CN110427097A - Voice data processing method, apparatus and system - Google Patents

Voice data processing method, apparatus and system Download PDF

Info

Publication number
CN110427097A
CN110427097A CN201910526214.5A CN201910526214A CN110427097A CN 110427097 A CN110427097 A CN 110427097A CN 201910526214 A CN201910526214 A CN 201910526214A CN 110427097 A CN110427097 A CN 110427097A
Authority
CN
China
Prior art keywords
processor
target service
server
voice data
handled
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201910526214.5A
Other languages
Chinese (zh)
Inventor
孙渊
屈伸
伍晓晖
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huawei Technologies Co Ltd
Original Assignee
Huawei Technologies Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huawei Technologies Co Ltd filed Critical Huawei Technologies Co Ltd
Priority to CN201910526214.5A priority Critical patent/CN110427097A/en
Publication of CN110427097A publication Critical patent/CN110427097A/en
Priority to PCT/CN2020/096545 priority patent/WO2020253715A1/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/26—Power supply means, e.g. regulation thereof
    • G06F1/32—Means for saving power
    • G06F1/3203—Power management, i.e. event-based initiation of a power-saving mode
    • G06F1/3234—Power saving characterised by the action undertaken
    • G06F1/3293—Power saving characterised by the action undertaken by switching to a less power-consuming processor, e.g. sub-CPU
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/28—Constructional details of speech recognition systems
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
    • G10L2015/223—Execution procedure of a spoken command

Abstract

The application provides a kind of voice data processing method, apparatus and system.Wherein, for include simultaneously first processor and second processor terminal device in, when the first processor for handling voice data is when being in low power consumpting state, the wake-up word in voice data can be detected by power consumption lower second processor；And in detecting target speech data include wake up word after, second processor further judges whether the corresponding target service of instruction in voice data is handled by second processor, if judging, target service is handled by second processor, and second processor directly handles the instruction.So that first processor is when being in low power consumpting state, still to voice data can wake up the detection of word by the lower second processor of power consumption, and judging that target service handles by second processor, then second processor directly handles the instruction, thus power consumption when reducing to language data process.

