Processing method and electronic equipment
A processing method and technology of electronic equipment, applied in the field of control, can solve the problems of reduced call quality, wrong target sound source, etc., and achieve the effect of improving call quality
Active Publication Date: 2019-09-20
LENOVO (BEIJING) CO LTD
15 Cites 3 Cited by
AI-Extracted Technical Summary
Problems solved by technology
[0003] However, when performing noise reduction processing, the loudest sound source is usually selected as the target sound source, and noise r...
Abstract
The invention discloses a processing method and electronic equipment. The processing method includes the steps of obtaining no less than one voiceprint information, identifying the no less than one voiceprint information, determining whether the no less than one voiceprint information includes voiceprint information matching voiceprint characteristics of a specific target voiceprint, and if first voiceprints in the no less than one voiceprint information match the voiceprint characteristics of the specific target voiceprint, and then carrying out a noise reduction process on voice information corresponding to other voiceprint information except the first voiceprints in the no less than one voiceprint information. By matching the identified no less than one voiceprint information to the voiceprint characteristics of the specific target voiceprint, the obtained specific target voiceprint can be confirmed so as to carry out the noise reduction process on the voice information corresponding to the other voiceprint information, and therefore, identification and selection for target voiceprints in voiceprints corresponding to obtained voice can be achieved, and call quality is improved.
Application Domain
Speech analysis
Technology Topic
Electric equipmentNoise reduction +2
Image
Examples
- Experimental program(1)
Example Embodiment
[0037] The technical solutions in the embodiments of the present application will be described clearly and completely in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of this application.
[0038] This application discloses a processing method, the flowchart is as figure 1 Shown, including:
[0039] Step S11: Obtain no less than one voiceprint information;
[0040] The radio equipment can obtain sound information, and the sound information includes voiceprint information. In addition to voiceprint information, the sound information also includes other information, such as volume information, content information, etc. After acquiring the sound information, determine the voiceprint from it information.
[0041] One voice message may include one voiceprint information or multiple voiceprint information. Different voiceprint information is issued by different users, and the same voiceprint information is issued by the same user.
[0042] Among them, one voice message includes one voiceprint information, which can be specifically: the radio device acquires one or more voice messages, and each voice message only corresponds to the voice of one user, that is, each voice message includes only one voiceprint information;
[0043] One voice message includes multiple voiceprint information, which can be specifically: the radio device acquires one or more voice messages, and each voice message corresponds to the voices of multiple users, that is, each voice message may include multiple voiceprints information.
[0044] Step S12: Identify no less than one voiceprint information, and determine whether no less than one voiceprint information has voiceprint information that matches the voiceprint feature of the specific target voiceprint;
[0045] Step S13: If the first voiceprint in no less than one voiceprint information matches the voiceprint feature of the specific target voiceprint, then no less than one voiceprint information other than the first voiceprint information is correspondingly matched The sound information is processed for noise reduction.
[0046] Among them, the radio equipment can be specifically a microphone. When the microphone collects sound, as long as the user is not in a quiet environment, then the microphone may collect sounds other than the user's sound, and the user's sound corresponds to The voiceprint information of is the specific target voiceprint, and the sound collected by the microphone other than the user's voice is the sound that does not need to be collected.
[0047] Therefore, it is necessary to determine whether there is a user's voice in the currently acquired voice information, that is, to determine whether there is no less than one voiceprint information that matches the voiceprint feature of the specific target voiceprint. If so, it indicates There are sounds made by the user in the acquired sound information. If not, it means that there are no sounds made by the user in the acquired sound information, and the sound information acquired by the radio equipment is all useless sounds;
[0048] Or, what needs to be determined is whether there is a voice that is not the user's voice in the currently acquired voice information, that is, whether there is voiceprint information in no less than one voiceprint information that does not match the voiceprint feature of the specific target voiceprint. It must be determined in advance that there must be a user’s voice in the acquired sound information. Therefore, it is only necessary to determine the sound of a person or device other than the user. If there is, it means that the acquired sound information In addition to the sound made by the user, there are other sounds made by other people or devices. If there is no sound, it means that only the sound made by the user is included in the acquired sound information, and there are no other sounds. Then, at this time, there is no need to respond to the acquired sound. The information is processed for noise reduction and can be directly applied or transmitted.
