Speech processing method, apparatus, device, and medium

Abstract

The application relates to the technical field of voice processing, and specifically provides a voice processing method and device, equipment and a medium, aiming to solve the problem that an existing voice processing method cannot balance voice interaction response speed and sound source positioning accuracy. To this end, the application acquires multiple audio data in a driving device; when a first audio data first triggers a wake-up word, a first energy value and a second energy value are acquired and compared, wherein the first energy value is an energy value corresponding to the first audio data, and the second energy value is an energy value corresponding to the rest of the audio data in the multiple audio data; if the first energy value is greater than or equal to the second energy value, the rest of the audio data does not need to trigger the wake-up word, otherwise the wake-up word triggering duration of the rest of the audio data is dynamically allocated. Such a setting can reduce the interaction waiting time as much as possible while ensuring the accuracy of sound source positioning.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of voice processing, and specifically provides a voice processing method, device, equipment and medium. BACKGROUND

[0002] A voice interaction system has become an indispensable module of an intelligent cockpit, but a vehicle-mounted voice interaction effect is seriously affected by in-vehicle noises such as electric motors, multimedia, air conditioners, human voices, and the like, greatly increasing the difficulty of voice interaction system recognition of user voice. Based on this, in order to guarantee the wake-up rate of users in a vehicle-mounted scenario, a front-end voice processing module (ECNR: Echo Cancellation and Noise Reduction) is usually added in the voice interaction system in the prior art, which can suppress environmental noise and enhance effective voice signals, and then the processed audio signals are sent to a wake-up engine for wake-up processing.

[0003] Limited by the limitations of the front-end voice processing module (ECNR) algorithm, the audio collected by the microphone in the direction of the non-human voice of the voice interaction system will still have some wake-up words after processing, and therefore in order to correctly locate the direction of the human voice, it is usually necessary to perform wake-up word triggering on all channels and then perform the next step of judgment. However, if wake-up word triggering is performed on all channels, it is easy to cause a long waiting time, thereby introducing a large delay, which seriously affects a good experience, and if the channel that first triggers the wake-up word is directly taken as the direction of the human voice, it will also cause a sharp decline in positioning accuracy.

[0004] Correspondingly, there is a need in the art for a new voice processing method to solve the above problems. SUMMARY

[0005] The present application aims to solve the above technical problems, i.e., to solve the problem that the existing voice processing method cannot balance the voice interaction response speed and the sound source positioning accuracy.

[0006] In order to achieve the above purpose, in a first aspect, the present application provides a voice processing method applied to a driving device, the method comprising the following steps:

[0007] obtaining multi-channel audio data in the driving device;

[0008] when a first wake-up word is first triggered by the first channel of audio data, obtaining a first energy value and a second energy value and comparing the first energy value with the second energy value, wherein the first energy value is an energy value corresponding to the first channel of audio data, and the second energy value is an energy value corresponding to the remaining channels of audio data in the multi-channel audio data;

[0009] if the first energy value is greater than or equal to the second energy value, the remaining audio data does not need to be used for the wake-up word triggering;

[0010] if the first energy value is less than the second energy value, the wake-up word triggering duration of the remaining audio data is dynamically allocated.

[0011] In the optional technical solution of the above voice processing method, before the first energy value and the second energy value are obtained, the method further comprises:

[0012] The first audio data and the remaining audio data are respectively filtered based on a preset rule.

[0013] In the optional technical solution of the above voice processing method, the first audio data and the remaining audio data respectively include a plurality of sampling points, and the step of "filtering the first audio data and the remaining audio data based on a preset rule" comprises:

[0014] The plurality of sampling points of the first audio data and the remaining audio data are respectively subjected to preset energy value filtering.

[0015] After the preset energy value filtering is completed, the first audio data and the remaining audio data are respectively subjected to preset duration filtering.

[0016] In the optional technical solution of the above voice processing method, the step of "filtering the plurality of sampling points of the first audio data and the remaining audio data based on a preset rule" comprises:

[0017] Each sampling point in the first audio data and the remaining audio data is respectively subjected to first preset energy value filtering.

[0018] Alternatively, a preset number of sampling points in the first audio data and the remaining audio data are respectively subjected to second preset energy value filtering.

[0019] In the optional technical solution of the above voice processing method, before the first audio data and the remaining audio data are respectively filtered based on a preset rule, the method further comprises:

[0020] The first audio data is filtered based on the wake-up word to obtain first audio data containing only the wake-up word;

[0021] A time region corresponding to the first audio data containing only the wake-up word is obtained, and the remaining audio data is filtered based on the time region.