Description

Voice data processing method, apparatus and system
Technical field
This application involves electronic technology more particularly to a kind of voice data processing methods, apparatus and system.
Background technique
With the development of electronic technology, more and more terminal devices, which all have, receives voice data and broadcasting voice data Etc. relevant data processing function, terminal device is engaged in the dialogue by way of interactive voice with user and is exchanged.With Family can be issued to terminal device by voice and be instructed, and after terminal device receives the voice data of user, handle voice number According to corresponding instruction.It therefore, can controlling terminal equipment reality by the instruction that voice issues when user is busy with other things It is existing for example, inquiry weather, the various functions such as listening to music or navigating, it is very strong which be provided with terminal device Practicability and interest.
In the prior art, when user and terminal device talk with, user needs first to say before saying instruction to terminal device It is specific to wake up word.Correspondingly, terminal device can be received constantly and detect received voice data, and only be called out in detection It wakes up after word, just continues with the instruction in voice data after the wake-up word.Due to terminal device in addition to provide and user into The function of row interactive voice, it is also necessary to meet the function such as communication, standby of the terminal device itself.Therefore in order to reduce language The power consumption of sound interactive function, when some terminal devices in a dormant state when, the higher central processing unit (central of power consumption Processing unit, CPU) it generally will not be used to receive and detect the wake-up word in voice data always, but pass through The processors such as the lower processor of power consumption such as Digital Signal Processing (digital signal processing, DSP) chip into The detection of word is waken up in row voice data.After it includes waking up word that dsp chip, which detects in voice data, dsp chip is again into one Step wakes up CPU and handles the instruction in voice data.
But the prior art is used, terminal device requires to call out in the instruction that user per treatment is issued by voice The CPU of awake terminal device and the display screen for waking up terminal device are bright screen state, so as to cause terminal device in processing language Power consumption when sound data is larger, and then reduces the stand-by time of terminal device, influences the user experience of terminal device.
Summary of the invention
The application provides a kind of voice data processing method, apparatus and system, to reduce terminal device to voice data Power consumption when processing to increase the stand-by time of terminal device, and then improves the user experience of terminal device.
One embodiment of the application first aspect provides a kind of voice data processing apparatus, comprising: first processor and second Processor；Wherein, the first processor connects the second processor, and the operation power consumption of the first processor is greater than described The operation power consumption of second processor；When the first processor is in low power consumpting state, the second processor is used for:
Voice data is received from outside by microphone；Determine the requested target service of the voice data；Judge institute State whether target service is handled by the second processor；When judging that the target service is handled by the second processor, The request for requesting the target service is sent to server, wherein the first processor maintains the low power consumpting state.
To sum up, the voice data processing apparatus provided in the present embodiment can be used to handle voice data in a device First processor is when being in low power consumpting state, additionally it is possible to be waken up by the lower second processor of power consumption to voice data The detection of word, and user per treatment issued by voice instruction when, if corresponding target service is instructed to judge by second Device processing is managed, then second processor directly handles the instruction, is identifying target speech data without second processor In wake-up word after, all go wake up first processor, but by the lesser second processor of power consumption can processing target business Corresponding instruction.To reduce power consumption of device when to language data process, especially when first processor is in low The power consumption of voice data is handled when power consumption state, to increase the stand-by time of above-mentioned apparatus, and then improves user experience.
In one embodiment of the application first aspect, the second processor is also used to: the transmission of Xiang Suoshu server is used for After the request for requesting the target service, the request results for the target service that the server is sent are received；According to institute The request results for stating target service handle the target service.
To sum up, the voice data processing apparatus provided in the present embodiment, when second processor processing target business is corresponding When instruction, if receiving the request results of the target service of server transmission, second processor directly carries out target service Processing, is again handled target service transmitted by server without first processor.It further reduces at device The power consumption of the corresponding target service of voice data is managed, and further increases stand-by time and improves user experience.
In one embodiment of the application first aspect, the second processor is also used to: when judging the target service not It is when being handled by the second processor, to wake up the first processor and be in normal operating conditions；
The first processor being waken up is used to send the request for requesting the target service to the server, And the request results for the target service that the server is sent are received, according to the request results of the target service to described Target service is handled.
To sum up, the voice data processing apparatus provided in the present embodiment, in second processor when to language data process, When judging target service not is handled by second processor, just directly wakes up first processor and be in normal operating conditions, and Voice data is handled by first processor.To provide a kind of more complete language data process mode, power consumption lesser the Two processors can wake up the biggish first processor of power consumption and handle target service when being unable to processing target business.
In one embodiment of the application first aspect, the second processor is also used to: when judging the target service not It is when being handled by the second processor, Xiang Suoshu server is sent for requesting the request of the target service, and indicates institute It states server and the request results of the target service is handed down to the first processor；
After the first processor is waken up by the second processor or the server, for according to the server The request results of the target service issued handle the target service.
To sum up, the voice data processing apparatus provided in the present embodiment is in normal work shape waking up first processor When state, by way of waking up indirectly, second processor then may be used due to having determined that corresponding target service in voice data To replace first processor directly to send the request of target service to server, and indicate request knot of the server by target service Fruit is sent to first processor processing, to improve efficiency when waking up first processor.
In one embodiment of the application first aspect, whether the second processor judges the target service by described Two processors processing, comprising: match the target service with pre-set business；According to matching result, the target is determined Whether business is handled by the second processor.
To sum up, the voice data processing apparatus provided in the present embodiment, second processor can be by matching pre-set business Mode, judge whether target service is handled by second processor.Then second processor can read store in advance it is pre- If business, by White List service or blacklist business, determine whether target service can be handled by second processor.
In one embodiment of the application first aspect, pre-set business processor operational capability required when including operation Less than the business of the first preset value；Alternatively, the pre-set business includes storage capacity required when running less than the second preset value Business.
In one embodiment of the application first aspect, the pre-set business includes below one or more: inquiry weather, Query time, control household, play music, setting alarm clock, play music, encyclopaedia question and answer, using schedule, use calculator, section Holiday inquires, translates, listening audiobook, listening cross-talk and listen radio station.
In one embodiment of the application first aspect, when the first processor is in low power consumpting state, described second Processor and the server keep long connection.
To sum up, the voice data processing apparatus provided in the present embodiment, due to being in low-power consumption shape in first processor When state, second processor needs to carry out voice data to wake up the detection of word, and the finger that user per treatment is issued by voice When enabling, if corresponding target service is instructed to judge to be handled by second processor, second processor needs to be maintained at the first processing Device keeps connecting with the long of server when being in low power consumpting state, so as to need to send mesh to server in second processor After the request of mark business, server can be sent to by long connection.
In one embodiment of the application first aspect, the second processor determines the requested target of the voice data Business includes: to determine the requested target service of the voice data when including waking up word in the voice data.
In one embodiment of the application first aspect, the first processor is primary processor, and the second processor is Coprocessor；Alternatively, the first processor is the first processor core in multi-core processor, the second processor is described Second processor core in multi-core processor；The operational capability of the first processor is greater than the operation energy of the second processor Power, alternatively, the storage capacity of the first processor is greater than the storage capacity of the second processor.
In one embodiment of the application first aspect, the low power consumpting state includes dormant state.
In one embodiment of the application first aspect, described device is chip or electronic equipment.
The application second aspect provides a kind of voice data processing system, including as described in the application any one of first aspect Voice data processing apparatus.
The application third aspect provides a kind of voice data processing system, comprising: terminal device and server；Wherein, institute Stating terminal device includes: first processor and second processor, and the first processor connects the second processor, and described the The operation power consumption of one processor is greater than the operation power consumption of the second processor；
The second processor is used for, and when the first processor is in low power consumpting state, passes through the terminal device Microphone from outside receive voice data；The second processor is also used to, and determines to include keyword in the voice data When, the voice data is sent to the server；
The server is used for, and is received and is determined the requested target service of the voice data；The server is also used In judging whether the target service is handled by the second processor；The server is also used to, when determining the target industry When business is handled by the second processor, Xiang Suoshu second processor handles the request results of the target service, wherein described First processor maintains the low power consumpting state.
To sum up, the voice data processing system provided in the present embodiment, when in terminal device for handling voice data First processor, can be by the lower second processor of power consumption in terminal device in voice data when being in low power consumpting state Wake-up word detected；And in detecting target speech data include after waking up word, second processor is by target voice number According to server is sent to, further judge the corresponding target service of instruction in voice data whether by second processing by server Device processing.If judging, target service is handled by second processor, and the request results of target service are returned to second processor, are made Second processor is obtained to handle target service according to the request results of target service.Without in terminal device As long as two processors identify the wake-up word in target speech data, the first processor for waking up terminal device can be all removed, but In the case where server judges that target service is handled by second processor, pass through the lesser second processor of terminal device power consumption It can the corresponding instruction of processing target business.Also, in the present embodiment, the finger after word is waken up in voice data is determined by server It enables corresponding destination service and destination service is matched, to further reduce terminal device at voice data Power consumption when reason, and can also have certain portability suitable for the lower terminal device of various operational capabilities.
In one embodiment of the application third aspect, the low power consumpting state includes dormant state.
In one embodiment of the application third aspect, the server is also used to, when determine the target service not and be by When second processor processing, wakes up the first processor and be in normal operating conditions；The server is also used to, to institute State the request results that first processor handles the target service；The first processor being waken up is used for, and receives the clothes The request results for the target service that business device is sent carry out the target service according to the request results of the target service Processing.
To sum up, the voice data processing system provided in the present embodiment, server can judge target service not and be by When second processor processing, the first processor for being in low power consumpting state is directly waken up, and the request results of target service are sent out First processor is given to be handled.
In one embodiment of the application third aspect, the server is also used to, when determine the target service not and be by When second processor processing, Xiang Suoshu second processor sends instruction information, be used to indicate the target service not and be by The second processor processing；The server is also used to, and Xiang Suoshu first processor handles the request knot of the target service Fruit；
The second processor is also used to, and according to the instruction information, is waken up the first processor and is in normal work State simultaneously establishes long connection with the server；
The first processor is used for, and the request results for the target service that the server is sent is received, according to institute The request results for stating target service handle the target service.
To sum up, voice data processing system provided in this embodiment, server can be by judging target service not When the processing of two processors, instruction information is sent indirectly to second processor, so that second processor wakes up first processor, with So that first processor is sent to the request results of the target service of first processor according to server, the target industry is handled Business.
In one embodiment of the application third aspect, whether the server judges the target service by described second Manage device processing, comprising: match the target service with pre-set business；According to matching result, the target service is determined Whether handled by the second processor.
To sum up, voice data processing system provided in this embodiment, server can by way of matching pre-set business, Judge whether target service is handled by second processor.Then server can read the pre-set business stored in advance, pass through White List service or blacklist business, determine whether target service can be handled by second processor.
In one embodiment of the application third aspect, pre-set business processor operational capability required when including operation Less than the business of the first preset value；Alternatively, the pre-set business includes storage capacity required when running less than the second preset value Business.
In one embodiment of the application third aspect, when the first processor is in low power consumpting state, described second Processor and the server keep long connection.
To sum up, voice data processing system provided in this embodiment, due to being in low power consumpting state in first processor When, second processor needs to carry out voice data to wake up the detection of word, and the instruction that user per treatment is issued by voice When, if corresponding target service is instructed to judge to be handled by second processor, second processor needs to be maintained at first processor It keeps connecting with the long of server when in low power consumpting state, so as to need to send target to server in second processor After the request of business, server can be sent to by long connection.
In one embodiment of the application third aspect, the first processor is primary processor, and the second processor is Coprocessor；Alternatively, the first processor is the first processor core in multi-core processor, the second processor is described Second processor core in multi-core processor；The operational capability of the first processor is greater than the operation energy of the second processor Power, alternatively, the storage capacity of the first processor is greater than the storage capacity of the second processor.
The application fourth aspect provides a kind of voice data processing method, is applied to include first processor and second processing The language data process device device of device, wherein the first processor connects the second processor, the first processor Run the operation power consumption that power consumption is greater than the second processor；The described method includes:
Voice data is received from outside by microphone by second processor；
The requested target service of the voice data is determined by the second processor；
Judge whether the target service is handled by the second processor by the second processor；
When judging that the target service is handled by the second processor, sent out by the second processor to server Send the request for requesting the target service, wherein the first processor maintains the low power consumpting state.
In one embodiment of the application fourth aspect, the method also includes:
The request results for the target service that the server is sent are received by the second processor；
By the second processor according to the request results of the target service, the target service is handled.
In one embodiment of the application fourth aspect, the method also includes:
When judging the target service not is handled by the second processor, institute is waken up by the second processor It states first processor and is in normal operating conditions；
By the first processor being waken up, Xiang Suoshu server is sent for requesting the target service Request, and the request results for the target service that the server is sent are received, according to the request results of the target service The target service is handled.
In one embodiment of the application fourth aspect, the second processor is also used to: when judging the target service not It is when being handled by the second processor, Xiang Suoshu server is sent for requesting the request of the target service, and indicates institute It states server and the request results of the target service is handed down to the first processor；
