A system and method for active noise cancellation in an open field

Abstract

An apparatus for actively canceling a target sound wavefront in an open space, the apparatus comprising a signal processing module comprising at least one processor operatively coupled with a data store, the at least one processor configured to: receive data comprising one or more geographic features, and one or more audio features generated by one or more receiving microphones in geographic relation to an array of receiving microphones in an area proximate to a user; process the data using a predictive model of a deep learning framework adapted to training; and provide an inverse sound wavefront output of a target sound at an area of the predictive microphone.

Description

Field of the invention

[0001] The present invention relates generally to a device for actively canceling a target acoustic wavefront in an open space. BACKGROUND

[0002] Sound is a pressure wave, which consists of alternating periods of compression and rarefaction. A noise-canceling speaker emits sound waves with the same amplitude but with opposite phase (also known as anti-phase) to the original sound. These waves combine in a process known as interference to form new waves and effectively cancel each other out - an effect known as destructive interference. Active noise cancellation (ANC), also known as noise control or active noise reduction (ANR), is a method of reducing unwanted sound by adding a second sound specifically designed to cancel the first.

[0003] ANC is typically implemented using analog circuitry or digital signal processing. Adaptive algorithms aim to analyze the waveform of the background audible or non-audible noise and then generate a signal that will either phase shift or invert the polarity of the original signal according to a specific algorithm. This inverse signal (anti-phase) is then amplified, and a transducer produces sound waves proportional to the amplitude of the original waveform, resulting in destructive interference. This effectively reduces the amount of perceivable noise.

[0004] A noise-canceling speaker can be located in the same place as the sound source for attenuation. Alternatively, the transducer emitting the cancellation signal can be located in the place where sound attenuation is desired (e.g., the user's ear). This requires much lower power levels to cancel, but is only effective for a single user. Attenuating noise in other locations is more difficult because the three-dimensional wavefront of the unwanted sound and the cancellation signal can match and create alternating regions of constructive and destructive interference, thereby reducing noise at some points and doubling it at others. SUMMARY

[0005] In one aspect, a device for actively canceling a target acoustic wavefront in an open space is provided, the device comprising: a signal processing module comprising at least one processor operatively coupled with a data store, the at least one processor configured to: receive data comprising one or more geographical features, and one or more audio features generated by one or more receiving microphones having a geographical relationship with a receiving microphone array in an area proximate to a user; process the data using a predictive model of an adaptive trained deep learning framework; and provide an inverse acoustic wavefront output of the target sound at an area of the receiving microphone.

[0006] In another aspect, a system of the present invention device is provided for actively canceling a target acoustic wavefront in an open space.

[0007] In yet another aspect, a method for actively cancelling a target sound wave front in an open space using the apparatus / system disclosed herein is provided.

[0008] INCORPORATION BY REFERENCE

[0009] All publications, patents, and patent applications mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent, or patent application was specifically and individually indicated to be incorporated by reference. BRIEF DESCRIPTION OF DRAWINGS

[0010] The novel features of the application are set forth with particularity in the appended claims. A better understanding of the features and advantages of the present application will be obtained by reference to the following detailed description that sets forth illustrative embodiments, in which the principles of the application are utilized, and the accompanying drawings of which:

[0011] Figure 1 A trained model is shown with respect to how data is collected.

[0012] Figure 2A An exemplary diagram is shown with respect to how exemplary receiving microphones 101 and 102 are used in the process to predict what the prediction microphone P-1 will receive.

[0013] Figure 2B A flowchart is provided showing the process 200 in the training mode that is processed by the GAN to predict artificial sound based on real sound.

[0014] Figure 2C A flowchart is provided showing the process 300 in the inference mode that is processed by the GAN.

[0015] Figure 2D A flowchart is provided showing the process 400 in the inference mode that is processed by the GAN.

[0016] Figure 2E An exemplary diagram is shown corresponding to Figure 2D that defines the distance and angle from an exemplary location (e.g., location 3) to receiving microphones 101 and 102.

[0017] Figure 3A An exemplary diagram is shown with respect to how exemplary receiving microphones 101 and 102 (shown externally to the speaker 10) are used to predict and provide the inverse sound (i.e., wave front) of a pre-identified sound N102, for example, at location 3.

