Wireless communication device

Abstract

Even when both ears are blocked with earphones or earplugs, it is possible to monitor ambient sounds from wireless communication devices. [Solution] The wireless communication device comprises an audio output device that outputs a demodulated audio signal and an audio input device that outputs an ambient sound audio signal. The wireless communication device comprises an audio signal switching means that selectively switches between the demodulated audio signal and an audio signal from the audio input device and outputs them to the audio output device, and a control circuit that controls the audio signal switching means to output the audio signal from the audio input device to the audio output device in place of the demodulated audio signal based on a predetermined audio switching signal.

Description

Technical Field

[0001] The present invention relates to a wireless communication device that performs a wireless call using earphones (including earphones of headphones, headsets or headphones, earphone microphones, etc.).

Background Art

[0002] For example, in a mobile radio or the like, listening to the content of a wireless call is performed using only earphones or using a bone conduction earphone and an earplug together (see, for example, Patent Documents 1 and 2).

[0003] Here, when performing a wireless call in a noisy environment, by using earphones on both ears, or using a bone conduction earphone and an earplug together, or using an earphone on one ear and an earplug on the other ear, noise entering from the eardrum can be suppressed, so that a voice call can be clearly heard.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, as shown in Fig. 9(A), when directly talking from this state, since both ears are blocked by earphones or earplugs, the content of the direct conversation cannot be heard. In order to respond to this direct conversation, it is necessary to visually recognize the partner of the direct conversation and remove the earphones or the like. In addition, there are problems that it is difficult to respond immediately because of the trouble of visual recognition judgment and attachment / detachment of earphones or the like, and it becomes difficult to attach / detach earphones or the like when wearing a handbag.

[0006] The objective of the present invention is to solve the above problems and provide a wireless communication device that can monitor ambient noise even when both ears are blocked with earphones or earplugs. [Means for solving the problem]

[0007] A wireless communication device according to one aspect of the present invention is: An audio output device that outputs a demodulated audio signal, An audio input device that outputs an audio signal of ambient sound, A wireless communication device equipped with, Audio signal switching means for selectively switching between the demodulated audio signal and the audio signal from the audio input device and outputting them to the audio output device, A control circuit controls the audio signal switching means to output the audio signal from the audio input device to the audio output device in place of the demodulated audio signal, based on a predetermined audio switching signal. It is equipped with. [Effects of the Invention]

[0008] Therefore, according to one aspect of the present invention, the wireless communication device can monitor ambient sounds even when both ears are blocked with earphones or earplugs. [Brief explanation of the drawing]

[0009] [Figure 1] This is a block diagram showing an example configuration of the wireless device 1 according to Embodiment 1. [Figure 2] Figure 1 shows a flowchart illustrating the transmission and reception control process by the wireless device 1. [Figure 3] This is a block diagram showing an example configuration of the wireless device 1A according to Modification 1. [Figure 4] This is a block diagram showing an example configuration of the wireless device 1B according to modified example 2. [Figure 5] This is a block diagram showing an example configuration of the wireless device 1C according to modified example 3. [Figure 6]This is a block diagram showing an example configuration of the wireless device 1D according to Modification 4. [Figure 7] Figure 6 is a block diagram showing an example configuration of the noise cancellation circuit 28. [Figure 8] This block diagram shows an example of the configuration of the deep learning model unit 35 in Figure 7. [Figure 9] (A) is a block diagram showing the operation of a conventional wireless device, and (B) is a block diagram showing the operation of a wireless device according to an embodiment. [Modes for carrying out the invention]

[0010] Embodiments and modified examples of the present invention will be described below with reference to the drawings. The same or similar components are denoted by the same reference numerals.

[0011] (Embodiment 1) Figure 1 is a block diagram showing an example configuration of the wireless device 1 according to Embodiment 1.

