TWI741841B - Wireless vibration audio conversion system and method - Google Patents

Abstract

The present invention relates to a wireless vibration audio conversion system and method, which sense a first vibration change result and an audio change result of a throat during a first sensing period to perform calculations and generate a corresponding audio reference According to the audio reference data, a second vibration change result sensed during the second sensing period is converted into an audio output signal for outputting a sound signal, and the sound signal corresponds to the audio change result. The present invention thus provides a sound signal closer to human voice.

Description

Wireless vibration audio conversion system and method

The invention relates to an audio conversion device and a method thereof, in particular to a wireless vibration audio conversion system and a method thereof.

Radio devices have become one of the most commonly used products in modern people’s daily life. For example, mobile communication equipment, voice recorders, music players with recording functions, etc., require good-quality radio devices to receive external sounds, especially when used The effect of the voice sent by the user, and various anti-noise methods are proposed, so that the received sound is not as unclear as the sound transmitted in the air, and especially during the user's movement, such as sports, driving, intense Activities or noisy environments will not be affected. Common radio devices include capacitive radio devices and piezoelectric radio devices. In piezoelectric radio devices, most of the piezoelectric radio devices use a piezoelectric material element that generates a piezoelectric signal through vibration and is attached to the human body to feel the sound generated by the human body. The vibration, and the pressure generated by the vibration is transferred to the piezoelectric material, so that the piezoelectric material element is subjected to external pressure to change the potential difference, thereby generating a voltage signal for further processing.

The traditional radio device is held by a human hand or held by the neck close to the user's mouth to facilitate the air microphone to receive the user's voice. However, the air-conducting radio device needs to be held by the user's mouth or neck-hung, which makes it difficult for users to free up their hands to handle affairs. Although the industry has further developed neck-hung radio devices or desktop radio devices that can be used People free their hands, but still need to adjust the setting position and angle of the radio frequently, and the air-conducting radio hanging on the chest of the user easily swings to the left and right with the user’s activities, which affects the user’s activities. It is quite inconvenient.

Furthermore, in order to overcome the problems of the above-mentioned air conduction type sound-receiving device, a throat vibration type sound-receiving device was developed. The sound-receiving device is mainly arranged in the user's throat, so that the sound-receiving device receives sound through the vibration generated by the throat cavity when the user speaks. , As the voice input source of the computing device, but the vibration There will still be situations where the voice reception is still unclear, so a throat radio device has been developed to further pick up the throat. However, the throat radio device still has unclear voices in the sound collection. The reason is that the throat sound is not emitted from the throat but transmitted to the mouth, so the volume will be low. Furthermore, the voice signal of the throat sound is related to the throat. The vibration signal of is a different type of signal, which is not easy to compensate.

Based on the above-mentioned problems, the present invention provides a wireless vibration audio conversion system and method thereof, which generates an audio reference data based on a first vibration change result and an audio change result in a first sensing period by a computing device, and then The audio reference material converts a second vibration change result during the second sensing period into an audio output signal, thereby providing an audio output signal close to human voice.

The main purpose of the present invention is to provide a wireless vibration audio conversion system and method thereof, which generates an audio reference from a first vibration variation result and an audio variation result input to the computing device by executing an application program in the computing device The data is further converted into an audio output signal from a second vibration change result to provide an audio output signal close to human voice.

The present invention discloses a wireless vibration audio conversion method with intelligent learning, which first uses a vibration sensor of a radio device to sense a first vibration change result generated by a throat during a first sensing period, and Use an audio sensor of the radio device to sense an audio change result generated by a mouth during the first sensing period; continue, wirelessly transmit the first vibration change result and the audio change result to an arithmetic device The arithmetic device executes an audio and vibration signal conversion program and converts the vibration change result and the audio change result to two corresponding features, so that the arithmetic device executes an application program, and the audio change result according to the two corresponding features Pair with the result of the vibration change, thereby generating a corresponding audio reference material. It can be seen from the above that the present invention can generate a corresponding audio reference data according to the first vibration change result and the audio change result through the computing device, for the artificial intelligence application program to learn audio vibration conversion.