After the first processor is waken up by the second processor or the server, for according to the server The request results of the target service issued handle the target service.
It is described that whether the target service is judged by the second processor in one embodiment of the application fourth aspect It is handled by the second processor, comprising:
The target service is matched with pre-set business by the second processor；
By the second processor according to matching result, determine the target service whether by the second processor Reason.
In one embodiment of the application fourth aspect, pre-set business processor operational capability required when including operation Less than the business of the first preset value；Alternatively,
Business of the pre-set business storage capacity required when including operation less than the second preset value.
In one embodiment of the application fourth aspect, the pre-set business includes below one or more:
Inquiry weather, control household, plays music, setting alarm clock, plays music, encyclopaedia question and answer, uses day query time Journey is inquired using calculator, festivals or holidays, translates, listens audiobook, listens cross-talk and listen radio station.
In one embodiment of the application fourth aspect, when the first processor is in low power consumpting state, described second Processor and the server keep long connection.
It is described to determine that the voice data is asked by the second processor in one embodiment of the application fourth aspect The target service asked, comprising:
When including waking up word in the voice data, determine that the voice data is requested by the second processor Target service.
In one embodiment of the application fourth aspect, the first processor is primary processor, and the second processor is Coprocessor；Alternatively, the first processor is the first processor core in multi-core processor, the second processor is described Second processor core in multi-core processor；
The operational capability of the first processor is greater than the operational capability of the second processor, alternatively, at described first The storage capacity for managing device is greater than the storage capacity of the second processor.
In one embodiment of the application fourth aspect, the low power consumpting state includes dormant state.
In one embodiment of the application fourth aspect, described device is chip or electronic equipment.
The 5th aspect of the application provides a kind of voice data processing method, is applied to voice data processing system, wherein institute The system of stating includes: terminal device and server；Wherein, the terminal device includes: first processor and second processor, described First processor connects the second processor, and the operation power consumption of the first processor is greater than the operation of the second processor Power consumption；The described method includes:
The second processor passes through the Mike of the terminal device when the first processor is in low power consumpting state Wind receives voice data from outside；
When the second processor is determined in the voice data including keyword, the voice data is sent to described Server；
The server receives and determines the requested target service of the voice data；
The server also judges whether the target service is handled by the second processor；
The server is when determining that the target service is handled by the second processor, at Xiang Suoshu second processor Manage the request results of the target service, wherein the first processor maintains the low power consumpting state.
In the 5th one embodiment of aspect of the application, the low power consumpting state includes dormant state.
In the 5th one embodiment of aspect of the application, the method also includes:
The server wakes up at described first when determining the target service not is handled by the second processor Reason device is in normal operating conditions；
The server handles the request results of the target service to the first processor；
The first processor being waken up receives the request results for the target service that the server is sent, according to The request results of the target service handle the target service.
In the 5th one embodiment of aspect of the application, the method also includes:
The server is when determining the target service not is handled by the second processor, Xiang Suoshu second processing Device sends instruction information, and being used to indicate the target service not is handled by the second processor；
The server handles the request results of the target service to the first processor；
The second processor according to the instruction information, wake up the first processor be in normal operating conditions and with The server establishes long connection；
The first processor receives the request results for the target service that the server is sent, according to the target The request results of business handle the target service.
In the 5th one embodiment of aspect of the application, whether the server judges the target service by described second Manage device processing, comprising: match the target service with pre-set business；According to matching result, the target service is determined Whether handled by the second processor.
In the 5th one embodiment of aspect of the application, pre-set business processor operational capability required when including operation Less than the business of the first preset value；Alternatively, the pre-set business includes storage capacity required when running less than the second preset value Business.
In the 5th one embodiment of aspect of the application, when the first processor is in low power consumpting state, described second Processor and the server keep long connection.
In the 5th one embodiment of aspect of the application, the first processor is primary processor, and the second processor is Coprocessor；Alternatively, the first processor is the first processor core in multi-core processor, the second processor is described Second processor core in multi-core processor；
The operational capability of the first processor is greater than the operational capability of the second processor, alternatively, at described first The storage capacity for managing device is greater than the storage capacity of the second processor.
The 6th aspect of the application provides a kind of terminal device, comprising: first processor, second processor and memory；
The memory is for storing program instruction and data；
The memory is coupled with the processor, and the first processor and the second processor can be called and be held The program instruction stored in the row memory, for realizing the function in the method for any one of above-mentioned fourth aspect description；
The terminal device can also include communication interface, the communication interface for the terminal device and other equipment into Row communication.
The 7th aspect of the application provides a kind of terminal device, comprising: first processor, second processor and memory；
The memory is for storing program instruction and data；
The memory is coupled with the processor, and the first processor and the second processor can be called and be held The program instruction stored in the row memory, for realizing the function in the method for any one of above-mentioned 5th aspect description；
The terminal device can also include communication interface, the communication interface for the terminal device and other equipment into Row communication.
The application eighth aspect provides a kind of computer readable storage medium, including instruction, when it runs on computers When, so that computer executes such as the described in any item methods of above-mentioned fourth aspect.
The 9th aspect of the application provides a kind of computer readable storage medium, including instruction, when it runs on computers When, so that described in any item methods in terms of computer executes the such as the above-mentioned 5th.
Detailed description of the invention
Fig. 1 for the application institute application scenarios schematic diagram；
Fig. 2 is the flow diagram of the method for terminal device processing voice data in the prior art；
Fig. 3 handles status diagram when voice data for terminal device in the prior art；
Fig. 4 is the structural schematic diagram of one embodiment of terminal device provided by the present application；
Fig. 5 is the structural schematic diagram of one embodiment of terminal device provided by the present application；
Fig. 6 is the flow diagram of voice data processing method embodiment one provided by the present application；
Fig. 7 is the status diagram of the corresponding terminal device of voice data processing method embodiment one provided by the present application；
Fig. 8 is the flow diagram of voice data processing method embodiment two provided by the present application；
Fig. 9 is the flow diagram of voice data processing method embodiment three provided by the present application；
Figure 10 is the flow diagram of voice data processing method example IV provided by the present application；
Figure 11 is the flow diagram of voice data processing method embodiment five provided by the present application；
Figure 12 is the flow diagram of voice data processing method embodiment six provided by the present application；
Figure 13 is the structural schematic diagram of one embodiment of terminal device provided by the present application；
Figure 14 is the structural schematic diagram of one embodiment of terminal device provided by the present application.
Specific embodiment
Fig. 1 for the application institute application scenarios schematic diagram.In scene as shown in Figure 1, user 1 can be with terminal device 2 are engaged in the dialogue exchange by way of interactive voice, and terminal device 2 has that receive voice data and broadcasting voice data etc. related Data processing function.Wherein, when user 1 and terminal device 2 are talked with, user 1 needs to need to send out saying to terminal device 2 Wake-up word is said before instruction out；Terminal device 2 can constantly receive and detect received voice data, only detect It include after waking up word, just continuing with the instruction in voice data after the wake-up word in received voice data.
For example, user 1 can say " ABCD, today, how is weather " to terminal device 2, then terminal device 2 is receiving To voice data " ABCD, today, how is weather " and identify its beginning wake-up word " ABCD " after, terminal device 2 just continues Handle the instruction that " today, how is weather " after word " ABCD " is waken up in voice data.Terminal device 2 can be by " today day Gas is how " instruction be sent to the server 3 of setting network side beyond the clouds so that server 3 determines that the instruction is requested After weather business, corresponding Weather information is sent to terminal device 2.Terminal device 2 is in the weather for receiving the transmission of server 3 After information, the voice of " today is fine, 15 to 25 degree " can be played out by the loudspeaker of terminal device according to Weather information, from And realize the interactive voice between terminal device 2 and user 1.
Optionally, in scene as shown in Figure 1, terminal device may is that mobile phone, wrist-watch, bracelet, TV, intelligent phase Frame, vehicular rear mirror, intelligent travelling crane recorder, tablet computer, laptop or desktop computer etc. are handed over related voice The smart machine of mutual function.
Meanwhile terminal device usually not merely also has other function for carrying out interactive voice with user, for example, When terminal device is mobile phone, the function of interactive voice is carried out with user in addition to providing, it is also necessary to meet daily standby, communication etc. Function, and mobile phone needs to power by battery.Therefore, consumption of the terminal devices such as mobile phone to voice interactive function provided by it Electricity requirement with higher, to improve the stand-by time of terminal device, reduce charge frequency.
Fig. 2 is the flow diagram of the method for terminal device processing voice data in the prior art, as shown in Figure 2 some To the language data process process for the terminal device that power consumption has higher requirements.Wherein, by taking the terminal device is mobile phone as an example, hand The higher central processing unit of power consumption (central processing unit, CPU) will not generally be used to always receive simultaneously in machine The wake-up word in voice data is detected, especially when terminal device is in dormant state, CPU is now in low power consumpting state.Then In order to meet the function of interactive voice, the Digital Signal Processing (digital signal processing, DSP) of interior of mobile phone Chip generally can detect the wake-up word in voice data when CPU is in low power consumpting state.
It is right then after dsp chip receives the voice data of audio stream form by microphone (microphone, MIC) Voice data is detected, if detecting includes waking up word in voice data, dsp chip further wakes up CPU to voice number It is handled according to the instruction after middle wake-up word.For example, CPU can be with if being the instruction for operating smart home device after waking up word The instruction is sent directly to corresponding smart home device by communication module；Alternatively, if after waking up word being inquiry weather Instruction, then the instruction can be sent to server by communication module by CPU, and receive the Weather information of server return Afterwards, voice data is generated according to Weather information by CPU, and is played out by loudspeaker.
For example, Fig. 3 be in the prior art terminal device handle voice data when status diagram, with as shown in Figure 2 For handling voice data, the terminal device takes the mobile phone as an example terminal device.
In the A condition of Fig. 3, when mobile phone in a dormant state, the CPU of interior of mobile phone is in low power consumpting state, Bu Huijin Row language data process, the dsp chip in mobile phone pass through MIC and receive voice data and wake up the detection of word.At this point, mobile phone Display interface be in blank screen, put out screen or go out under screen state.
It include waking up word in voice data when dsp chip detects, then after waking up CPU to wake-up word in voice data Instruction is further processed, and the state of mobile phone switches to B state from the A condition in Fig. 3 at this time.When the CPU of mobile phone is waken up, CPU exits low power consumpting state, starts to work with normal operating conditions, correspondingly, mobile phone B exits standby mode simultaneously, mobile phone Display screen is also at bright screen state.
Then, in C-state shown in Fig. 3, the CPU of mobile phone is further processed the instruction waken up after word in voice data " today, weather was how ", and the content is sent to by server by communication module.The communication module includes: cellular communication Module or Wireless Fidelity (wireless fidelity, WiFi) module etc..Then, CPU receives server by communication module After the Weather information of return, the voice of " today is fine, 15 to 25 degree " is played out by loudspeaker.
To sum up, in the above prior art, the part of label 1. is only used for receiving voice in terminal device as shown in Figure 2 Data simultaneously detect the wake-up word in voice data, and once detect the wake-up word in voice data, it is necessary to wake up The part of label 2. in terminal device, and the instruction after waking up word in voice data is carried out into one by CPU and communication module Step processing.And in state as shown in Figure 3, terminal device every time receive and detect in voice data include wake up word after, Requiring to wake up the CPU of terminal device and wake up the display screen of terminal device is bright screen state.Therefore, terminal device is to language When sound data are handled, need continually to exit suspend mode and continually bright screen, so as to cause in the prior art, eventually Power consumption of the end equipment when handling voice data is larger, and then reduces the stand-by time of terminal device, influences terminal device User experience.
The application is based on above-mentioned deficiency in the prior art, provides a kind of voice data processing method and device, to reduce Power consumption of terminal device when to language data process, to increase the stand-by time of terminal device, and then terminal device mentions High user experience.
Wherein, Fig. 4 is the structural schematic diagram of one embodiment of terminal device provided by the present application.In example as shown in Figure 4 Shown in terminal device can be used in application scenarios as shown in Figure 1, have to receive and voice data and voice data be known Ability that is other and being further processed.Wherein, terminal device 2 as shown in Figure 4 specifically includes: first processor 21 and second processing Device 22.First processor 21 and second processor 22 can communicate can for example be led to by communication mode between processor core Letter, and power consumption when first processor 21 operates normally is greater than power consumption when second processor 22 operates normally, i.e., at first The operation power consumption for managing device is greater than the operation power consumption of second processor.
Optionally, first processor 21 includes the CPU of the terminal device 2；Second processor 22 includes: that the terminal is set Standby 2 micro-control unit (microcontroller unit, MCU), dsp chip or universal intelligent senses hub (sensorhub).Alternatively, optionally, the first processor 21 and second processor 22 can also be the same of terminal device 2 Different kernel in multi-core processor.Wherein, two or more kernels be can integrate in multi-core processor (kernel again can quilt Referred to as: computing engines), the calculating of processor can be individually performed in each kernel.
In example as shown in Figure 4, first communication module 23 can be the communication module (packet in first processor 21 Include: cellular communication module, WiFi module or other communication modules), second communication module 24 can be in second processor 22 Communication module (including: cellular communication module, WiFi module or other communication modules).It is logical that first processor 21 can be used first Letter module 23 is communicated with server 3；Second communication module 24 and server communication can be used in second processor 22.
Optionally, second processor 22 as shown in Figure 4 and second communication module 24 are plotted in 2 location of terminal device It is interior, the inclusion relation of the two in logic is referred only to, and in concrete implementation, second processor 22 and second communication module 24 are also It can be set in the accessory 4 being connect with terminal device 2.For example, if the earphone that accessory 4 is connected by terminal device, at this time Second processor 22 may include the dsp chip in earphone, and second communication module 24 may include the network communication mould in earphone Block.
Alternatively, Fig. 5 is that the structure of one embodiment of terminal device provided by the present application is shown on basis as shown in Figure 4 It is intended to.In embodiment as shown in Figure 5, first processor 21 and second processor 22 in terminal device 2 can common terminal set Communication module 25 (including: cellular communication module, WiFi module or other communication modules) in standby 2.That is, first processor 21 can To use communication module 25 to communicate with server 3, second processor 22 also can be used communication module 25 and communicate with server 3.
Similarly, in the concrete realization, second processor 22 is also possible in logic in terminal device 2, terminal The dsp chip in accessory 4 that equipment 2 is connected.After accessory 4 connects terminal device 2, the second processor 22 in accessory 4 can To be communicated by the communication module 25 in terminal device 2 with server 3.
With reference to the accompanying drawing, voice data processing method provided by the present application is illustrated.In each embodiment of the application Method can be executed by terminal device 2 as shown in Figure 4 or terminal device 2 as shown in Figure 5.