[0012] In Figure 1, the radio 1 is an example of a wireless communication device and comprises a receiving antenna 11, a receiving demodulation unit 12, an audio signal switching unit 13, an audio signal amplification unit 14, an earphone 15, a control unit 20, an operation unit including a PTT (Push To Talk) key 21A and a monitor key 21B, a microphone 22, an audio signal amplification unit 23, a modulation transmission unit 24, and a transmitting antenna 25. Here, the PTT key 21A is turned on when it is desired to transmit spoken audio, and the monitor key 21B is turned on when ambient sound monitored by the microphone 22 is listened to through the earphone 15. When the monitor key 21B is turned on, the audio signal switching unit 13 switches its switch SW1 to side c under control from the control unit 20. The control unit 20 controls the overall operation of the radio 1.

[0013] Here, the wireless device 1 according to Embodiment 1 is an example of a wireless communication device of a specific low-power wireless station for a specific low-power wireless communication system. The wireless device 1 according to this Embodiment 1 is characterized by including a monitor key 21B and an audio signal switching unit 13. Also, a plurality of wireless devices 1 constitute a wireless communication system.

[0014] In FIG. 1, the radio signal received by the receiving antenna 11 is input to the reception demodulation unit 12. The reception demodulation unit 12 performs low-noise amplification, low-frequency conversion, intermediate-frequency amplification, etc. on the received radio signal, and then demodulates it into an audio signal by a predetermined demodulation method such as FM (frequency modulation) or PM (phase modulation), and outputs it to the earphone 15 via the audio signal switching unit 13 and the audio signal amplification unit 14. Here, the audio signal switching unit 13 is controlled by the control unit 20 and is selectively switched as follows. (1) When not transmitting or receiving, it is switched to the a side. (2) When receiving, it is switched to the b side. At this time, the demodulated audio signal from the reception demodulation unit 12 is output to the earphone 15 via the audio signal switching unit 13 and the audio signal amplification unit 14. (3) When the monitor key 21B is turned on, it is switched to the c side. At this time, an audio signal such as ambient sound monitored by the microphone 22 is output to the earphone 15 via the audio signal amplification unit 23, the audio signal switching unit 13, and the audio signal amplification unit 14.

[0015] When transmitting, the microphone 22 converts the input audio into an audio signal and outputs it to the audio signal switching unit 13 and the modulation transmission unit 24 via the audio signal amplification unit 23.

[0016] The control unit 20 is composed of a processor such as a CPU (Central Processing Unit) or a DSP (Digital Signal Processor). When the PTT key 21A is turned on, it operates the modulation transmission unit 24. The modulation transmission unit 24 modulates the radio carrier wave according to the input audio signal using the predetermined modulation method. After that, it converts the modulated radio carrier wave, which is the radio signal, to a high frequency and amplifies its power before transmitting it from the transmitting antenna 25. When the monitor key 21B is turned on, the control unit 20 uses an audio switching signal to switch SW1 of the audio signal switching unit 13 to side c, and the audio signal such as ambient sound monitored by the microphone 22 is output to the earphone 15 via the audio signal amplification unit 23, the audio signal switching unit 13, and the audio signal amplification unit 14.

[0017] Figure 2 is a flowchart showing the transmission and reception control process by the wireless device 1 in Figure 1.

[0018] In step 1 of Figure 2, the initialization process is performed, stopping the modulation transmission unit 24 from transmitting, setting the reception demodulation unit 12 to no signal, and switching the audio signal switching unit 13 to side a to stop the output of the audio signal. In addition, both the PTT key 21A and the monitor key 21B are turned off.

[0019] In step S2, it is determined whether the monitor key 21B is turned on or not. If it is YES, the process proceeds to step S3; otherwise, it proceeds to step S6. In step S3, the audio signal switching unit 13 switches switch SW1 to the c side, i.e., to the microphone 22 (ambient sound monitor), and outputs an audio signal. Next, in step S4, it is determined whether the monitor key 21B is turned on or not. If it is YES, the process returns to step S4; otherwise, the process proceeds to step S5. In step S5, the audio signal switching unit 13 switches switch SW1 to the a side to stop outputting the audio signal, and the process returns to step S2.