The present invention provides an embodiment, wherein the application program includes an artificial intelligence algorithm, and the artificial intelligence algorithm is a Deep Neural Networks (DNN).

The present invention provides an embodiment, wherein in the step of executing an audio and vibration signal conversion program by the arithmetic device and converting the vibration variation result and the audio variation result to two corresponding features, the arithmetic device converts the audio variation result into Corresponding to an audio corresponding feature and converting the vibration change result into a corresponding vibration corresponding feature. The audio corresponding feature and the vibration corresponding feature are logarithmic power spectrum, Mel-Frequency Cepstrum (MFC) or linear prediction Analyze the conversion result of the signal processing of the LPC Spectrum.

The present invention provides an embodiment, wherein the vibration sensor is an acceleration sensor or a piezoelectric sensor.

The present invention also discloses a wireless vibration audio conversion system with intelligent learning, which includes a radio device and an arithmetic device, and the radio device includes a vibration sensor, an audio sensor and a first wireless transmission unit. The device is provided with a processing unit, a storage unit, and a second wireless transmission unit. The vibration sensor senses a first vibration change result generated by a throat during a first sensing period and senses the throat at A second vibration change result generated during a second sensing period; and the audio sensor senses an audio change result generated by the mouth during the first sensing period, and the first wireless transmission unit is connected to the An arithmetic device, the vibration sensor and the audio sensor; and the arithmetic device is provided with a processing unit, a storage unit and a second wireless transmission unit, the storage unit stores an application program, and the second wireless transmission unit is connected to the A first wireless transmission unit, the processing unit executes the application program and receives the first vibration change result and the audio change result via the first wireless transmission unit and the second wireless transmission unit, and based on the first vibration change result and The audio change results in an audio reference material. It can be seen from the above that the present invention can generate a corresponding audio reference data according to the first vibration change result and the audio change result through the computing device, for the artificial intelligence application program to learn audio vibration conversion.

The present invention provides another embodiment, wherein the application program includes an artificial intelligence algorithm and an audio vibration conversion program, and the artificial intelligence algorithm is a Deep Neural Networks (DNN).

The present invention provides another embodiment, wherein the computing device converts the audio change result into a corresponding audio corresponding feature and converts the vibration change result into a corresponding vibration corresponding feature, The audio corresponding feature and the vibration corresponding feature are conversion results of logarithmic power spectrum, Mel-Frequency Cepstrum (MFC) or linear prediction analysis spectrum (LPC Spectrum) signal processing.

The present invention provides another embodiment, wherein the vibration sensor is an acceleration sensor or a piezoelectric sensor.

The present invention discloses a wireless vibration audio conversion method, which first uses a vibration sensor to sense a vibration change result generated by a throat during a sensing period; then, the vibration change result is wirelessly transmitted to an operation A device in which the computing device executes an audio vibration conversion program to convert the vibration change result to a corresponding feature for the computing device to execute an artificial intelligence application program, and the corresponding feature is based on an audio reference data pre-stored in a storage unit The result of the vibration change is converted into an audio mapping signal having a reference audio sound field effect feature, and the computing device executes the audio vibration conversion program to convert the audio mapping signal of the corresponding feature into an audio output in an output format Signal. It can be seen from the above that the present invention can use an arithmetic device based on an audio reference data, and then after the arithmetic device receives the vibration change result, refer to the audio reference data to convert the second vibration change result and interpolate it into an audio output close to the human voice. Signal.

The present invention provides an embodiment, wherein the application program includes an artificial intelligence algorithm and an audio vibration conversion program, and the artificial intelligence algorithm is a Deep Neural Networks (DNN).