[0049] No matter which of the above methods is adopted, as long as it is determined that the voiceprint information corresponding to the acquired sound contains voiceprint information other than the first voiceprint corresponding to the voice of the user, then the voiceprint information other than the first voiceprint The sound information corresponding to the information is subjected to noise reduction processing, so that the final voice is a sound that only has no noise reduction processing on the specific target voiceprint, which improves the call quality.
[0050] By determining the specific target voiceprint, so that after determining that the acquired voiceprint information includes the voiceprint information matching the voiceprint feature of the specific target voiceprint, only the voiceprint information is not denoised, that is, the first voice is not reduced. Noise reduction processing is performed on the grain, and noise reduction processing is performed on all the acquired voiceprint information except the first voiceprint.
[0051] Among them, the specific target voiceprint may be pre-stored in the electronic device, that is, the specific target voiceprint is pre-recorded through the radio device, so that in the follow-up, as long as the sound information acquired by the radio device has a voiceprint corresponding to the specific target voiceprint For voiceprint information with matching features, noise reduction processing is performed on the voice information corresponding to other voiceprint information simultaneously acquired by the electronic device;
[0052] In addition, the specific target voiceprint can also be automatically recognized by the electronic device.
[0053] Specifically, the electronic device acquires the first sound information, the content information corresponding to the first sound information is the first content, and the voiceprint information corresponding to the first sound information is the second voiceprint; the electronic device acquires the second sound information, the first content Second, the content information corresponding to the sound information is the second content, and the voiceprint information of the second sound information is the first voiceprint, where the first content and the second content are related content, then the second content corresponding to the second sound information A voiceprint is determined as a specific target voiceprint, so that noise reduction processing is performed on other voice information except the voice information corresponding to the first voiceprint in the acquired second voice information.
[0054] In this embodiment, the specific target voiceprint is determined according to the content of the sound information, so that noise reduction processing is performed on the voice information corresponding to other voiceprint information except the specific target voiceprint.
[0055] For example, in a conference call, content related to topic A is being discussed, then the specific target voiceprint should be the voiceprint of the sound information corresponding to the content related to topic A collected by the radio equipment. If user a is participating in the conference call at this time, user a is participating in the conference call in the office, and user a is not the only person in the office, and there are other people talking, when user a is in the conference call where topic A is discussed When speaking in the media, the content of his speech must be related to topic A, and the content of other people in the same office as user a must not be related to topic A. Therefore, the radio equipment collects user a and user a When other people in the same office say something, since only what user a said is related to topic A, the voiceprint information of the voice of user a is determined as the specific target voiceprint, and the electronic device collects the radio equipment All voices except the voice corresponding to the specific target voiceprint are processed for noise reduction; after the voiceprint information of user a’s voice is determined to be the specific target voiceprint, when user a speaks again, he can directly The voiceprint information corresponding to the voice uttered by the user a is locked, so that noise reduction processing is performed on the voice information corresponding to other voiceprint information.
[0056] The processing method disclosed in this embodiment obtains no less than one voiceprint information, recognizes no less than one voiceprint information, and determines whether no less than one voiceprint information has a voiceprint that matches the voiceprint feature of a specific target voiceprint Information, if the first voiceprint in no less than one voiceprint matches the voiceprint feature of the specific target voiceprint, then no less than one voiceprint information corresponding to other voiceprint information except the first voiceprint Perform noise reduction processing. This solution matches no less than one of the recognized voiceprint information with the voiceprint features of the specific target voiceprint to determine the acquired specific target voiceprint, and then perform noise reduction processing on the voice information corresponding to other voiceprint information , Realize the recognition and selection of the target voiceprint in the voiceprint corresponding to the acquired voice, and improve the call quality.
[0057] This embodiment discloses a processing method, and the flowchart is as figure 2 Shown, including:
[0058] Step S21: Acquire no less than one voice information through no less than one set of radio equipment, and determine that no less than one sound information corresponds to each voiceprint information in no less than one voiceprint information, the radio equipment and the sound One-to-one correspondence of pattern information;
[0059] One voice information contains only one voiceprint information, and the relationship between voice information and voiceprint information is one-to-one. Then, if there are N voiceprint information, there are corresponding N voice information; in addition, each group of radio equipment uses To obtain a piece of sound information, that is, there is a one-to-one correspondence between the radio equipment and the sound information.
[0060] Only one sound information corresponding to one voiceprint information is acquired through a group of radio equipment. Then, multiple radio equipment should be set in the same electronic device. During a sound collection process, each radio equipment is only used to collect one sound print. The sound message corresponding to the message.