[0020] In step S6, it is determined whether the PTT key 21A is turned on or not. If it is YES, the process proceeds to step S7; otherwise, it proceeds to step S10. In step S7, the modulation transmitter 24 starts wirelessly transmitting the voice signal. In step S8, it is determined whether the PTT key 21A is turned on or not. If it is YES, the process returns to step S8; otherwise, the process proceeds to step S9. In step S9, the modulation transmitter 24 stops wirelessly transmitting the voice signal and returns to step S2.

[0021] In step S10, it is determined whether or not there is a signal at the receiver / demodulator 12. If the answer is YES, the process proceeds to step S11; otherwise, the process returns to step S2. In step S11, the audio signal switching unit 13 switches switch SW1 to side b and outputs the audio signal from the receiver / demodulator 12 (received audio). Next, in step S12, it is determined whether or not there is a signal at the receiver / demodulator 12. If the answer is YES, the process returns to step S12; otherwise, the process proceeds to step S13. In step S13, the audio signal switching unit 13 switches to side a to stop outputting the audio signal and returns to step S2.

[0022] Figure 9(A) is a block diagram showing the operation of a conventional wireless device, and Figure 9(B) is a block diagram showing the operation of a wireless device according to an embodiment.

[0023] As shown in Figure 9(A), when engaging in direct conversation, the content of the conversation cannot be heard because both ears are blocked by earphones or earplugs. To engage in direct conversation, it is necessary to visually identify the person being spoken to and remove the earphones, etc. Furthermore, the difficulty in visually identifying the person, the effort required to put on and take off the earphones, etc., and the difficulty of putting on and taking off earphones, etc., when wearing gloves, meant that emergency responses became impossible.

[0024] In contrast, as shown in Figure 9(B) relating to Embodiment 1, ambient sound monitored by the microphone 22 is output to the earphone 15. The receiver can monitor the speech of the person they are talking to.

[0025] As described above, according to Embodiment 1, the wireless device 1 is equipped with a microphone 22 and an earphone 15 on its main unit (the earphone 15 may be connected to an earphone jack or be built into the main unit), and during reception, it outputs an audio signal from the earphone 15, and during other reception standby, the above-mentioned trigger allows the main unit's microphone 22 to monitor ambient sound and output the sound from the earphone 15. Alternatively, ambient monitoring can be forced even during reception using the monitor key 21B.

[0026] In the above embodiment 1, the antennas 11 and 25 may be configured to connect the input terminal of the receiving / demodulating unit 12 and the output terminal of the modulation / transmitting unit 24 to the same antenna via a circulator or the like.

[0027] (Variation 1) Figure 3 is a block diagram showing an example configuration of the wireless device 1A according to Modification 1. Modification 1 in Figure 3 differs from Embodiment 1 in the following ways. (1) A VOX (Voice Operated Transmitter) circuit 26 was inserted between the audio signal amplification unit 23 and the audio signal switching unit 13. The differences are explained below.

[0028] In Figure 3, the VOX circuit 26 is an example of an audio signal detection unit. It compares the input audio signal with a predetermined threshold value and outputs an H-level audio detection signal (comparison result signal) as a control signal for the switch SW1 of the audio signal switching unit 13 when the input audio signal is equal to or greater than the threshold value. Conversely, when the input audio signal is less than the threshold value, it outputs an L-level audio detection signal (comparison result signal) to the switch SW1. This allows the earphone 15 to monitor ambient noise (including audio signals) when ambient noise above a predetermined level is input to the microphone 22.

[0029] Alternatively, the control signal from the VOX circuit 26 may be input to the control terminal of the switch SW1 via the control unit 20.

[0030] (Modification 2) Figure 4 is a block diagram showing an example configuration of the wireless device 1B according to Modification 2. Modification 2 in Figure 4 differs from Embodiment 1 in Figure 1 in the following ways. (1) The wireless device 1 is equipped with an acceleration sensor 27 (or vibration sensor) for detecting vibrations. The differences are explained below.