The present invention provides an embodiment, wherein the arithmetic device converts the result of the vibration change into a corresponding vibration characteristic, and the vibration corresponding characteristic is a logarithmic power spectrum, a Mel-Frequency Cepstrum (MFC) or a line The conversion result of the signal processing of the LPC Spectrum (LPC Spectrum).

The present invention provides an embodiment, wherein the vibration sensor is an acceleration sensor or a piezoelectric sensor.

The present invention provides an embodiment, wherein the throat is a position on the body surface corresponding to the vocal cords or the throat.

The present invention also discloses a wireless vibration audio conversion system, which includes a radio device and an arithmetic device, and the radio device includes a vibration sensor, an audio sensor and a first wireless transmission Unit, the arithmetic device is provided with a processing unit, a storage unit and a second wireless transmission unit. The vibration sensor senses a first vibration change result generated by a throat during a first sensing period and senses the A second vibration change result generated by the throat during a second sensing period; and the audio sensor senses an audio change result generated by the throat during the first sensing period, and the first wireless transmission The unit is connected to the arithmetic device and the audio sensor; and the arithmetic device is provided with a processing unit, a storage unit and a second wireless transmission unit, the storage unit stores an application program, and the second wireless transmission unit is wirelessly connected to the first A wireless transmission unit, the processing unit receives the first vibration change result and the audio change result via the first wireless transmission unit and the second wireless transmission unit, and the arithmetic device executes an audio vibration conversion program to convert the vibration change result to a Corresponding to the feature, the processing unit executes an artificial intelligence application to convert the vibration change result of the corresponding feature into an audio mapping signal having a reference audio sound field effect feature according to an audio reference data pre-stored in a storage unit, the The processing unit executes the audio vibration conversion program to convert the audio mapping signal of the corresponding feature into an audio output signal in an output format. It can be seen from the above that the present invention can generate an audio reference data corresponding to the first vibration change result and the audio change result through the arithmetic device, and then after the arithmetic device receives the second vibration change result, refer to the audio reference data to change the second vibration As a result, an audio output signal close to the human voice is converted.

The present invention provides another embodiment, further comprising an output device connected to the computing device, receiving and outputting an audio signal according to the audio output signal of the outputable format, the audio signal corresponding to the audio change result.

The present invention provides another embodiment, wherein the application program includes an artificial intelligence algorithm and an audio vibration conversion program, and the artificial intelligence algorithm is a Deep Neural Networks (DNN).

The present invention provides another embodiment, wherein the arithmetic device converts the vibration change result into a corresponding vibration corresponding feature, and the vibration corresponding feature is a logarithmic power spectrum, a Mel-Frequency Cepstrum (MFC) or A conversion result of the signal processing of the LPC Spectrum.

The present invention provides another embodiment, wherein the vibration sensor is an acceleration sensor or a piezoelectric sensor.

1: Wireless vibration audio conversion system

10: Radio device

12: Vibration sensor

14: Audio sensor

16: The first wireless transmission unit

20: Computing device

22: processing unit

24: storage unit

26: The second wireless transmission unit

30: output unit

AI: artificial intelligence algorithm

DNN: Deep Neural Network

F _V1 : First vibration characteristic

F _V2 : Second vibration feature

IFT: Inverse Fourier Transformation Module

M: Mouth

OUT: output signal

P: Application

P1: Audio vibration conversion program

P2: Artificial Intelligence Module

Pd1: During the first sensing period

Pd2: the second sensing period

RAM: storage unit

REF: Audio reference material

ST: Fourier transform module

S _V1 : The first vibration change result

S _V2 : Second vibration change result

S _W : Audio change result

T: Throat

U: User

V1: Vibration

V2: Vibration

VF1: Corresponding characteristics of the first vibration

VF2: Corresponding characteristics of the second vibration

W: sound

WF: Audio corresponding features

WI: Audio mapping signal

WO: Audio output signal

WT: Audio conversion module

S10-S50: steps

Figure 1: It is a flowchart of an embodiment of the present invention; Figure A: It is a schematic diagram of synchronously sensing audio and vibration in an embodiment of the present invention; Figure B: It is one of the present invention The schematic diagram of calculating the audio reference data of the embodiment; the third figure: it is a flowchart of another embodiment of the present invention; the fourth figure A: it is the schematic diagram of sensing vibration of another embodiment of the present invention; Figure B: It is a schematic diagram of vibration-converted audio according to another embodiment of the present invention; and Figure C: It is a schematic diagram of audio output according to another embodiment of the present invention.