[0061] When sound collection is required through the radio equipment, if the sound information corresponding to the multiple voiceprint information collected by the multiple groups of radio equipment in the electronic device contains the sound information required by the electronic device, then the collection electronic The sound information required by the equipment must be a group of radio equipment. Therefore, when the first voiceprint is determined according to the specific target voiceprint, that is, the radio equipment that collects the sound information corresponding to the first voiceprint is determined. If the radio equipment is determined to be the first radio equipment, it needs to be no less than The sound information corresponding to the voiceprint information acquired by the cash register devices other than the first radio device in a group of radio devices is subjected to noise reduction processing.
[0062] If it is determined that the radio device of the sound information corresponding to the first voiceprint is the first radio device, then the voiceprint information of the sound information collected by all radio devices except the first radio device is the same as the specific If the target voiceprint does not match, the sound information collected by other radio equipment except the first radio equipment is not the sound information required by the electronic device, but is noise. Therefore, other radio equipment other than the first radio equipment The collected sound information is processed for noise reduction.
[0063] Only the sound information corresponding to one voiceprint information is collected through a group of radio equipment, so that after the first voiceprint matching the specific target voiceprint is determined, there is no need to separate the voiceprint, and the sound information collected by different radio equipment can be directly used The voiceprint is different. Noise reduction processing is performed on the sounds corresponding to other voiceprints, which improves the accuracy and efficiency of noise reduction processing, and avoids the use of a set of radio equipment to collect all the voice information corresponding to the voiceprint information that can be obtained. Realizing the noise reduction processing requires a process of separating the voiceprint matching the specific target voiceprint.
[0064] Specifically, the L-shaped multi-stream denoising scheme can be used to capture the voices of multiple people, that is, the voices of each person are obtained through a set of radio equipment, and different human voices are obtained through different radio equipment, and then through different The audio channel recognizes the different audio streams obtained by different radio equipment, so that after determining the audio stream corresponding to the first voiceprint matching the specific target voiceprint, the audio channel corresponding to the audio stream can be directly locked, and then the The audio streams in other unlocked audio channels undergo noise reduction processing.
[0065] For example: such as image 3 As shown, there are three groups of Mic, each group of Mic is used to obtain one set of voiceprints, different Mic obtains different voiceprints, the first group of Mic obtains the first voiceprint of the first user, and the second group of Mic obtains the second The user’s second voiceprint and the third group of Mic acquire the third user’s third voiceprint. After the three groups of Mic acquire the voiceprint, they are transmitted and processed through the corresponding audio channels. The first voiceprint obtained by the group Mic is processed and recognized, the second voiceprint obtained by the second group of Mic is processed and recognized through the second channel 32, and the third voice obtained by the third group of Mic is processed and recognized through the third channel 33. The pattern is processed and identified.
[0066] After the three voiceprints are processed, the processor 34 first determines whether the first voiceprint matches the specific target voiceprint. If they match, the audio stream corresponding to the first voiceprint is directly fixed. The audio stream is subjected to noise reduction processing; if it is determined that the first voiceprint does not match the specific target voiceprint, it is determined whether the second voiceprint matches the specific target voiceprint, and if it matches, the audio stream corresponding to the second voiceprint is directly fixed. Perform noise reduction processing on audio streams corresponding to other voiceprints; if it is determined that the second voiceprint does not match the specific target voiceprint, determine whether the third voiceprint matches the specific target voiceprint, and if it matches, directly fix the third voice For the audio stream corresponding to the pattern, noise reduction is performed on the audio streams corresponding to other voiceprints; if it is determined that the third voiceprint does not match the specific target voiceprint, different voiceprints are obtained through three sets of Mic, so that the The voiceprint performs the above processing.
[0067] Step S22: Identify no less than one voiceprint information, and determine whether no less than one voiceprint information has voiceprint information that matches the voiceprint feature of the specific target voiceprint;
[0068] Step S23: If the first voiceprint in no less than one voiceprint information matches the voiceprint feature of the specific target voiceprint, then no less than one voiceprint information other than the first voiceprint information is correspondingly matched The sound information is processed for noise reduction.