[0022] In the optional technical solution of the voice processing method, the step of dynamically allocating the wake-up word trigger time length of the remaining channel audio data comprises:

[0023] obtaining a difference value between the first energy value and the second energy value;

[0024] dynamically allocating the wake-up word trigger time length of the remaining channel audio data based on the difference value.

[0025] In the optional technical solution of the voice processing method, the remaining channel audio data comprises X channel audio data, where X is greater than 1 and X is a positive integer, and after the step of dynamically allocating the wake-up word trigger time length of the remaining channel audio data, the method further comprises:

[0026] when Y channel audio data in the remaining channel audio data triggers the wake-up word, comparing the energy values corresponding to the Y channel audio data, where Y is greater than 1 and less than or equal to X, and Y is a positive integer;

[0027] determining the user's orientation based on the comparison result of the energy values corresponding to the Y channel audio data.

[0028] In a second aspect, the present application further provides a voice processing device applied to a driving device, the device comprising:

[0029] an audio data acquisition module configured to acquire multi-channel audio data in the driving device;

[0030] an energy value acquisition and comparison module configured to, when first channel audio data first triggers a wake-up word, acquire a first energy value and a second energy value and compare the first energy value with the second energy value, where the first energy value is an energy value corresponding to the first channel audio data, and the second energy value is an energy value corresponding to remaining channel audio data in the multi-channel audio data;

[0031] a first processing module configured to, if the first energy value is greater than or equal to the second energy value, not require the remaining channel audio data to trigger a wake-up word;

[0032] a second processing module configured to, if the first energy value is less than the second energy value, dynamically allocate a wake-up word trigger time length of the remaining channel audio data.

[0033] In a third aspect, the present application further provides a computer device comprising a processor and a storage device, the storage device being adapted to store a plurality of program codes, the program codes being adapted to be loaded and run by the processor to execute the voice processing method of any one of the above.

[0034] In a fourth aspect, the present application also provides a computer readable storage medium, wherein a plurality of program codes are stored in the computer readable storage medium, and the program codes are adapted to be loaded and run by a processor to execute the voice processing method according to any one of the preceding aspects.

[0035] Those skilled in the art can understand that, in the technical solution of the present application, the multi-path audio data in the driving device is acquired; when the first-path audio data first triggers the wake-up word, the first energy value and the second energy value are acquired and compared, wherein the first energy value is the energy value corresponding to the first-path audio data, and the second energy value is the energy value corresponding to the remaining-path audio data in the multi-path audio data; if the first energy value is greater than or equal to the second energy value, the remaining-path audio data does not need to trigger the wake-up word; if the first energy value is less than the second energy value, the wake-up word triggering duration of the remaining-path audio data is dynamically allocated. Such a setting can reduce the interaction waiting time as much as possible under the premise of ensuring the accuracy of sound source positioning, thereby improving the user experience.

[0036] Further, the first-path audio data and the remaining-path audio data respectively include a plurality of sampling points, and the filtering of the first-path audio data and the remaining-path audio data based on the preset rule respectively includes: performing preset energy value filtering on the plurality of sampling points of the first-path audio data and the remaining-path audio data respectively; after the preset energy value filtering is completed, performing preset duration filtering on the first-path audio data and the remaining-path audio data respectively. Such a setting can avoid the influence of the user's pronunciation method on the calculation of the audio energy value, thereby avoiding interference with the allocation of the wake-up word triggering duration based on the energy value, and further improving the user experience. BRIEF DESCRIPTION OF DRAWINGS

[0037] The disclosure of the present application will become more apparent from the following description in conjunction with the accompanying drawings. Those skilled in the art will readily understand that the drawings are merely for illustrative purposes and are not intended to limit the scope of protection of the present application. In addition, similar numbers in the figures are used to represent similar components, wherein:

[0038] Figure 1 is a main step flow diagram of a voice processing method according to an embodiment of the present application;

[0039] Figure 2 is a main step flow diagram of filtering of the first-path audio data and the remaining-path audio data based on a preset rule according to an embodiment of the present application;

[0040] Figure 3 is a main step flow diagram before filtering of the first-path audio data and the remaining-path audio data based on a preset rule according to an embodiment of the present application;

[0041] Figure 4 is a main step flowchart diagram of a dynamic allocation of a wake-up word trigger duration of remaining road audio data according to an embodiment of the present application;

[0042] Figure 5 is a detailed step flowchart diagram of a voice processing method according to an embodiment of the present application;

[0043] Figure 6 is a flowchart diagram of a voice processing method according to an embodiment of the present application in an actual application scenario;

[0044] Figure 7 is a main structure block diagram of a voice device according to an embodiment of the present application;

[0045] Figure 8 is a main structure diagram of a computer device according to an embodiment of the present application.