In order for your review committee to have a further understanding and understanding of the features of the present invention and the effects achieved, the examples and accompanying descriptions are provided here. The description is as follows: In view of the fact that the existing vibrating radios are not manufactured to provide better quality output. According to the signal problem, the present invention proposes a wireless vibration audio conversion system and method to solve the problem that the prior art cannot provide a better quality output signal.

Hereinafter, the characteristics and structure of a wireless vibration audio conversion system and method disclosed in the present invention will be further explained: First, please refer to the first figure, which is a flowchart of an embodiment of the present invention. As shown in the figure, the steps of the wireless vibration audio conversion method of the present invention include: Step S10: Use a vibration sensor of a radio device to sense a throat during a first sensing period and generate a first vibration change As a result, using an audio sensor of the radio device to sense a mouth during the first sensing period and generate an audio change result; Step S20: Wirelessly transmit the first vibration change result and the audio change result to an arithmetic device; Step S25: The arithmetic device executes an audio and vibration signal conversion program and converts the vibration change result and the audio change result to corresponding features; and Step S30 : The computing device executes an application program based on the audio change result to compare the first vibration change result to generate a corresponding audio reference data.

Please refer to Figures 2A to 2B, which are schematic diagrams of synchronously sensing audio and vibration during the first sensing period and a schematic diagram of computing audio reference data in an embodiment of the present invention. As shown in the figure, the wireless vibration audio conversion system 1 of the present invention includes a radio device 10 and an arithmetic device 20. The radio device 10 is provided with a vibration sensor 12, an audio sensor 14 and a first wireless device. The transmission unit 16 and the computing device 20 are provided with a processing unit 22, a storage unit 24 and a second wireless transmission unit 26. The storage unit 24 stores an application program P. The first wireless transmission unit 16 is connected to the second wireless transmission unit 26.

In step S10, as shown in Fig. 2A, a user U wears the radio device 10 on a throat T in the form of hanging, neck strap or neck loop. When the user U makes a sound The throat T generates a vibration V1 correspondingly, and is transmitted to the mouth M to emit a sound W. The vibration sensor 12 on the radio device 10 senses the vibration V1 generated by the throat T during a first sensing period Pd1. A vibration change result S _V1 . At the same time, the audio sensor 14 of the radio device 10 senses the sound W from the mouth M during the first sensing period Pd1, and accordingly generates an audio change The result is S _W. In step S20, as shown in Figure 2A, the radio device 10 is connected to the wireless transmission interface formed by the second wireless transmission unit 26 through the first wireless transmission unit 16 (for example: Bluetooth, wireless local area network) (WIFI), cellular network (ZigBee), LoRa long-distance transmission technology), and transmit the first vibration change result S _V1 and the audio change result S _W to the computing device 20, especially the processing unit 22 The first vibration change result S _V1 and the audio change result _{SW are} temporarily stored in the storage unit 24.

In step S25, as shown in FIG. 2B, the computing device 20 reads the application program P in the storage unit 24 through the processing unit 22 to compute the first vibration change result S _V1 and the audio frequency Change the result S _W to generate an audio reference data REF, where the application P includes an audio vibration conversion program P1 and an artificial intelligence module P2, and the audio vibration conversion program P1 includes a Fourier conversion module ST and an audio conversion Module WT, Fourier transform module ST is to run Fourier transform function (Fourier Transform) to transform the first vibration change result S _V1 into a first vibration corresponding characteristic VF1, the audio conversion module WT converts the audio change The result S _W is an audio corresponding feature. The audio corresponding feature WF and the first vibration corresponding feature VF1 of this embodiment are Log-Power Spectrum (LPS). In addition, the audio corresponding feature WF is related to The first vibration corresponding characteristic VF1 may be a conversion result of signal processing of Mel-Frequency Cepstrum (MFC) or LPC Spectrum (LPC Spectrum).