[0018] Figure 3B A flowchart is provided showing the process 500 in the inference mode that is processed by the GAN corresponding to Figure 3A . DETAILED DESCRIPTION

[0019] Reference in the specification to "certain embodiments" or similar language means that a particular feature, structure, or characteristic described is included in at least one embodiment of the present application. Thus, appearances of the phrase "in certain embodiments" in various places in the specification are not necessarily referring to the same embodiment.

[0020] In small enclosed spaces (e.g., a passenger compartment of a car), overall noise reduction can be achieved by multiple loudspeakers and feedback microphones and measurements of the modal response of the enclosed space. Typically, as previously disclosed, known open field ANC systems include multiple directional microphones and loudspeakers forming an array to produce a noise cancellation wavefront that actively cancels ambient sound wavefronts. This design has its limitations; for example, since the loudspeakers are fixed in the open field, the cancellation of the wavefront is limited to a specific area; however, the user can be moving and not limited to the above-mentioned area. Also, this design is only suitable for low frequency sounds, such as mechanical noise, and not for highway or airplane environments where other frequencies cannot be cancelled.

[0021] Thus, there is a need for ANC devices / systems for use in non-fixed transferable open fields.

[0022] In some embodiments, there is provided an apparatus for actively cancelling a target sound wave front in an open space, the apparatus comprising: a signal processing module comprising at least one processor operatively coupled with a data store, the at least one processor configured to: receive data comprising one or more geographical features, and one or more audio features generated by one or more receiving microphones having a geographical relationship with a receiving microphone array in a region proximate to a user; process the data using a predictive model of an adaptive trained deep learning framework; and provide an inverse sound wave front output of the target sound at a region of the predictive microphone. In some embodiments, the apparatus is a speaker. In some embodiments, the one or more receiving microphones are located on opposite sides of the user. In some embodiments, the deep learning framework is a generative adversarial network or a conditional generative adversarial network. In certain embodiments, the deep learning framework is a conditional generative adversarial network. In some embodiments, the predictive microphone array has between 1 and n microphones. The region of the predictive microphone array is within 30 cm, 25 cm, 20 cm, 15 cm, 10 cm or 5 cm from the user (e.g., from the user’s ear). In some embodiments, the region of the predictive microphone array is between 1 cm and 50 cm, 1 cm and 40 cm, 1 cm and 30 cm, 1 cm and 25 cm, 1 cm and 20 cm or 1 cm and 10 cm from the user (e.g., from the user’s ear). In certain embodiments, the predictive microphone array is between 5 cm and 10 cm from the user. In some embodiments, the apparatus further comprises a monitoring device to monitor the movement of the user. In certain embodiments, the monitoring device is a camera. In certain embodiments, the monitoring device provides geographical location feedback of the user’s movement to the apparatus, thereby allowing the apparatus to automatically generate a noise cancelling wave front. In certain embodiments, the geographical location feedback comprises data of the geographical features and the audio features. In certain embodiments, the geographical features comprise distance and angle from the receiving microphones to the selected location of the predictive microphone.

[0023] In some embodiments, the target sound wave front is ambient noise of the open space. In some embodiments, the target sound wave front is pre-identified through a database or through a pre-recorded device. In some embodiments, the pre-identified target sound wave front is isolated from all sounds received by the receiving microphones. In certain embodiments, the apparatus generates a noise cancelling wave front at a region selected by the user.

[0024] In some embodiments, a system is provided that includes the device disclosed herein and optionally an array of prediction microphones to provide accuracy feedback after training a deep learning framework. In certain embodiments, the signal processing module uses a prediction model to provide a pattern of sound in each location of the prediction microphones. In some embodiments, the deep learning framework is a conditional generative adversarial network. In some embodiments, the array of prediction microphones has 1 to n microphones. In some embodiments, the area of the array of prediction microphones is within 30 cm, 25 cm, 20 cm, 15 cm, 10 cm, or 5 cm from the user (e.g., from the user’s ear). In some embodiments, the area of the array of prediction microphones is between 1 cm to 50 cm, 1 cm to 40 cm, 1 cm to 30 cm, 1 cm to 25 cm, 1 cm to 20 cm, or 1 cm to 10 cm from the user (e.g., from the user’s ear). In certain embodiments, the array of prediction microphones is between 5 cm to 10 cm from the user. In some embodiments, the device further includes a monitoring device to monitor the movement of the user. In certain embodiments, the monitoring device is a camera. In certain embodiments, the monitoring device provides geographical location feedback of the user’s movement to the device, allowing the loudspeaker to automatically generate noise-cancelling wavefronts. In certain embodiments, the geographical location feedback includes data of geographical features and audio features. In certain embodiments, the geographical features include distance and angle from the receiving microphone to the selected location of the prediction microphones.