[0069] The processing method disclosed in this embodiment obtains no less than one voiceprint information, recognizes no less than one voiceprint information, and determines whether no less than one voiceprint information has a voiceprint that matches the voiceprint feature of a specific target voiceprint Information, if the first voiceprint in no less than one voiceprint matches the voiceprint feature of the specific target voiceprint, then no less than one voiceprint information corresponding to other voiceprint information except the first voiceprint Perform noise reduction processing. This solution matches no less than one of the recognized voiceprint information with the voiceprint features of the specific target voiceprint to determine the acquired specific target voiceprint, and then perform noise reduction processing on the voice information corresponding to other voiceprint information , Realize the recognition and selection of the target voiceprint in the voiceprint corresponding to the acquired voice, and improve the call quality.
[0070] This embodiment discloses a processing method, and the flowchart is as Figure 4 Shown, including:
[0071] Step S41: Obtain at least one voiceprint information;
[0072] Step S42: Pre-store the voiceprint feature of the specific target voiceprint, and the specific target voiceprint is not less than one;
[0073] Step S43: Identify no less than one voiceprint information, and determine whether no less than one voiceprint information has voiceprint information that matches the voiceprint feature of a specific target voiceprint stored in advance;
[0074] Step S44: If at least two voiceprints in no less than one voiceprint information match the pre-stored voiceprint features of no less than one specific target voiceprint, determine at least two voiceprints that match the at least two voiceprints. Pre-stored priority order of specific target voiceprint;
[0075] Step S45: According to the priority order of the at least two specific target voiceprints, determine the first voiceprint corresponding to the first target voiceprint with higher priority among the at least two specific target voiceprints as the final target voiceprint, so as to correct No less than one voiceprint information corresponding to other voiceprint information except the first voiceprint information is subjected to noise reduction processing.
[0076] If the specific target voiceprint is a specific target voiceprint pre-stored in the electronic device, the specific target voiceprint pre-stored in the electronic device can be one, two, or more, and it will not be described here. The specific number is limited.
[0077] When there is one specific target voiceprint stored in advance, it can directly determine the voiceprint that matches the specific target voiceprint, and then perform noise reduction processing on the sound information corresponding to other unmatched voiceprints;
[0078] When there are multiple specific target voiceprints stored in advance, the final target voiceprint can be determined according to the priority order of the specific target voiceprints stored at the same time when the multiple specific target voiceprints are stored. Other voice information corresponding to the voice information is processed for noise reduction;
[0079] When there are multiple pre-stored specific target voiceprints, it can also be: multiple voiceprint information recognized by the electronic device matches the specific target voiceprint, then it can be determined that the current electronic device is in the current call. This time the radio equipment collects the voiceprint information corresponding to the sound information emitted before the sound information, determines the voiceprint information as the final target voiceprint, and performs sound information corresponding to other voiceprint information except the final target voiceprint. Noise reduction processing;
[0080] When there are multiple pre-stored specific target voiceprints, it may also be: multiple voiceprint information recognized by the electronic device matches the specific target voiceprint, then the topic can be determined by determining the topic being discussed in the current call Content, select the voiceprint information corresponding to the multiple voiceprint information that matches the specific target voiceprint and the voiceprint information corresponding to the content that matches the content of the topic currently under discussion as the final target voiceprint, and remove the final The voice information corresponding to the voiceprint information other than the target voiceprint is processed for noise reduction.
[0081] In addition, if the specific target voiceprint is automatically recognized by the electronic device, the recognized specific target voiceprint may also be one or more.
[0082] When there is one specific target voiceprint automatically identified, the voiceprint that matches the specific target voiceprint can be directly determined, and then the voice information corresponding to the other unmatched voiceprints is processed for noise reduction;
[0083] When there are multiple specific target voiceprints automatically identified, the final target voiceprint can be determined according to the time sequence of the acquired voice information, and the noise information corresponding to other voiceprint information except the final target voiceprint can be denoised. ;
[0084] Or, when there are multiple specific target voiceprints automatically identified, the voice information corresponding to the multiple specific target voiceprints can also be randomly arranged, and the voice information corresponding to the different specific target voiceprints are sequentially compared to other voiceprints. The sound information undergoes noise reduction processing, and then the electronic device sequentially outputs the multiple sound information in the sequence.