In step S30, as shown in Figure 2B, the artificial intelligence module P2 runs at least one artificial intelligence algorithm AI, for example: the artificial intelligence algorithm AI is a deep neural network DNN (Deep Neural Networks, DNN) Based on the same format, let the artificial intelligence algorithm AI learn the comparison between the audio corresponding feature WF and the first vibration corresponding feature VF1, that is, between the audio corresponding feature WF and the first vibration corresponding feature VF1 Therefore, the corresponding audio reference material REF is generated, that is, the weight relationship between the audio corresponding feature WF and the first vibration corresponding feature VF1 is used as the audio reference material REF.

The above is the wireless vibration audio conversion method with intelligent learning, that is, the artificial intelligence application is executed through the computing device, and the corresponding weight relationship between the audio corresponding feature and the first vibration corresponding feature is learned by the artificial intelligence algorithm. The artificial intelligence algorithm converts the vibration change result into the weight reference of the audio output data. The following embodiment is a wireless vibration audio conversion method, that is, the learned audio reference data is referred to by an artificial intelligence algorithm to convert the received vibration change result into the corresponding audio output signal. The following is a detailed description: please refer to the third figure , Which is a flowchart of another embodiment of the present invention. As shown in the figure, the steps of the wireless vibration audio conversion method of the present invention include: Step S40: Use the vibration sensor to sense the throat during a second sensing period and generate a second vibration change result; Step S42 : Wirelessly transmit the second vibration change result to the computing device; step S45: the computing device executes the audio and vibration signal conversion program and converts the vibration change result to the corresponding feature; and Step S50: The computing device executes the application program and converts the second vibration change result into an audio output signal with a reference audio sound field effect according to the audio reference data pre-stored in a storage unit.

In step S40, as shown in FIG. 4A, the vibration sensor 12 of the radio device 10 senses the vibration V2 from the throat T during the second sensing period Pd2, thereby obtaining a second vibration change result S _{V2 , and in step S42, as shown in Figure 4A, the second vibration change result S V2 is} transmitted to the wireless transmission interface constructed by the first wireless transmission unit 16 and the second wireless transmission unit 26 The computing device 20, further, the processing unit 22 stores the second vibration change result S _V2 received by the computing device 20 in the storage unit 24.

In step S45, as shown in FIG. 4B, the processing unit 22 reads and executes the application program P stored in the storage unit 24, and also reads the second vibration change result S _V2 to use the application program The second vibration change result S _V2 is calculated in P, where the artificial intelligence algorithm AI executed by the processing unit 22 is to read the second vibration change result S _V2 converted by the Fourier transform module to Convert the second vibration change result S _V2 into a corresponding feature, that is, a second vibration corresponding feature VF2. The second vibration corresponding feature VF2 of this embodiment is a log-power spectrum (Log-Power Spectrum, LPS), except for this In addition, the second vibration corresponding feature VF2 can be a conversion result of a Mel-Frequency Cepstrum (MFC) or a linear predictive analysis spectrum (LPC Spectrum) signal processing. Then in step S50, as shown in FIG. 4B, the processing unit 22 according to the artificial intelligence algorithm AI and the audio reference data REF pre-stored in the storage unit RAM corresponding to the processing unit 22, for example: stored in In the memory, the second vibration change result S _{V2 is} converted into an audio mapping signal WI, and the audio mapping signal WI is converted into an audio output signal WO of an output format through an inverse Fourier transform module IFT for Subsequent output to an output device 30, such as speakers or earphones. As shown in Figure 4C, the audio output signal WO in an output format is output from the computing device 20 to the output unit 30, thus outputting an output signal OUT close to human voice. .