Communication tool if being executed by Fig. 4, in subsequent each embodiment, between described first processor and server Body is that first processor 21 passes through between first communication module 23 and server communication, described second processor and server Communication to be specially second processor 22 pass through second communication module 24 and server communication.If terminal device as shown in Figure 5 Execute, then the communication between described first processor and server be specially first processor 21 by communication module 25 with Communication between server communication, described second processor and server is specially that second processor 22 passes through communication module 25 and server communication, it repeats no more.In addition, processor can refer to existing skill by way of communication module and server communication Art, the present embodiment do not limit this.
Fig. 6 is the flow diagram of voice data processing method embodiment one provided by the present application.Voice as shown in FIG. 6 In data processing method, using terminal device as executing subject for be illustrated, rather than it is defined.As shown in Figure 6 Voice data processing method can also by other it is any there are at least two processors electronic equipment execute, such as: speaker, Mobile phone, TV etc., alternatively, voice data processing method as shown in FIG. 6 can also be executed by the chip in electronic equipment.
As shown in fig. 6, voice data processing method provided in this embodiment includes:
S100: the first processor of terminal device enters low power consumpting state.
Specifically, when the first processor described in the terminal device is in normal operating conditions, terminal device can be used for In application scenarios as shown in Figure 1, the first processor in terminal device is used for by outside microphone receiving terminal apparatus Voice data, and in voice data whether include wake up word detect.If detecting includes waking up word in voice data, First processor continues to handle the instruction after wake-up word in voice data.
And in the application embodiment as shown in Figure 6, enter low power consumpting state operation for the first processor of terminal device Afterwards, the application scenarios that terminal device handles voice data.
Wherein, the working condition of first processor includes at least: normal operating condition and the low power consumpting state.When first Processor is under normal operating condition, and first processor can execute times that all first processors in terminal device should execute Business, will not enough power supply big because of power consumption or terminal device without executing certain tasks or reducing execute certain tasks Frequency.And under low power consumpting state, first processor can be abandoned executing the partial task and reduction that first processor should execute First processor executes the frequency of task, so that terminal device reduces electricity of the battery to first processor of a part of terminal device Pressure output, to reduce the electricity of the consumed terminal device of first processor.It is understood that first processor is in low-power consumption Consumed electricity is less than its consumed electricity under normal operating conditions when running under state.The low power consumpting state can also To include dormant state at first processor, then when first processor in a dormant state when will not handle task.Or Person, the low power consumpting state may not be dormant state, but not handle voice data (voice command), but handle other The function of small power consumption such as screen locking or the CPU task to be treated at backstage.Wherein, first processor exists in the present embodiment Under normal operating condition, it can keep long between server and connect.Optionally, when first processor is in low power consumpting state When, the long connection between first processor and server can be disconnected.Wherein, refer on which can be continuous for the long connection Multiple data packets are sent, during the long connection between first processor and server is kept, if sent without data packet, are needed Both sides are wanted to send out link detecting packet, to nonexpondable can connect after realizing primary establish.For example, the long connection can be with It include: transmission control protocol (transmission control protocol, TCP) connection, hypertext transfer protocol (hyper Text transfer protocol, HTTP) connection, User Datagram Protocol (user datagram protocol, UDP) company It connects or Hyper text transfer security protocol (hyper text transfer protocol over Secure Socket Layer, HTTPS) agreement connection.
Optionally, when terminal device enters suspend mode, first processor enters low power consumpting state operation.Wherein, eventually End equipment can be after the screen locking operation for detecting user, into suspend mode；Alternatively, terminal device can be within a preset time After not detecting user's operation, it is advanced into suspend mode certainly.
Optionally, after the instruction for the state switching module that first processor can be set in receiving terminal device, Into low power consumpting state；Wherein, the state switching module is used to switch first processor and the according to the state of terminal device The state of two processors.Alternatively, first processor can also be determined voluntarily after entering low power consumpting state, and logical to second processor Know that it has entered low power consumpting state, so that second processor executes subsequent step.
S101: the second processor and server of terminal device establish long connection.
Specifically, in the present embodiment second processor after determining that first processor has entered low power consumpting state state, with Server establishes long connection.The long connection for example may include: that TCP connection, UDP connection, HTTPS connection or HTTP connect It connects.Also, in the embodiment illustrated in fig. 6, second processor is after establishing the long connection with server, in needs and clothes When business device communication, the long connection and server transport data can be used.
Wherein, the operation power consumption of each second processor as described in the examples of the application is less than the operation function of first processor Consumption.When the power consumption of the processor can be run by processor, the quantity of the consumed energy quantifies in the unit time It measures.When the unit of the quantity of the energy can be watt (W), milliampere hour (mAh) or microampere (μ Ah) etc..The present embodiment In for first processor and second processor specific implementation without limitation, need to only meet the power consumption of second processor less than the One processor.Alternatively, in the possibility that other specifically distinguish first processor and second processor, at first The operational capability for managing device can be greater than the operational capability of second processor；Alternatively, the storage capacity of first processor is greater than second The storage capacity of processor.The storage capacity can pass through the random access memory (random of the terminal device Access memory, RAM) size measured.
For example, such as: CPU, second processor can be terminal device if first processor is the primary processor of terminal device Coprocessor, such as: MCU, dsp chip or universal intelligent sense hub；Alternatively, if first processor is terminal device In certain dsp chip, then second processor can be in terminal device run power consumption be less than first processor other DSP cores Piece；Or the first processor can be the first processor core of multi-core processor in terminal device, second processor can To be the second processor core in the multi-core processor.
Optionally, after the instruction for the state switching module that second processor can be set in receiving terminal device, It requests to establish long connection to server；Wherein, the state switching module is used for according at the state of terminal device switching first Manage the state of device and second processor.Alternatively, second processor can be when receiving first processor and entering low power consumpting state After the notice of transmission, request to establish long connection to server.
S102: second processor receives voice data from outside by microphone.
Specifically, low power consumpting state is entered by S100 when the first processor in terminal device, second processor passes through S101 and server are established after long connection, at this time the voice data where the MIC acquisition terminal equipment of terminal device in environment Afterwards, voice data collected is sent to second processor, by second processor in voice data whether include wake up word It is detected.
S103: whether it includes waking up word that second processor identifies in voice data.
Wherein, whether voice data acquired in second processor identification S102 includes waking up word.It will in the present embodiment Target speech data is denoted as including waking up the voice data of word acquired in second processor in S102.Then when second processor is known It Chu not include after waking up the target speech data of word, S104 being executed according to target speech data and is further processed；And when the It includes that after waking up word, will not continue to handle voice data, and return in S102 that two processors, which identify voice data not, New voice data is obtained again.
Such as: if waking up word is " ABCD ", when the voice data that second processor obtains is that " ABCD, today, weather was how When the target speech data of sample ", then it includes " ABCD " that second processor, which detects in voice data, then continues to target voice number According to being handled；When the voice data that second processor obtains is " hello ", identify not include waking up word in the voice data, The voice data is not further processed then.Wherein, it is called out in second processor identification voice data in the present embodiment S103 The technology of awake word is without limitation.
Further, the application, which also provides, a kind of can be applied to wake up word in light weight level processor identification voice data Method can be used for second processor identification in S103 and wake up word, so that the lesser second processor energy of power consumption in terminal device It is enough that the wake-up word in voice data is identified faster with less calculation amount.
Specifically, the MIC being usually arranged in existing terminal device is the array MIC of array format.For example, a MIC Array includes 4 MIC, 5 MIC or 6 MIC etc..And each MIC in MIC array can ring where acquisition terminal equipment The voice data in border, and the processor for being sent to terminal device jointly carries out waking up word identification.Therefore, it is identified in second processor Voice data wake up word identification model be also to be obtained jointly by the voice data of MIC each in MIC array, model compared with Greatly, result in processor when wake up word identification calculation amount is larger, calculating speed is slower.And then it results in when terminal is set When standby second processor computing capability is poor, second processor is slower to the recognition speed for waking up word in voice data, causes Terminal device can not be waken up immediately after receiving the voice data of user, when influencing the voice data interaction of terminal device User experience.
Therefore, the wake-up word recognition method of lightweight can be set in second processor provided in this embodiment.Specifically, The identification model for waking up word in second processor for identification is obtained by the voice data of a MIC；In identification process, the The voice data that two processors obtain MIC multiple in MIC array carries out the processing of multichannel pickup, then selects one of those The voice data that MIC is obtained, the wake-up word in voice data obtained according to identification model to a MIC identify.By This, reduces calculation amount when waking up word identification, so that the lesser second processor core of computing capability can also be identified faster and be called out Awake word.
In a kind of concrete implementation mode, second processor can select optimal according to terminal device placement position MIC.For example, if leaning on voice data received by the MIC of wall displacement in MIC array is reflection when terminal device is close to metope Sound wave obtains, then second processor can carry out the voice data of the MIC far from wall locations waking up word identification；Alternatively, the After two processors can also determine very noisy source direction, recorded according to sound source position as a result, selecting the highest sound source position of frequency The voice data of the MIC in direction carries out waking up word identification.
S104: second processor further determines that target service corresponding to the target speech data determined in S103.
Specifically, second processor can determine the finger waken up after word in voice data by way of speech recognition It enables, and further target service corresponding to determine instruction.For example, if target speech data is that " ABCD, today, how is weather Sample ", then second processor is in S104 further to " today, how is weather " after wake-up word " ABCD " in voice data Speech recognition is carried out, determines that the corresponding target service of the instruction is " weather lookup ".
Optionally, in the present embodiment second processor can by target speech data wake up word after instruction into The mode of row semantic analysis determines the corresponding target service of the instruction.For example, instruction " today, how is weather " passes through semantic point The corresponding target service of the available instruction is analysed to be " weather lookup ", " what day is it today " is instructed can to obtain by semantic analysis It is the corresponding relationship of " date inquiries " to the corresponding target service of the instruction.Wherein, the application carries out second processor semantic The concrete mode of analysis is not specifically limited.
S105: second processor judges whether identified target service should be carried out by the second processor in S104 Reason.
Wherein, in voice data processing method provided in this embodiment, second processor is in addition to it needs to be determined that target voice Target service corresponding to instruction in data, it is also necessary to target service further be judged, only identified target In the case that business should be handled by second processor, just further the instruction in target speech data is handled； Otherwise, when the target service corresponding to the instruction in target speech data is handled by second processor, at second Reason device will not continue to handle target service corresponding to the instruction in target speech data.
Optionally, in the present embodiment, second processor can be especially by by identified target service and pre-set business After being matched, determined whether to handle the target service by second processor according to matching result.Wherein, the pre-set business can To be stored in second processor, alternatively, be stored in the storage equipment of the terminal device, and can by second processor into Row calls.
In one possible implementation, pre-set business includes the business white list handled by second processor, i.e., should Business in business white list can be handled by second processor, and the business other than the business white list is then by first processor Processing.After second processor instructs corresponding target service in determining target speech data, by target service and the white name of business It is singly matched, if matching, target service are handled by second processor；If mismatching, target service is not by second Device processing is managed, i.e., target service should be handled by first processor.For example, business white list includes: " weather lookup ", " date Inquiry " and " control smart machine ", then when second processor is after abovementioned steps determine target service for " weather lookup ", Being fitted in business white list includes " weather lookup " identical with target service, it is determined that target service is by second processor Reason, and execute subsequent step.
In alternatively possible specific implementation, pre-set business includes the black name of business handled by second processor Single, i.e., the business other than the business blacklist can be handled by second processor, and the business in the business blacklist is then by the The processing of one processor.Second processor determines corresponding target service is instructed in target speech data after, by target service and industry Business blacklist is matched, if matching, target service is not handled by second processor, and is handled by first processor；If no Matching, then target service is handled by second processor.For example, business blacklist includes: " address navigation " and " video calling ", then When second processor by abovementioned steps determine target service be " weather lookup " after, match business blacklist in do not include with Target service is identical " weather lookup ", it is determined that target service is handled by second processor, and executes subsequent step.
Optionally, in the present embodiment, business handled by second processor, i.e. business white list, including terminal are needed Needed for equipment it is to be processed it is relatively simple, lesser business is consumed to processor resource.Its measurement standard can be, and processor exists When handling first pre-set business, business of the required processor operational capability less than the first preset value；Alternatively, required deposits Business of the energy storage power less than the second preset value.The storage capacity can pass through the random access memory of the terminal device (random access memory, RAM) size is measured.Such as: the business white list includes one below or more : inquiry weather, control home equipment, plays music, setting alarm clock, plays music, encyclopaedia question and answer, uses day query time The business such as radio station are inquired, translate, listen audiobook, listen cross-talk and listened to journey using calculator, festivals or holidays.
And the business for needing first processor to be stored, i.e. business blacklist, including processor required for terminal device It is complex, biggish business is consumed to processor resource, measurement standard can be, and processor is to handle described second default When business, required processor operational capability is greater than or equal to the business of the first preset value；Alternatively, required storage capacity is greater than Or the business equal to the second preset value.Such as: navigation, video, the setting business such as mobile phone and incoming call sound, these business can be by First processor is handled.
It is understood that if target service of the second processor by determination is " address navigation " in S105, according to upper Stating the example target service is handled by second processor, then the subsequent of S105 as shown in Figure 6 is not carried out in second processor Step, at this point, second processor can refer to embodiment illustrated in fig. 8 for the subsequent step of the target service.
S106: if judge that target service is handled by second processor, second processor sends target industry to server The request of business.
Specifically, if second processor judges that target service is handled by second processor in S105, it is determined that can be with Subsequent processing is carried out to the instruction after waking up word in target speech data by second processor.Second processor can be according to target Business sends the request of target service to server.For example, if the instruction in target speech data is that " today, how is weather Sample ", target service corresponding to the instruction are " inquiry weather ", then second processor is in judgement " inquiry weather " and pre-set business After matching, determination can be handled the service of inquiry weather by second processor.