[0085] The processing method disclosed in this embodiment obtains no less than one voiceprint information, recognizes no less than one voiceprint information, and determines whether no less than one voiceprint information has a voiceprint that matches the voiceprint feature of a specific target voiceprint Information, if the first voiceprint in no less than one voiceprint matches the voiceprint feature of the specific target voiceprint, then no less than one voiceprint information corresponding to other voiceprint information except the first voiceprint Perform noise reduction processing. This solution matches no less than one of the recognized voiceprint information with the voiceprint features of the specific target voiceprint to determine the acquired specific target voiceprint, and then perform noise reduction processing on the voice information corresponding to other voiceprint information , Realize the recognition and selection of the target voiceprint in the voiceprint corresponding to the acquired voice, and improve the call quality.
[0086] This embodiment discloses an electronic device, and its structure diagram is as Figure 5 Shown, including:
[0087] The processor 51 and the radio equipment 52.
[0088] Wherein, the radio equipment 52 is used to obtain not less than one piece of voice information, and not less than one piece of voice information corresponds to not less than one voiceprint information one to one;
[0089] The processor 51 is used to identify not less than one voiceprint information in no less than one voice information, and determine whether there is no less than one voiceprint information that matches the voiceprint feature of a specific target voiceprint. If the first voiceprint in less than one voiceprint information matches the voiceprint feature of the specific target voiceprint, the voice information corresponding to the voiceprint information other than the first voiceprint in no less than one voiceprint information is reduced. Noise processing.
[0090] The radio equipment can obtain sound information, and the sound information includes voiceprint information. In addition to voiceprint information, the sound information also includes other information, such as volume information, content information, etc. After acquiring the sound information, determine the voiceprint from it information.
[0091] One voice message may include one voiceprint information or multiple voiceprint information. Different voiceprint information is issued by different users, and the same voiceprint information is issued by the same user.
[0092] Among them, one voice message includes one voiceprint information, which can be specifically: the radio device acquires one or more voice messages, and each voice message only corresponds to the voice of one user, that is, each voice message includes only one voiceprint information;
[0093] One voice message includes multiple voiceprint information, which can be specifically: the radio device acquires one or more voice messages, and each voice message corresponds to the voices of multiple users, that is, each voice message may include multiple voiceprints information.
[0094] Among them, the radio equipment can be specifically a microphone. When the microphone collects sound, as long as the user is not in a quiet environment, then the microphone may collect sounds other than the user's sound, and the user's sound corresponds to The voiceprint information of is the specific target voiceprint, and the sound collected by the microphone other than the user's voice is the sound that does not need to be collected.
[0095] Therefore, it is necessary to determine whether there is a user's voice in the currently acquired voice information, that is, to determine whether there is no less than one voiceprint information that matches the voiceprint feature of the specific target voiceprint. If so, it indicates There are sounds made by the user in the acquired sound information. If not, it means that there are no sounds made by the user in the acquired sound information, and the sound information acquired by the radio equipment is all useless sounds;
[0096] Or, what needs to be determined is whether there is a voice that is not the user's voice in the currently acquired voice information, that is, whether there is voiceprint information in no less than one voiceprint information that does not match the voiceprint feature of the specific target voiceprint. It must be determined in advance that there must be a user’s voice in the acquired sound information. Therefore, it is only necessary to determine the sound of a person or device other than the user. If there is, it means that the acquired sound information In addition to the sound made by the user, there are other sounds made by other people or devices. If there is no sound, it means that only the sound made by the user is included in the acquired sound information, and there are no other sounds. Then, at this time, there is no need to respond to the acquired sound. The information is processed for noise reduction and can be directly applied or transmitted.
[0097] No matter which of the above methods is adopted, as long as it is determined that the voiceprint information corresponding to the acquired sound contains voiceprint information other than the first voiceprint corresponding to the voice of the user, then the voiceprint information other than the first voiceprint The sound information corresponding to the information is subjected to noise reduction processing, so that the final voice is a sound that only has no noise reduction processing on the specific target voiceprint, which improves the call quality.
[0098] By determining the specific target voiceprint, so that after determining that the acquired voiceprint information includes the voiceprint information matching the voiceprint feature of the specific target voiceprint, only the voiceprint information is not denoised, that is, the first voice is not reduced. Noise reduction processing is performed on the grain, and noise reduction processing is performed on all the acquired voiceprint information except the first voiceprint.
[0099] Among them, the specific target voiceprint may be pre-stored in the electronic device, that is, the specific target voiceprint is pre-recorded through the radio device, so that in the follow-up, as long as the sound information acquired by the radio device has a voiceprint corresponding to the specific target voiceprint For voiceprint information with matching features, noise reduction processing is performed on the voice information corresponding to other voiceprint information simultaneously acquired by the electronic device;
[0100] In addition, the specific target voiceprint can also be automatically recognized by the electronic device.
