

[0071] In step S15, keyword detection unit 103 and playback start / stop time generation unit 105 determine whether or not they have received keyword detection stop signal 111 output by external sound playback unit 107. If keyword detection stop signal 111 has not been received, the process returns to step S12. On the other hand, if keyword detection stop signal 111 has been received, the process proceeds to step S16. Note that keyword detection stop signal 111 is transmitted when a cancel command 114 is input during playback of external sound, or when playback of external sound reaches the playback stop time (details will be described later).

[0072] In step S16, the keyword detection unit 103 performs processing to stop the keyword detection that has been being performed, and stops outputting the keyword detection data 104. Furthermore, the playback start / stop time generation unit 105 performs processing to stop generating the playback start time and playback stop time related to the playback of the external sound, and stops outputting the external sound playback command 106.

[0073] In step S17, keyword detection unit 103 and playback start / stop time generation unit 105 determine whether or not they have received keyword detection resume signal 112 output by external sound playback unit 107. If keyword detection resume signal 112 has been received, the process proceeds to step S18.

[0074] In step S18, the keyword detection unit 103 updates the keyword detection start time Tkst based on the time information included in the keyword detection resume signal 112 received in step S17. If the keyword detection resume signal 112 does not include time information, the Tkst updated in the process of step S12 is used as is. Then, the process proceeds to step S12.

[0075] The time information included in keyword detection resume signal 112 corresponds to the time when cancel command 114 was input during playback of external sound, and is used to resume keyword detection from the beginning of external sound that the user has not yet heard. If the cancel command 114 is input after the user has listened to playback of external sound up to the playback stop time, keyword detection resume signal 112 that does not include time information is input to keyword detection unit 103, and keyword detection unit 103 resumes keyword detection from the oldest external sound data that has not been subject to the matching process in step S12 (details will be described later).

[0076] 9 is a flowchart showing the operation of the external sound reproducing unit 107 and the noise canceling headphones 10 in the noise cancellation function control device according to embodiment 1. This operation is performed in parallel as a separate process from the operation of the external sound recording unit 102 shown in FIG. 6 described above, and the operation of the keyword detecting unit 103 and the reproduction start / stop time generating unit 105 shown in FIG.

[0077] The operation shown in FIG. 9 starts when the noise removal function control device according to the first embodiment is powered on.

[0078] In step S21, the noise reduction function control device initializes itself. Specifically, the noise reduction function control device performs the following process to initialize the external sound reproducing unit 107. The playback start / stop time generation unit 105 starts accepting input of the external sound playback command 106 output therefrom.

[0079] The noise cancellation function control device also performs the following process to initialize the noise cancellation headphones 10. A noise elimination function is performed based on signals from the external sound signal input microphone 17 and the error signal input microphone 18. The function of reproducing a sound source signal such as a music source input from the audio signal input terminal 11 through the driver unit 15 in the headphones is performed. The D / A converter 6, the audio amplifier 7, and the audio mixer 8 that play back external sounds are activated.

[0080] In step S22, the external sound playback unit 107 determines whether or not the external sound playback command 106 has been received from the playback start / stop time generation unit 105. If not received, the process of step S22 is looped. On the other hand, if received, the process proceeds to step S23.

[0081] In step S23, the external sound playback unit 107 starts reading out the external sound data stored in the external sound memory 5 at the corresponding playback start time included in the received external sound playback command 106, and sequentially outputs the read external sound data to the D / A converter 6. The external sound data converted into an analog signal by the D / A converter 6 is amplified by the audio amplifier 7 and input to the audio mixer 8 as an external sound playback signal 108. The audio mixer 8 mixes the input external sound playback signal 108 with a sound source such as music input from a playback sound source input terminal 9, and outputs the mixed signal (audio signal) to the audio signal input terminal 11 of the noise canceling headphones 10. The noise canceling headphones 10 play back the audio signal input from the audio signal input terminal 11 within the headphones via the driver unit 15.

[0082] In step S24, the external sound playback unit 107 determines whether the playback of the external sound started in step S23 has progressed to the playback stop time included in the external sound playback command 106. If the playback has not progressed to the playback stop time, the process proceeds to step S28. On the other hand, if the playback has progressed to the playback stop time, the process proceeds to step S25.

[0083] Step S28 is a process that is executed when it is determined in step S24 that the playback of the external sound has not progressed to the playback stop time. In step S28, the external sound playback unit 107 determines whether or not the user has operated the cancel command input switch 13 and received a cancel command 114 input via the cancel command input unit 113. If the cancel command 114 has been received, the process proceeds to step S31. On the other hand, if the cancel command 114 has not been received, the process proceeds to step S29.

[0084] In step S31, the external sound playback unit 107 stops the external sound playback process, i.e., the process of reading out the external sound data from the external sound memory 5 and outputting it to the D / A converter 6. The external sound playback unit 107 also outputs a keyword detection stop signal 111 to the keyword detection unit 103 and the playback start / stop time generation unit 105, and proceeds to step S32.

[0085] In step S32, the external sound playback unit 107 outputs to the keyword detection unit 103 the keyword detection resume signal 112 including, as time information, the end time of the playback of the external sound stopped in step S31 (the final time of the played external sound data).

[0086] The processes in steps S31 and S32 correspond to the case where the user determines that the keyword detection was incorrect when listening to the reproduced external sound partway through, or determines that the detected keyword does not require stopping the noise cancellation function of the noise-canceling headphones 10. This process is for restarting detection of a new keyword immediately after the end of the external sound listened to by the user, that is, starting from external sound data that the user has not yet listened to.

[0087] After step S32, the process returns to step S22, and the external sound reproducing unit 107 waits for the next external sound reproducing command 106 to be received.

[0088] In step S29, the external sound playback unit 107 determines whether or not a new external sound playback command 106 has been received during playback of the external sound. If a new external sound playback command 106 has not been received, the process returns to step S24. On the other hand, if a new external sound playback command 106 has been received, the process proceeds to step S30.

[0089] In step S30, the external sound playback unit 107 overwrites the playback stop time included in the previously received external sound playback command 106 with the playback stop time included in the newly received external sound playback command 106, and returns to step S24. As a result, in accordance with the detection result of the keyword that caused the newly received external sound playback command 106 to be generated, playback of the external sound continues until the end of the keyword, and when multiple keywords are uttered consecutively, the external sound is played so as to include these multiple keywords.

[0090] Step S25 is a process that is executed when it is determined in step S24 that the playback of the external sound has progressed to the playback stop time. In step S25, the external sound playback unit 107 performs the following three processes. Then, the process proceeds to step S26. A noise cancellation control signal 110 is output to the noise cancellation headphones 10 to stop the noise cancellation function. A keyword detection stop signal 111 is output to the keyword detection unit 103 and the playback start / stop time generation unit 105. The external sound reproduction process in the external sound reproduction unit 107, that is, the process of reading the external sound data from the external sound memory 5 and outputting it to the D / A converter 6, is stopped.

[0091] In step S26, the external sound reproducing unit 107 determines whether or not it has received a cancel command 114 input via the cancel command input unit 113 by the user operating the cancel command input switch 13. The cancel command 114 is a command for turning on again the noise cancellation function of the noise canceling headphones 10 that was stopped in the processing in step S25. If it has not received the cancel command 114, it loops the processing of step S26 (i.e., it waits for reception of the cancel command 114). On the other hand, if it has received the cancel command 114, it proceeds to step S27.

[0092] In step S27, the external sound reproducing unit 107 performs the following two processes. The noise cancellation control signal 110 for restarting the noise cancellation function of the noise cancellation headphones 10 that was stopped in step S25 is output to the noise cancellation control signal input terminal 12 of the noise cancellation headphones 10. This restarts the processing of the noise cancellation unit 109 in the noise cancellation headphones 10. In step S25, a keyword detection resume signal 112 is output to the keyword detection unit 103 as a signal for canceling the keyword detection stop signal 111 output to the keyword detection unit 103 and playback start / stop time generation unit 105. The keyword detection resume signal 112 output at this time does not include time information. This ends the processing loop in step S17 of Fig. 8, and the process proceeds to step S18, where new keyword detection is performed.

[0093] After step S27, the process returns to step S22, and the external sound reproducing unit 107 waits for a new external sound reproducing command 106 to be received.

[0094] <Effects> From the above, according to the noise cancellation function control device of embodiment 1, when a keyword set to stop the noise cancellation function of the noise cancellation headphones 10 is detected as being contained in an external sound, the external sound containing that keyword is played back within the noise cancellation headphones 10, allowing the user to determine whether the keyword detection was correct or not.

[0095] Furthermore, if the user determines that the keyword detection was incorrect after listening to the reproduced external sound, the user can cancel the process of stopping the noise removal function, which is triggered by the keyword detection, by operating the cancel command input switch 13.

[0096] In addition, the above description has been given of the generation rules read by the playback start / stop time generation unit 105. Playback start time = Keyword start time - 0.5 seconds Playback stop time = Keyword end time + 0.5 seconds is used as an example, but it is not limited to this. By changing the generation rule according to the keyword you set, it is possible to play appropriate external sounds. For example, the generation rule for the keyword "Mr. / Ms. X" (calling the user's name) can be changed to Playback start time = Keyword start time - 0.5 seconds Playback stop time = Keyword end time + 5 seconds By doing so, the utterances such as requirements following the calling utterance "Mr. / Ms. XX" are also played back as external sounds, making it possible for the user to hear the content of the utterance following the keyword.

[0097] In addition, the correction times for the start time and end time in the generation rules may be dynamically changed taking into account the level of background noise, the spectral characteristics of the background noise, or the level of the voice of the detected keyword. As an example, in a situation where one or both of the following conditions are met: the level of background noise is low and the level of the voice of the detected keyword is high, the SNR (Signal to Noise Ratio) of the voice of the keyword is increased, allowing for accurate keyword detection. Therefore, by applying a correction time shorter than that specified in the above generation rule, the time the user hears external sounds other than the keyword before and after the utterance of the keyword can be shortened. In other words, the following generation rule is applied. ·Playback start time = Keyword start time - Ofst_b Playback stop time = Keyword end time + Ofst_e However, Ofst_b=0.1 seconds and Ofst_e=0.1 seconds.

[0098] Conversely, for keywords detected in situations where either or both of the following conditions are met: high background noise level, low keyword audio level (low SNR of the keyword audio), if the characteristics of the detection algorithm cause the beginning and ending of the word to sound more in-depth than the actual speech, you can set Ofst_b and Ofst_e to a time longer than the above 0.1 seconds. This will prevent the beginning and ending of the keyword from being missed when it is spoken.

[0099] In the above description, the number N of external sound data arrays stored in the external sound memory 5 is set to a value sufficient to store external sound data during one use of the noise elimination function control device according to embodiment 1 (one use is defined as from power-on to power-off), but this is not limited to this. For example, in order to save memory space, a ring buffer configuration may be adopted in which, when external sound data is written to the end of the external sound data array, the oldest written data (i.e., the beginning of the external sound data array) is deleted and replaced with new data. In this case, the leading data index included in the management information in the external sound memory 5 is not fixed to "1," but becomes the index value of the array in which the oldest external sound data is stored in the external sound data array, and accordingly, the reading and writing of external sound data is also performed in accordance with the ring buffer.