Thereby, the audio output signal WO provided by the present invention corresponds to the audio change result S _W captured in step S10, that is to say, the present invention uses the arithmetic device to convert the first vibration obtained in step S10 The change result S _V1 and the audio change result S _W are calculated to obtain the audio reference data for the computing device 20 as a reference basis, so that the second vibration change result S _V2 obtained by the connection is converted into the audio output The signal WO, and the audio output signal WO is an output signal OUT close to the human voice, so it can provide an audio signal that is less susceptible to distortion in the application of converting the throat vibration signal into an audio signal.

In summary, the wireless vibration audio conversion system and method of the present invention provide a computing device for computing the first vibration change result and the audio change result sensed by the radio device during the first sensing period, thereby generating the corresponding The audio reference data allows the computing device to learn, so that the second vibration change result sensed during the second sensing period can be converted into an audio output signal corresponding to the audio change result, thereby providing an output signal close to human voice.

Therefore, the present invention is truly novel, progressive, and available for industrial use. It should meet the patent application requirements of China's patent law. Undoubtedly, I filed an invention patent application in accordance with the law. I pray that the Bureau will grant the patent as soon as possible.

However, the above are only the preferred embodiments of the present invention, and are not used to limit the scope of implementation of the present invention. For example, the shapes, structures, features and spirits described in the scope of the patent application of the present invention are equally changed and modified. , Should be included in the scope of patent application of the present invention.

S40-S50: steps

Claims (17)