Therefore, in S106, second processor can send weather lookup request to server, to request day to server The corresponding Weather information data of gas inquiry business；Alternatively, second processor can also be directly by the instruction in target speech data " today, weather was how " is sent to server, after determining the corresponding target service of the instruction for " inquiry weather " by server, Weather information data are returned to second processor.
S107: server returns to the request results of target service to second processor.For second processor, then receive From the request results of the target service transmitted by server.
After server receives the request of target service transmitted by second processor, target service is determined according to the request Request results, and the request results of target service are sent to second processor.For example, if target service request is looked into for weather Ask request, then after server determines real-time Weather information, using acquired weather lookup request corresponding Weather information as The request results of target service are sent to second processor.
S108:, can after second processor receives the request results of target service transmitted by server by S107 To handle target service according to the request results of target service.
For example, if when the request results of target service are Weather information, second processor can be according to being connect in S108 The loudspeaker of the request results of the target service received, controlling terminal equipment plays the voice of the Weather information.
Alternatively, optionally, if the instruction in voice data is to control the instruction of smart machine, such as " turning on light ", then second Processor sends " turning on light " to server after judging that the corresponding target service of the instruction is handled by second processor, through S106 Request, then server can be opened according to lamp from the request to needs sends the control signal turned on light, without passing through again S107 returns to the request results of target service to second processor, and also there is no need to target service for second processor Reason.
It is understood that first processor maintains its low in the whole flow process of S101-S108 as described in Figure 6 Power consumption state.And when first processor is in low power consumpting state, second processor can keep the length established with server Connection.So that second processor can carry out communication faster with server.
To sum up, in the voice data processing method provided by the present embodiment, for handling voice data in terminal device First processor be in low power consumpting state when, can be by the lower second processor of power consumption in terminal device to voice data In wake-up word detected；And in detecting target speech data include wake up word after, second processor is further to language Whether the corresponding target service of instruction in sound data is handled by second processor is judged, if judging target service by second Processor processing, then second processor directly handles the instruction.So that the first processor of terminal device is being in low-power consumption shape When state, terminal device can wake up to voice data the detection of word by the lower second processor of power consumption, and locate every time When the instruction that reason user is issued by voice, if instructing corresponding target service to judge to be handled by second processor, at second Reason device directly handles the instruction, without second processor after identifying the wake-up word in target speech data, all goes The first processor of terminal device is waken up, but can the corresponding finger of processing target business by the lesser second processor of power consumption It enables.To reduce at power consumption of terminal device when to language data process, the especially first processor in terminal device Power consumption when low power consumpting state to increase the stand-by time of terminal device, and then improves the user experience of terminal device.
For example, the state that Fig. 7 is the corresponding terminal device of voice data processing method embodiment one provided by the present application is shown Be intended to, with terminal device using method as shown in FIG. 6 to language data process during, carried out for the state of terminal device Illustrate, the terminal device takes the mobile phone as an example, and first processor is CPU in mobile phone, second processor is DSP in mobile phone.
Wherein, in normal operation, the CPU of interior of mobile phone is used to carry out waking up to voice data the inspection of word to mobile phone It surveys, and detects that is instructed in the voice data comprising wake-up word is further processed.And in the A1 state of Fig. 6, work as mobile phone In a dormant state, the CPU of interior of mobile phone is in low power consumpting state, not will do it language data process, and power consumption is smaller in mobile phone DSP receive voice data and carry out wake up word detection.At this point, the display interface of mobile phone is in blank screen, puts out screen or the screen that goes out Under state.
Include waking up word " ABCD " in voice data when DSP is detected, then further determines that after waking up word in voice data The corresponding target service of instruction, and determine target service whether handled by DSP.At this point, as in the B1 state in Fig. 7, mobile phone CPU be not waken up, display interface be still in blank screen, put out screen or go out under screen state.
If DSP determines that target service is handled by DSP, DSP does not need to wake up CPU, but directly process instruction is corresponding Target service.Wherein, DSP can send the request of target service by communication module to server, and pass through communication module After the request results for receiving the target service that server returns, by DSP processing target business.For example, DSP can be directly according to mesh The request results of mark business, the loudspeaker for controlling mobile phone play the voice of " today is fine, 15 to 25 degree ".And in entire DSP processing In the instruction process of target service, CPU is not waken up and is constantly in low power consumpting state, similarly, the C1 shape in Fig. 7 In state, since CPU is not waken up, the display interface of mobile phone is also constantly in blank screen without bright screen, puts out screen or go out and shield shape Under state.
Therefore, by the status diagram of mobile phone shown in Fig. 7 and Fig. 3 it can be concluded that, use is provided in this embodiment The terminal device of voice data processing method, can be in the case where not waking up the display screen of terminal device, can be to judgement The instruction of the target service handled by second processor is handled, and completes the function of the interactive voice of terminal device.To Reduce terminal device in a dormant state or power consumption when low power consumpting state when processing voice data, is set to increase terminal Standby stand-by time, the user experience that can be improved terminal device.
Further, Fig. 8 is the flow diagram of voice data processing method embodiment two provided by the present application.Such as scheming In embodiment two shown in 8, on the basis of showing embodiment one shown in Fig. 6, if second processor is by target industry in S105 Business carries out after it fails to match with pre-set business, to the subsequent processing of target speech data.
As shown in figure 8, S100-S105 can refer to the description in embodiment one as shown in Figure 6, implementation and principle phase Together.
In S206, it is when being handled by second processor if second processor by S105 determines target service not, for example, Target service is " address navigation ", then second processor can wake up first processor, so that first processor processing is to target Business is further processed.
Optionally, second processor sends to first processor especially by the mode of intercore communication and wakes up thing in S206 Part exits low power consumpting state, is switched to normal operating conditions after first processor receives the wake events.Wherein, first Processor establishes long connection when exiting low power consumpting state, with server.So that first processor is exiting low-power consumption shape After state, pass through the long connection established and server communication.
Optionally, after S206 second processor wakes up first processor, second processor can disconnect and server Long connection.
In S207, when first processor is waken up and is in normal operating conditions, second processor will be acquired Target speech data is sent to first processor, so that first processor carries out subsequent processing to target speech data.
It optionally, can be to the wake-up word in target speech data after first processor receives target speech data It is detected again.And after the wake-up word in detection target speech data, the finger waken up after word in target speech data is determined Enable corresponding target service.
In S208, first processor sends the request of target service to server to according to target service.For example, if Instruction in target speech data is " address A is gone in navigation ", and target service corresponding to the instruction is " address navigation ", then first Processor can send the navigation requests of the address A to server, to request corresponding navigation data to server；Alternatively, first Instruction " address A is gone in navigation " in target speech data directly can also be sent to server by processor, be determined by server After the corresponding target service of the instruction is " address navigation ", corresponding navigation data is returned to server.
S209: server is then received to the request results that first processor returns to target service then for first processor The request results of the target service transmitted by the server.
After server receives the request of target service transmitted by first processor, target service is determined according to the request Request results, and the request results of target service are sent to first processor.For example, if target service request is the address A Navigation requests, then after server determines the navigation data of the address A, using acquired navigation data information as the target industry The request results of business are sent to first processor.
S210: after first processor receives the request results of target service transmitted by server by S209, root According to the request results of target service, target service is handled.
For example, if when the request results of target service are navigation data, first processor can be according to being connect in S210 The request results of the target service received show guidance path by the display interface of terminal device, and are played by loudspeaker Guidance path voice prompting etc..
Optionally, in S210 after the request results of first processor processing target business, at this time at first processor In normal operating conditions, then the voice number of environment where can continuing through MIC receiving terminal apparatus by first processor According to, and in voice data whether include wake up word detect.If detecting includes waking up word in voice data, directly by the One processor continues to handle the instruction after wake-up word in voice data.
And in the preset time (such as: 30 minutes) after S210, terminal device is not all in the voice number received Wake-up word is detected in, then illustrates the function that interactive voice is not all reused in the user preset time.Therefore, in order to save The power consumption of terminal device is saved, first processor is again introduced into low power consumpting state, and second processor can establish and server Between long connection, S102 as shown in FIG. 6 is continued to execute by second processor.
Further, Fig. 9 is the flow diagram of voice data processing method embodiment three provided by the present application.
In embodiment three as shown in Figure 9, after S306 second processor wakes up first processor, second processor is not Target speech data can be transmitted directly to first processor, but according to the processor in S102-S105 as a result, by second It manages device and target service request is sent to server by S307.And mesh is returned from server in subsequent embodiment to first processor The request results of mark business.
Since second processor is after waking up first processor, withouts waiting for first processor and exit low power consumpting state Time, therefore, S306 and S307 can be executed by second processor simultaneously.
Optionally, second processor sends to first processor especially by the mode of intercore communication and wakes up thing in S306 Part exits low power consumpting state, is switched to normal operating conditions after first processor receives the wake events.Wherein, first Processor establishes long connection when exiting low power consumpting state, with server.So that first processor is exiting low-power consumption shape After state, pass through the long connection established and server communication.
Optionally, after S306 and S307, second processor can disconnect the long connection between server.
In S308, after server receives the request of the target service of second processor transmission, need further exist for really Whether the business that sets the goal is handled by second processor, needs the request results of target service being back to first processor with determination Or second processor.
In one possible implementation, business white list and/or business blacklist also be can store in server.Clothes Device be engaged in after receiving the request of destination service, it is also necessary to according to the business white list or business blacklist stored, to mesh Whether mark service is handled by second processor is judged.If judging target service not is incited somebody to action when being handled by second processor The request results of target service are back to first processor, so that first processor processing target business.If judging target service When being handled by second processor, then the request results of target service are back to second processor, so that second processor is handled Target service.
In alternatively possible implementation, when second processor sends target service request to server in S307, The identification information that first processor can also be carried is enabled the server to according to the identification information, and determination is needed target industry The request results of business are back to first processor.
Then, in S308, if it is when being handled by second processor that server, which judges target service not, it is determined that by first Therefore processor processing target business establishes long connection in S309, between server and first processor.Wherein, at first The heartbeat that reason device can establish low-power consumption between first processor first is connect, and then, is connected by the heartbeat established logical Know that first processor is established and the long connection between server.
In S310, server according to established in S308 it is long connect, the request results of target service are back to the The process that one processor and subsequent first processor handle target service can refer in S209-210 as shown in Figure 8 Description, implementation is identical as principle.
Optionally, in S310 after the request results of first processor processing target business, at this time at first processor It, then can be by the voice data of environment where first processor receiving terminal apparatus, and to voice in normal operating conditions It whether include waking up word to be detected in data.If detect in voice data include wake up word, directly by first processor after The continuous instruction to after wake-up word in voice data is handled.And in the preset time (such as: 30 minutes) after S310, eventually End equipment does not all detect wake-up word in the voice data received, then illustrates all not make again in the user preset time With the function of interactive voice.Therefore, in order to save the power consumption of terminal device, first processor is again introduced into low power consumpting state, and And second processor can establish the long connection between server, continue to execute S102 as shown in FIG. 6 by second processor.
Further, in embodiment as shown in Figure 9, due to being also required to storage and phase in second processor in server Same pre-set business, therefore, when the pre-set business stored in server updates, server, which can send to update to second processor, disappear Breath, is used to indicate second processor and synchronizes and be updated to the pre-set business stored in second processor.
In embodiment as illustrated in figures 6-10, when second processor can judge that target service is handled by second processor, It determines whether by second processor oneself processing target business, or first processor is waken up by second processor and handles mesh Mark business.
And in the alternatively possible implementation of the application, also provide it is a kind of by server determine target service whether by Second processor processing, so that server instruction second processor processing target business or server wake up first processor The mode of processing target business.
Specifically, Figure 10 is the flow diagram of voice data processing method example IV provided by the present application, is such as being schemed In embodiment shown in 10, S400-S403 can refer to S100-S103 described in embodiment as shown in Figure 6, implementation It is identical as principle, it repeats no more.
After second processor identifies that target speech data includes waking up word by S403, second processor then exists In S404, acquired target speech data is sent to server, target speech data is carried out further by server Reason.
In S405, after server receives target speech data, the corresponding mesh of target speech data can be determined Mark business, specifically, server can determine the instruction waken up after word in voice data by way of speech recognition, and Further determine that the corresponding target service of instruction.For example, if target speech data is " ABCD, today, how is weather ", then Server carries out speech recognition to obtain instruction being " today, how is weather " to target speech data, and further determines that this refers to Enabling corresponding target service is " weather lookup ".It optionally, can be by way of semantics recognition in server in the present embodiment The corresponding target service of determine instruction.
In S406, server can further judge whether target service is handled by second processor.Wherein, in one kind In possible specific implementation, after server instructs corresponding target service in determining target speech data, pass through service The matching result of the first pre-set business of at least one stored in device judges whether target service is handled by second processor；Its In, if matching includes target service at least one first pre-set business, target service is handled by second processor；If matching It does not include target service at least one first pre-set business, then target service is whether to be handled by second processor.Alternatively, logical Whether the matching result for crossing at least one the second pre-set business stored in server judges target service by second processor Reason；Wherein, if matching at least one second pre-set business includes target service, whether target service is by second processor Processing；If matching at least one second pre-set business does not include target service, target service is handled by second processor.
In S407, server sends the judging result in S406 to second processor.