[0100] Furthermore, the audio mixer 8 (external sound superimposing unit) is not limited to mixing the external sound reproduction signal 108 with a signal intended for listening by the user, such as an audio output signal from a smartphone input from the playback sound source input terminal 9, at the same level as it is. For example, when the audio mixer 8 detects the input of the external sound reproduction signal 108, it may attenuate the signal input from the playback sound source input terminal 9 by -6 dB or -10 dB before mixing. This makes the external sound reproduction signal 108 a relatively louder sound, making it easier for the user to hear. Note that the attenuation of the signal input from the playback sound source input terminal 9 may be greater than -10 dB.

[0101] <Embodiment 2> <Configuration> The noise removal function control device according to the second embodiment will be described with reference to FIGS. 1, 3, 10, and 11. FIG.

[0102] The hardware configuration (FIGS. 1, 2, 3, and 4) of the noise removal function control device according to the second embodiment is the same as that of the first embodiment. The only difference from the first embodiment is the data stored in the memory 14 shown in FIGS. 1 and 3. That is, in the first embodiment, there is only one keyword that is preset to stop the noise removal function, and when that keyword is detected, the playback start / stop time generating unit 105 reads only one generation rule related to the keyword. In the second embodiment, multiple keywords and multiple generation rules corresponding to each keyword are stored in the memory 14.

[0103] FIG. 10 is a functional block diagram of a noise removal function control device according to the second embodiment.

[0104] The following describes the differences from the functional block diagram (FIG. 5) of the noise removal function control device according to the first embodiment.

[0105] Through the initialization process, the keyword detection unit 120 starts a process of matching the external sound data written to the external sound memory 5 by the external sound recording unit 102 without delay with a plurality of preset keywords (Kw keywords). As an example of the matching process in this case, Kw matching processes for one keyword are run in parallel, and when the matching score exceeds a threshold value set for each keyword, it is determined that the keyword has been detected. When the keyword detection unit 120 detects a keyword in the external sound data, it outputs information on the start and end times of the keyword and the ID of the detected keyword (= 1···Kw) as keyword detection data 104. The keyword detection unit 120 outputs keyword detection data 104 every time it detects a keyword until it receives a keyword detection stop signal 111, which will be described later.

[0106] When the keyword detection data 104 is input, the playback start / stop time generation unit 121 generates the playback start time and playback stop time for the external sound data stored in the external sound memory 5 in accordance with the keyword detection data 104. An example of the generation rule 122 at this time is shown below. The ID of the detected keyword is assumed to be k. Playback start time = start time of keyword (ID=k) + Ofst_b(k) seconds Playback stop time = End time of keyword (ID=k) + Ofst_e(k) seconds

[0107] The generation rules 122 are stored in the memory 14, and Ofst_b(k) (k=1 to Kw) and Ofst_e(k) (k=1 to Kw) are set for each keyword ID (1 to Kw).

[0108] For example, for the keyword "Dangerous" (ID=1), Ofst_b(1) = -0.5 and Ofst_e(1) = 0.5 are set, and for the keyword "Mr. / Ms. XX" (calling out a user name, (ID=2), Ofst_b(1) = -0.5 and Ofst_e(1) = 5 are set. In this case, when the emergency keyword "Dangerous" is detected, a short external sound containing that keyword is played back, and when the keyword "Mr. / Ms. XX" calling out a user name is detected, not only that keyword but also the subsequent speech is played back, allowing the user to hear the requirements accompanying the call.

[0109] The generation rule 122 as described above is stored in the memory 14 and is read out from the memory 14 when the playback start / stop time generation unit 121 is initialized.

[0110] In the playback start / stop time generation unit 121, the operation after outputting the external sound playback command 106 to the external sound playback unit 107 based on the playback start time and playback stop time of the external sound data generated as described above is the same as in embodiment 1.

[0111] <Operation> Next, a detailed description will be given of the operation of the noise elimination function control device according to embodiment 2. The operations of the external sound recording unit 102, the external sound reproducing unit 107, and the noise canceling headphones 10 are the same as those of the noise elimination control device according to embodiment 1 (FIGS. 6 and 9), and therefore the description will be omitted.

[0112] 11 is a flowchart showing the operations of the keyword detection unit 120 and the playback start / stop time generation unit 121 in the noise removal function control device according to Embodiment 2. The following describes the differences from the operations of the noise removal function control device according to Embodiment 1 (FIG. 8).

[0113] In step S41, the noise cancellation function control device initializes itself, specifically, the keyword detection unit 120 and the playback start / stop time generation unit 121.

[0114] The keyword detection unit 120 performs the following three initializations. · Obtain time information and set that time as the keyword detection start time Tkst. ·Read the data of multiple (Kw) keywords to be detected and the matching score threshold Th(k) (k=1···Kw) corresponding to each keyword. -Initialize the matching process to detect keywords.

[0115] Furthermore, the playback start / stop time generating unit 121 performs the following initialization. The generation rules 122 {Ofst_b(k), Ofst_e(k)} (k=1···Kw) for playback start times and playback stop times corresponding to the Kw keywords are read from the memory 14 .

[0116] In step S42, the keyword detection unit 120 executes a matching process to detect a keyword. As an example of the matching process in this case, the keyword detection unit 120 runs Kw matching processes in parallel when there is one keyword, and determines that the keyword (ID=k) has been detected when the score S(k) (k=1···Kw) of each keyword being matched exceeds the threshold Th(k) (k=1···Kw) for the keyword (ID=k). The keyword detection unit 120 reads one or more samples of external sound data from the external sound memory 5, starting from the external sound data corresponding to the keyword detection start time Tkst up to the most recently written external sound data, and performs a matching process to detect a keyword. The keyword detection start time Tkst is updated by a time corresponding to the number of samples of the read external sound data. With this update, the keyword detection start time Tkst corresponds to the last (latest) external sound data input to the matching process.

[0117] In step S13, the keyword detection unit 120 determines whether or not a keyword has been detected. If a keyword has not been detected, the process proceeds to step S15. On the other hand, if a keyword has been detected, the process proceeds to step S43.

[0118] In step S43, the playback start / stop time generation unit 121 uses information on the start time and end time of the detected keyword (ID=k) and the playback start time and playback stop time generation rule 122{Ofst_b(k), Ofst_e(k)} (k=1···Kw) read in step S41 to generate the playback start time and playback stop time of the external sound data in the external sound memory 5 as follows, and outputs an external sound playback command 106 including this time information to the external sound playback unit 107. Playback start time = start time of keyword (ID=k) + Ofst_b(k) seconds Playback stop time = End time of keyword (ID=k) + Ofst_e(k) seconds

[0119] The processing from step S15 onwards is the same as that in the first embodiment (FIG. 8).

[0120] <Effects> From the above, according to the noise cancellation function control device of embodiment 2, when a keyword set to stop the noise cancellation function of the noise cancellation headphones 10 is detected in an external sound, the external sound containing the keyword is reproduced within the noise cancellation headphones 10, allowing the user to determine whether the keyword detection was correct or incorrect.

[0121] Furthermore, if the user determines that the keyword detection was incorrect after listening to the reproduced external sound, the user can cancel the process of stopping the noise removal function, which is triggered by the keyword detection, by operating the cancel command input switch 13.

[0122] Furthermore, by setting multiple keywords and corresponding generation rules, the degree of freedom in keyword selection is improved. Also, since external sounds are played at appropriate playback start times and playback stop times for each keyword, the user is less likely to miss external sounds.

[0123] Furthermore, the user can determine which keywords have been detected in the external sound by listening to the reproduced external sound.

[0124] <Third Embodiment> <Configuration> The noise removal function control device according to the third embodiment will be described with reference to FIGS. 12, 13, 14, 15, and 16. FIG.

[0125] 12 is a hardware configuration diagram of a noise cancellation function control device according to embodiment 3. The noise cancellation function control device according to embodiment 3 is characterized by including a command input switch 21. Other configurations are the same as those of the noise cancellation function control device according to embodiment 1 (FIG. 1), so detailed description will be omitted here. Furthermore, the hardware configuration of the noise cancellation headphones 10 according to embodiment 3 is the same as that of the noise cancellation headphones 10 according to embodiment 1 (FIG. 2).

[0126] The command input switch 21 is a switch that the user operates to interrupt or cancel a process (playback of external sound, stopping of noise reduction function) that is being executed or will be executed after the detection of a keyword as a trigger. Unlike the cancel command input switch 13 in other embodiments, the command input switch 21 is capable of inputting a cancel command when the user listens to the playback of external sound corresponding to the detected keyword and finds that the keyword detection is incorrect, and a noise reduction resume command when the keyword detection is correct. For example, a cancel command is input by short-pressing the command input switch 21, and a noise reduction resume command is input by long-pressing the command input switch 21.

[0127] The output of the command input switch 21 is input to the CPU 4. When the CPU 4 receives a cancel command or a noise cancellation resume command, it stops the playback of the external sound if it is currently playing the external sound, and also cancels the output of a noise cancellation control signal that stops the noise cancellation function, which would have been output to the noise-canceling headphones 10 after the playback of the external sound has finished. Furthermore, when the CPU 4 receives a cancel command or a noise cancellation command, if the playback of the external sound has already finished and the CPU 4 is outputting a noise cancellation function control signal that stops the noise cancellation function of the noise-canceling headphones 10, it outputs a noise cancellation function control signal that executes the noise cancellation function of the noise-canceling headphones 10.

[0128] Moreover, only when a cancel command is received, the CPU 4 executes adaptive learning of the matching process in keyword detection in order to prevent repeated erroneous keyword detection.

[0129] 13 is a hardware configuration diagram showing an example of another configuration of the noise cancellation function control device according to embodiment 3. The noise cancellation function control device according to embodiment 3 is characterized by including a command input switch 21. The other configurations are the same as those of the noise cancellation function control device according to embodiment 1 (FIG. 3). Furthermore, the hardware configuration of the noise cancellation headphones 10 in FIG. 13 is the same as that of the noise cancellation headphones 10 according to embodiment 1 (FIG. 4).

[0130] 14 is a functional block diagram of a noise elimination function control device according to embodiment 3. The noise elimination function control device according to embodiment 3 is characterized by including a learning-type keyword detection unit 130, a command input unit 132, and an external sound playback unit 131. A learning flag is added to the keyword detection resume signal 112 to control the execution of adaptive learning of the matching algorithm used to detect keywords in the learning-type keyword detection unit 130. The other configurations are the same as those of the noise elimination function control device according to embodiment 2 (FIG. 10).

[0131] The command input unit 132 receives the cancel command 114 or the noise reduction restart command 133 input from the command input switch 21. The received cancel command 114 or the noise reduction restart command 133 is output to the external sound reproducing unit 131.