[0101] Specifically, the electronic device acquires the first sound information, the content information corresponding to the first sound information is the first content, and the voiceprint information corresponding to the first sound information is the second voiceprint; the electronic device acquires the second sound information, the first content Second, the content information corresponding to the sound information is the second content, and the voiceprint information of the second sound information is the first voiceprint, where the first content and the second content are related content, then the second content corresponding to the second sound information A voiceprint is determined as a specific target voiceprint, so that noise reduction processing is performed on other voice information except the voice information corresponding to the first voiceprint in the acquired second voice information.
[0102] In this embodiment, the specific target voiceprint is determined according to the content of the sound information, so that noise reduction processing is performed on the voice information corresponding to other voiceprint information except the specific target voiceprint.
[0103] For example, in a conference call, content related to topic A is being discussed, then the specific target voiceprint should be the voiceprint of the sound information corresponding to the content related to topic A collected by the radio equipment. If user a is participating in the conference call at this time, user a is participating in the conference call in the office, and user a is not the only person in the office, and there are other people talking, when user a is in the conference call where topic A is discussed When speaking in the media, the content of his speech must be related to topic A, and the content of other people in the same office as user a must not be related to topic A. Therefore, the radio equipment collects user a and user a When other people in the same office say something, since only what user a said is related to topic A, the voiceprint information of the voice of user a is determined as the specific target voiceprint, and the electronic device collects the radio equipment All voices except the voice corresponding to the specific target voiceprint are processed for noise reduction; after the voiceprint information of user a’s voice is determined to be the specific target voiceprint, when user a speaks again, he can directly The voiceprint information corresponding to the voice uttered by the user a is locked, so that noise reduction processing is performed on the voice information corresponding to other voiceprint information.
[0104] Further, the processor 51 recognizes not less than one voiceprint information in not less than one voice information, including:
[0105] No less than one piece of sound information is obtained through no less than one set of radio equipment, and the processor determines that no less than one piece of sound information corresponds to each of no less than one piece of voiceprint information.
[0106] One voice information contains only one voiceprint information, and the relationship between voice information and voiceprint information is one-to-one. Then, if there are N voiceprint information, there are corresponding N voice information; in addition, each group of radio equipment uses To obtain a piece of sound information, that is, there is a one-to-one correspondence between the radio equipment and the sound information.
[0107] Only one sound information corresponding to one voiceprint information is acquired through a group of radio equipment. Then, multiple radio equipment should be set in the same electronic device. During a sound collection process, each radio equipment is only used to collect one sound print. The sound message corresponding to the message.
[0108] When sound collection is required through the radio equipment, if the sound information corresponding to the multiple voiceprint information collected by the multiple groups of radio equipment in the electronic device contains the sound information required by the electronic device, then the collection electronic The sound information required by the equipment must be a group of radio equipment. Therefore, when the first voiceprint is determined according to the specific target voiceprint, that is, the radio equipment that collects the sound information corresponding to the first voiceprint is determined. If the radio equipment is determined to be the first radio equipment, it needs to be no less than The sound information corresponding to the voiceprint information acquired by the cash register devices other than the first radio device in a group of radio devices is subjected to noise reduction processing.
[0109] If it is determined that the radio device of the sound information corresponding to the first voiceprint is the first radio device, then the voiceprint information of the sound information collected by all radio devices except the first radio device is the same as the specific If the target voiceprint does not match, the sound information collected by other radio equipment except the first radio equipment is not the sound information required by the electronic device, but is noise. Therefore, other radio equipment other than the first radio equipment The collected sound information is processed for noise reduction.
[0110] Only the sound information corresponding to one voiceprint information is collected through a group of radio equipment, so that after the first voiceprint matching the specific target voiceprint is determined, there is no need to separate the voiceprint, and the sound information collected by different radio equipment can be directly used The voiceprint is different. Noise reduction processing is performed on the sounds corresponding to other voiceprints, which improves the accuracy and efficiency of noise reduction processing, and avoids the use of a set of radio equipment to collect all the voice information corresponding to the voiceprint information that can be obtained. Realizing the noise reduction processing requires a process of separating the voiceprint matching the specific target voiceprint.