[0046] List of reference signs :

[0047] 11: an audio data acquisition module; 12: an energy value acquisition and comparison module; 13: a first processing module; 14: a second processing module. DETAILED DESCRIPTION

[0048] Some embodiments of the present application will be described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present application, and are not intended to limit the protection scope of the present application.

[0049] In the description of the present application, "module" and "processor" can include hardware, software or a combination of both. A module can include hardware circuitry, various suitable sensors, communication ports, memories, and can also include a software part such as program code, and can be a combination of software and hardware. The processor can be a central processor, a microprocessor, an image processor, a digital signal processor or any other suitable processor. The processor has data and / or signal processing functions. The processor can be implemented in software, hardware or a combination of both. The non-transitory computer readable storage medium includes any suitable medium that can store program code, such as a magnetic disk, a hard disk, an optical disk, a flash memory, a read-only memory, a random access memory, etc. The term "A and / or B" means all possible combinations of A and B, such as only A, only B or both A and B. The term "at least one of A or B" or "at least one of A and B" has a similar meaning as "A and / or B", and can include only A, only B or both A and B. The singular form of the term "one", "this" can also include the plural form.

[0050] As described in the background section, in order to balance the response speed of voice interaction and the accuracy of sound source positioning, the present application provides a voice processing method.

[0051] Referring to the accompanying Figure 1 , Figure 1 is a main step flow diagram of a voice processing method according to an embodiment of the present application. The voice processing method can be executed by a driving device, can be executed by a server, or can also be executed by a server and a driving device together. As shown in Figure 1 , the voice processing method of the present application is applied to a driving device, and the method comprises the following steps:

[0052] Step S101: Obtain multi-channel audio data in the driving device.

[0053] Specifically, a sound collecting device is arranged at each position in the driving device, which can be a microphone for example. When the user speaks in the driving device, the sound collecting device at each position can collect sound to obtain multi-channel audio data in the driving device.

[0054] Step S102: When the first channel audio data first triggers the wake-up word, obtain the first energy value and the second energy value and compare the first energy value with the second energy value, wherein the first energy value is the energy value corresponding to the first channel audio data, and the second energy value is the energy value corresponding to the remaining channel audio data in the multi-channel audio data.

[0055] Specifically, the wake-up word is a technology based on intelligent speech recognition technology, which can wake up the voice interaction system on the device by recognizing a specific spoken wake-up word, so that the voice interaction system can receive the user's instruction and complete the corresponding action. In the use process of the voice interaction system, the user can wake up the voice interaction system by speaking the wake-up word, and perform corresponding operation or query according to the instruction issued by the wake-up word. The energy value can represent the difference between different audio data, and the energy value can have various expressions, for example, the energy value can be represented by at least one of amplitude, amplitude square and mean value, and decibel.

[0056] Step S103: If the first energy value is greater than or equal to the second energy value, the remaining channel audio data is not required to trigger the wake-up word.

[0057] Specifically, if the energy value of the first channel audio data that first triggers the wake-up word is greater than or equal to the energy value of the remaining channel audio data that does not trigger the wake-up word, the remaining channel audio data is no longer allocated with the wake-up word trigger duration, and the position of the first channel is taken as the position of the user, so as to avoid the situation that more wake-up word trigger duration is spent due to the low energy value of the audio data, and the voice trigger efficiency is improved.

[0058] Step S104: if the first energy value is less than the second energy value, dynamically allocating the wake-up word trigger duration of the rest of the audio data.

[0059] Specifically, if the energy value of the first audio data that first triggers the wake-up word is less than the energy value of the rest of the audio data that does not trigger the wake-up word, the wake-up word trigger duration of the rest of the audio data is dynamically allocated based on the first energy value and the second energy value, so as to facilitate positioning of the user's sound source based on the wake-up word trigger result of the rest of the audio data, and the accuracy of positioning of the user's sound source is ensured as much as possible.

[0060] In some embodiments, the rest of the audio data includes multiple audio data, wherein the rest of the audio data includes audio data with an energy value less than or equal to the first audio data, and also includes audio data with an energy value greater than the first audio data, and the method is: without the audio data with an energy value less than or equal to the first audio data performing the wake-up word trigger; dynamically allocating the wake-up word trigger duration of the audio data with an energy value greater than the first audio data.