[0132] The learning-type keyword detection unit 130 performs the same operation as the keyword detection unit 120 in the second embodiment until it receives the keyword detection resume signal 112 from the external sound playback unit 131. That is, the learning-type keyword detection unit 130 reads the external sound data that the external sound recording unit 102 has written to the external sound memory 5 without delay through the initialization process, and starts a process of matching the data with a plurality of (Kw) preset keywords. When a keyword is detected in the external sound data, the learning-type keyword detection unit 130 outputs information on the start time and end time of the keyword and the ID (=1···Kw) of the detected keyword as keyword detection data 104.

[0133] Until the learning-type keyword detection unit 130 receives the keyword detection stop signal 111, it outputs the keyword detection data 104 every time it detects a keyword.

[0134] Until receiving a cancel command 114 or a noise removal resume command 133, the external sound playback unit 131 performs the same operation as the external sound playback unit 107 in the second embodiment. That is, the external sound playback unit 131 receives an external sound playback command 106 as input, and in accordance with the information on the playback start time and playback stop time included therein, reads the external sound data for the corresponding time from the external sound memory 5 and outputs it to the D / A converter 6. When the external sound playback unit 131 receives a new external sound playback command 106 before the playback of the external sound corresponding to the playback start time and playback stop time of the external sound playback command 106 ends, the external sound playback unit 131 continues playing the external sound until the playback stop time included in the new external sound playback command 106.

[0135] Upon receiving the cancel command 114 or the noise removal restart command 133, the external sound playback unit 131 outputs a keyword detection stop signal 111 or a keyword detection restart signal 112 depending on the state of the external sound playback (playing or playback completed) and the type of command received (cancel command 114 or noise removal restart command 133).

[0136] Upon receiving the keyword detection resume signal 112, the learning-type keyword detection unit 130 performs adaptive learning of the matching algorithm for keyword detection in accordance with the learning flag included in the keyword detection resume signal 112, and applies the adaptively learned matching algorithm to subsequent keyword detection processes, thereby suppressing erroneous keyword detection.

[0137] <Operation> Next, a detailed description will be given of the operation of the noise reduction function control device according to embodiment 3. Note that the operation of the external sound recording unit 102 is the same as the operation of the noise reduction function control device according to embodiment 3 (FIG. 6), and therefore the description will be omitted.

[0138] 15 is a flowchart showing the operations of the learning-type keyword detection unit 130 and the playback start / stop time generation unit 121 in the noise removal function control device according to embodiment 3. The following describes the differences from the operations of the noise removal function control device according to embodiment 2 (FIG. 11).

[0139] In step S51, the noise removal function control device according to embodiment 3 is initialized. Specifically, the learning-type keyword detection unit 130 and the playback start / stop time generation unit 121 are initialized. The learning-type keyword detection unit 130 performs the following three initialization processes. · Obtain time information and set that time as the keyword detection start time Tkst. · Read the data of multiple (Kw) keywords to be detected and the matching score threshold Th(k) (k=1···Kw) corresponding to each keyword. -Initialize the matching process to detect keywords.

[0140] The above three initialization processes are the same as those of the above keyword detection unit 120. The learning-type keyword detection unit 130 further performs the following two initialization processes. A counter array C(k) (k=1···Kw) that stores the number of times the cancel command 114 has been input by the user for the Kw keywords to be detected is cleared to zero. · An array Kh(i) (i=1···Imax, the number of Imax arrays) is secured to store the history of detected keywords, and a counter Ckd that stores the number of times the keyword is detected is cleared to zero.

[0141] The array Kh(i) (i=1...Imax) stores, as a history, the IDs of keywords detected from the time the learning-type keyword detection unit 130 starts keyword detection due to initialization until the external sound replay unit 131 outputs the keyword detection stop signal 111. The learning-type keyword detection unit 130 uses this history to perform adaptive learning. In addition, this history is cleared when the learning-type keyword detection unit 130 resumes keyword detection (details will be described later).

[0142] The playback start / stop time generating unit 121 performs the following initialization process. The generation rules 122 {Ofst_b(k), Ofst_e(k)} (k=1···Kw) for playback start times and playback stop times corresponding to the Kw keywords are read from the memory 14 .

[0143] In step S52, the learning keyword detection unit 130 executes a matching process to detect a keyword. Specifically, the learning keyword detection unit 130 reads the external sound data written by the external sound recording unit 102 to the external sound memory 5, starting from the external sound data corresponding to the keyword detection start time Tkst up to the most recently written external sound data, and performs a matching process using only the Kw keywords (ID=k) (where k=1···Kw) for which C(k)=0. As an example of the matching process in this case, the learning keyword detection unit 130 runs the matching process for one keyword in parallel for the Kw keywords, and determines that the keyword (ID=k) has been detected when the score S(k) (k=1···Kw) of each keyword to be matched exceeds the threshold Th(k) (k=1···Kw) for the keyword (ID=k). The keyword detection start time Tkst is updated by the time corresponding to the number of samples of the external sound data read from the external sound memory 5.

[0144] In step S13, the learning-type keyword detection unit 130 determines whether or not a keyword is detected. If a keyword is not detected, the process proceeds to step S15. On the other hand, if a keyword is detected, the process proceeds to step S53.

[0145] In step S53, the playback start / stop time generation unit 121 uses the information on the start time and end time of the detected keyword (ID=k) and the playback start time and playback stop time generation rule 122{Ofst_b(k), Ofst_e(k)} (k=1···Kw) read in step S51 (initialization) to generate the playback start time and playback stop time of the external sound in the external sound memory 5 as follows, and outputs an external sound playback command 106 including this time information to the external sound playback unit 107. Playback start time = start time of keyword (ID=k) + Ofst_b(k) seconds Playback stop time = End time of keyword (ID=k) + Ofst_e(k) seconds

[0146] In addition, the ID (ID=1···Kw) of the detected keyword is stored in the array Kh(Ckd+1) that stores the history of detected keywords, and "1" is added to the counter Ckd for the number of times the keyword has been detected.

[0147] The processes of steps S15, S16, and S17 are the same as those of steps S15, S16, and S17 in Fig. 11 in embodiment 2. When the learning-type keyword detection unit 130 receives the keyword detection resume signal 112 in step S17, the process proceeds to step S54.

[0148] In step S54, the learning-type keyword detection unit 130 performs processing according to the value of the learning flag included in the received keyword detection resume signal 112. If the external sound replay unit 131 receives the cancel command 114, the value of the learning flag is set to "1." If the external sound replay unit 131 receives the noise removal resume command 133, the value of the learning flag is set to "0." If the value of the learning flag is "1," the process proceeds to step S55. On the other hand, if the value of the learning flag is "0," the process proceeds to step S56.

[0149] In step S55, the learning-type keyword detection unit 130 performs adaptive learning of the matching algorithm. That is, in accordance with the user's judgment that the keyword detection in step S52 was incorrect, the learning-type keyword detection unit 130 adds "1" to C(Kh(1)) in the counter array C(k) (k=1···Kw) that stores the number of times the cancel command 114 has been input, in order to exclude from matching targets the keyword whose ID is the value of the first Kh(1) in the array Kh(i) (i=1···Ckd) that stores the history of detected keywords. Then, the process proceeds to step S56.

[0150] In step S56, the learning-type keyword detection unit 130 clears Kh(i) (i=1···Imax) and Ckd to zero in order to clear the history of keywords detected from the start of the keyword detection process until the user inputs the cancel command 114. The learning-type keyword detection unit 130 also updates the keyword detection start time Tkst according to the time information included in the keyword detection resume signal 112 received in step S17. If the keyword detection resume signal 112 does not include time information, the Tkst updated in the process in step S12 is used as is. Then, the process returns to step S52.

[0151] 16 is a flowchart showing the operation of the external sound reproducing unit 131 and the noise-canceling headphones 10 according to embodiment 3. The following describes the differences from the operation of the external sound reproducing unit 107 and the noise-canceling headphones 10 according to embodiments 1 and 2 (FIG. 9).

[0152] In steps S26 and S28 of FIG. 9, processing is performed depending on whether or not a cancel command 114 input by the user has been received, but in embodiment 3, in steps S57 and S59 of FIG. 16, processing is performed depending on whether or not a noise removal resume command 133 or a cancel command 114 input by the user has been received.

[0153] 16 also includes additional steps S58 and S60 for setting the value of a learning flag included in keyword detection resume signal 112 output in steps S27 and S32. As a result, if the user inputs cancel command 114, keyword detection resume signal 112 including a learning flag value of "1" is output from external sound replay unit 131, and adaptive learning of the detection algorithm is performed in step S55 of Fig. 15 described above. If the user inputs noise removal resume command 133, the learning flag is set to "0" in steps S58 and S60, and adaptive learning of the detection algorithm is not performed.

[0154] <Effects> From the above, according to the noise cancellation function control device of embodiment 3, when a keyword set to stop the noise cancellation function of the noise cancellation headphones 10 is detected in an external sound, the external sound containing the keyword is played back within the noise cancellation headphones 10, allowing the user to determine whether the keyword detection was correct or incorrect.

[0155] Furthermore, the degree of freedom in keyword selection is improved by setting a plurality of keywords and corresponding generation rules 122. Furthermore, external sounds are played back at appropriate playback start times and playback stop times for each keyword, which reduces the chance of the user missing an external sound.

[0156] Furthermore, if the user determines that the keyword detection was incorrect after listening to the reproduced external sound, the user can cancel the process of stopping the noise removal function triggered by the keyword detection by operating the command input switch 21 and inputting the cancel command 114. Furthermore, adaptive learning of the matching algorithm in keyword detection makes it possible to suppress similar erroneous keyword detection in subsequent keyword detections.

[0157] Furthermore, when the user who has listened to the reproduced external sound determines that the correct keyword has been detected and that the necessary information has been heard, the user can cancel the reproduction of the external sound that was performed as a result of the keyword detection and the subsequent suspension of the noise reduction function by inputting a noise reduction resume command 133. In this case, adaptive learning of the matching algorithm is not performed.

[0158] In the third embodiment, in step S55 of FIG. 15, the keywords excluded from matching in the matching process for detecting the next keyword are limited to those satisfying keyword ID=Kh(1). However, it is possible that multiple keywords are detected before the user inputs the cancel command 114 (the number of detected keywords is stored in Ckd). In that case, the keywords excluded from matching in the matching process for detecting the next keyword may be all keywords matching keyword ID=Kh(i) (where i=1...Ckd) rather than just keyword ID=Kh(1). In this case, this is achieved by expanding the targets for which "1" is added in the counter array C(k) to those satisfying C(Kh(i)) (where i=1...Ckd).