[0025] In some embodiments, the receiving microphone is located at the same location as the one or more loudspeakers. In some embodiments, the one or more loudspeakers include the signal processing module.

[0026] In some embodiments, the signal processing module includes at least one processor operably coupled with a data store, the at least one processor configured to: receive data including one or more geographical features and one or more audio features; process the data using a prediction model; and provide an output pattern of sound signals at each location of the prediction microphones.

[0027] In certain embodiments, the prediction model is adapted to a trained deep learning framework, etc. In certain embodiments, the deep learning framework is a generative adversarial network (GAN) or a conditional generative adversarial network (GAN).

[0028] A generative adversarial network (GAN) is a class of machine learning frameworks in which two neural networks compete with each other. Given a training set, the technique will learn to generate new data with the same statistics as the training set. For example, a GAN trained on photographs can generate new photographs that, to a casual observer, appear authentic, with many realistic features. Although originally proposed as a form of generative model for unsupervised learning, GANs have also proven useful for semi-supervised learning, fully supervised learning, and reinforcement learning. The core idea of GANs is based on "indirect" training through a discriminator, which itself is also dynamically updated. Essentially, this means that rather than training a generator (e.g., a model for creating new data based on original data) to minimize distance to a particular image, a discriminator (for identifying data patterns and determining whether input data is original data or fake data generated from a generator) is fooled. This enables the model to learn in an unsupervised manner. The generative network generates candidates, while the discriminative network evaluates the candidates. The competition proceeds according to the data distribution. Typically, the generative network learns to map from a latent space to a data distribution of interest, while the discriminative network distinguishes candidates produced by the generator from the true data distribution. The training objective for the generative network is to improve the error rate of the discriminative network (i.e., to "fool" the discriminator network by producing new candidates that the discriminator believes are not synthetic (part of the true data distribution)). A known dataset is used as initial training data for the discriminator. Training it involves presenting it with samples from the training dataset until acceptable accuracy is reached. The generator is trained according to whether it successfully fooled the discriminator. Typically, the generator uses random inputs sampled from a predefined latent space (e.g., a multivariate normal distribution) as seeds. Thereafter, the candidates synthesized by the generator are evaluated by the discriminator.

[0029] Generative adversarial networks (GANs) were recently introduced as a new approach to training generative models. Conditional versions of generative adversarial networks are constructed by simply feeding the data y with the condition into the generator and discriminator. This model can generate MNIST digits conditioned on class labels and can be used to learn a multi-modal model that generates descriptive labels that are not part of the training labels.

[0030] In some embodiments, the device / system for actively canceling a target acoustic wavefront (i.e., a sound of interest) in an open space utilizes a deep learning framework to implement APN in non-stationary open field. For example, a conditional GAN can be used, in which two neural networks compete with each other and learn from each other from a particular sound source. Given a training set, the technique will learn to generate new data with the same statistics as the training set.

[0031] Figure 1How data is collected during training mode is shown. Noise N1 is generated and collected at various locations by n number of prediction microphones (e.g., P-1, P-2, P-3,... to P-n), where each microphone produces a signal that is processed by the signal processing module using a deep learning framework (e.g., conditional GAN) to learn the sound patterns at different locations.

[0032] In some embodiments, the array of prediction microphones has 1 to n number of prediction microphones. In certain embodiments, the area of the array of prediction microphones is within 30 cm, 25 cm, 20 cm, 15 cm, 10 cm, or 5 cm from the user (i.e., from each ear of the user). In certain embodiments, the area of the array of prediction microphones is between 1 cm to 50 cm, 1 cm to 40 cm, 1 cm to 30 cm, 1 cm to 25 cm, 1 cm to 20 cm, or 1 cm to 10 cm from the user. In certain embodiments, the array of prediction microphones is between 5 cm to 10 cm from the user.

[0033] Once the deep learning framework is trained (shown in Figure 2B ), the prediction model can predict the sound received by the prediction microphones (e.g., microphones P-1, P-2, P-3, P-4, … , etc.) at the set locations based on the sound / noise received by the receiving microphones (e.g., microphones 101 and 102 located in speaker 10). See Figure 2A . In some embodiments, the receiving microphones are located on opposite sides of the user. In some embodiments, the system includes one or more speakers. In certain embodiments, the system includes two, three, four, or more speakers. In certain embodiments, the system includes two speakers located on opposite sides of the user.