[0061] Based on the above steps S101 to S104, the present application acquires multiple audio data in the driving device; when the first audio data first triggers the wake-up word, acquires the first energy value and the second energy value and compares the first energy value with the second energy value, wherein the first energy value is the energy value corresponding to the first audio data, and the second energy value is the energy value corresponding to the rest of the audio data in the multiple audio data; if the first energy value is greater than or equal to the second energy value, the rest of the audio data does not need to perform the wake-up word trigger; if the first energy value is less than the second energy value, the wake-up word trigger duration of the rest of the audio data is dynamically allocated. Such a setting can reduce the interaction waiting time as much as possible while ensuring the accuracy of sound source positioning, and improves the user's experience.

[0062] Next, steps S102 and S104 are further described.

[0063] In some embodiments, before acquiring the first energy value and the second energy value, the method further includes: screening the first audio data and the rest of the audio data based on a preset rule, respectively.

[0064] Referring to the accompanying drawings Figure 2 , Figure 2 is the main step flowchart diagram of screening the first audio data and the rest of the audio data based on a preset rule according to an embodiment of the present application. As shown in Figure 2In some embodiments, the first audio data and the rest of the audio data each include a plurality of sampling points, and the filtering of the first audio data and the rest of the audio data based on the preset rule includes the following steps:

[0065] Step S201: performing preset energy value filtering on the plurality of sampling points of the first audio data and the rest of the audio data, respectively.

[0066] Step S202: after the preset energy value filtering is completed, performing preset time length filtering on the first audio data and the rest of the audio data, respectively.

[0067] Specifically, after the audio data is received by the receiving device, the audio data is converted into a discrete digital signal, i.e., a waveform file, through sampling, quantization, and encoding. Different sampling rates will cause the received audio data to include different sampling points. For example, when the sampling rate is 16K, audio data with a time length of 1 second includes 16000 sampling points; when the sampling rate is 8K, audio data with a time length of 1 second includes 8000 sampling points. Performing preset energy value filtering on the plurality of sampling points of the first audio data and the rest of the audio data can avoid the influence of different speaking manners on the calculation of the energy value. For example, if the audio data includes some segments in which the user pauses or prolongs the tail of the speech, the overall energy value of the audio data will be affected, and therefore, the audio data needs to be subjected to preset energy value filtering. After the preset energy value filtering is completed, the audio data needs to be further subjected to preset time length filtering. The preset time length is related to the length of the wake-up word, i.e., the longer the wake-up word, the longer the preset time length, and the shorter the wake-up word, the shorter the preset time length, so as to ensure that the audio data includes the entire wake-up word. After the preset energy value filtering and the preset time length filtering of the first audio data and the rest of the audio data are completed, respectively, the energy value of the first audio data and the energy value of the rest of the audio data are obtained, respectively.

[0068] In some embodiments, the preset energy value filtering on the plurality of sampling points of the first audio data and the rest of the audio data includes the following steps:

[0069] Step S301: performing first preset energy value filtering on each sampling point in the first audio data and the rest of the audio data, respectively.

[0070] Step S302: alternatively, performing second preset energy value filtering on a preset number of sampling points in the first audio data and the rest of the audio data, respectively.

[0071] Specifically, when the first preset energy value screening is performed on each sample point in the first channel audio data and the rest channel audio data respectively, the energy value of each sample point in the first channel audio data and the rest channel audio data is obtained respectively, and each sample point in the first channel audio data and the rest channel audio data is sorted based on the order of the energy value from large to small, so as to facilitate the subsequent first preset energy value screening. When the second preset energy value screening is performed on the preset number of sample points in the first channel audio data and the rest channel audio data respectively, the energy value of the preset number of sample points in the first channel audio data and the rest channel audio data is obtained respectively, and the preset number of sample points in the first channel audio data and the rest channel audio data is sorted based on the order of the energy value from large to small, so as to facilitate the subsequent second preset energy value screening.

[0072] Exemplarily, the energy value of each sample point can be represented by the amplitude of each sample point, and the first preset energy value screening is performed on each sample point in the first channel audio data and the rest channel audio data respectively, that is, it is judged whether the amplitude of each sample point in the first channel audio data is greater than the first preset amplitude, and if so, the sample point is retained in the first channel audio data, otherwise, the sample point is deleted from the first channel audio data. The first preset energy value screening of the rest channel audio data is similar to the first energy value screening of the first channel audio data, and will not be described here.

[0073] The energy value of the preset number of sample points can be represented by the square sum of the amplitudes of the preset number of sample points and the average value, and the second preset energy value screening is performed on the preset number of sample points in the first channel audio data and the rest channel audio data respectively, that is, it is judged whether the average value of the square sum of the amplitudes of each preset number of sample points in the first channel audio data is greater than the second preset amplitude square sum average value, and if so, the preset number of sample points is retained in the first channel audio data, otherwise, the preset number of sample points is deleted from the first channel audio data. The second preset energy value screening of the rest channel audio data is similar to the second energy value screening of the first channel audio data, and will not be described here.