[0159] Furthermore, in the third embodiment, the matching algorithm is learned by excluding a keyword for which the user has entered a cancel command 114 from a matching target in detecting the next keyword, for a plurality of preset keywords, but the learning method is not limited to this example. For example, instead of determining whether to exclude a keyword from matching targets based on the value of C(k), a method may be used in which, for keywords for which the user has entered a cancel command 114, the threshold value Th(k) (k=1···Kw) for detection determination is increased in proportion to the number of times C(k) (k=1···Kw) that the cancel command 114 is entered. In this case, Kw may be 1. Because the determination to exclude a keyword from matching targets based on the value of C(k) is not performed, keyword detection is possible even if only one preset keyword is used.

[0160] Furthermore, if multiple keywords are detected before the user inputs the cancel command 114, the threshold for keywords that match the keyword ID = Kh(i) (where i = 1···Ckd) may be increased (learning is performed for all detected keywords).

[0161] Furthermore, in the third embodiment, the counter array C(k) (k=1···Kw) that stores the number of times the cancel command 114 is input by the user is cleared to zero only in the initialization process of step S51 in FIG. 15, but this is not limited to this. A process of periodically (at a time period T) clearing C(k) to zero may be added as the process of step S56 in FIG. 15. This makes it possible to return to the initial state the matching algorithm that has been adaptively trained due to erroneous detection of keywords that depend on the usage environment of the noise removal function control device according to the third embodiment.

[0162] Also, in step S52 of FIG. 15, for the counter array C(k) (k=1···Kw) that stores the number of times the cancel command 114 was input by the user, keywords (ID=k) for which C(k)>C (where C=0) are excluded from matching targets, but a value of "1" or greater may be used as the value of C. For example, if C=1, when the user determines that a keyword has been erroneously detected two or more times, the keyword is excluded from matching targets. This makes it possible to prevent inappropriate matching algorithm learning due to accidental erroneous keyword detection.

[0163] <Fourth Embodiment> <Configuration> A noise elimination function control device according to embodiment 4 will be described with reference to Figures 1, 3, 17, and 18. The hardware configuration of the noise elimination function control device according to embodiment 4 is the same as that of embodiment 1 (Figures 1, 2, 3, and 4). The noise elimination function control device according to embodiment 4 differs from the noise elimination function control device according to embodiment 1 in that the noise elimination function control device according to embodiment 4 includes an external sound playback unit 140 that performs playback of external sound executed by CPU 4 shown in Figures 1 and 3 at double speed, which is faster than the normal playback speed, and the trigger for stopping playback of external sound is not the playback stop time included in external sound playback command 106, but rather the trigger is when playback of external sound stored in external sound memory 5 has caught up with the current time.

[0164] 17 is a functional block diagram of a noise elimination function control device according to embodiment 4. The noise elimination function control device according to embodiment 4 is characterized by including an external sound reproducing unit 140. Other configurations are the same as those of the noise elimination function control device according to embodiment 1 (FIG. 5), and therefore detailed description thereof will be omitted here.

[0165] The external sound playback unit 140 receives the external sound playback command 106 as input, and reads the external sound data for the corresponding time from the external sound memory 5 according to the playback start time information included in the external sound playback command 106, and outputs the data to the D / A converter 6. The external sound playback unit 140 does not use the playback stop time information included in the external sound playback command 106 to stop the playback of the external sound. The stop of the playback of the external sound is triggered when the external sound data stored in the external sound memory 5 that is read out becomes the external sound data for the current time (when the playback of the external sound stored in the external sound memory 5 has caught up with the current time).

[0166] That is, when the external sound data stored in the external sound memory 5 to be read becomes the external sound data at the current time, the external sound reproducing unit 140 performs the following three processes. A noise cancellation control signal 110 for stopping the noise cancellation function is output to the noise cancellation unit 109 via the noise cancellation control signal input terminal 12 of the noise cancellation headphones 10. A keyword detection stop signal 111 is output to the keyword detection unit 103 and the playback start / stop time generation unit 105. The output of the external sound data stored in the external sound memory 5 to the D / A converter 6 is stopped (the playback of the external sound is stopped).

[0167] When the external sound reproducing unit 140 receives the cancel command 114, it performs processing according to the following condition 1 or condition 2. Condition 1 "External sound is being played": The output of external sound data to the D / A converter 6 is stopped, and a keyword detection stop signal 111 is temporarily output. Next, a keyword detection resume signal 112 containing information on the final end time of external sound playback is output, and input of the next external sound playback command 106 is awaited. Condition 2 "External sound playback has already completed": A noise removal control signal 110 that turns on the noise removal function is output to the noise removal unit 109. Next, a keyword detection restart signal 112 (the time information is the end time of external sound playback) that causes the keyword detection unit 103 to restart keyword detection is output, and input of the next external sound playback command 106 is awaited.

[0168] Furthermore, if the external sound reproducing unit 140 receives a new external sound reproducing command 106 before the reproducing of the external sound ends, the external sound reproducing unit 140 rejects the new external sound reproducing command 106 .

[0169] <Operation> Next, a detailed description will be given of the operation of the noise elimination function control device according to embodiment 4. The operations of the external sound recording unit 102, the keyword detection unit 103, and the playback start / stop time generation unit 105 are the same as those of the noise elimination function control device according to embodiment 1 (FIGS. 6 and 8), and therefore the description will be omitted.

[0170] 18 is a flowchart showing the operation of the external sound reproducing unit 140 and the noise canceling headphones 10 in the noise cancellation function control device according to embodiment 4. The following describes the differences from the operation of the noise cancellation function control device according to embodiment 1 (FIG. 9).

[0171] In step S61, the noise reduction function control device initializes the noise reduction function control device. Specifically, in addition to the initialization process in step S21 in Fig. 9, the external sound reproducing unit 140 performs the following initialization process. -Set the playback speed for double speed playback. The process for starting reception of the external sound playback command 106 output by the playback start / stop time generation unit 105 is performed. If necessary for double-speed playback, the D / A converter 6 is set for double-speed playback.

[0172] In step S62, the external sound playback unit 140 starts reading the external sound data stored in the external sound memory 5 at the corresponding playback start time included in the received external sound playback command 106, and sequentially outputs the external sound data to the D / A converter 6. However, the external sound playback unit 140 performs playback at a speed faster than the normal playback speed (double speed playback).

[0173] In step S28, the external sound playback unit 140 determines whether or not a cancel command 114 has been received during double-speed playback of the external sound. The operation when a cancel command 114 has been received is the same as the operation in FIG. 9, and the process proceeds to step S31. The operation when a cancel command 114 has not been received is different from the operation in FIG. 9, and the process returns to step S24. This is because in the fourth embodiment, even if a keyword is detected during external sound playback and the external sound playback unit 140 receives a new external sound playback command 106, there is no need to update the stop time of double-speed playback of the external sound with the playback stop time information included in the new external sound playback command 106.

[0174] Operations other than those described above are the same as those in FIG. 9, and therefore will not be described here.

[0175] <Effects> From the above, according to the noise cancellation function control device of embodiment 4, when a keyword set to stop the noise cancellation function of the noise cancellation headphones 10 is detected in external sound, the external sound containing the keyword is played back at double speed within the noise cancellation headphones 10, allowing the user to determine whether the keyword detection was correct or not.

[0176] When the double-speed external sound playback catches up with the external sound at the current time, the noise cancellation function of the noise-canceling headphones 10 is stopped, and the playback of the external sound also stops. If the user hears the played external sound and determines that the keyword detection was incorrect, he or she can cancel the process of stopping the noise cancellation function, which was triggered by the keyword detection, by operating the cancel command input switch 13.

[0177] In the operation of the fourth embodiment, the external sound is played back at double speed, but the playback speed is not limited to double speed as long as it is faster than normal playback. As long as it is within a speed range that the user can hear, playback at a speed faster than normal playback may be adopted.

[0178] <Fifth Embodiment> <Configuration> A noise removal function control device according to the fifth embodiment will be described with reference to FIGS. 1, 3, 19, 20, and 21. FIG.

[0179] The hardware configuration of the noise cancellation function control device according to embodiment 5 is the same as the hardware configuration (FIGS. 1, 2, 3, and 4) of the noise cancellation function control device according to embodiment 1. The difference between the noise cancellation function control device according to embodiment 5 and the noise cancellation function control device according to embodiment 1 is that the noise cancellation function control device according to embodiment 5 has a subsequent speech detection unit as a new process executed by CPU 4 shown in FIGS.

[0180] 19 is a functional block diagram of a noise removal function control device according to embodiment 5. The difference from embodiment 1 is that it has a subsequent utterance detection unit 150 that performs a process of detecting an utterance following a keyword utterance detected in external sound data based on keyword detection data 104 and outputs subsequent utterance detection data 151 to a playback start / stop time generation unit 152, and a playback start / stop time generation unit 152 that receives keyword detection data 104 and subsequent utterance detection data 151 as input and generates a playback start time and a playback stop time for playing back external sound including the keyword utterance and the subsequent utterance.

[0181] Triggered by receiving keyword detection data 104 output by keyword detection unit 103, subsequent speech detection unit 150 starts determining whether or not there is any speech (subsequent speech) following the keyword detected by keyword detection unit 103, using the external sound data written to external sound memory 5 by external sound recording unit 102 as input. When subsequent speech detection unit 150 determines that there is a subsequent speech, it outputs subsequent speech detection data 151, including time information of the external sound data at the end of the subsequent speech at that time, to playback start / stop time generation unit 152. This operation is repeatedly performed until the subsequent speech ends or until keyword detection stop signal 111 output by external sound playback unit 107 is received. Note that if new keyword detection data 104 is input before the subsequent speech ends, subsequent speech detection unit 150 starts the above process anew, using the new keyword detection data 104 as a new trigger.

[0182] When playback start / stop time generation unit 152 receives keyword detection data 104 output by keyword detection unit 103, it generates a playback start time and a playback stop time of the external sound using the start time and end time of the keyword included in the keyword detection data, and outputs external sound playback command 106 including information on these times to external sound playback unit 107. When playback start / stop time generation unit 152 receives new keyword detection data 104 thereafter, it similarly generates a playback start time and a playback stop time of the external sound using the start time and end time of the keyword included in the keyword detection data, and outputs external sound playback command 106 including information on these times to external sound playback unit 107. The processing up to this point is the same as that of playback start / stop time generation unit 105 according to embodiment 1. In addition to these operations, the playback start / stop time generation unit 152 according to the fifth embodiment updates the playback stop time generated using the keyword detection data 104 as a trigger, triggered by receiving subsequent utterance detection data 151 output from the subsequent utterance detection unit 150, with the time information of the end of the subsequent utterance included in the subsequent utterance detection data 151, and outputs the updated playback stop time to the external sound playback unit 107 as a new external sound playback command 106.

[0183] <Operation> Next, a detailed description will be given of the operation of the noise elimination function control device according to embodiment 5. The operations of the external sound recording unit 102, the external sound reproducing unit 107, and the noise canceling headphones 10 are the same as those of the noise elimination control device according to embodiment 1 (FIGS. 6 and 9), and therefore will not be described here.