And in embodiment as shown in Figure 10, if server determines that target service is handled by second processor in S406 When, server further passes through the request results that S408 returns to target service to second processor.That is, if judge target service by Second processor processing, then server obtains the request results of target service and returns according to target service identified in S405 Back to second processor.
It, can after receiving the request results of target service of server transmission for second processor then in S409 To handle target service according to the request results of target service.For example, if the request results of target service are navigation number According to when, then second processor can target service based on the received request results, pass through the display interface of terminal device It shows guidance path, and guidance path voice prompting etc. is played by loudspeaker, with processing target business.
To sum up, it in the voice data processing method provided by the present embodiment as shown in Figure 10, is used in terminal device The first processor of voice data is handled when being in low power consumpting state, it can be by the lower second processing of power consumption in terminal device Device detects the wake-up word in voice data；It and include second processing after waking up word in detecting target speech data Target speech data is sent to server by device, and the corresponding target service of instruction in voice data is further judged by server Whether handled by second processor.If judging, target service is handled by second processor, returns to target industry to second processor The request results of business, so that second processor is handled target service according to the request results of target service.From without As long as second processor is wanted to identify the wake-up word in target speech data, the first processor for waking up terminal device can be all removed, But in the case where server judges that target service is handled by second processor, at terminal device power consumption lesser second Managing device can the corresponding instruction of processing target business.Also, in the present embodiment, determined by server after waking up word in voice data The corresponding destination service of instruction and destination service is matched, to further reduce terminal device in voice number Power consumption when according to processing, and can also have certain portable suitable for the lower terminal device of various operational capabilities Property.
Further, Figure 11 is the flow diagram of voice data processing method embodiment five provided by the present application.Such as It in embodiment five shown in Figure 11, shows on the basis of embodiment illustrated in fig. 10 four, if server judges target in S406 Business is not after being handled by second processor, to the subsequent processing of target speech data.
As shown in figure 11, S400-S407 can refer to the description in example IV as shown in Figure 10, implementation and principle It is identical.
And in S508, if second processor determines that the judging result that server returns in S407 determines that i.e. target service is not To be handled by second processor, then second processor can wake up first processor so that first processor to target service into Row is further processed.
Optionally, second processor can send wake events to first processor especially by the mode of intercore communication, After first processor receives the wake events, low power consumpting state is exited, normal operating conditions is switched to.Wherein, at first Device is managed when exiting low power consumpting state, establishes long connection with server.So that first processor is exiting low power consumpting state Later, pass through the long connection established and server communication.Optionally, after second processor wakes up first processor, the Two processors can be disconnected to be connected with the long of server.
In S509, after second processor is waken up by first processor, long connection is established with server.
Then, in S510, server returns to the request results of target service according to the long connection established in S509 To first processor, then for first processor, then the request results of the target service transmitted by the server are received.
It is right after request results of the first processor by receiving target service transmitted by server in S511 Target service is handled.The processing can refer to the description in S210, repeat no more.
Optionally, in S510 after the request results of first processor processing target business, at this time at first processor In normal operating conditions, then the voice number of environment where can continuing through MIC receiving terminal apparatus by first processor According to, and in voice data whether include wake up word detect.If detecting includes waking up word in voice data, directly by the One processor continues to handle the instruction after wake-up word in voice data.
And in the preset time (such as: 30 minutes) after S510, terminal device is not all in the voice number received Wake-up word is detected in, then illustrates that user does not reuse the function of interactive voice within a preset time.Therefore, in order to The power consumption of terminal device is saved, first processor is again introduced into low power consumpting state, and second processor can establish and service Long connection between device, S402 as shown in Figure 10 is continued to execute by second processor.
Further, Figure 12 is the flow diagram of voice data processing method embodiment six provided by the present application.Such as It in embodiment six shown in Figure 12, shows on the basis of embodiment illustrated in fig. 10 four, if server judges target in S406 Business is not another mode that subsequent processing is carried out to target speech data after being handled by second processor.
Wherein, as shown in figure 12, S400-S406 can refer to the description in example IV as shown in Figure 10, implementation It is identical as principle.
And in S607, it, can be with if the target service that server determines in S406 is not when being handled by second processor It determines by first processor processing target business, therefore, between server and first processor establishes long connection.Wherein, first The heartbeat that processor can establish low-power consumption between first processor first is connect, and then, is connected by the heartbeat established First processor is notified to establish and the long connection between server.
Then, in S608, server obtains the request results of target service according to target service identified in S405 And return to first processor.
In S609, for first processor, when receiving asking for the target service transmitted by the server in S607 After seeking result, low power consumpting state is exited, and be switched to normal operating conditions.
Optionally, after S609 second processor first processor is waken up, server can be disconnected with second processor Open long connection；Alternatively, first processor core can notify second processor core is disconnected to connect with the long of server.
And in subsequent S610, after first processor exits low power consumpting state, sent out according to server is received The request results for the target service sent, handle target service.The processing can refer to the description in S210, no longer superfluous It states.
Further, Figure 13 is the structural schematic diagram of one embodiment of terminal device provided by the present application, as shown in fig. 13 that In embodiment, a kind of terminal device that can be used for executing above-described embodiment is shown, in the terminal device, in addition to the first processing Device and second processor further include: wake up control module and pre-processing algorithm module.
Optionally, in a kind of concrete implementation mode, the wake-up control module and the pre-processing algorithm module are Two sections of independent program codes being stored in the storage equipment of terminal device, are locating for first processor and second processor It is called when managing voice data.Wherein, it when first processor is used to carry out waking up word identification to voice data, calls and wakes up control After module and pre-processing algorithm module are successively handled the voice data that MIC is received, then to treated voice data It carries out waking up word identification；When second processor be used for voice data carry out wake up word identification when, call wake up control module and After pre-processing algorithm module is successively handled the voice data that MIC is received, then voice data is called out to treated Word of waking up identifies.
Wherein, it is described wake up control module be used for working condition according to the state of terminal device, to first processor and The working condition of second processor switches over.
The pre-processing algorithm module, for the audio stream form received of the MIC to terminal device voice data into Row processing, and by treated, voice data is sent to second processor, is carried out waking up word identification by second processor.Before described The processing that Processing Algorithm module carries out voice data includes below one or more: voice speed adaption algorithm, frequency are adaptive It answers, orient pickup enhancing, phonetic feature aging algorithm, model ratio optimization, waking up model and MIC array and preferentially calculate automatically Method.
Specifically, when the state of terminal device is low power consumpting state, then first processor can be indicated by waking up control module It disconnects and being connected with the long of server, and indicate that second processor is established and connected with the long of the server；When terminal device exits Low power consumpting state, then wake-up module can indicate that first processor establishes the long connection with the server, and indicate at second Device disconnection is managed to connect with the long of the server.
It include: normal identification model and lightweight identification model in the pre-processing algorithm module, wherein the positive common sense Other model is obtained jointly by the voice data of each MIC in MIC array, and the lightweight identification model passes through in MIC array The voice data of one MIC obtains.It is understood that the occupied memory space of lightweight identification model be less than it is described just The other occupied memory space of model of common sense.
Then when the state of terminal device is low power consumpting state, control module is waken up by the lightweight in pre-processing algorithm module Identification model is loaded into second processor, so that second processor meets multiple MIC in MIC array by waking up control module After the voice data received carries out the processing of multichannel pickup, the wake-up word in voice data is known according to lightweight identification model Not.And after terminal device exits low power consumpting state, wake-up module loads the normal identification model in pre-processing algorithm module Into first processor, so that first processor obtains the voice that multiple MIC are received in MIC array by waking up control module After wake-up word in data is identified, the wake-up word in voice data is identified according to normal identification model.
In above-mentioned embodiment provided by the present application, handed over respectively between the network equipment, terminal and the network equipment and terminal Mutual angle is described method provided by the embodiments of the present application.In order to realize above-mentioned method provided by the embodiments of the present application In each function, the network equipment and terminal may include hardware configuration and/or software module, with hardware configuration, software module or Hardware configuration adds the form of software module to realize above-mentioned each function.Some function in above-mentioned each function is with hardware configuration, soft Part module or hardware configuration add the mode of software module to execute, the specific application and design constraint depending on technical solution Condition.
For example, can be used for realizing above-mentioned such as the structural schematic diagram that Figure 14 is one embodiment of terminal device provided by the present application The function of terminal device in any embodiment.Wherein, which can be the chip system in terminal device.The chip System can be made of chip, also may include chip and other discrete devices.Terminal device 1000 includes at least one processing Device, for example, first processor 1021 and second processor 1022.Processor in terminal device 1000 can be used for realizing the application The function of processor in the method that any of the above-described embodiment provides.
Terminal device 1000 can also include at least one processor 1030, for storing program instruction and/or data.It deposits Reservoir 1030 is coupled with first processor 1021, second processor 1022.Coupling in the embodiment of the present application be device, unit or Indirect coupling or communication connection between module, can be electrical property, and mechanical or other forms are used for device, unit or module Between information exchange.First processor 1021 may be possible with 1030 cooperating of memory, such as first processor 1021 Execute the program instruction stored in memory 1030.Second processor 1022 may be with 1030 cooperating of memory, such as One processor 1022 may execute the program instruction stored in memory 1030.In at least one processor 1030 at least One may include in first processor 1021 and/or at least one of at least one processor 1030 may include In second processor 1022.
Terminal device 1000 can also include the first communication interface 1011 and the second communication interface 1012, for passing through transmission Medium and other equipment are communicated, for the device in terminal device 1000 can be communicated with other equipment.Show Example property, which can be server.First processor 1021 can use 1011 sending and receiving data of the first communication interface, And for realizing method performed by first processor described in the aforementioned any embodiment of the application.Second processor 1022 can To utilize 1012 sending and receiving data of the second communication interface, and for realizing second processing described in the aforementioned any embodiment of the application Method performed by device.First communication interface 1011 and the second communication interface 1012 can be same in terminal device 1000 One communication interface.
Illustratively, if the terminal device 1000 can be used for executing terminal device in the embodiment as shown in Fig. 6-9 Performed method, then when first processor 1021 is in low power consumpting state, second processor 1022 can be used for passing through Mike Wind receives voice data from outside；Determine the requested target service of the voice data；Judge the target service whether by The second processor processing；When judging that the target service is handled by the second processor, pass through the second communication interface 1012 send the request for requesting the target service to server.Alternatively, second processor 1022 can be also used for passing through Second communication interface 1012 receives the request results of target service, and handles target service.Referring specifically to aforementioned implementation Exemplary detailed description in example, is not repeated herein.
Again illustratively, if the terminal device 1000 can be used for executing terminal in embodiment as illustrated in figs. 10-12 Method performed by equipment, then when first processor 1021 is in low power consumpting state, second processor 1022 can be used for passing through Microphone receives voice data from outside；If it is determined that including keyword in the voice data, then pass through the second communication interface Voice data is sent to server by 1012.Alternatively, second processor 1022 can be also used for through the second communication interface 1012 The request results of target service are received, and target service is handled.For details, reference can be made to exemplary detailed in previous embodiment Description, is not repeated herein.
The specific connection medium between above-mentioned communication interface, processor and memory is not limited in the embodiment of the present application. The embodiment of the present application is being connected in Figure 14 with passing through bus 1040 between memory, processor and communication interface, and bus is being schemed It is indicated in 14 with thick line, the connection type between other components is only to be schematically illustrated, does not regard it as and be limited.It is described total Line can be divided into address bus, data/address bus, control bus etc..Only to be indicated with a thick line in Figure 14, but simultaneously convenient for indicating Only a bus or a type of bus are not indicated.
In the embodiment of the present application, processor can be general processor, digital signal processor, specific integrated circuit, Field programmable gate array or other programmable logic device, discrete gate or transistor logic, discrete hardware components, It may be implemented or execute disclosed each method, step and the logic diagram in the embodiment of the present application.General processor can be Microprocessor or any conventional processor etc..The step of method in conjunction with disclosed in the embodiment of the present application, can directly embody Execute completion for hardware processor, or in processor hardware and software module combination execute completion.
In the embodiment of the present application, memory can be nonvolatile memory, such as hard disk (hard disk drive, HDD) or solid state hard disk (solid-state drive, SSD) etc., it can also be volatile memory (volatile ), such as random access memory (random-access memory, RAM) memory.Memory can be used for carrying or deposit Store up the desired program code with instruction or data structure form and can be by any other medium of computer access, but not It is limited to this.Memory in the embodiment of the present application can also be circuit or other devices that arbitrarily can be realized store function, For storing program instruction and/or data.
It, can be wholly or partly by software, hardware, firmware or it is any in method provided by the embodiments of the present application Combination is to realize.When implemented in software, it can entirely or partly realize in the form of a computer program product.The meter Calculation machine program product includes one or more computer instructions.Load and execute on computers the computer program instructions When, it entirely or partly generates according to process or function described in the embodiment of the present invention.The computer can be general-purpose computations Machine, special purpose computer, computer network, the network equipment, user equipment or other programmable devices.The computer instruction can To store in a computer-readable storage medium, or computer-readable deposit from a computer readable storage medium to another Storage media transmission, for example, the computer instruction can pass through from a web-site, computer, server or data center Wired (such as coaxial cable, optical fiber, Digital Subscriber Line (digital subscriber line, abbreviation DSL)) or wireless (example Such as infrared, wireless, microwave) mode transmitted to another web-site, computer, server or data center.It is described Computer readable storage medium can be any usable medium that computer can access or include one or more available The data storage devices such as medium integrated server, data center.The usable medium can be magnetic medium (for example, floppy disk, Hard disk, tape), optical medium (for example, digital video disk (digital video disc, abbreviation DVD)) or semiconductor be situated between Matter (such as SSD) etc..
Obviously, those skilled in the art can carry out various modification and variations without departing from the model of the application to the application It encloses.In this way, if these modifications and variations of the application belong within the scope of the claim of this application and its equivalent technologies, then The application is also intended to include these modifications and variations.