[0074] The setting mode of the first preset energy value screening and the setting mode of the second preset energy value screening described above are only exemplary, and can be selected according to actual needs in actual application.

[0075] Referring to the accompanying drawings Figure 3 , Figure 3 is a schematic diagram of the main step flow before the first channel audio data and the rest channel audio data are screened based on the preset rule according to an embodiment of the present application. As shown in Figure 3 , in some embodiments, before the first channel audio data and the rest channel audio data are screened based on the preset rule, the method further comprises the following steps:

[0076] Step S401: filtering the first audio data based on the wake-up word to obtain the first audio data containing only the wake-up word.

[0077] Step S402: obtaining the time region corresponding to the first audio data containing only the wake-up word, and filtering the remaining audio data based on the time region.

[0078] Specifically, in addition to containing the wake-up word, the first audio data can also contain other audio, so it is necessary to filter the first audio data to obtain the first audio data containing only the wake-up word. The corresponding time region of the wake-up word in the first audio data, i.e. the generation time of the wake-up word in the first audio data, is obtained, and the remaining audio data is filtered to obtain the remaining audio data containing only the corresponding time region, so as to facilitate the subsequent triggering of the wake-up word on the remaining audio data containing only the corresponding time region. After obtaining the first audio data containing only the wake-up word, the first audio data containing only the wake-up word is filtered based on a preset rule; similarly, after obtaining the remaining audio data containing only the corresponding time region, the remaining audio data containing only the corresponding time region is filtered based on the preset rule.

[0079] Referring to FIG. 4, Figure 4 Figure 4 FIG. 4 is a main step flow diagram of dynamically allocating the wake-up word triggering duration of the remaining audio data according to an embodiment of the present application. As shown in FIG. 4, Figure 4 in some embodiments, dynamically allocating the wake-up word triggering duration of the remaining audio data includes the following steps:

[0080] Step 501: obtaining the difference between the first energy value and the second energy value.

[0081] Step 502: dynamically allocating the wake-up word triggering duration of the remaining audio data based on the difference.

[0082] Specifically, the greater the difference between the first energy value and the second energy value, the longer the wake-up word triggering duration allocated to the remaining audio data; on the contrary, the smaller the difference between the first energy value and the second energy value, the shorter the wake-up word triggering duration allocated to the remaining audio data. Such a setting can allocate sufficient wake-up word triggering time to the audio with a large energy value in the remaining audio data to ensure the accuracy of subsequent positioning based on the user sound source; at the same time, it can also allocate appropriate wake-up word triggering time to the audio with a large energy value in the remaining audio data to avoid increasing the delay degree of voice interaction.

[0083] ​In some embodiments, the rest of the X-channel audio data, where X is greater than 1 and X is a positive integer, after the dynamic allocation of the wake-up word trigger duration of the rest of the X-channel audio data, the method further comprises the following steps:

[0084] Step S601: When Y-channel audio data in the rest of the X-channel audio data triggers a wake-up word, compare the energy values corresponding to the Y-channel audio data, where Y is greater than 1 and less than or equal to X, and Y is a positive integer.

[0085] Step S602: Based on the comparison result of the energy values corresponding to the Y-channel audio data, determine the orientation of the user.

[0086] Specifically, by comparing the energy values corresponding to the Y-channel audio data, the Y-channel audio data with the maximum energy value is obtained, and the orientation where the Y-channel audio data with the maximum energy value is located is taken as the orientation of the user.

[0087] Referring to FIG. 8, Figure 5 , Figure 5 is a detailed step flow diagram of a voice processing method according to an embodiment of the present application. The voice processing method can be executed by a driving device, can be executed by a server, or can also be executed by a server and a driving device together. As shown in FIG. 8, Figure 5 the voice processing method of the present application is applied to a driving device, and the method comprises the following steps:

[0088] Step S701: Obtain multi-channel audio data in the driving device.

[0089] Step S702: When the first-channel audio data first triggers a wake-up word, filter the first-channel audio data based on the wake-up word to obtain first-channel audio data containing only the wake-up word; obtain the time region corresponding to the first-channel audio data containing only the wake-up word, and filter the rest of the X-channel audio data based on the time region to obtain the rest of the X-channel audio data containing only the corresponding time region.

[0090] Step S703: Filter the first-channel audio data containing only the wake-up word and the rest of the X-channel audio data containing only the corresponding time region based on a preset rule, respectively, to obtain first-channel valid audio and the rest of the X-channel valid audio.