[0184] 20 is a flowchart showing the operations of the keyword detection unit 103, the subsequent speech detection unit 150, and the playback start / stop time generation unit 152. The following describes the differences from the operation of the noise removal control device according to embodiment 1 (FIG. 8).

[0185] In step S71, the noise cancellation function control device initializes itself. Specifically, in addition to the initialization process in step S11 of Fig. 8, the noise cancellation function control device reads parameters required for the operation of subsequent utterance detection unit 150, resets subsequent utterance detection execution flag Fkf to zero, and resets buffer memory and the like required for executing other processes.

[0186] Step S72 is a process that is executed following step S14, triggered by the detection of a keyword in step S13. Specifically, when subsequent utterance detection unit 150 receives keyword detection data 104 output by keyword detection unit 103, subsequent utterance detection unit 150 sets a subsequent utterance detection execution flag Fkf to "1" for later performing subsequent utterance detection processing.

[0187] Step S73 is a process performed when no keyword is detected in step S13 or after the process in step S72. Specifically, the subsequent utterance detection unit 150 checks the subsequent utterance detection execution flag Fkf, and if Fkf=0, skips the process in step S74 and proceeds to step S15. On the other hand, if Fkf=1, proceeds to step S74.

[0188] In step S74, the subsequent speech detection unit 150 reads the external sound data written to the external sound memory 5 by the external sound recording unit 102, and performs a process of determining whether or not there is a speech following the keyword detected by the keyword detection unit 103. The specific determination process is described below.

[0189] 21 shows an example of a rule for detecting subsequent speech. Using the amplitude power of the input external sound data, the subsequent speech detection unit 150 determines a section that is, for example, +6 dB higher than the average power of the background noise as a speech section, and a section that is lower than +6 dB as a silent section.

[0190] As long as the time Ts of the silent interval following the end time of the utterance of the keyword "Mr. / Ms. XX" does not exceed the threshold value Tsmax (corresponding to the interval (A) in Figure 21), there is a possibility that a subsequent utterance will be uttered, so the subsequent utterance detection execution flag Fkf is kept at "1" and the process proceeds to step S15 in order to continue the subsequent utterance detection process.

[0191] On the other hand, if the duration Ts of the silent period following the end of the utterance of the keyword "Mr. / Ms. XX" exceeds the threshold Tsmax, it is determined that there is no subsequent utterance (the utterance has ended), the subsequent utterance detection execution flag Fkf is cleared to zero, and the process proceeds to step S15.

[0192] Furthermore, if the amplitude power becomes +6 dB or more higher than the average power of the background noise before the time Ts of the silent section following the end time of the utterance of the keyword "Mr. / Ms." exceeds the threshold Tsmax and it is determined that the subsequent utterance section is in the subsequent utterance section (the section of the subsequent utterance "The train is coming soon" in FIG. 21), the last time of the utterance section at that time is output to the playback start / stop time generation unit 152 as subsequent utterance detection data 151. Upon receiving the subsequent utterance detection data 151, the playback start / stop time generation unit 152 updates the playback stop time of the external sound playback start time and playback stop time previously generated in step S14 with the last time of the subsequent utterance included in the subsequent utterance detection data 151 (the time of the external sound determined to be the subsequent utterance section), and outputs the updated playback stop time to the external sound playback unit 107 as a new external sound playback command 106. In this case, since the subsequent utterance is continuing in this section, the subsequent utterance detection execution flag Fkf is maintained at "1" and the process proceeds to step S15.

[0193] If the time determined to be a silent section continues for more than Tsmax while the processing of step S74 is being repeatedly executed (corresponding to the section (B) in Figure 21), it is determined that the subsequent speech has ended, and the time of the external sound data that was last determined to be a speech section is output to the playback start / stop time generation unit 152 as subsequent speech detection data 151, the subsequent speech detection execution flag Fkf is cleared to zero, and the process proceeds to step S15.

[0194] If a new keyword is detected in step S12 while the process of step S74 is being repeatedly executed, a new process of detecting a subsequent utterance targeting the new keyword is started.

[0195] In step S75, in addition to the process of step S16 in FIG. 8 (stopping the keyword detection process), the subsequent utterance detection unit 150 stops the subsequent utterance detection process, and the process proceeds to step S17.

[0196] Step S76 is a step that is executed after exiting step S17, which is a loop process that waits for reception of keyword detection resume signal 112. In addition to the process in step S18 in Fig. 8 (updating keyword detection start time Tkst if time information is included in keyword detection resume signal 112), subsequent utterance detection unit 150 sets subsequent utterance detection execution flag Fkf to "0" and returns to step S12.

[0197] <Effects> As described above, the noise reduction function control device according to the fifth embodiment enables the user to hear not only the keyword but also the utterance following the keyword (subsequent utterance). The user can hear, without missing a single utterance, for example, "Mr. / Ms. XX (keyword), it's almost time to get off at the station (subsequent utterance)." On the other hand, in the first embodiment, the playback stop time is fixedly extended later according to the generation rule used by the playback start / stop time generating unit 105. However, if the subsequent utterance does not end within the extended time, the playback of the external sound ends in the middle of the subsequent utterance, which may cause the user to miss it.

[0198] The subsequent speech detection rule described in embodiment 5 is an example, and the detection rule may use other means, algorithms, or parameters obtained from the audio signal to detect subsequent speech, or may use a combination of these.

[0199] <Sixth Embodiment> <Configuration> The noise removal function control device according to the sixth embodiment will be described with reference to FIGS. 22, 23, 24, 25, and 26. FIG.

[0200] 22 is a hardware configuration diagram of a noise cancellation function control device according to embodiment 6. The noise cancellation function control device according to embodiment 6 is characterized by including an external sound playback operation command input switch 22 and a noise cancellation control signal input switch 23. Other configurations are the same as those of the noise cancellation function control device according to embodiment 1 (FIG. 1), so detailed description will be omitted here. Furthermore, the hardware configuration of the noise cancellation headphones 10 according to embodiment 6 is the same as that of the noise cancellation headphones 10 according to embodiment 1 (FIG. 2).

[0201] The CPU 4 performs a matching process on the input external sound data to detect whether a preset keyword is included. When a keyword is detected in the external sound data, the CPU 4 reads the external sound data stored in the external sound memory 5 based on the information on the start time of the keyword and outputs the read data to the D / A converter 6.

[0202] The external sound playback operation command input switch 22 is a switch that allows the user to perform an external sound playback operation. The external sound playback operation command input switch 22 includes buttons having the following functions, for example. When the user operates these buttons, an external sound playback operation command 205 (FIG. 24) is output to the CPU 4. - Button to stop playback. - Button to play at normal speed. - A button to play back the currently playing external sound. A button to fast forward the currently playing external sound.

[0203] The CPU 4 reads the external sound data stored in the external sound memory 5 in accordance with the external sound playback operation command 205, and performs playback as desired by the user. However, in the case of fast-forward playback, playback resumes at normal speed once the external sound being played catches up with the current time.

[0204] The noise cancellation control signal input switch 23 has the same function as a switch in conventional noise-canceling headphones that is manually operated by the user to turn the noise cancellation function on or off. The noise cancellation control signal 110 (a signal that turns the noise cancellation function on or off) output from the noise cancellation control signal input switch 23 is input not only to the noise-canceling headphones 10 but also to the CPU 4, and serves as a trigger to stop the playback of external sound, resume keyword detection, etc.

[0205] The noise cancellation headphones 10 receive an audio signal output from an audio mixer 8 as input, and reproduce the audio signal using a driver unit 15 provided in the noise cancellation headphones 10. The noise cancellation headphones 10 also switch the noise cancellation function of the noise cancellation headphones 10 ON or OFF in accordance with a noise cancellation control signal 110 input from a noise cancellation control signal input switch 23.

[0206] 23 is a hardware configuration diagram showing another example of the configuration of the noise cancellation function control device according to embodiment 6. The noise cancellation function control device according to embodiment 6 is characterized by including an external sound playback operation command input switch 22 and a noise cancellation control signal input switch 23. The other configurations are the same as those of the noise cancellation function control device according to embodiment 1 (FIG. 3). Furthermore, the hardware configuration of the noise cancellation headphones 10 in FIG. 23 is the same as that of the noise cancellation headphones 10 according to embodiment 1 (FIG. 4).

[0207] FIG. 24 is a functional block diagram of a noise removal function control device according to the sixth embodiment.

[0208] The external sound data input unit 101 receives the external sound data output from the A / D converter 3 .

[0209] The external sound recording unit 102 receives external sound data as input and writes the external sound data and information about the time when the external sound data was input into the external sound memory 5 without delay.

[0210] The keyword detection unit 200 receives the external sound data written to the external sound memory 5 by the external sound recording unit 102 as an input through the initialization process and starts a process of matching the data with a preset keyword. When a keyword is detected in the external sound data, the keyword detection unit 200 outputs information on the start time of the keyword and the ID of the detected keyword as keyword detection data 104. The keyword detection unit 200 outputs keyword detection data 104 every time it detects a keyword until it receives a keyword detection stop signal 111, which will be described later.

[0211] When the keyword detection data 104 is input, the playback start time generation unit 201 generates a playback start time for the external sound data stored in the external sound memory 5 in accordance with the keyword detection data 104. An example of a generation rule for generating the playback start time is shown below. Playback start time = Keyword start time - 0.5 seconds

[0212] As a result, even if the matching algorithm in the matching process of the keyword detection unit 200 has a tendency to determine the start time of the keyword to be a time slightly later than the actual utterance, the start part of the keyword is reproduced without being missing.

[0213] The playback start time generating unit 201 outputs the generated playback start time to the external sound playback unit 203 as an external sound playback command 202 .

[0214] The external sound playback unit 203 receives the external sound playback command 202 as input, and reads the external sound data for the corresponding time from the external sound memory 5 in accordance with the playback start time information included in the external sound playback command 202, and outputs the read data to the D / A converter 6. The external sound playback unit 203 also outputs a keyword detection stop signal 111 to the keyword detection unit 200 and the playback start time generation unit 201.

[0215] Upon receiving keyword detection stop signal 111, keyword detection unit 200 and playback start time generation unit 201 respectively stop the matching process for detecting a keyword and the process of generating a playback start time. Furthermore, upon receiving keyword detection resume signal 112 from external sound playback unit 203, keyword detection unit 200 and playback start time generation unit 201 respectively release the stopped state of the matching process for detecting a keyword and the process of generating a playback start time, and start a new process for detecting a keyword and a new process of generating a playback start time. However, the start of the external sound data from which keyword detection unit 200 is to detect a keyword follows the time information included in keyword detection resume signal 112.

[0216] The external sound reproduction signal 108 obtained via the D / A converter 6 and the audio amplifier 7 is input to the audio mixer 8. The audio mixer 8 mixes the external sound reproduction signal 108 with a signal that the user intends to listen to, such as an audio output signal from a smartphone that is input from a playback sound source input terminal 9, and outputs the mixed signal to the audio signal input terminal 11 of the noise cancellation headphones 10.