[0034] In the case where the prediction sound is at P-1 location, its inverse wave is produced and added by the speaker (e.g., speaker 10 with receiving microphones 101 and 102) that includes the signal processing module to offset the noise N1 to achieve ANC. The receiving microphones are configured to receive the sound signals produced by the array of n number of prediction microphones in the area, where the prediction microphones have a geographical relationship (e.g., geographical feature) with the receiving microphones, as shown in Figure 2A .

[0035] Figure 2Bis a flowchart showing the process 200 in a training mode by a GAN to predict artificial sound based on real sound. First, data including geographical features 201 and audio features 202 are input to a generator network 204. At step 206, a discriminator network 205 compares the output from the generator network with an input 203 of real audio to predict a label (e.g., true or false). Finally, at step 207, the output parameters of the predicted label are updated in the generator and discriminator. In some implementations, the geographical features include a distance between a prediction microphone (e.g., microphone P-1) and a receiving microphone (e.g., microphone 101).

[0036] Figure 2C is a flowchart showing the process 300 in an inference mode by a GAN. For example, after given geographical features 301 (e.g., distance and angle (e.g., from microphones P-1, P-2, etc. to receiving microphone 101)) and audio features 302 (i.e., audio data) to a generator network 303, an output 304 is generated.

[0037] In the particular example shown in process 400, the particular data 401 includes a distance and angle to a selected location (e.g., location 3), and the data set includes prediction microphone (e.g., P-1, P-2) audio data and, for example, a distance at location 3, which particular data 401 is fed to a generator network 403 at step 404 to prepare inverse surround sound at location 3, as shown. Figure 2D Figure 2E An example geographical feature of process 400 (corresponding to Figure 2D ) is shown, in particular, the distance and angle from the selected location 3 to receiving microphones 101 and 102 located in speakers 10.

[0038] Although a prediction microphone is used in the training mode, it is not necessary to remove the prediction microphone after the training mode. In some implementations, after training a deep learning framework, a prediction microphone is used in the system of the present invention to provide accuracy feedback. For example, after training a GAN, if the prediction microphone remains at the selected location 3, the microphone can provide data received at location 3 to a prediction model in the system to adjust and prepare a more accurate inverse sound wave front for ANC.

[0039] The device / system also includes a signal processing module configured to receive sound from the array of prediction microphones to learn sound patterns in each location of the prediction microphones in the area and transmit control signals to one or more speakers. (including a signal processing module) configured to produce a noise cancellation wave front, wherein the noise cancellation wave front is equal in size and opposite in polarity to a target sound wave.​

[0040] In the same way, noise cancellation can be applied to other predictive microphones (e.g., microphones P-1 to P-n shown) with fixed positions (providing geographical relationship) as the exemplary microphones 101 / 102. This will effectively allow the user to move around the area where the predictive microphones are located. In some embodiments, in a practical way, the position close to the user’s ear will be the most useful and effective for cancelling noise, as the further away from the ear, more variables (e.g., echo, ambient sound, etc.) will interfere with the training data and noise cancellation effect. In some embodiments, using directional microphones has better effect due to the stronger directionality of the sound.

[0041] Enhanced noise reduction of specific sound

[0042] Similarly, the system and method of ANC in an open field without fixation is applicable to specific sound cancellation, e.g., to specific environmental noise. Such application can be based on the teachings of WO2019228329A1.

[0043] In contrast to the general noise reduction as shown in Figure 2A In the present application, specific sound / noise of interest can be determined based on sound patterns.

[0044] As shown in Figure 3A The “general noise” includes, for example, N101, N102, N103, etc. The sound N102 (“noise N102”) is pre-identified (known) by known sources in the database, for example, or can be pre-identified by pre-recorded and processed means (e.g., by previous recording devices before applying ANC) in addition to other unknown or non-targeted sounds (e.g., N101, N103 and other sounds). The known sound is processed by the same or similar way as disclosed in WO2019228329A1.

[0045] In the case of applying the pre-identified sound N102 (received by one or more receiving microphones (e.g., 101 and 102) at a fixed position (e.g., at position 3)), the device / system first isolates the sound N102 from all known and unknown collected sounds, and then the signal processing module generates a signal for the predictive model to process and predict. The loudspeaker 10 then generates and adds the inverse phase N102 of the sound (inverse wavefront of N102) based on the prediction to implement ANC of, for example, any selected area of position 3 based on the known sound N102. In the same way, noise cancellation can be applied to other areas selected by the user (e.g., areas 1 to 6 shown or any area / position around the user).