[0111] Specifically, the L-shaped multi-stream denoising scheme can be used to capture the voices of multiple people, that is, the voices of each person are obtained through a set of radio equipment, and different human voices are obtained through different radio equipment, and then through different The audio channel recognizes the different audio streams obtained by different radio equipment, so that after determining the audio stream corresponding to the first voiceprint matching the specific target voiceprint, the audio channel corresponding to the audio stream can be directly locked, and then the The audio streams in other unlocked audio channels undergo noise reduction processing.
[0112] For example: such as image 3 As shown, there are three groups of Mic, each group of Mic is used to obtain one set of voiceprints, different Mic obtains different voiceprints, the first group of Mic obtains the first voiceprint of the first user, and the second group of Mic obtains the second The user’s second voiceprint and the third group of Mic acquire the third user’s third voiceprint. After the three groups of Mic acquire the voiceprint, they are transmitted and processed through the corresponding audio channels. The first voiceprint obtained by the group Mic is processed and recognized, the second voiceprint obtained by the second group of Mic is processed and recognized through the second channel 32, and the third voice obtained by the third group of Mic is processed and recognized through the third channel 33. The pattern is processed and identified.
[0113] After the three voiceprints are processed, the processor 34 first determines whether the first voiceprint matches the specific target voiceprint. If they match, the audio stream corresponding to the first voiceprint is directly fixed. The audio stream is subjected to noise reduction processing; if it is determined that the first voiceprint does not match the specific target voiceprint, it is determined whether the second voiceprint matches the specific target voiceprint, and if it matches, the audio stream corresponding to the second voiceprint is directly fixed. Perform noise reduction processing on audio streams corresponding to other voiceprints; if it is determined that the second voiceprint does not match the specific target voiceprint, determine whether the third voiceprint matches the specific target voiceprint, and if it matches, directly fix the third voice For the audio stream corresponding to the pattern, noise reduction is performed on the audio streams corresponding to other voiceprints; if it is determined that the third voiceprint does not match the specific target voiceprint, different voiceprints are obtained through three sets of Mic, so that the The voiceprint performs the above processing.
[0114] Further, the processor 51 determines that the first voiceprint in no less than one voiceprint information matches with the voiceprint feature of the specific target voiceprint, and then it determines that no less than one voiceprint information other than the first voiceprint The sound information corresponding to the pattern information is processed for noise reduction, including:
[0115] If the processor 51 determines that at least two voiceprints in no less than one voiceprint information match the pre-stored voiceprint features of no less than one specific target voiceprint, it determines that at least two voiceprints match the at least two voiceprints. According to the priority order of at least two specific target voiceprints, the first voiceprint corresponding to the first target voiceprint with higher priority among the at least two specific target voiceprints It is determined as the final target voiceprint, so that no less than one voiceprint information corresponding to other voiceprint information except the first voiceprint information can be denoised.
[0116] If the specific target voiceprint is a specific target voiceprint pre-stored in the electronic device, the specific target voiceprint pre-stored in the electronic device can be one, two, or more, and it will not be described here. The specific number is limited.
[0117] When there is one specific target voiceprint stored in advance, it can directly determine the voiceprint that matches the specific target voiceprint, and then perform noise reduction processing on the sound information corresponding to other unmatched voiceprints;
[0118] When there are multiple specific target voiceprints stored in advance, the final target voiceprint can be determined according to the priority order of the specific target voiceprints stored at the same time when the multiple specific target voiceprints are stored. Other voice information corresponding to the voice information is processed for noise reduction;
[0119] When there are multiple pre-stored specific target voiceprints, it can also be: multiple voiceprint information recognized by the electronic device matches the specific target voiceprint, then it can be determined that the current electronic device is in the current call. This time the radio equipment collects the voiceprint information corresponding to the sound information emitted before the sound information, determines the voiceprint information as the final target voiceprint, and performs sound information corresponding to other voiceprint information except the final target voiceprint. Noise reduction processing;
[0120] When there are multiple pre-stored specific target voiceprints, it may also be: multiple voiceprint information recognized by the electronic device matches the specific target voiceprint, then the topic can be determined by determining the topic being discussed in the current call Content, select the voiceprint information corresponding to the multiple voiceprint information that matches the specific target voiceprint and the voiceprint information corresponding to the content that matches the content of the topic currently under discussion as the final target voiceprint, and remove the final The voice information corresponding to the voiceprint information other than the target voiceprint is processed for noise reduction.