[0031] In Figure 4, the acceleration sensor 27 is an example of a vibration sensor. When the wireless device 1 detects vibrations above a predetermined threshold, it outputs an H-level vibration detection signal (comparison result signal) as a control signal to the switch SW1 of the audio signal switching unit 13. Conversely, when the vibration is below the threshold, it outputs an L-level vibration detection signal (comparison result signal) to the switch SW1. This allows the earphone 15 to monitor ambient sounds when vibrations above a predetermined level are applied to the wireless device 1. In particular, the ambient sound monitoring state can be activated without requiring the user to perform fine operations (such as pressing the monitor key 21B) when wearing gloves.

[0032] Alternatively, the control signal from the acceleration sensor 27 may be input to the control terminal of the switch SW1 via the control unit 20.

[0033] (Variation 3) Figure 5 is a block diagram showing an example configuration of the wireless device 1C according to Modification 3. Modification 3 in Figure 5 differs from Modification 1 in Figure 3 in the following ways. (1) The microphone 22 and the earphone 15 were configured as an earphone microphone 15A. The differences are explained below.

[0034] In Figure 5, ambient sound monitoring is automatically initiated by the user's own voice, triggered by ambient sound detection by the VOX circuit 26 in the earphone microphone 15A. At that time, by prioritizing ambient sound monitoring using the user's own voice over the incoming call audio, it becomes possible to automatically prioritize direct conversations such as customer service.

[0035] Alternatively, the control signal from the VOX circuit 26 may be input to the control terminal of the switch SW1 via the control unit 20.

[0036] (Modification 4) Figure 6 is a block diagram showing an example configuration of the wireless device 1D according to Modification 4. Modification 4 in Figure 6 differs from Embodiment 1 in Figure 2 in the following ways. (1) A noise cancellation circuit 28 was inserted between the audio signal amplification unit 23 and the audio signal switching unit 13. The differences are explained below.

[0037] The microphone 22 converts the input sound into an audio signal and outputs it to the noise cancellation circuit 28 via the audio signal amplification unit 23. The noise cancellation circuit 28 uses a deep learning model unit 35 (see Figures 7 and 8) that has been deeply learned using human speech to process the audio signal so as to cancel noise by allowing only the audio signal period to pass through the input audio signal, and then outputs it to the audio signal switching unit 13 via the audio signal detection unit 29. The audio signal detection unit 29 is an example of a VOX circuit including a comparator, which compares the input audio signal with a predetermined threshold, and when the input audio signal is greater than or equal to the threshold, outputs an H-level audio detection signal (comparison result signal) to the control terminal of switch SW1 of the audio signal switching unit 13, thereby switching switch SW1 to the c side. Conversely, when the input audio signal is less than the threshold, outputs an L-level audio detection signal (comparison result signal) to the control terminal of switch SW1 of the audio signal switching unit 13, thereby switching switch SW1 to the a side.

[0038] Alternatively, the audio detection signal from the audio signal detection unit 29 may be output to the control terminal of the switch SW1 of the audio signal switching unit 13 via the control unit 20.

[0039] Next, referring to Figure 7, the configuration and operation of the noise cancellation circuit 28 in Figure 1 using the deep learning model unit 35 will be described below.

[0040] Figure 7 is a block diagram showing an example configuration of the noise cancellation circuit 28 in Figure 1.

[0041] Here, the term "phoneme" refers to a unit of sound that distinguishes one word from another in a particular language, the term "vibration rate" refers to the number of movements between 0 and 1 in the digitized vibration data in each second, and the term "vibration count value (VC)" refers to the sum of the values ​​of the digitized vibration data in each frame. Furthermore, "vibration pattern" refers to the data distribution of the sum of vibration frequencies calculated for a predetermined number of frames along the time axis. The deep learning model unit 35 performs noise cancellation processing by considering the differences in data distribution of different vibration patterns, i.e., different sums of vibration count values ​​(VS values). The vibration rate is similar to the vibration count value, but the larger the vibration rate, the larger the vibration count value.