[0046] Figure 3BFurther illustrating an example geographic feature of the process 500, where the specific data 501 includes distance and angle to a selected location (e.g., region 3), the data set 502 including predicted microphone audio data (e.g., microphones P-1, P-2, etc.) and certain data 503 of distance from, for example, the selected region 3 and pre-identified features are fed to a generator network 504 for processing to generate at step 505 an inverse pre-identified sound (i.e., sound N102) in location 3 to cancel the N102 sound.

[0047] Alternatively, a monitoring system (e.g., camera, video recording device, etc.) can be used to feed the location feedback (e.g., distance and angle between the predictive microphone and the user; distance and angle from the receiving microphone to the predictive microphone) to a predictive model used in the device / system about the movement of the user, thereby automatically adjusting the noise-cancellation region.

[0048] The predictive model is used to predict and / or generate a processed sound wavefront at different locations. In some embodiments, the effective range of the ANC is within 30 cm, 25 cm, 20 cm, 15 cm, 10 cm, or 5 cm from the user. In some embodiments, the effective range is between 1 cm to 50 cm, 1 cm to 40 cm, 1 cm to 30 cm, 1 cm to 25 cm, 1 cm to 20 cm, or 1 cm to 10 cm from the user. In certain embodiments, the effective range is between 5 cm to 10 cm.

[0049] In some embodiments, the target sound wavefront is an ambient noise of an open space. In other embodiments, the target sound wavefront is pre-identified through a database or through a pre-recording device. In some embodiments, the pre-identified target sound wavefront is isolated from all sounds received by the receiving microphone. In certain embodiments, the speaker generates a noise-cancellation wavefront at a region selected by the user.

[0050] In some embodiments, the device further comprises a monitoring device to monitor the movement of the user. In certain embodiments, the monitoring device provides location feedback of the user movement to the device, thereby allowing the speaker to automatically generate a noise-cancellation wavefront.

[0051] It should be understood that the pre-identified sound for processing can be inputted through a wireless communication device between the device / system of the present invention and an external sound processing device, such as Bluetooth, infrared, or Wi-Fi. In some embodiments, the communication between the device / system of the present invention and the external sound processing device is not limited to direct point-to-point communication. In some embodiments, it can also be through a local area network, a mobile phone network, or the Internet.

[0052] Those of ordinary skill in the art will readily recognize that the present application can be implemented as a device / system including a computer system / device, method, or computer readable medium. Thus, the present application can be implemented in various forms, for example, an entirely hardware implementation, an implementation including a complete software embodiment (including firmware, resident software, micro-code, etc.), or an implementation combining software and hardware. The following will refer to these collectively as "circuitry," "modules," or "systems." Additionally, the present application can also be implemented as a computer program product on any tangible medium having stored thereon computer usable program code.

[0053] Any combination of one or more computer usable or computer readable medium(s) can be utilized. For example, a computer usable or computer readable medium can be, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, device, or propagation medium. More specific examples (a non-exhaustive list) of the computer readable medium include the following: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a transmission medium such as the Internet or Intranet, or magnetic storage device. Note that the computer readable medium can even be paper or any other suitable medium upon which the program is printed, as the program can be electronically captured, for example via optical scanning of the paper or other medium, then compiled, interpreted, or otherwise processed in a suitable manner, if necessary, and then stored in a computer memory. As used herein, a computer usable or computer readable medium is any medium that can store, temporarily or permanently, the program for use by or in connection with the instruction execution system, apparatus, or device. The computer usable medium can include any medium that can store data signals in the form of instructions, such as baseband or carrier wave signals. The computer can use any suitable medium, including but not limited to wireless, wireline, optical, fiber cable, and / or RF, to transmit the program code.

[0054] The description of the application can include flow diagrams and / or block diagrams that illustrate the systems, devices, methods and computer program products according to the particular embodiments of the application. It should be understood that each block of the flow diagrams and / or block diagrams and any combination thereof can be implemented by computer program instructions. These computer program instructions can be executed by a processor of a general purpose computer or special purpose computer or by a machine comprising other programmable data processing equipment. These computer program instructions can also be stored on a computer readable medium to instruct a computer or other programmable data processing equipment to perform specific functions, including instructions for implementing the functions or operations described in the flow diagrams and / or block diagrams. Computer program instructions can also be loaded onto a computer or other programmable data processing equipment to facilitate system operation steps on the computer or other programmable equipment and execute instructions on the computer or other programmable equipment. Sometimes programs executed by computers are generated to implement the functions or operations shown in the flow diagrams and / or block diagrams.