[0217] The noise cancellation unit 109 in the noise cancellation headphones 10 is set to execute a noise cancellation function as a result of the initialization process, and generates a signal that cancels external noise that penetrates the main body and earmuffs of the noise cancellation headphones 10 and enters the noise cancellation headphones 10, using the output signals from the external sound signal input microphone 17 and the error signal input microphone 18. The noise cancellation unit 109 then superimposes the generated signal on the signal input from the audio signal input terminal 11, and outputs the signal to the driver unit 15 as a noise-free playback signal 19.

[0218] The driver unit 15 reproduces the noise-reduced playback signal 19 in the noise-canceling headphones 10.

[0219] The external sound playback unit 203 plays back external sound data stored in the external sound memory 5 in accordance with an external sound playback operation command 205 input from the command input unit 204 when the user operates the external sound playback operation command input switch 22. An example of the external sound playback operation command 205 is shown below. -Stop external sound playback. - Play external sounds at normal speed. -Reverse playback to go back to past external sounds. Fast forward playback (playing at a faster than normal speed).

[0220] In accordance with these commands input by the user, the external sound playback unit 203 reads the external sound data stored in the external sound memory 5 and performs the above-mentioned playback operation, i.e., stopping playback, normal playback, reverse playback, or fast-forward playback. However, in the case of fast-forward playback, if the external sound being played catches up with the current time as a result of the playback, the external sound thereafter is played at normal speed.

[0221] While external sound is being played back in accordance with external sound playback operation command 205 input by the user, external sound playback unit 203 holds the latest time of the played external sound, and when the user inputs external sound playback operation command 205 to stop the playback of external sound, outputs information about the latest time of the played external sound as part of keyword detection resume signal 112. As a result, the matching process for detecting a keyword to be resumed in keyword detection unit 200 starts from external sound data that the user has not yet listened to.

[0222] In the sixth embodiment, the noise cancellation function of the noise cancellation headphones 10 is turned on or off in accordance with the noise cancellation control signal 110 that is input to the noise cancellation control signal input unit 206 when the user operates the noise cancellation control signal input switch 23. This is the same operation as commercially available noise cancellation headphones in which the user manually turns the noise cancellation function on or off. However, the noise cancellation control signal 110 is also input to the external sound reproduction unit 203, which performs the following processing. When a noise removal control signal 110 that turns off the noise removal function is input, a keyword detection stop signal 111 is output to stop keyword detection in the keyword detection unit 200. Also, if an external sound is being played at this time, the playback of the external sound is also stopped. When the noise removal control signal 110 that turns on the noise removal function is input, the current time is acquired and a keyword detection restart signal 112 including the time information is output to the keyword detection unit 200. In this case, the keyword detection unit 200 starts a matching process to detect a keyword from the external sound data corresponding to the time information included in the keyword detection restart signal 112.

[0223] In this way, if the external sound playback operation command 205 input by the user is one that stops playback of external sound, keyword detection is resumed starting from the external sound data at the latest time that the user has heard it (i.e., the time closest to the present.) Also, if the user operates the noise cancellation control signal input switch 23 to turn off the noise cancellation function that is being executed in the noise cancellation headphones 10 and then turn the noise cancellation function on, keyword detection is resumed starting from the external sound data corresponding to the time when the noise cancellation control signal 110 for turning on the noise cancellation function was input.

[0224] <Operation> Next, a detailed description will be given of the operation of the noise reduction function control device according to embodiment 6. The operation of the external sound recording unit 102 is the same as the operation of the noise reduction function control device according to embodiment 1 (FIG. 6), and therefore a description thereof will be omitted.

[0225] FIG. 25 is a flowchart showing the operations of the keyword detecting unit 200 and the playback start time generating unit 201 according to the sixth embodiment.

[0226] The operation of FIG. 25 starts when the power supply of the noise removal function control device according to the sixth embodiment is turned on.

[0227] In step S101, the noise reduction function control device initializes itself. Specifically, the noise reduction function control device initializes the keyword detection unit 200 by performing the following process. · Obtain time information and set that time as the keyword detection start time Tkst. Load the data of keywords to be detected. -Initialize the matching process to detect keywords.

[0228] Furthermore, the noise removal function control device performs the following process to initialize the playback start time generation unit 201. Load the rules for generating the playback start time.

[0229] In step S12, the keyword detection unit 200 executes a matching process to detect a keyword. Specifically, the keyword detection unit 200 reads one sample or a predetermined number of samples of external sound data stored in the external sound memory 5 by the external sound recording unit 102, starting from the external sound data corresponding to the keyword detection start time Tkst up to the most recently written external sound data, and performs a matching process to detect a keyword. The keyword detection unit 200 updates the keyword detection start time Tkst by a time corresponding to the number of samples of the read external sound data.

[0230] In step S13, the keyword detection unit 200 determines whether or not a keyword has been detected. If a keyword has not been detected, the process proceeds to step S15. On the other hand, if a keyword has been detected, the process proceeds to step S102.

[0231] In step S102, the playback start time generation unit 201 generates a playback start time for the external sound data stored in the external sound memory 5 using information on the start time of the detected keyword and the playback start time generation rule read in step S101, and outputs an external sound playback command 202 including the generated playback start time to the external sound playback unit 203.

[0232] In step S15, the keyword detection unit 200 and the playback start time generation unit 201 determine whether or not they have received the keyword detection stop signal 111 from the external sound playback unit 203. If they have not received the keyword detection stop signal, they return to step S12. On the other hand, if they have received the keyword detection stop signal 111, they proceed to step S16.

[0233] In step S16, the keyword detection unit 200 performs processing to stop the keyword detection processing that has been executed.

[0234] In step S17, the keyword detection unit 200 and the playback start time generation unit 201 determine whether or not they have received the keyword detection resume signal 112 from the external sound playback unit 203. If they have not received the keyword detection resume signal 112, they loop the processing of step S17. If they have received the keyword detection resume signal 112, they proceed to step S18.

[0235] In step S18, the keyword detection unit 200 sets the time information included in the received keyword detection restart signal 112 as the keyword detection start time Tkst, and then proceeds to step S12.

[0236] As will be described later, the time information included in the keyword detection resume signal 112 is the latest time in the external sound at which the external sound playback unit 203 has performed fast-forward playback or reverse playback in accordance with the external sound playback operation command 205. The keyword detection unit 200 sets this time as the keyword detection start time Tkst and starts the keyword detection process in step S12, thereby performing the keyword detection process from external sound data that the user has not heard, thereby preventing missed keyword detection. Furthermore, if the keyword detection resume signal 112 is triggered by the noise reduction control signal 110 (noise reduction function On) input by the user, the current time at which the user inputs the noise reduction control signal 110 (noise reduction function On) serves as the time information, and the keyword detection resume time in this case is the current time.

[0237] FIG. 26 is a flowchart showing the operations of the external sound reproducing unit 203 and the noise canceling headphones 10 in the noise cancellation function control device according to the sixth embodiment.

[0238] The operation shown in FIG. 26 starts when the noise cancellation function control device is powered on.

[0239] In step S110, the noise reduction function control device initializes itself. Specifically, the noise reduction function control device performs the following process to initialize the external sound reproducing unit 203. The external sound playback command 202 output by the playback start time generation unit 201 starts to be received. Starts accepting the noise removal control signal 110 input by user operation. Start accepting the external sound playback operation command 205 input by the user operation.

[0240] The noise cancellation function control device also performs the following process to initialize the noise cancellation headphones 10. Executes a noise elimination function based on signals from the external sound signal input microphone 17 and the error signal input microphone 18. The function of reproducing a sound source signal such as a music source input from the audio signal input terminal 11 within the noise canceling headphones 10 via the driver unit 15 is performed.

[0241] Furthermore, the noise cancellation function control device activates a D / A converter 6 that reproduces external sounds, an audio amplifier 7, and an audio mixer 8.

[0242] In step S111, the external sound reproducing unit 203 determines whether or not the user has operated the noise cancellation control signal input switch 23 to receive the noise cancellation control signal 110 that stops the noise cancellation function of the noise canceling headphones 10. If the noise cancellation control signal 110 that stops the noise cancellation function has been received, the process proceeds to step S119. On the other hand, if the noise cancellation control signal 110 that stops the noise cancellation function has not been received, the process proceeds to step S112.

[0243] In step S119, the external sound reproducing unit 203 outputs the keyword detection stop signal 111 to stop the keyword detection process in the keyword detecting unit 200.

[0244] In step S120, the external sound reproducing unit 203 determines whether or not it has received the noise reduction control signal 110 that turns on the noise reduction function from the user. That is, the external sound reproducing unit 203 waits until it receives the noise reduction control signal 110 that turns on the noise reduction function by operating the noise reduction control signal input switch 23. When it has received the noise reduction control signal 110 that turns on the noise reduction function, it proceeds to step S121.

[0245] In step S121, external sound replay unit 203 acquires the current time and outputs keyword detection resume signal 112 including this time information. This causes the loop processing of step S17 in Fig. 25 to be exited, and step S18 in the same figure is executed, thereby resuming keyword detection processing. In Fig. 26, the processing returns to step S111.

[0246] In step S111, if the external sound reproducing unit 203 has not received the noise reduction control signal 110 that stops the noise reduction function, the process proceeds to step S112.

[0247] In step S112, the external sound playback unit 203 determines whether or not it has received the external sound playback command 202 output by the playback start time generation unit 201. If it has not received the external sound playback command 202, it returns to step S111. On the other hand, if it has received the external sound playback command 202, it proceeds to step S113.

[0248] In step S113, the external sound playback unit 203 starts playing the external sound data stored in the external sound memory 5 in accordance with the playback start time information included in the external sound playback command 202. The external sound playback unit 203 also outputs a keyword detection stop signal 111 to stop the keyword detection process. Next, the process proceeds to step S114.

[0249] In step S114, similar to step S111, the external sound reproducing unit 203 determines whether or not it has received the noise reduction control signal 110 that stops the noise reduction function of the noise canceling headphones 10 and that has been input by the user operating the noise reduction control signal input switch 23. If it has not received the noise reduction control signal 110 that stops the noise reduction function, it proceeds to step S115. On the other hand, if it has received the noise reduction control signal 110 that stops the sound reduction function, it proceeds to step S122.

[0250] In step S115, the external sound reproducing unit 203 determines whether or not it has received an external sound reproducing operation command 205 input by the user operating the external sound reproducing operation command input switch 22. If it has received the external sound reproducing operation command 205, it proceeds to step S116. On the other hand, if it has not received the external sound reproducing operation command 205, it returns to step S114.

[0251] In step S116, the external sound playback unit 203 performs one of the following four processes in accordance with the content of the received external sound playback operation command 205. Then, the process proceeds to step S117. Note that if the external sound data being played back catches up with the current time as a result of continuing fast-forward playback, the subsequent external sound data is played back at normal speed. Stop external sounds from playing. - Play external sounds at normal speed. -Reverse playback to go back to past external sounds. Fast forward playback (playback at a faster speed than normal).