[0042] Both the amplitude and vibration rate of the audio signal are observable. The noise cancellation circuit 28 is characterized by detecting audio events according to the amplitude and vibration rate of the audio signal. Another feature is that it distinguishes between audio and non-audio / silence by measuring the sum of vibration count values ​​of the digitized vibration data for a predetermined number of frames. Yet another feature is that it classifies the input audio signal data stream into different phonemes based on its vibration pattern. Another feature is that it correctly identifies the first trigger phoneme from the input audio signal data stream in order to trigger the downstream processing unit, thereby saving computational costs such as power consumption of the computing system including the processing unit.

[0043] In FIG. 7, the noise canceling circuit 28 performs noise canceling processing using voice event detection, and includes a voice signal preprocessing unit 38, an AD converter 39, and a voice signal processing unit 30. Here, the voice signal preprocessing unit 38 performs voice signal preprocessing on the analog voice signal, including high-pass filtering, low-pass filtering, amplification, or a combination thereof, etc., and outputs the processed analog voice signal to the AD converter 39. That is, the voice signal preprocessing unit 38 allows only a predetermined bandwidth within a predetermined level range of the human voice signal to pass through the voice signal from the microphone 22. Next, the AD converter 39 AD-converts the analog voice signal into a digital voice signal according to a predetermined reference voltage Vref and an allowable voltage Vadm (<Vref), and outputs it to the input interface 36 of the voice signal processing unit 30.

[0044] In this embodiment, the AD converter 39 uses a tolerance voltage Vadm, which is smaller than the reference voltage Vref, in combination with the reference voltage Vref to form a first threshold voltage Vth1 (=Vref+Vadm) and a second threshold voltage Vth2 (=Vref-Vadm). Based on the first threshold voltage Vth1 and the second threshold voltage Vth2, the AD converter 39 does not perform AD conversion on noise that is greater than or equal to the first threshold voltage Vth1 or less than or equal to the second threshold voltage Vth2, but performs AD conversion on the audio signal in between, thereby removing noise and interference from the input analog audio signal. Here, for example, if Vref=1.0V and Vadm=0.01V, it can be understood that the frequency of vibration data is low in a quiet environment and high in an audio environment. In this embodiment, "frame size" means the number of sampling points corresponding to the digitized vibration data in each frame, and "phoneme window Tw" means the time for collecting the audio features of each phoneme. In a preferred embodiment, the duration Tf of each frame is, for example, 0.1 to 1 millisecond (ms), and the phoneme window Tw is, for example, about 0.3 seconds. In a more preferred embodiment, the number of sampling points corresponding to the digitized vibration data within each frame is, for example, in the range of 1 to 16.

[0045] When analyzing audio signals, short-term analysis methods are usually employed because most audio signals stabilize within a short period. For example, assuming that the sampling frequency fs used in the AD converter 39 is 16000 and the duration Tf of each frame is 1 ms, the frame size will be fs × 1 / 1000 = 16 sample points.

[0046] In Figure 7, the audio signal processing unit 30 is composed of, for example, a computer device. (1) A CPU (Central Processing Unit) 31 that performs predetermined audio signal processing such as noise cancellation, (2) An operating system that performs the basic processing of the CPU 31 and the program for audio signal processing, and a ROM (Read Only Memory) 32 that stores data necessary for executing the program, (3) When the operating system that executes the basic processing of the CPU 31 and the program for audio signal processing are executed, the RAM (Read Access Memory) 33 stores the data being processed, (4) A non-volatile EEPROM (Electrically Erasable Programmable Memory) 34 that stores the setting data and the like necessary to perform the audio signal processing described later, (5) For example, a deep learning model unit 35 which is composed of a neural network or the like and is deeply learned based on human speech signal data, removes noise from the input speech signal data and extracts essentially only the speech signal for output, (6) An input interface 36 that performs a predetermined signal conversion process to convert the audio signal data input from the AD converter 39 into a signal specification value for the subsequent stage and outputs it to the CPU 31, (7) An output interface 37 that performs a predetermined signal conversion process to convert the audio signal data, from which noise has been removed by the deep learning model unit 35, into a signal specification value for the subsequent stage, and outputs it to the audio signal detection unit 29, It is composed of the following features.