[0055] In some embodiments, a method for actively canceling a target sound wavefront in an open space is provided, comprising: receiving a target sound wavefront; and performing target sound cancellation using a predictive model trained using two or more receiving microphones configured to receive sound signals produced by a predictive microphone array in a user area to receive a target wavefront, wherein the predictive microphones have a geographical relationship with the receiving microphones, generating a noise cancellation wavefront that is the same size and opposite polarity as the target sound by a signal processing module configured to receive sound from the predictive microphone array in each location of the predictive microphones in the area to learn the sound pattern, and transmitting control signals to one or more loudspeakers configured to produce the noise cancellation wavefront. In some embodiments, the predictive model is adapted to a deep learning framework. In some embodiments, the deep learning framework is a generative adversarial network or a conditional generative adversarial network. In some embodiments, the method further comprises monitoring the user’s movement by a monitoring device. In certain embodiments, the monitoring device provides geographical location feedback of the user’s movement to the signal processing module, allowing the signal processing module to automatically produce the noise cancellation wavefront. In certain embodiments, the geographical location feedback comprises data of geographical features and audio features. In some embodiments, the predictive microphone array has 1 to n microphones. In some embodiments, the area of the predictive microphone array is within 30 cm, 25 cm, 20 cm, 15 cm, 10 cm, or 5 cm from the user (e.g., from the user’s ear). In some embodiments, the area of the predictive microphone array is between 1 cm to 50 cm, 1 cm to 40 cm, 1 cm to 30 cm, 1 cm to 25 cm, 1 cm to 20 cm, or 1 cm to 10 cm from the user (e.g., from the user’s ear). In certain embodiments, the predictive microphone array is between 5 cm to 10 cm from the user. In some embodiments, the device further comprises a monitoring device to monitor the user’s movement. In certain embodiments, the monitoring device is a camera. In certain embodiments, the monitoring device provides geographical location feedback of the user’s movement to the device, allowing the device to automatically produce the noise cancellation wavefront. In certain embodiments, the geographical location feedback comprises data of geographical features and audio features. In certain embodiments, the geographical features comprise distance and angle from the receiving microphones to the selected locations of the predictive microphones.

[0056] In some embodiments, the target acoustic wavefront is ambient noise of an open space. In some embodiments, the target acoustic wavefront is pre-identified through a database or through a pre-recorded device. In some embodiments, the pre-identified target acoustic wavefront is isolated from all sounds received by the receiving microphone. In certain embodiments, the device generates a noise-cancelling wavefront at a selected area of the user.

[0057] While preferred embodiments of the application have been shown and described herein, it is to be understood that the application is not limited to these embodiments. Since many changes and modifications of the application can be made, it is intended that only such changes and modifications that are claimed by the applicant be protected.

Claims

1. An apparatus for actively cancelling a target acoustic wavefront in an open space, the apparatus comprising: a signal processing module comprising at least one processor operatively coupled with a data store, the at least one processor configured to: receive data comprising one or more geographical features, and one or more audio features generated by one or more receiving microphones having a geographical relationship with a predictive microphone array in an area proximate to a user; process the data using a predictive model of an adaptively trained deep learning framework; and provide an inverse acoustic wavefront output of the target acoustic wavefront at an area of the predictive microphone, wherein the target acoustic wavefront is isolated from all sound received by the receiving microphones and is pre-identified through a database or through a pre-recorded device.

2. The apparatus of claim 1, wherein the apparatus is a speaker.

3. The apparatus of claim 1, wherein the one or more receiving microphones are located on opposite sides of the user.

4. The apparatus of claim 1, wherein the deep learning framework is a generative adversarial network or a conditional generative adversarial network.

5. The apparatus of claim 1, wherein the predictive microphone array has between 1 and n microphones.

6. The apparatus of claim 5, wherein the area of the predictive microphone array is located within 30 cm, 25 cm, 20 cm, 15 cm, 10 cm or 5 cm from the user.