[0252] In step S117, the external sound reproducing unit 203 determines whether the process executed in step S116 was "stop reproducing" to stop reproducing the external sound. If the process was not to stop reproducing, the process returns to step S114. On the other hand, if the process was to stop reproducing, the process proceeds to step S118.

[0253] In step S118, the external sound reproducing unit 203 outputs the keyword detection resume signal 112 including information on the latest time in the external sound data reproduced in the reproduction process in step S116. Then, the process returns to step S111.

[0254] In step S114, if the external sound reproducing unit 203 receives the noise removal control signal 110 to stop the noise removal function, the process proceeds to step S122. In step S122, the external sound reproducing unit 203 stops the currently executing reproduction of the external sound, and the process proceeds to step S119.

[0255] <Effects> As described above, according to the noise reduction function control device of embodiment 6, by detecting a preset keyword in an external sound, playback of the external sound containing the keyword begins, and the user can determine whether the keyword was detected correctly or incorrectly. Also, by operating the external sound playback operation command input switch 22 to perform fast-forward playback or rewind playback while the external sound is being played back, the user can listen to not only the keyword but also the external sounds before and after it.

[0256] Furthermore, since the noise reduction function can be turned on or off only by the user operating the noise reduction control signal input switch 23 (because keyword detection is used as a trigger for playing external sounds), the noise reduction function will not be turned off unexpectedly by the user due to false keyword detection.

[0257] In the above operation, the user can perform operations such as playing or stopping external sound using the external sound playback operation command input switch 22 only after the keyword is detected and playback of external sound has started, but the operation is not limited to this. The user may also perform operations such as playing external sound using the external sound playback operation command input switch 22 regardless of whether a keyword is detected or not.

[0258] <Seventh Embodiment> <Configuration> The noise elimination function control device according to the seventh embodiment will be described with reference to Fig. 27 and Fig. 28. The hardware configuration of the noise elimination function control device according to the seventh embodiment is the same as the hardware configuration of the noise elimination function control device according to the sixth embodiment (Figs. 22, 2, 23, and 4), and therefore description thereof will be omitted.

[0259] 27 is a functional block diagram of a noise cancellation function control device according to embodiment 7. The difference between the noise cancellation function control device according to embodiment 7 and the noise cancellation function control device according to embodiment 6 is that the noise cancellation function control device according to embodiment 7 includes an external sound reproducing unit 210. The other configurations are the same as those of embodiment 6, and therefore description thereof will be omitted.

[0260] The external sound playback unit 210 receives an external sound playback command 202 as input, and in accordance with the playback start time information contained therein, reads out from the external sound memory 5 only the amount of external sound data at the corresponding time required for double-speed playback, and outputs this to the D / A converter 6. The external sound playback unit 210 also outputs a keyword detection stop signal 111 to the keyword detection unit 200 and the playback start time generation unit 201. Furthermore, when the time of the external sound data read from the external sound memory 5 catches up with the current time as double-speed playback progresses, the external sound playback unit 210 transitions to normal-speed playback from that point onward.

[0261] The external sound playback unit 210 plays back external sound data stored in the external sound memory 5 in accordance with an external sound playback operation command 205 input from the command input unit 204 when the user operates the external sound playback operation command input switch 22. An example of the external sound playback operation command 205 is shown below. Stop external sounds from playing. - Play external sounds at normal speed. -Reverse playback to go back to past external sounds. Fast forward playback (playback at a faster speed than normal).

[0262] In accordance with these commands input by the user, the external sound playback unit 210 reads the external sound data stored in the external sound memory 5 and performs the above-mentioned playback operations, i.e., stopping playback, normal playback, reverse playback, or fast-forward playback. However, in the case of fast-forward playback, when the playback progresses and the external sound data to be played catches up with the current time, the external sound data thereafter is played at normal speed (constant speed).

[0263] While external sound is being played back in accordance with external sound playback operation command 205 input by the user, external sound playback unit 210 holds the latest time of the played external sound, and when the user inputs external sound playback operation command 205 to stop the playback of external sound, outputs information about the latest time of the played external sound as part of keyword detection resume signal 112. As a result, the matching process for detecting a keyword to be resumed in keyword detection unit 200 starts from external sound data that the user has not yet listened to.

[0264] Furthermore, when the external sound reproducing unit 210 receives the noise removal control signal 110, it performs the following process. When a noise removal control signal 110 that turns off the noise removal function is received, a keyword detection stop signal 111 is output to stop keyword detection in the keyword detection unit 200. At this time, if an external sound is being played back, the playback of the external sound is also stopped. When the noise removal control signal 110 that turns on the noise removal function is received, the current time is acquired and a keyword detection restart signal 112 including the time information is output to the keyword detection unit 200. In this case, the keyword detection unit 200 starts a matching process to detect a keyword from the external sound data corresponding to the time information included in the keyword detection restart signal 112.

[0265] In this way, if the external sound playback operation command 205 input by the user is one that stops playback of external sound, keyword detection is resumed starting from the external sound data at the latest time that the user has heard it (i.e., the time closest to the present.) Also, if the user operates the noise cancellation control signal input switch 23 to turn off the noise cancellation function that is being executed in the noise cancellation headphones 10 and then turn the noise cancellation function on, keyword detection is resumed starting from the external sound data at the time that the noise cancellation control signal 110 for turning on the noise cancellation function is input.

[0266] <Operation> FIG. 28 is a flowchart showing the operations of the external sound reproducing unit 210 and the noise canceling headphones 10 in the noise cancellation function control device according to the seventh embodiment.

[0267] In the operation of FIG. 28, only the processes of steps S130 and S131 are different from those of the sixth embodiment (FIG. 26), and a description of the operations of the other steps will be omitted.

[0268] Step S130 is a process performed when the external sound playback command 202 is received in step S112. The external sound playback unit 210 starts double-speed playback of the external sound data stored in the external sound memory 5, based on the playback start time information included in the external sound playback command 202. The external sound playback unit 210 also outputs a keyword detection stop signal 111.

[0269] Step S131 is processing performed when the external sound reproducing unit 210 does not receive the noise removal control signal 110 that turns off the noise removal function in step S114. As the data of the external sound being reproduced continues at double speed, the external sound reproducing unit 210 reaches the external sound data at the current time, and subsequent reproduction cannot be performed at double speed, so the external sound reproducing unit 210 performs processing to change the reproduction speed from double speed to normal speed.

[0270] <Effects> As described above, the noise reduction function control device according to the seventh embodiment detects a preset keyword in an external sound and plays back the external sound containing the keyword at double speed, allowing the user to determine whether the keyword was detected correctly or incorrectly in a shorter time than during normal speed playback. Also, by operating the external sound playback operation command input switch 22 to perform fast-forward playback or rewind playback while the external sound is being played back, the user can hear not only the keyword but also the external sound before and after it.

[0271] Furthermore, the noise reduction function can be turned on or off only by the user operating the noise reduction control signal input switch 23 (keyword detection is used as a trigger for playing external sounds), so the noise reduction function will not be turned off unexpectedly by the user due to incorrect keyword detection.

[0272] It should be noted that the double-speed playback performed by the external sound playback unit 210 is not limited to double speed. The external sound playback unit 210 may play back at a speed faster than normal speed playback (constant speed playback) and at a speed within the range that the user can hear.

[0273] <Modification> In the first to seventh embodiments, the operation after detecting a keyword is not limited to that of each of the first to seventh embodiments. For example, it may be configured to perform any of the operations of the first to seventh embodiments for each detected keyword.

[0274] In the first to seventh embodiments, the application to the noise-canceling headphones 10 has been described as an example, but the application is not limited to the noise-canceling headphones 10. For example, the first to seventh embodiments may be applied to hearing aids or other audiovisual equipment with a noise cancellation function. The sixth and seventh embodiments may be applied to ordinary headphones with high sound insulation, or VR equipment including headphones.

[0275] The first to seventh embodiments are not limited to being realized by hardware, but may be realized as an application on a smartphone, for example.

[0276] In the first to seventh embodiments, only the external sound stored in the external sound memory 5 is played, but the user may be notified of the start and end of playback by sounding a short beep or the like at the timing when playback of the external sound starts and ends. Similarly, when the user operates various switches (cancel command input switch 13, command input switch 21, external sound playback operation command input switch 22, noise removal control signal input switch 23), the same beep or a different beep for each switch may be heard by the user.

[0277] The learning-type keyword detection unit 130 with the supervised learning function and its associated functions in the third embodiment are not limited to application to the first or second embodiment, but may also be applied to the other fourth to seventh embodiments. Similarly, the setting of multiple keywords realized in the second embodiment may be applied to the fourth embodiment in which external sound is played back at double speed up to the current time, the fifth embodiment in which utterances following a keyword (subsequent utterances) are also played back, or the sixth and seventh embodiments in which detection of a keyword triggers transition to hear-through mode. A combination of two or more of the embodiments may also be employed.

[0278] In the operation of the keyword detection unit in embodiments 1 to 7, for the processing after receiving the keyword detection resume signal output by the external sound playback unit, i.e., step S18 in FIG. 8, step S18 in FIG. 11, step S56 in FIG. 15, step S76 in FIG. 20, and step S18 in FIG. 25, it is possible to select either to use the value of Tkst that was updated in the matching processing of the keyword detection unit as is, or to set the time information included in the keyword detection resume signal to Tkst, and then add the following processing to return Tkst to a time in the past by a predetermined time. Tkst = Tkst - Tkst_Ofst For example, Tkst_Ofst=0.2 seconds.

[0279] As a result, the matching process in the keyword detection unit employs an algorithm that requires that external sound data be included as background noise before the keyword is spoken, and even if the start time of the keyword is the same as the original Tkst, the matching process starts with the external sound data going back by Tkst_Ofst, making it possible to accurately detect the voice of this keyword.

[0280] Furthermore, even if the time information contained in the keyword detection resume signal output by the external sound playback unit overlaps with the beginning of the keyword (for example, if the user's operation to stop external sound playback is delayed), the keyword detection unit starts the matching process starting from the earlier time, thereby preventing missed keyword detections.

[0281] Within the scope of the present disclosure, the embodiments can be freely combined, modified, or omitted as appropriate.