[0091] Step S704: Obtain the energy value of the first-channel valid audio and the energy value of the rest of the X-channel valid audio.

[0092] Step S705: Compare the energy value of the first-channel valid audio with the energy value of the rest of the X-channel valid audio.

[0093] Step S706: If the energy value of the first valid audio is greater than or equal to the energy value of the rest of the valid audio, the wake-up word trigger only containing the rest of the audio data of the corresponding time region is not needed.

[0094] Step S707: If the energy value of the first valid audio is less than the energy value of the rest of the valid audio, the wake-up word trigger duration only containing the rest of the audio data of the corresponding time region is dynamically allocated.

[0095] Step S708: The rest of the valid audio contains X valid audio, where X is greater than 1 and X is a positive integer. When Y valid audio in the rest of the valid audio triggers the wake-up word, the energy value corresponding to the Y valid audio is compared, where Y is greater than 1 and less than or equal to X, and Y is a positive integer.

[0096] Step S709: Based on the comparison result of the energy value corresponding to the Y valid audio, the user's orientation is determined.

[0097] Specifically, the flowchart of the voice processing method involved in the actual application scenario can be as shown in Figure 6 The N orientations in the driving device are set to receive audio with microphones, where N is greater than 1 and N is a positive integer. The audio data of the N orientations obtained is sent to the front-end voice processing module (ECNR) for processing, and M channel audio data is obtained based on the processed N orientation audio data, where M is greater than 1 and M is a positive integer. Start M wake-up engines to process M channel audio data synchronously, when the first channel audio data first triggers the wake-up word, obtain the first channel audio data containing only the wake-up word and the rest of the (M-1) channel audio data containing only the corresponding time region; based on a predetermined rule, the first channel audio data containing only the wake-up word and the rest of the (M-1) channel audio data containing only the corresponding time region are screened to obtain the first channel valid audio and the rest of the (M-1) channel valid audio. The energy values of the first channel valid audio and the rest of the (M-1) channel valid audio are obtained to dynamically allocate the wake-up word trigger duration, thereby ensuring the accuracy of the positioning based on the user's sound source.

[0098] Further, the voice processing method of the present application can be applied to the V2X scene. Specifically, the Internet of Vehicles refers to the network communication technology applied to vehicles. This technology is based on in-vehicle network, inter-vehicle network and vehicle mobile Internet, and in accordance with the agreed communication protocol and data exchange standard, wireless communication and information exchange between vehicle-X (vehicle, road, person and cloud, etc.) (Vehicle to X, V2X) is realized. The big system network, that is, the real-time online communication between vehicles, facilities, and cloud can be realized.

[0099] It should be noted that the multi-channel audio data involved in the embodiments of the present disclosure are all audio data authorized by the user or fully authorized by all parties. The actions such as obtaining of audio data involved in the embodiments of the present disclosure are all performed after being authorized by the user, the object or fully authorized by all parties.

[0100] It should be noted that although the above embodiments describe the steps in a specific order, those skilled in the art can understand that, in order to achieve the effects of the present application, the different steps do not have to be executed in such an order, and they can be executed simultaneously (in parallel) or in other orders, and these changes are within the protection scope of the present application. At the same time, all the above embodiment schemes can be combined to form optional embodiments of the present application, which will not be repeated here.

[0101] Further, the present application also provides a voice processing device.

[0102] Reference is made to the accompanying Figure 7 , Figure 7 is the main structural block diagram of the voice device according to an embodiment of the present application. As shown in Figure 7 , the voice processing device in the embodiments of the present application mainly includes an audio data acquisition module 11, an energy value acquisition and comparison module 12, a first processing module 13 and a second processing module 14. In some embodiments, one or more of the audio data acquisition module 11, the energy value acquisition and comparison module 12, the first processing module 13 and the second processing module 14 can be combined together to become one module. In some embodiments, the audio data acquisition module 11 can be configured to acquire multi-channel audio data in a driving device; the energy value acquisition and comparison module 12 can be configured to acquire a first energy value and a second energy value and compare the first energy value with the second energy value when a first channel audio data first triggers a wake-up word, wherein the first energy value is an energy value corresponding to the first channel audio data, and the second energy value is an energy value corresponding to the remaining channel audio data in the multi-channel audio data; the first processing module 13 can be configured to not require the remaining channel audio data to trigger the wake-up word if the first energy value is greater than or equal to the second energy value; and the second processing module 14 can be configured to dynamically allocate a wake-up word triggering duration of the remaining channel audio data if the first energy value is less than the second energy value.