7. The apparatus of claim 5, wherein the area of the predictive microphone array is located between 1 cm and 50 cm, 1 cm and 40 cm, 1 cm and 30 cm, 1 cm and 25 cm, 1 cm and 20 cm or 1 cm and 10 cm from the user.

8. The apparatus of claim 7, wherein the predictive microphone array is located between 5 cm and 10 cm from the user.

9. The apparatus of claim 1, wherein the target acoustic wavefront is ambient noise of the open space.

10. The apparatus of claim 1, wherein the apparatus produces a noise cancelling wavefront at an area selected by the user.

11. The apparatus of claim 1, wherein the apparatus further comprises a monitoring device to monitor movement of the user.

12. The apparatus of claim 11, wherein the monitoring device provides geographical location feedback of user movement to the apparatus, thereby allowing the apparatus to automatically produce a noise cancelling wavefront.

13. The apparatus of claim 12, wherein the geographical location feedback comprises data of one or more geographical features and audio features.

14. The apparatus of claim 1, wherein the one or more geographical features comprise a distance and an angle from the receiving microphones to a selected location of the predictive microphone.

15. A system for actively cancelling a target acoustic wavefront in an open space, comprising an apparatus corresponding to claim 1, comprising a predictive microphone array to provide accuracy feedback after training the deep learning framework.

16. The system of claim 15, wherein the signal processing module uses a predictive model to provide a pattern of the sound in each location of the predictive microphone.

17. The system of claim 15, wherein the deep learning framework is a conditional generative adversarial network.

18. The system of claim 15, wherein the array of predictive microphones has a quantity of predictive microphones from 1 to n.

19. The system of claim 18, wherein the area of the array of predictive microphones is within 30 cm, 25 cm, 20 cm, 15 cm, 10 cm or 5 cm from the user.

20. The system of claim 18, wherein the area of the array of predictive microphones is between 1 cm to 50 cm, 1 cm to 40 cm, 1 cm to 30 cm, 1 cm to 25 cm, 1 cm to 20 cm or 1 cm to 10 cm from the user.

21. The system of claim 15, wherein the target sound wave front is an ambient noise of the open space.

22. The system of claim 15, wherein the device produces a noise cancellation wave front at a selected area of the user.

23. The system of claim 22, wherein the system further comprises a monitoring device to monitor the movement of the user.

24. The system of claim 23, wherein the monitoring device provides a geographical location feedback of the user movement to the device, allowing the loudspeakers to automatically produce a noise cancellation wave front.

25. A method for actively cancelling a target sound wave front in an open space, the method comprising: receiving the target sound wave front; performing target sound cancellation using a predictive model trained using two or more receiving microphones configured to receive a sound signal produced by an array of predictive microphones in a user area to receive the target sound wave front, wherein the predictive microphones have a geographical relationship with the receiving microphones, and generating a noise cancellation wave front of the same magnitude and opposite polarity of the target sound by a signal processing module configured to receive sound from the array of predictive microphones in each location of the predictive microphones in the area to learn a pattern of sound, and transmit a control signal to one or more loudspeakers configured to produce a noise cancellation wave front, wherein the target sound wave front is isolated from all sounds received by the receiving microphones and is pre-identified through a database or through a pre-recorded device.

26. The method of claim 25, wherein the predictive model is adapted to a deep learning framework.

27. The method of claim 26, wherein the deep learning framework is a generative adversarial network or a conditional generative adversarial network.

28. The method of claim 25, wherein the array of predictive microphones has a quantity of microphones from 1 to n.

29. The method of claim 25, wherein the area of the predictive microphone array is located within 30 cm, 25 cm, 20 cm, 15 cm, 10 cm, or 5 cm from the user.

30. The method of claim 25, wherein the area of the predictive microphone array is located between 1 cm and 50 cm, 1 cm and 40 cm, 1 cm and 30 cm, 1 cm and 25 cm, 1 cm and 20 cm, or 1 cm and 10 cm from the user.

31. The method of claim 25, wherein the target acoustic wavefront is ambient noise of the open space.

32. The method of claim 25, wherein the signal processing module generates noise-cancellation wavefronts at a selected area of the user.

33. The method of claim 25, wherein the method further comprises monitoring the movement of the user by a monitoring device.

34. The method of claim 33, wherein the monitoring device provides geographical location feedback of the user movement to the signal processing module, thereby allowing the signal processing module to automatically generate noise-cancellation wavefronts.

35. The method of claim 34, wherein the geographical location feedback comprises data of geographical features and audio features.

Images