[0282] Although the present disclosure has been described in detail, the above description is illustrative in all respects and is not restrictive. It is understood that countless variations not illustrated can be envisioned. [Explanation of symbols]

[0283] 1 external sound input microphone, 2 microphone amplifier, 3 A / D converter, 4 CPU, 5 external sound memory, 6 D / A converter, 7 audio amplifier, 8 audio mixer, 9 playback sound source input terminal, 10 noise cancellation headphones, 11 audio signal input terminal, 12 noise cancellation control signal input terminal, 13 cancellation command input switch, 14 memory, 15 driver unit, 16 noise cancellation circuit, 17 external sound signal input microphone, 18 error signal input microphone, 19 noise-canceled playback signal, 20 external sound microphone signal output terminal, 21 command input switch, 22 external sound playback operation command input switch, 23 noise cancellation control signal input switch, 101 external sound data input unit, 102 external sound recording unit, 103 keyword detection unit, 104 keyword detection data, 105 playback start / stop time generation unit, 106 external sound playback command, 107 external sound playback unit, 108 external sound playback signal, 109 Noise removal unit, 110 noise removal control signal, 111 keyword detection stop signal, 112 keyword detection resume signal, 113 cancel command input unit, 114 cancel command, 120 keyword detection unit, 121 playback start / stop time generation unit, 122 generation rule, 130 learning-type keyword detection unit, 131 external sound playback unit, 132 command input unit, 133 noise removal resume command, 140 external sound playback unit, 150 subsequent speech detection unit, 151 subsequent speech detection data, 152 playback start / stop time generation unit, 200 keyword detection unit, 201 playback start time generation unit, 202 external sound playback command, 203 external sound playback unit, 204 command input unit, 205 external sound playback operation command, 206 noise removal control signal input unit, 210 external sound playback unit.

Claims

1. an external sound recording unit that records external sound data input from the outside and information on the time when the external sound data was input in an external sound memory; a keyword detection unit that performs a matching process to detect whether or not the external sound data recorded in the external sound memory includes one or more preset keywords, and when the external sound data includes the keyword, outputs keyword detection data including information on the start time and end time of the keyword; an external sound reproducing unit that reproduces external sounds including the keyword from the external sound data recorded in the external sound memory based on the keyword detection data; Equipped with a playback start / stop time generating unit that generates a playback start time and a playback stop time of the external sound based on the keyword detection data output by the keyword detecting unit and outputs the generated external sound playback command; an external sound superimposing unit that superimposes an external sound playback signal corresponding to the external sound played back by the external sound playback unit on a playback sound source signal input from a playback sound source and outputs the superimposed signal to a playback device with a noise reduction function; a cancel command input unit that, when a user of the playback device determines that erroneous detection of the keyword has occurred, receives a cancel command from the user indicating that the erroneous detection has occurred; Furthermore, the external sound playback unit plays the external sound based on the external sound playback command, and when playback of the external sound ends, outputs a keyword detection stop signal that stops detection of the keyword to the keyword detection unit and the playback start / stop time generation unit, and outputs a noise removal control signal that stops a noise removal function to the playback device; The noise reduction function control device, wherein the cancellation commands include a command to cancel the playback of the external sound, a command to cancel the stop of keyword detection, and a command to cancel the stop of a noise reduction function.

2. a memory for storing a rule for generating a playback start / stop time set for each of the IDs of the plurality of keywords; the keyword detection unit performs a matching process to detect whether or not at least one of the plurality of keywords is included in the external sound data recorded in the external sound memory, and when the keyword is included in the external sound data, outputs the keyword detection data including the ID of the keyword and information on the start time and end time; The noise removal function control device according to claim 1 , wherein the playback start / stop time generation unit generates the playback start time and the playback stop time of the external sound based on the generation rule corresponding to the ID.

3. the external sound replay unit outputs a keyword detection restart signal, including a learning flag, for restarting detection of the keyword, based on the cancel command received by the cancel command input unit; 3. The noise removal function control device according to claim 2, further comprising a learning type keyword detection unit that performs adaptive learning of a matching algorithm using the learning flag included in the keyword detection resume signal as a teacher signal in a matching process that detects whether the external sound data recorded in the external sound memory contains at least one of the plurality of keywords.

4. An external sound recording unit that records external sound data input from the outside and information on the time when the external sound data was input in an external sound memory; a keyword detection unit that performs a matching process to detect whether or not the external sound data recorded in the external sound memory includes one or more preset keywords, and when the external sound data includes the keyword, outputs keyword detection data including information on the start time and end time of the keyword; an external sound reproducing unit that reproduces external sounds including the keyword from the external sound data recorded in the external sound memory based on the keyword detection data; Equipped with an external sound superimposing unit that superimposes an external sound playback signal corresponding to the external sound played back by the external sound playback unit on a playback sound source signal input from a playback sound source and outputs the superimposed signal to a playback device with a noise reduction function; a cancel command input unit that, when a user of the playback device determines that erroneous detection of the keyword has occurred, receives a cancel command from the user indicating that the erroneous detection has occurred; Furthermore, the external sound playback unit performs double-speed playback of the external sound recorded in the external sound memory from a playback start time of the external sound based on information on the start time of the keyword included in the keyword detection data, and when the double-speed playback of the external sound catches up with the current time, outputs a keyword detection stop signal to the keyword detection unit to stop detection of the keyword, and outputs a noise removal control signal to the playback device to stop a noise removal function; The noise reduction function control device, wherein the cancellation commands include a command to cancel the playback of the external sound, a command to cancel the stop of keyword detection, and a command to cancel the stop of a noise reduction function.

5. An external sound recording unit that records external sound data input from the outside and information on the time when the external sound data was input in an external sound memory; a keyword detection unit that performs a matching process to detect whether or not the external sound data recorded in the external sound memory includes one or more preset keywords, and when the external sound data includes the keyword, outputs keyword detection data including information on the start time and end time of the keyword; an external sound reproducing unit that reproduces external sounds including the keyword from the external sound data recorded in the external sound memory based on the keyword detection data; Equipped with a subsequent utterance detection unit that determines whether or not there is a subsequent utterance that is an utterance following the keyword based on the keyword detection data output by the keyword detection unit, and if there is a subsequent utterance, outputs subsequent utterance detection data including an end time of the subsequent utterance at that time; a playback start / stop time generation unit that generates a playback start time and a playback stop time of the external sound based on the keyword detection data output by the keyword detection unit and outputs the generated external sound playback command, and when the subsequent utterance detection data is input from the subsequent utterance detection unit, outputs the external sound playback command with the playback stop time updated to the end time of the subsequent utterance included in the subsequent utterance detection data; an external sound superimposing unit that superimposes an external sound playback signal corresponding to the external sound played back by the external sound playback unit on a playback sound source signal input from a playback sound source and outputs the superimposed signal to a playback device with a noise reduction function; a cancel command input unit that, when a user of the playback device determines that erroneous detection of the keyword has occurred, receives from the user an input of a cancel command indicating the erroneous detection; Furthermore, the external sound playback unit plays the external sound based on the external sound playback command, and when playback of the external sound ends, outputs a keyword detection stop signal that stops detection of the keyword to the keyword detection unit and the playback start / stop time generation unit, and outputs a noise removal control signal that stops a noise removal function to the playback device; The noise reduction function control device, wherein the cancellation commands include a command to cancel the playback of the external sound, a command to cancel the stop of keyword detection, and a command to cancel the stop of a noise reduction function.

6. the external sound replay unit outputs a keyword detection restart signal, including a learning flag, for restarting detection of the keyword, based on the cancel command received by the cancel command input unit; 2. The noise removal function control device according to claim 1, further comprising a learning-type keyword detection unit that performs adaptive learning of a matching algorithm using the learning flag included in the keyword detection resume signal as a teacher signal in a matching process that detects whether the external sound data recorded in the external sound memory contains one or more of the keywords.

7. Recording external sound data input from the outside and information on the time when the external sound data was input in an external sound memory; performing a matching process to detect whether or not the external sound data recorded in the external sound memory contains one or more preset keywords, and if the external sound data contains the keyword, outputting keyword detection data including information on the start time and end time of the keyword; playing an external sound including the keyword from the external sound data recorded in the external sound memory based on the keyword detection data; generating a playback start time and a playback stop time for the external sound based on the output keyword detection data and outputting the generated external sound playback command; an external sound reproduction signal corresponding to the reproduced external sound is superimposed on a reproduction sound source signal input from a reproduction sound source, and the superimposed signal is output to a reproduction device with a noise reduction function; receiving a cancel command from the user of the playback device, when the user determines that an erroneous detection of the keyword has occurred, indicating that the erroneous detection has occurred; reproducing the external sound based on the external sound reproduction command, and outputting a keyword detection stop signal for stopping the detection of the keyword and a noise removal control signal for stopping a noise removal function to the reproduction device when the reproduction of the external sound has ended; The noise reduction function control method, wherein the cancel command includes a command to cancel the playback of the external sound, a command to cancel the stop of keyword detection, and a command to cancel the stop of a noise reduction function.

8. Recording external sound data input from the outside and information on the time when the external sound data was input in an external sound memory, performing a matching process to detect whether or not the external sound data recorded in the external sound memory contains one or more preset keywords, and if the external sound data contains the keyword, outputting keyword detection data including information on the start time and end time of the keyword; playing an external sound including the keyword from the external sound data recorded in the external sound memory based on the keyword detection data; an external sound reproduction signal corresponding to the reproduced external sound is superimposed on a reproduction sound source signal input from a reproduction sound source, and the superimposed signal is output to a reproduction device with a noise reduction function; receiving a cancel command from the user of the playback device, when the user determines that an erroneous detection of the keyword has occurred, indicating that the erroneous detection has occurred; performing double-speed playback of the external sound recorded in the external sound memory from a playback start time of the external sound based on information on the start time of the keyword included in the keyword detection data, and outputting a keyword detection stop signal for stopping detection of the keyword and a noise removal control signal for stopping a noise removal function to the playback device when the double-speed playback of the external sound catches up with the current time; The noise reduction function control method, wherein the cancel command includes a command to cancel the playback of the external sound, a command to cancel the stop of keyword detection, and a command to cancel the stop of a noise reduction function.

9. Recording external sound data input from the outside and information on the time when the external sound data was input in an external sound memory, performing a matching process to detect whether or not the external sound data recorded in the external sound memory contains one or more preset keywords, and if the external sound data contains the keyword, outputting keyword detection data including information on the start time and end time of the keyword; playing an external sound including the keyword from the external sound data recorded in the external sound memory based on the keyword detection data; Based on the output keyword detection data, it is determined whether or not there is a subsequent utterance that is an utterance following the keyword, and if there is a subsequent utterance, it outputs subsequent utterance detection data including an end time of the subsequent utterance at that time; generating a playback start time and a playback stop time of the external sound based on the output keyword detection data and outputting them as an external sound playback command, and when the subsequent utterance detection data is input, outputting the external sound playback command updated with the end time of the subsequent utterance included in the subsequent utterance detection data as the playback stop time; an external sound reproduction signal corresponding to the reproduced external sound is superimposed on a reproduction sound source signal input from a reproduction sound source, and the superimposed signal is output to a reproduction device with a noise reduction function; when a user of the playback device determines that an erroneous detection of the keyword has occurred, an input of a cancel command indicating the erroneous detection is received from the user; reproducing the external sound based on the external sound reproduction command, and outputting a keyword detection stop signal for stopping the detection of the keyword and a noise removal control signal for stopping a noise removal function to the reproduction device when the reproduction of the external sound has ended; The noise reduction function control method, wherein the cancel command includes a command to cancel the playback of the external sound, a command to cancel the stop of keyword detection, and a command to cancel the stop of a noise reduction function.

Images