[0047] Here, the EEPROM 34 stores, for example, a series of vibration count values ​​VC, the sum of vibration count values ​​VS, the sum of vibration count values ​​VSF, the sum of vibration count values ​​VSP (described later), and the audio feature values ​​of all feature vectors. Note that the EEPROM 34 may be an external memory or other storage device. The audio event detection method applied to the audio signal processing unit 30 is executed by the CPU 31 during runtime to capture audio events. Audio event detection is performed assuming fs=16000, Tf=1ms, and Tw=0.3s.

[0048] Specifically, the CPU 31 calculates the sum of vibration data values ​​for the current frame (i.e., within 1ms) to obtain the vibration count value VC, and then stores the VC value of the current frame at time Tj in the EEPROM 34. Here, the vibration count values ​​VC of x frames are added together to obtain the sum VS of the vibration count values ​​of the current frame at time Tj. The x frames include the current frame. In one embodiment, the CPU 31 adds the vibration count value VC of the current frame at time Tj to the sum VSp of the vibration count values ​​of the (x-1) frames immediately preceding it to obtain the sum VS (=VC+VSp) of the vibration count values ​​of x frames at time Tj.

[0049] In the modified example, the CPU 31 adds the vibration count value VC of the current frame at time Tj, the sum of the vibration count values ​​VSf of the y frames immediately following it, and the sum of the vibration count values ​​VSp of the (xy-1) frames immediately preceding it, to obtain the sum of the vibration count values ​​VS (=VC+VSf+VSp) of x frames at time Tj, where y is greater than or equal to zero. The CPU 31 stores the values ​​of VS, VSf, and VSp in the EEPROM 34. In a preferred embodiment, the duration (x×Tf) of x frames (phoneme window Tw) is approximately 0.3 seconds. In a more preferred embodiment, the number of sampling points corresponding to the digitized vibration data of x frames is in the range of x~16x.

[0050] Generally, for audio signal data, the vibration patterns of the vibration count value VC are similar for the same phoneme, but the vibration patterns of the VS value are completely different for different phonemes. Therefore, phonemes can be distinguished using the vibration pattern of the vibration count value VC. In particular, the frequency distribution of the vibration count value VC for sounds such as chicken or cat meows is completely different from that of human speech, and it is known that most of the vibration count values ​​VC for human speech are distributed below 40.

[0051] In the learning phase, the CPU 31 of the speech signal processing unit 30 first executes a predetermined speech signal data collection method multiple times to collect multiple feature vectors for multiple phonemes, and adds labels corresponding to the multiple feature vectors to form multiple labeled learning examples. Subsequently, multiple labeled learning examples for different phonemes, including the activated phoneme, are applied to the training of the deep learning model unit 35. Finally, the trained deep learning model unit 35 (which constitutes a prediction model for speech signal data) is created to classify whether the input stream of speech signal data contains the activated phoneme. If a predetermined phoneme is specified as the activated phoneme for the speech signal processing unit 30, the deep learning model unit 35 is trained with multiple labeled learning examples for different phonemes, including at least the specified phoneme.

[0052] In other words, during the training phase, the deep learning model 35 is trained using a set of labeled training examples, so that the deep learning model 35 recognizes a predetermined activation phoneme between j=0 and 299 based on three speech features (e.g., (VSj, TDj, TGj)) of each frame of the labeled training examples. At the end of the training phase, the trained deep learning model 35 provides a trained score corresponding to the activation phoneme, and this trained score is then used as a criterion for classifying the input stream of speech signal data at runtime. VSj, TDj, and TGj are defined as follows. (1) VSj: The sum of the vibration counter values ​​of frame j (VS value); (2) TDj: The time period of the sum of non-zero vibration count values ​​(VS value) in frame j; and (3) TGj; the time gap between the sum of non-zero vibration count values ​​(VS values) in frame j.