[0103] In some embodiments, the energy value acquisition and comparison module 12 filters the first channel audio data and the remaining channel audio data based on a preset rule respectively.

[0104] In some embodiments, the energy value acquisition and comparison module 12 performs preset energy value screening on a plurality of sampling points of the first channel audio data and the rest channel audio data respectively; after the preset energy value screening is completed, preset time length screening is performed on the first channel audio data and the rest channel audio data respectively.

[0105] In some embodiments, the energy value acquisition and comparison module 12 performs first preset energy value screening on each sampling point of the first channel audio data and the rest channel audio data respectively; or, performs second preset energy value screening on a preset number of sampling points of the first channel audio data and the rest channel audio data respectively.

[0106] In some embodiments, the energy value acquisition and comparison module 12 screens the first channel audio data based on the wake-up word to obtain the first channel audio data containing only the wake-up word; obtains a time region corresponding to the first channel audio data containing only the wake-up word, and screens the rest channel audio data based on the time region.

[0107] In some embodiments, the second processing module 14 obtains a difference value between the first energy value and the second energy value; based on the difference value, dynamically allocates the wake-up word triggering time length of the rest channel audio data.

[0108] In some embodiments, the device further comprises a direction determination module, when Y channel audio data in the rest channel audio data triggers the wake-up word, the direction determination module compares the energy values corresponding to the Y channel audio data, where Y is greater than 1 and less than or equal to X, and Y is a positive integer; based on the comparison result of the energy values corresponding to the Y channel audio data, determines the direction where the user is located.

[0109] Those skilled in the art can understand that all or part of the processes in the method of the above-mentioned embodiment can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. When the computer program is executed by a processor, the steps of the above-mentioned various method embodiments can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable files or some intermediate forms, etc. The computer readable storage medium can include any entity or device, medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory, random access memory, electric carrier signal, telecommunication signal and software distribution medium, etc. that can carry the computer program code. It should be noted that the contents included in the computer readable storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to the legislation and patent practice, the computer readable storage medium does not include electric carrier signals and telecommunication signals.

[0110] Furthermore, this application also provides a computer device.

[0111] See appendix Figure 8 , Figure 8 This is a schematic diagram of the main structure of a computer device according to an embodiment of this application. Figure 8 As shown, the computer device in this embodiment mainly includes a storage device 21 and a processor 22. The storage device 21 can be configured to store a program for executing the speech processing method of the above-described method embodiments, and the processor 22 can be configured to execute the program in the storage device. This program includes, but is not limited to, a program for executing the speech processing method of the above-described method embodiments. For ease of explanation, only the parts related to the embodiments of this application are shown. For specific technical details not disclosed, please refer to the method section of the embodiments of this application.

[0112] In the embodiments of this application, the computer device may be a control device comprising various electronic devices. In some possible implementations, the computer device may include multiple storage devices 21 and multiple processors 22. The program executing the speech processing method of the above method embodiments may be divided into multiple subroutines, each subroutine may be loaded and run by a processor to execute different steps of the speech processing method of the above method embodiments. Specifically, each subroutine may be stored in different storage devices 21, and each processor 22 may be configured to execute programs in one or more storage devices 21 to jointly implement the speech processing method of the above method embodiments, that is, each processor 22 executes different steps of the speech processing method of the above method embodiments to jointly implement the speech processing method of the above method embodiments.

[0113] The aforementioned multiple processors 22 can be processors deployed on the same device. For example, the aforementioned computer device can be a high-performance device composed of multiple processors, and the aforementioned multiple processors 22 can be processors configured on that high-performance device. Alternatively, the aforementioned multiple processors 22 can also be processors deployed on different devices. For example, the aforementioned computer device can be a server cluster, and the aforementioned multiple processors 22 can be processors on different servers within the server cluster.

[0114] Further, the present application also provides a computer readable storage medium. In an embodiment of the computer readable storage medium according to the present application, the computer readable storage medium can be configured to store a program for executing the voice processing method of the above-mentioned method embodiment, which can be loaded and run by a processor to implement the above-mentioned voice processing method. For the convenience of illustration, only the parts related to the embodiments of the present application are shown, and the specific technical details not disclosed are referred to the method part of the embodiments of the present application. The computer readable storage medium can be a storage device formed by various electronic devices, and optionally, the computer readable storage medium in the embodiments of the present application is a non-transitory computer readable storage medium.

[0115] Further, it should be understood that, since the setting of each module is only for illustrating the functional units of the device of the present application, the corresponding physical device of the module can be the processor itself, or a part of the software, a part of the hardware, or a part of the combination of the software and the hardware in the processor. Therefore, the number of each module in the figure is only illustrative.