A wireless vibration audio conversion method with intelligent learning, the steps include: using a vibration sensor of a radio device to sense a throat during a first sensing period and generate a vibration change result, and using the radio device An audio sensor senses a mouth during the sensing period and generates an audio change result; wirelessly transmits the vibration change result and the audio change result to an arithmetic device; the arithmetic device executes an audio and vibration signal conversion program and The vibration change result and the audio change result are converted to two corresponding characteristics; and the computing device executes an artificial intelligence program to pair the audio change result of the two corresponding characteristics with the first vibration change result to generate a corresponding audio reference material. The wireless vibration audio conversion method according to claim 1, wherein the artificial intelligence program includes an artificial intelligence algorithm, and the artificial intelligence algorithm is a Deep Neural Networks (DNN). The wireless vibration audio conversion method according to claim 1, wherein in the step of executing the audio and vibration signal conversion program by the computing device and converting the vibration change result and the audio change result to two corresponding features, the computing device will The audio change result is converted into an audio corresponding feature corresponding to the corresponding feature and the vibration change result is converted into a corresponding vibration corresponding feature. The audio corresponding feature and the vibration corresponding feature are logarithmic power spectrum and Mel cepstrum (Mel cepstrum). -Frequency Cepstrum (MFC) or linear predictive analysis spectrum (LPC Spectrum) signal processing conversion results. The wireless vibration audio conversion method according to claim 1, wherein the vibration sensor is an acceleration sensor or a piezoelectric sensor. A wireless vibration audio conversion system with intelligent learning, which includes: a radio device, which includes: A vibration sensor that senses a vibration change result of a throat during a sensing period; an audio sensor that senses an audio change result of a mouth during the sensing period; and a first A wireless transmission unit, connected to the vibration sensor and the audio sensor; an arithmetic device, including: a second wireless transmission unit, wirelessly connected to the first wireless transmission unit; a processing unit, electrically connected to the first wireless transmission unit A wireless transmission unit; and a storage unit storing an artificial intelligence application program and an audio vibration conversion program, the processing unit receives the vibration change result and the audio change result via the first wireless transmission unit and the second wireless transmission unit , And execute the audio vibration conversion program to convert the vibration change result and the audio change result to two corresponding characteristics, so as to generate an audio reference data according to the first vibration change result and the audio change result of the two corresponding characteristics. The wireless vibration audio conversion system according to claim 5, wherein the application includes an artificial intelligence algorithm, and the artificial intelligence algorithm is a Deep Neural Networks (DNN). The wireless vibration audio conversion system according to claim 5, wherein the processing unit converts the result of the audio change into a corresponding audio feature and converts the result of the vibration change into a corresponding vibration feature, and the audio corresponding feature is The corresponding characteristic of the vibration is the conversion result of the signal processing of the logarithmic power spectrum, Mel-Frequency Cepstrum (MFC) or linear prediction analysis spectrum (LPC Spectrum). The wireless vibration audio conversion system according to claim 5, wherein the vibration sensor is an acceleration sensor or a piezoelectric sensor. A wireless vibration audio conversion method, the steps include: Use a vibration sensor to directly sense a throat during a sensing period and generate a vibration change result; wirelessly transmit the vibration change result to an arithmetic device; the arithmetic device executes an audio vibration conversion program to convert the vibration change result to A corresponding feature; the computing device executes an artificial intelligence application program and converts the vibration change result of the corresponding feature into an audio mapping signal with a reference audio sound field effect feature according to an audio reference data pre-stored in a storage unit And the computing device executes the audio vibration conversion program to reversely convert the corresponding feature of the audio mapping signal into an audio output signal in an output format. The wireless vibration audio conversion method according to claim 9, wherein the application program includes an artificial intelligence algorithm and an audio vibration conversion program, and the artificial intelligence algorithm is a Deep Neural Networks (DNN). The wireless vibration audio conversion method according to claim 9, wherein in the step of converting the vibration change result to a corresponding feature when the arithmetic device executes an audio vibration conversion program, the processing unit converts the vibration change result into a corresponding one Vibration corresponding feature, the vibration corresponding feature is a conversion result of a logarithmic power spectrum, a Mel-Frequency Cepstrum (MFC) or a linear predictive analysis spectrum (LPC Spectrum) signal processing. The wireless vibration audio conversion method according to claim 9, wherein the vibration sensor is an acceleration sensor or a piezoelectric sensor. A wireless vibration audio conversion system, which includes: a radio device, which includes: a vibration sensor that senses a vibration change result of a throat during a sensing period; and a first wireless transmission unit connected to The vibration sensor; an arithmetic device, which includes: A second wireless transmission unit, wirelessly connected to the first wireless transmission unit; a processing unit, electrically connected to the second wireless transmission unit; and a storage unit, storing an artificial intelligence application program and an audio vibration conversion program, the processing The unit receives the vibration change result via the first wireless transmission unit and the second wireless transmission unit and executes the audio vibration conversion program to convert the vibration change result to a corresponding feature. The processing unit executes the artificial intelligence application according to pre-stored An audio reference data in a storage unit converts the vibration change result of the corresponding feature into an audio mapping signal having a reference audio sound field effect feature, and the processing unit executes the audio vibration conversion program to convert the audio of the corresponding feature The mapping signal is converted into an audio output signal in an output format. The wireless vibration audio conversion system according to claim 13, further comprising an output device connected to the computing device, receiving and outputting an audio signal according to the audio output signal in the outputable format. The wireless vibration audio conversion system according to claim 13, wherein the application program includes an artificial intelligence algorithm and an audio vibration conversion program, and the artificial intelligence algorithm is a Deep Neural Networks (DNN). The wireless vibration audio conversion system according to claim 13, wherein the processing unit converts the vibration change result into a corresponding vibration characteristic, and the vibration corresponding characteristic is a logarithmic power spectrum and a Mel-Frequency frequency spectrum (Mel-Frequency Cepstrum, MFC) or a linear predictive analysis spectrum (LPC Spectrum) signal processing conversion results. The wireless vibration audio conversion system according to claim 13, wherein the vibration sensor is an acceleration sensor or a piezoelectric sensor.

Images