[0053] To train the deep learning model unit 35, various machine learning techniques related to supervised learning can be used, such as support vector machines (SVM), random forests, and convolutional neural networks. In supervised learning, a function computation unit (i.e., the deep learning model unit 35) is created using multiple labeled training examples, each example consisting of an input feature vector and a labeled output. Once trained, the deep learning model unit 35 can be applied to new unlabeled examples to generate corresponding scores or predictions.

[0054] Figure 8 is a block diagram showing a detailed configuration example of the deep learning model unit 35 in Figure 7.

[0055] The deep learning model unit 35 is implemented using a neural network, for example, as shown in Figure 8. Here, the neural network includes one input layer 41, at least one, preferably more, hidden layers 42, and one output layer 43. The input layer 41 has three input neurons 51, 52, and 53, each of which corresponds to three audio feature values ​​(i.e., VSj, TDj, TGj) for each frame of the feature vector. The hidden layer 42 consists of neurons 61 to 74, each having a weight coefficient associated with each input neuron 51, 52, and 53, and a bias coefficient for each neuron. By changing the weight coefficients and bias coefficients of each neuron 61 to 74 in the hidden layer 42 throughout the learning phase cycle, the neural network can be trained to report predicted values ​​for a given type of input. Furthermore, the output layer 43 includes one output neuron 81 that provides one predicted value corresponding to a phoneme (specifically, indicating whether it is a speech period or a non-speech period containing noise).

[0056] As explained above, in the noise cancellation circuit 28, the deep learning model unit 35 is trained using human speech characteristic parameters and determines whether or not there is a non-speech period containing noise in the input speech signal. Then, the CPU 31 of the speech signal processing unit 30 performs noise cancellation processing based on the determination of the deep learning model unit 35 to prevent the non-speech period containing noise from passing through the input speech signal, and outputs the speech signal after the noise cancellation processing. Here, the deep learning model unit 35 is composed of the neural network shown in Figure 8, which takes human speech characteristic parameters as input and outputs the determination result of whether or not there is a non-speech period containing noise in the input speech signal.

[0057] As explained above, according to Modification 4, by passing the input sound from the microphone 22 through the noise cancellation circuit 28 that removes non-voice sounds, ambient sound monitoring of only human voices becomes possible. Furthermore, if the ambient sound monitoring is triggered by the voice of the other party detected by the audio signal detection unit 29 (VOX circuit) of the main unit microphone 22 (in the Modification, by the VOX circuit 26 of the earphone microphone 15A, by using it in conjunction with the noise cancellation circuit 28, ambient sound monitoring will start with only the other party's voice, only your own voice, or both, making it possible to prevent false detection due to ambient noise.

[0058] Furthermore, when monitoring ambient sound, the distance between the other party and the microphone 22 on the main unit is longer compared to normal wireless transmission. Therefore, the system may be configured to automatically increase the gain of the microphone 22 (the amplification level of the audio signal amplification unit 23) when monitoring ambient sound. Also, by using it in conjunction with the noise cancellation circuit 28, it becomes possible to amplify only human voices and monitor ambient sound with ambient noise at a low level. This prevents missed conversations when using the earphones 15 and eliminates the need to remove them, enabling emergency responses. Moreover, when bone conduction earphones and earplugs are used together, direct conversations are more difficult to hear, so the benefits of this ambient sound monitoring system are greatly enhanced.