[0121] In addition, if the specific target voiceprint is automatically recognized by the electronic device, the recognized specific target voiceprint may also be one or more.
[0122] When there is one specific target voiceprint automatically identified, the voiceprint that matches the specific target voiceprint can be directly determined, and then the voice information corresponding to the other unmatched voiceprints is processed for noise reduction;
[0123] When there are multiple specific target voiceprints automatically identified, the final target voiceprint can be determined according to the time sequence of the acquired voice information, and the noise information corresponding to other voiceprint information except the final target voiceprint can be denoised. ;
[0124] Or, when there are multiple specific target voiceprints automatically identified, the voice information corresponding to the multiple specific target voiceprints can also be randomly arranged, and the voice information corresponding to the different specific target voiceprints are sequentially compared to other voiceprints. The sound information undergoes noise reduction processing, and then the electronic device sequentially outputs the multiple sound information in the sequence.
[0125] The electronic device disclosed in this embodiment obtains no less than one voiceprint information, recognizes no less than one voiceprint information, and determines whether no less than one voiceprint information has a voiceprint that matches the voiceprint feature of a specific target voiceprint Information, if the first voiceprint in no less than one voiceprint matches the voiceprint feature of the specific target voiceprint, then no less than one voiceprint information corresponding to other voiceprint information except the first voiceprint Perform noise reduction processing. This solution matches no less than one of the recognized voiceprint information with the voiceprint features of the specific target voiceprint to determine the acquired specific target voiceprint, and then perform noise reduction processing on the voice information corresponding to other voiceprint information , Realize the recognition and selection of the target voiceprint in the voiceprint corresponding to the acquired voice, and improve the call quality.
[0126] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments, and the same or similar parts between the various embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method part.
[0127] Professionals can further realize that the units and algorithm steps of the examples described in the embodiments disclosed in this article can be implemented by electronic hardware, computer software, or a combination of both, in order to clearly illustrate the possibilities of hardware and software. Interchangeability, in the above description, the composition and steps of each example have been described generally in terms of function. Whether these functions are executed by hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods for each specific application to implement the described functions, but such implementation should not be considered beyond the scope of this application.
[0128] The steps of the method or algorithm described in combination with the embodiments disclosed herein can be directly implemented by hardware, a software module executed by a processor, or a combination of the two. The software module can be placed in random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, removable disks, CD-ROMs, or all areas in the technical field. Any other known storage media.
[0129] The above description of the disclosed embodiments enables those skilled in the art to implement or use this application. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined in this document can be implemented in other embodiments without departing from the spirit or scope of the application. Therefore, this application will not be limited to the embodiments shown in this text, but should conform to the widest scope consistent with the principles and novel features disclosed in this text.
PUM


Description & Claims & Application Information
We can also present the details of the Description, Claims and Application information to help users get a comprehensive understanding of the technical details of the patent, such as background art, summary of invention, brief description of drawings, description of embodiments, and other original content. On the other hand, users can also determine the specific scope of protection of the technology through the list of claims; as well as understand the changes in the life cycle of the technology with the presentation of the patent timeline. Login to view more.
Similar technology patents
Method and apparatus for improving call quality of hands-free call device, and hands-free call device
ActiveUS9648154B1improve call qualitytransmission signal to noise ratio be low
Owner:QINGDAO GOERTEK
A call processing method and a mobile terminal
ActiveCN109788142AImprove call quality
Owner:VIVO MOBILE COMM CO LTD
Skin touch type capacitor vibration pickup device
ActiveCN1946249AImprove noise immunityImprove call quality
Owner:山西太微电声科技有限公司
Classification and recommendation of technical efficacy words
- Improve call quality
Method for synchronously displaying subtitle in video telephone call
InactiveCN101500127AImprove call qualityimprove communication
Owner:TECHFAITH INTELLIGENT HANDSET TECH BEIJING
Apparatus and method for generating three-dimensional stereo sound in a mobile communication system
InactiveUS20050135629A1improve call quality
Owner:SAMSUNG ELECTRONICS CO LTD
Method and mobile terminal for amplifying conversation volume
Owner:GUANGDONG OPPO MOBILE TELECOMM CORP LTD
Mobile terminal and call switching method for multi-card multi-standby mobile terminal
InactiveCN104994487Areduce radiationImprove call quality
Owner:SHANGHAI YUDE COMM TECH