Claims (21)

1. a kind of voice data processing apparatus characterized by comprising first processor and second processor；Wherein, described One processor connects the second processor, and the operation power consumption of the first processor is greater than the operation function of the second processor Consumption；When the first processor is in low power consumpting state, the second processor is used for:
Voice data is received from outside by microphone；
Determine the requested target service of the voice data；
Judge whether the target service is handled by the second processor；
When judging that the target service is handled by the second processor, send to server for requesting the target service Request, wherein the first processor maintains the low power consumpting state.
2. the apparatus according to claim 1, which is characterized in that the second processor is also used to:
After sending the request for requesting the target service to the server, the mesh that the server is sent is received The request results of mark business；
According to the request results of the target service, the target service is handled.
3. device according to claim 1 or 2, which is characterized in that
The second processor is also used to: when judging the target service not is handled by the second processor, waking up institute It states first processor and is in normal operating conditions；
The first processor being waken up is used to send to the server for requesting the request of the target service, and connects The request results for receiving the target service that the server is sent, according to the request results of the target service to the target Business is handled.
4. device according to claim 1 or 2, which is characterized in that
The second processor is also used to: when judging the target service not is handled by the second processor, Xiang Suoshu Server is sent for requesting the request of the target service, and indicates the server by the request results of the target service It is handed down to the first processor；
After the first processor is waken up by the second processor or the server, for being issued according to the server The request results of the target service target service is handled.
5. device according to claim 1-4, which is characterized in that the second processor judges the target industry Whether business is handled by the second processor, comprising:
The target service is matched with pre-set business；
According to matching result, determine whether the target service is handled by the second processor.
6. device according to claim 5, which is characterized in that
Business of the pre-set business processor operational capability required when including operation less than the first preset value；Alternatively,
Business of the pre-set business storage capacity required when including operation less than the second preset value.
7. device according to claim 5 or 6, which is characterized in that the pre-set business includes below one or more:
Inquire weather, query time, control household, play music, setting alarm clock, play music, encyclopaedia question and answer, using schedule, It inquired using calculator, festivals or holidays, translate, listen audiobook, listen cross-talk and listen radio station.
8. device according to claim 1-7, which is characterized in that
When the first processor is in low power consumpting state, the second processor and the server keep long connection.
9. device according to claim 1-7, which is characterized in that the second processor determines the voice number Include: according to requested target service
When including waking up word in the voice data, the requested target service of the voice data is determined.
10. -9 described in any item devices according to claim 1, which is characterized in that
The first processor is primary processor, and the second processor is coprocessor；Alternatively, the first processor is more First processor core in core processor, the second processor are the second processor core in the multi-core processor；
The operational capability of the first processor is greater than the operational capability of the second processor, alternatively, the first processor Storage capacity be greater than the second processor storage capacity.
11. -10 described in any item devices according to claim 1, which is characterized in that the low power consumpting state includes suspend mode shape State.
12. -11 described in any item devices according to claim 1, which is characterized in that described device is that chip or electronics are set It is standby.
13. a kind of voice data processing system characterized by comprising server and as described in claim any one of 1-11 Device.
14. a kind of voice data processing system characterized by comprising terminal device and server；Wherein, the terminal is set Standby includes: first processor and second processor, and the first processor connects the second processor, the first processor Operation power consumption be greater than the second processor operation power consumption；
The second processor is used for, and when the first processor is in low power consumpting state, passes through the wheat of the terminal device Gram wind receives voice data from outside；
The second processor is also used to, and when determining in the voice data including keyword, the voice data is sent to The server；
The server is used for, and is received and is determined the requested target service of the voice data；
The server is also used to, and judges whether the target service is handled by the second processor；
The server is also used to, when determining that the target service is handled by the second processor, Xiang Suoshu second processing Device handles the request results of the target service, wherein the first processor maintains the low power consumpting state.
15. system according to claim 14, which is characterized in that the low power consumpting state includes dormant state.
16. system according to claim 14 or 15, which is characterized in that
The server is also used to, when determine the target service not and be handled by the second processor when, wake up described the One processor is in normal operating conditions；
The server is also used to, and Xiang Suoshu first processor handles the request results of the target service；
The first processor being waken up is used for, and receives the request results for the target service that the server is sent, root The target service is handled according to the request results of the target service.
17. system according to claim 14 or 15, which is characterized in that
The server is also used to, when determining the target service not is handled by the second processor, Xiang Suoshu second Processor sends instruction information, and being used to indicate the target service not is handled by the second processor；
The server is also used to, and Xiang Suoshu first processor handles the request results of the target service；
The second processor is also used to, and according to the instruction information, is waken up the first processor and is in normal operating conditions And long connection is established with the server；
The first processor is used for, and the request results for the target service that the server is sent is received, according to the mesh The request results of mark business handle the target service.
18. the described in any item systems of 4-17 according to claim 1, which is characterized in that the server judges the target industry Whether business is handled by the second processor, comprising:
The target service is matched with pre-set business；
According to matching result, determine whether the target service is handled by the second processor.
19. system according to claim 18, which is characterized in that
Business of the pre-set business processor operational capability required when including operation less than the first preset value；Alternatively,
Business of the pre-set business storage capacity required when including operation less than the second preset value.
20. the described in any item systems of 4-19 according to claim 1, which is characterized in that
When the first processor is in low power consumpting state, the second processor and the server keep long connection.
21. the described in any item systems of 4-20 according to claim 1, which is characterized in that
The first processor is primary processor, and the second processor is coprocessor；Alternatively, the first processor is more First processor core in core processor, the second processor are the second processor core in the multi-core processor；
The operational capability of the first processor is greater than the operational capability of the second processor, alternatively, the first processor Storage capacity be greater than the second processor storage capacity.
CN201910526214.5A 2019-06-18 2019-06-18 Voice data processing method, apparatus and system Pending CN110427097A (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201910526214.5A CN110427097A (en) 2019-06-18 2019-06-18 Voice data processing method, apparatus and system
PCT/CN2020/096545 WO2020253715A1 (en) 2019-06-18 2020-06-17 Voice data processing method, device and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910526214.5A CN110427097A (en) 2019-06-18 2019-06-18 Voice data processing method, apparatus and system