[0116] Those skilled in the art can understand that each module in the device can be adaptively split or combined. Such splitting or combining of the specific module will not cause the technical solution to deviate from the principles of the present application, therefore, the technical solution after splitting or combining will fall within the protection scope of the present application.

[0117] So far, the technical solution of the present application has been described in combination with the preferred embodiments shown in the drawings, but those skilled in the art can easily understand that the protection scope of the present application is obviously not limited to these specific embodiments. Those skilled in the art can make equivalent changes or replacements to the related technical features without deviating from the principles of the present application, and the technical solution after the changes or replacements will fall within the protection scope of the present application.

Claims

1. A voice processing method applied to a driving device, characterized by, The method comprises the following steps: acquiring multi-channel audio data in the driving device; when the first channel audio data first triggers the wake-up word, acquiring a first energy value and a second energy value and comparing the first energy value with the second energy value, wherein the first energy value is an energy value corresponding to the first channel audio data, and the second energy value is an energy value corresponding to the remaining channel audio data in the multi-channel audio data; if the first energy value is greater than or equal to the second energy value, the remaining channel audio data is not required to trigger the wake-up word; if the first energy value is less than the second energy value, the wake-up word triggering time length of the remaining channel audio data is dynamically allocated.

2. The voice processing method of claim 1, wherein, Before the step of acquiring the first energy value and the second energy value, the method further comprises: screening the first channel audio data and the remaining channel audio data based on a preset rule respectively.

3. The voice processing method of claim 2, wherein, The first channel audio data and the remaining channel audio data respectively comprise a plurality of sampling points, and the step of screening the first channel audio data and the remaining channel audio data based on a preset rule respectively comprises: performing preset energy value screening on the plurality of sampling points of the first channel audio data and the remaining channel audio data respectively; after the preset energy value screening is completed, performing preset time length screening on the first channel audio data and the remaining channel audio data respectively.

4. The voice processing method of claim 3, wherein, The step of performing preset energy value screening on the plurality of sampling points of the first channel audio data and the remaining channel audio data respectively comprises: performing first preset energy value screening on each sampling point in the first channel audio data and the remaining channel audio data respectively; or, performing second preset energy value screening on a preset number of sampling points in the first channel audio data and the remaining channel audio data respectively.

5. The voice processing method of claim 2, wherein, Before the step of screening the first channel audio data and the remaining channel audio data based on a preset rule respectively, the method further comprises: screening the first channel audio data based on the wake-up word to obtain first channel audio data containing only the wake-up word; acquiring a time region corresponding to the first channel audio data containing only the wake-up word, and screening the remaining channel audio data based on the time region.

6. The voice processing method of claim 1, wherein, The step of dynamically allocating the wake-up word triggering time length of the remaining channel audio data comprises: acquiring a difference value between the first energy value and the second energy value; based on the difference value, dynamically allocating the wake-up word triggering time length of the remaining channel audio data.

7. The voice processing method of claim 1, wherein, The remaining channel audio data comprises X channel audio data, wherein X is greater than 1 and X is a positive integer, and after the step of dynamically allocating the wake-up word triggering time length of the remaining channel audio data, the method further comprises: when Y channel audio data in the remaining channel audio data triggers the wake-up word, comparing energy values corresponding to the Y channel audio data, wherein Y is greater than 1 and less than or equal to X, and Y is a positive integer; based on the comparison result of the energy values corresponding to the Y channel audio data, determining a user's orientation.

8. A voice processing apparatus applied to a driving device, characterized by, The device comprises: an audio data acquisition module configured to acquire multi-channel audio data in the driving device; The energy value obtaining and comparing module is configured to obtain a first energy value and a second energy value and compare the first energy value with the second energy value when the first audio data first triggers the wake-up word, wherein the first energy value is an energy value corresponding to the first audio data, and the second energy value is an energy value corresponding to the rest of the audio data in the multi-channel audio data; The first processing module is configured to, if the first energy value is greater than or equal to the second energy value, not require the rest of the audio data to trigger the wake-up word; The second processing module is configured to, if the first energy value is less than the second energy value, dynamically allocate a wake-up word triggering duration of the rest of the audio data.

9. A computer device comprising a processor and a storage device, said storage device being adapted to store a plurality of program codes, characterized in that, The program code is adapted to be loaded and run by the processor to execute the voice processing method in any one of claims 1 to 7.

10. A computer readable storage medium having stored therein a plurality of program codes, characterized in that, The program code is adapted to be loaded and run by the processor to execute the voice processing method in any one of claims 1 to 7.

Images