[0059] (Another variation) In the embodiments and modifications described above, the earphone 15 is an example of an audio output device having a structure that seals the user's ears so that the user has difficulty hearing ambient sounds, or can hardly hear them at all. For example, it may be headphones, earphones of a headset, or headphones with an earphone microphone. The structure that seals the user's ears so that the user has difficulty hearing ambient sounds, or can hardly hear them at all, may include other sealing means. Here, the earphone 15 (including other modifications) may be mounted on the main unit, connected to an external option via a wired connection, or connected to an external option wirelessly using, for example, Bluetooth.

[0060] In the above embodiments and modifications, the transmitting and receiving functions of the radio 1 may be separate devices or an integrated transmitting and receiving device. In this case, the microphone 22 for ambient sound monitoring in the separate radios may be shared with the microphone for voice transmission. The microphone 22 is just one example of a voice input device and may be an earphone microphone or the like, as mentioned above. [Industrial applicability]

[0061] As detailed above, the wireless communication device according to the present invention allows for monitoring of ambient sounds even when both ears are blocked by earphones or earplugs. [Explanation of symbols]

[0062] 1,1A,1B,1C,1D,1-1,1-2 Radio 11 Antennas 12 Receiving and demodulating section 13. Audio signal switching section 14. Audio signal amplification section 15 earphones 15A Earphone Microphone 20 Control Unit 21A PTT key on the control panel 21B Monitor key on the control panel 22 Microphones 23 Audio signal amplification section 24 Modulation Transmitter 25 Antennas 26 VOX circuit 27. Accelerometer 28. Noise Cancellation Circuit 29 Audio signal detection unit 30 Audio signal processing unit 31 CPU 32 ROM 33 RAM 34 EEPROM 35 Deep Learning Model Section 36 Input Interfaces 37 Output Interface 38 Audio signal preprocessing unit 39 AD Converter 41 Input Layer 42 Middle Class 43 Output Layer 51-81 neurons SW1 Switch

Claims

1. An audio output device that outputs a demodulated audio signal, An audio input device that outputs an audio signal of ambient sound, A wireless communication device equipped with, The aforementioned wireless communication device further, Audio signal switching means for selectively switching between the demodulated audio signal and the audio signal from the audio input device and outputting them to the audio output device, A control circuit controls the audio signal switching means to output the audio signal from the audio input device to the audio output device in place of the demodulated audio signal, based on a predetermined audio switching signal. A noise cancellation unit that processes the audio signal from the audio input device to cancel out noise and outputs the result, A sound detection unit compares the audio signal from the noise cancellation unit with a predetermined threshold and outputs the audio switching signal, which is the comparison result signal, when the audio signal from the noise cancellation unit is greater than or equal to the threshold. Equipped with, The noise cancellation unit performs noise cancellation processing using a deep learning model unit that is trained using human speech characteristic parameters and determines whether or not a given speech signal contains noise. The noise cancellation unit includes an audio signal processing unit that, based on the determination of the deep learning model unit, performs noise cancellation processing on the input audio signal to prevent the passage of non-audio periods containing noise, and outputs the audio signal after the noise cancellation processing. Wireless communication device.

2. The aforementioned wireless communication device is The system further includes an operation unit that includes a monitor key for instructing the audio input device to output an audio signal to the audio output device, The control circuit outputs the audio switching signal to the audio signal switching means when the monitor key is turned on. The wireless communication device according to claim 1.

3. The control circuit controls the audio signal switching means using the audio switching signal to output the audio signal from the audio input device to the audio output device in place of the demodulated audio signal when the audio signal input to the audio input device is above a predetermined threshold. The wireless communication device according to claim 1.

4. The wireless communication device further comprises a vibration sensor for detecting vibrations of the wireless communication device, The control circuit controls the audio signal switching means using the audio switching signal so that when the vibration sensor detects the vibration, it outputs sound from the audio input device to the audio output device instead of the demodulated audio signal. The wireless communication device according to claim 1.

5. The aforementioned audio output device is headphones, earphones of a headset, or headphones or earphone microphone. The wireless communication device according to claim 1.

6. The aforementioned audio input device is a microphone or an earphone microphone. The wireless communication device according to claim 1.

Images