Publications (1)

Publication Number Publication Date
CN110427097A true CN110427097A (en) 2019-11-08

Family

ID=68407754

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910526214.5A Pending CN110427097A (en) 2019-06-18 2019-06-18 Voice data processing method, apparatus and system

Country Status (2)

Country Link
CN (1) CN110427097A (en)
WO (1) WO2020253715A1 (en)

Cited By (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111429911A (en) * 2020-03-11 2020-07-17 云知声智能科技股份有限公司 Method and device for reducing power consumption of speech recognition engine in noise scene
CN111755002A (en) * 2020-06-19 2020-10-09 北京百度网讯科技有限公司 Speech recognition device, electronic apparatus, and speech recognition method
WO2020253715A1 (en) * 2019-06-18 2020-12-24 华为技术有限公司 Voice data processing method, device and system
CN112382281A (en) * 2020-11-05 2021-02-19 北京百度网讯科技有限公司 Voice recognition method and device, electronic equipment and readable storage medium
CN112506331A (en) * 2020-12-11 2021-03-16 北京搜狗科技发展有限公司 Data processing method and earphone accommodating device
CN112581956A (en) * 2020-12-04 2021-03-30 海能达通信股份有限公司 Voice recognition method of dual-mode terminal and dual-mode terminal
CN112835826A (en) * 2021-03-04 2021-05-25 深圳市广和通无线股份有限公司 Communication method, device, equipment and readable storage medium
CN112968783A (en) * 2021-01-20 2021-06-15 广州技象科技有限公司 Low-power-consumption processing method and device based on transmitted data
CN112996089A (en) * 2019-12-17 2021-06-18 Oppo广东移动通信有限公司 Data transmission method, device, storage medium and electronic equipment
CN112992135A (en) * 2019-12-17 2021-06-18 Oppo广东移动通信有限公司 Electronic equipment and voice control display method
CN113269318A (en) * 2021-06-04 2021-08-17 安谋科技（中国）有限公司 Electronic device, neural network model operation method thereof and storage medium
CN114222062A (en) * 2021-12-13 2022-03-22 杭州萤石软件有限公司 Stream taking method, low-power-consumption battery equipment, client, stream taking system and equipment
CN114285892A (en) * 2021-08-26 2022-04-05 海信视像科技股份有限公司 Server, intelligent device and method for awakening intelligent device with screen
CN116828007A (en) * 2023-05-24 2023-09-29 广州汽车集团股份有限公司 Service issuing method and device, electronic equipment and storage medium

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140278435A1 (en) * 2013-03-12 2014-09-18 Nuance Communications, Inc. Methods and apparatus for detecting a voice command
CN105493180A (en) * 2013-08-26 2016-04-13 三星电子株式会社 Electronic device and method for voice recognition
CN108600219A (en) * 2018-04-23 2018-09-28 海信（广东）空调有限公司 A kind of sound control method and equipment
CN108877805A (en) * 2018-06-29 2018-11-23 上海与德通讯技术有限公司 Speech processes mould group and terminal with phonetic function

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110427097A (en) * 2019-06-18 2019-11-08 华为技术有限公司 Voice data processing method, apparatus and system

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140278435A1 (en) * 2013-03-12 2014-09-18 Nuance Communications, Inc. Methods and apparatus for detecting a voice command
CN105493180A (en) * 2013-08-26 2016-04-13 三星电子株式会社 Electronic device and method for voice recognition
CN108600219A (en) * 2018-04-23 2018-09-28 海信（广东）空调有限公司 A kind of sound control method and equipment
CN108877805A (en) * 2018-06-29 2018-11-23 上海与德通讯技术有限公司 Speech processes mould group and terminal with phonetic function

Cited By (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2020253715A1 (en) * 2019-06-18 2020-12-24 华为技术有限公司 Voice data processing method, device and system
CN112996089B (en) * 2019-12-17 2022-10-21 Oppo广东移动通信有限公司 Data transmission method, device, storage medium and electronic equipment
CN112996089A (en) * 2019-12-17 2021-06-18 Oppo广东移动通信有限公司 Data transmission method, device, storage medium and electronic equipment
CN112992135A (en) * 2019-12-17 2021-06-18 Oppo广东移动通信有限公司 Electronic equipment and voice control display method
CN111429911A (en) * 2020-03-11 2020-07-17 云知声智能科技股份有限公司 Method and device for reducing power consumption of speech recognition engine in noise scene
CN111755002B (en) * 2020-06-19 2021-08-10 北京百度网讯科技有限公司 Speech recognition device, electronic apparatus, and speech recognition method
CN111755002A (en) * 2020-06-19 2020-10-09 北京百度网讯科技有限公司 Speech recognition device, electronic apparatus, and speech recognition method
CN112382281A (en) * 2020-11-05 2021-02-19 北京百度网讯科技有限公司 Voice recognition method and device, electronic equipment and readable storage medium
CN112382281B (en) * 2020-11-05 2023-11-21 北京百度网讯科技有限公司 Voice recognition method, device, electronic equipment and readable storage medium
CN112581956A (en) * 2020-12-04 2021-03-30 海能达通信股份有限公司 Voice recognition method of dual-mode terminal and dual-mode terminal
CN112506331A (en) * 2020-12-11 2021-03-16 北京搜狗科技发展有限公司 Data processing method and earphone accommodating device
CN112968783A (en) * 2021-01-20 2021-06-15 广州技象科技有限公司 Low-power-consumption processing method and device based on transmitted data
CN112835826A (en) * 2021-03-04 2021-05-25 深圳市广和通无线股份有限公司 Communication method, device, equipment and readable storage medium
CN113269318A (en) * 2021-06-04 2021-08-17 安谋科技（中国）有限公司 Electronic device, neural network model operation method thereof and storage medium
CN114285892A (en) * 2021-08-26 2022-04-05 海信视像科技股份有限公司 Server, intelligent device and method for awakening intelligent device with screen
CN114285892B (en) * 2021-08-26 2023-10-31 海信视像科技股份有限公司 Server, intelligent device and awakening method of intelligent device with screen
CN114222062A (en) * 2021-12-13 2022-03-22 杭州萤石软件有限公司 Stream taking method, low-power-consumption battery equipment, client, stream taking system and equipment
CN116828007A (en) * 2023-05-24 2023-09-29 广州汽车集团股份有限公司 Service issuing method and device, electronic equipment and storage medium

Also Published As

Publication number Publication date
WO2020253715A1 (en) 2020-12-24

Similar Documents

Publication Publication Date Title
CN110427097A (en) Voice data processing method, apparatus and system
CN103716905B (en) Managing connectivity between wireless devices
CN107027141B (en) Information processing method, device and mobile terminal
CN107145329A (en) Apparatus control method, device and smart machine
CN107809793A (en) The wake-up control method and device of intelligent terminal
CN108566634A (en) Reduce method, apparatus and Baffle Box of Bluetooth that Baffle Box of Bluetooth continuously wakes up delay
US9891698B2 (en) Audio processing during low-power operation
WO2015081664A1 (en) Method, apparatus, device and system for controlling wireless network to be switched on/off
CN107145425B (en) Information processing method, device and mobile terminal
CN109949801A (en) A kind of smart home device sound control method and system based on earphone
CN109544183A (en) A kind of business consultation method and device
CN108922524A (en) Control method, system, device, Cloud Server and the medium of intelligent sound equipment
CN107731231A (en) A kind of method for supporting more high in the clouds voice services and a kind of storage device
CN112230877A (en) Voice operation method and device, storage medium and electronic equipment
CN110853644B (en) Voice wake-up method, device, equipment and storage medium
CN109741740A (en) Voice interactive method and device based on external trigger
CN109298775A (en) A kind of terminal device and task processing method
CN108566706A (en) flash lamp control method, device, terminal device and storage medium
CN107193707B (en) Information processing method, device and mobile terminal
CN106954191B (en) Broadcast transmission method, apparatus and terminal device
CN108563468A (en) A kind of method, apparatus and Baffle Box of Bluetooth of Baffle Box of Bluetooth data processing
CN107027160A (en) Information processing method, device and mobile terminal
CN109511139A (en) WIFI control method, device, mobile device, computer readable storage medium
CN110046033A (en) Applied program processing method and device, electronic equipment, computer readable storage medium
CN109819297A (en) A kind of method of controlling operation thereof and set-top box

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination