A volume adaptive method and device in speech recognition scenario

Abstract

A volume adaptation method and system in a speech recognition scenario, extracting voiceprint features from user voices, distinguishing user objects based on the voiceprint features; obtaining attribute information of a microphone, and saving the attribute information of the microphone to a first database; constructing a user account-microphone information-volume optimal value table in the first database, and recording the relationship between the user account, microphone information, and volume optimal value through the user account-microphone information-volume optimal value table; adapting the microphone recording volume according to the user's voiceprint features, user account, microphone information, and volume optimal value. At the same time, the microphone recording volume is trained for a given user to obtain the optimal value of the microphone recording volume. The present invention forms relational data of users, microphones, and optimal values ​​through voiceprint technology and microphone recording volume training, thereby improving the accuracy of speech recognition; saving a large number of technicians from repeated setting work; and realizing microphone data update and accumulation.

Description

Technical Field

[0001] The present invention relates to the field of speech recognition technology, and in particular to a volume adaptation method and device in a speech recognition scenario. Background Art

[0002] The microphone recording volume determines the amplitude and quality of the voice data, which in turn affects the speech recognition rate. Setting the microphone recording volume too low will result in a low amplitude voice data. The microphone's own device noise and environmental noise will then be superimposed on the valid voice data, reducing the recognition rate. Setting the microphone recording volume too high can cause two problems: First, it increases the sound reception distance, collecting distant noise and intermittent sounds, reducing the recognition rate; second, excessively high microphone recording volume can produce reverberation and plosive sounds.

[0003] Therefore, when users use the voice recognition system, the following problems and inconveniences may occur:

[0004] First, when users first use the voice recognition system, they need to manually adjust the microphone recording volume, usually starting from a low recording volume value and adjusting it upwards while testing the voice recognition effect until a satisfactory result is achieved. However, users often ignore this step, resulting in low voice recognition rates.

[0005] Second, when rolling out speech recognition systems in batches, technicians often manually adjust each system individually, resulting in a significant amount of repetitive adjustments. Furthermore, because technicians' voice levels and microphone positioning differ from actual users, the recording volume preset by the technicians may not be suitable for actual users.

[0006] Third, when a microphone is used by multiple users, it needs to be adjusted frequently according to the users. Summary of the Invention

[0007] To this end, the present invention provides a volume adaptation method and device in a speech recognition scenario to solve the problems of low recognition rate caused by users ignoring microphone recording volume adjustment, a large amount of repetitive work required by technicians to initialize the recording volume, and frequent adjustments required for different users and different microphones.

[0008] In order to achieve the above object, the present invention provides the following technical solution: a volume adaptation method in a speech recognition scenario, comprising the following steps:

[0009] Extracting voiceprint features from user speech and distinguishing user objects based on the voiceprint features;

[0010] Acquiring attribute information of the microphone, and saving the attribute information of the microphone to a first database;

[0011] A user account-microphone information-volume optimal value table is constructed in the first database, and the relationship between the user account, microphone information and volume optimal value is recorded through the user account-microphone information-volume optimal value table;

[0012] Adapt the microphone recording volume based on the user's voiceprint characteristics, user account, microphone information and optimal volume value.

[0013] As a preferred solution of the volume adaptation method in the speech recognition scenario, the differentiated user objects are marked, a user account is assigned to the marked user objects, and the voiceprint features and user accounts are stored in a second database;

[0014] A user account-voiceprint information table is constructed in the second database, and the user account and the corresponding voiceprint feature matrix are stored in the user account-voiceprint information table.

[0015] As a preferred solution for the volume adaptation method in the speech recognition scenario, the attribute information includes the microphone name, model, and device manufacturer, and the microphone information in the user account-microphone information-volume optimal value table is the attribute information of the microphone.

[0016] As a preferred solution of the volume adaptation method in the speech recognition scenario, it also includes constructing a microphone information-volume optimal value table in a third database, and recording the microphone information and the corresponding volume optimal value through the microphone information-volume optimal value table.

[0017] As a preferred solution for volume adaptation in speech recognition scenarios, when the microphone and user are known, the optimal microphone recording volume is adapted through the user account-microphone information-volume optimal value table;

[0018] When the microphone is known and the user is unknown, perform voiceprint feature extraction and user account creation for the unknown user;

[0019] When it is an unknown microphone + a known user, register the microphone into the microphone information-volume optimal value table and update the user account-microphone information-volume optimal value table;

[0020] When it is an unknown microphone + an unknown user, the user account-voiceprint information table, the microphone information-volume optimal value table and the user account-microphone information-volume optimal value table are updated.

[0021] As a preferred solution for the volume adaptation method in the speech recognition scenario, microphone recording volume training is performed on a given user to obtain the optimal microphone recording volume. The microphone recording volume training steps include:

[0022] Preset training text, which is part of the speech recognition model training set;

[0023] The training user orally recites the training text at a preset recording distance to obtain the recognition result, and compares it with the standard text to obtain the recognition rate;

[0024] When the recognition rate reaches a preset value, the corresponding microphone recording volume is taken as the optimal value to complete the microphone recording volume training.

[0025] The present invention also provides a volume adaptation device in a speech recognition scenario, comprising:

[0026] A voiceprint recognition module is used to extract voiceprint features from user speech and distinguish user objects based on the voiceprint features;

[0027] A microphone information acquisition module, configured to acquire attribute information of the microphone and save the attribute information of the microphone to a first database;

[0028] An information management module is configured to construct a user account-microphone information-volume optimal value table in the first database, and record the relationship between the user account, microphone information, and volume optimal value through the user account-microphone information-volume optimal value table;

[0029] The volume adaptation module is used to adapt the microphone recording volume according to the user's voiceprint characteristics, user account, microphone information and optimal volume value.

[0030] As a preferred solution of the volume adaptation device in the speech recognition scenario, it also includes an account allocation module for marking the distinguished user objects and assigning user accounts to the marked user objects;

[0031] The information management module saves the voiceprint features and user accounts through the second database; a user account-voiceprint information table is constructed in the second database, and the user account and the corresponding voiceprint feature matrix are saved through the user account-voiceprint information table;

[0032] In the microphone information acquisition module, the attribute information includes the microphone name, model, and device manufacturer, and the microphone information in the user account-microphone information-volume optimal value table is the attribute information of the microphone.

[0033] As a preferred solution for the volume adaptation device in the speech recognition scenario, the information management module further constructs a microphone information-volume optimal value table through a third database, and records the microphone information and the corresponding volume optimal value through the microphone information-volume optimal value table;

[0034] When the microphone and user are known, the optimal microphone recording volume is adapted based on the user account-microphone information-volume optimal value table;

[0035] When the microphone is known and the user is unknown, perform voiceprint feature extraction and user account creation for the unknown user;

[0036] When it is an unknown microphone + a known user, register the microphone into the microphone information-volume optimal value table and update the user account-microphone information-volume optimal value table;

[0037] When it is an unknown microphone + an unknown user, the user account-voiceprint information table, the microphone information-volume optimal value table and the user account-microphone information-volume optimal value table are updated.

[0038] As a preferred solution for the volume adaptation device in the speech recognition scenario, a recording volume training module is also included for training the microphone recording volume of a given user to obtain the optimal microphone recording volume. The recording volume training module includes:

[0039] The training text presetting submodule is used to pre-set training text, which belongs to the content of the speech recognition model training set;

[0040] The recognition rate acquisition submodule is used to enable the training user to orally recite the training text at a preset recording distance to obtain the recognition result, and compare it with the standard text to obtain the recognition rate;

[0041] In the optimal value acquisition submodule, when the user's recognition rate reaches a preset value, the corresponding microphone recording volume is used as the optimal value to complete the microphone recording volume training.

[0042] The present invention has the following advantages: extracting voiceprint features from user voices and distinguishing user objects based on the voiceprint features; obtaining attribute information of the microphone and saving the attribute information of the microphone to a first database; constructing a user account-microphone information-volume optimal value table in the first database, and recording the relationship between the user account, microphone information and volume optimal value through the user account-microphone information-volume optimal value table; adapting the microphone recording volume according to the user's voiceprint features, user account, microphone information and volume optimal value. At the same time, the microphone recording volume is trained for a given user to obtain the microphone recording volume optimal value. The present invention forms relational data of users, microphones and optimal values ​​through voiceprint technology and microphone recording volume training, allowing users to realize the automation of microphone volume optimal value setting without having to understand professional knowledge of speech recognition, thereby improving the accuracy of speech recognition; saving a large amount of repeated setting work of technical personnel; and completing the automatic update and accumulation of a large amount of microphone data. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the embodiments or the description of the prior art. Obviously, the drawings described below are merely exemplary, and those skilled in the art can derive other implementation drawings based on the provided drawings without inventive effort.

[0044] The structures, proportions, sizes, etc. illustrated in this specification are intended solely to complement the contents disclosed herein and to facilitate understanding and reading by persons skilled in the art. They are not intended to limit the conditions under which the present invention may be implemented and therefore have no substantive technical significance. Any structural modifications, changes in proportions, or adjustments in sizes, without affecting the efficacy and objectives of the present invention, shall remain within the scope of the technical contents disclosed herein.

[0045] Figure 1 A flow chart of a volume adaptation method in a speech recognition scenario provided in an embodiment of the present invention;

[0046] Figure 2 A schematic diagram of determining the training recognition rate in the volume adaptation method in a speech recognition scenario provided in an embodiment of the present invention;

[0047] Figure 3 Schematic diagram of a volume adaptation device in a speech recognition scenario provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The following describes the implementation of the present invention using specific embodiments. Those skilled in the art will readily understand the other advantages and benefits of the present invention from the disclosure herein. Obviously, the embodiments described are only a portion of the present invention, not all of it. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort are intended to fall within the scope of protection of the present invention.

[0049] Example 1

[0050] See also Figure 1 and Figure 2 Embodiment 1 of the present invention provides a volume adaptation method in a speech recognition scenario, comprising the following steps:

[0051] S1. Extracting voiceprint features from user speech and distinguishing user objects based on the voiceprint features;

[0052] S2. Obtain attribute information of the microphone, and save the attribute information of the microphone to a first database;

[0053] S3. A user account-microphone information-volume optimal value table is constructed in the first database, and the relationship between the user account, microphone information and volume optimal value is recorded through the user account-microphone information-volume optimal value table;

[0054] S4. Adapt the recording volume of the microphone according to the user's voiceprint characteristics, user account, microphone information and optimal volume value.

[0055] In this embodiment, the differentiated user objects are marked, a user account is assigned to the marked user objects, and the voiceprint features and the user account are stored in a second database;

[0056] A user account-voiceprint information table is constructed in the second database, and the user account and the corresponding voiceprint feature matrix are stored in the user account-voiceprint information table.

[0057] Specifically, voiceprint features are the sound wave spectra that carry speech information, as detected by electroacoustic instruments. Voiceprints are not only specific but also relatively stable. Voiceprint features vary from user to user, and by extracting voiceprint features from user speech, different speakers can be distinguished. After voiceprint recognition is performed on different speakers, the user objects (i.e., speakers) are labeled and assigned unique user accounts, such as A001 or A002. The voiceprint features and user accounts are then stored in a second database.

[0058] In this embodiment, the attribute information includes the microphone name, model, and device manufacturer. The microphone information in the user account-microphone information-volume optimal value table is the attribute information of the microphone.

[0059] Specifically, the microphone information acquisition process can obtain the corresponding information of the microphone, including the microphone name, model, equipment manufacturer, etc., through the interface provided by the operating system, and save the microphone information to the user account-microphone information-volume optimal value table in the first database.

[0060] In this embodiment, the method further includes constructing a microphone information-volume optimal value table in the third database, and recording the microphone information and the corresponding volume optimal value in the microphone information-volume optimal value table.

[0061] Specifically, the User Account - Voiceprint Information Table, the User Account - Microphone Information - Optimal Volume Value Table, and the Microphone Information - Optimal Volume Value Table need to be managed and updated. For the User Account - Microphone Information - Optimal Volume Value Table, as shown in Table 1, user A001 uses a Huawei microphone (model i43), and the optimal volume value is 60%. User A002 uses a Philips microphone (model 3325), and the optimal volume value is 70%.

[0062] User Account Microphone Information Optimal volume A001 HUAWEI-i43 60％ A002 Philips-3325 70％

[0063] Table 1 User account-microphone information-volume optimal value table

[0064] The Microphone Information - Volume Optimization Table records microphone information and the corresponding optimal volume values, as shown in Table 2. The "Preset" field records whether the microphone's empirical optimal value is preset by the system (written during system deployment). The optimal value for OPPO microphones is captured and calculated after the system is used. As the system is deployed and used, more and more microphone data accumulates, forming a microphone knowledge base. This means that more and more microphones can be pre-set when new users are deployed.

[0065] Microphone Information Optimal volume Whether to preset HUAWEI-i43 50％ yes OPPO-1220 60％ no

[0066] Table 2 Microphone information - optimal volume value table

[0067] In this embodiment, when it is a known microphone + a known user, the optimal microphone recording volume is adapted through the user account-microphone information-volume optimal value table;

[0068] When the microphone is known and the user is unknown, perform voiceprint feature extraction and user account creation for the unknown user;

[0069] When it is an unknown microphone + a known user, register the microphone into the microphone information-volume optimal value table and update the user account-microphone information-volume optimal value table;

[0070] When it is an unknown microphone + an unknown user, the user account-voiceprint information table, the microphone information-volume optimal value table and the user account-microphone information-volume optimal value table are updated.

[0071] Specifically, the accumulated microphone information and optimal recording volume values ​​are written into the third database as the system's preset microphones. When the speech recognition system is activated, the microphone information is obtained. The user dictates the speech content to obtain the user's voiceprint information. This user's voiceprint information is then retrieved from the second database to obtain the user's account. Four possible combinations exist: known microphone + known user, known microphone + unknown user, unknown microphone + known user, and unknown microphone + unknown user.

[0072] If it is a known microphone + a known user, the optimal recording volume value is retrieved from the user account-microphone information-volume optimal value table in the first database, and the value is set to the microphone.

[0073] If the microphone is known and the user is unknown, a new user account is registered for the user and recorded in the User Account - Voiceprint Information Table and the User Account - Microphone Information - Optimal Volume Value Table. Next, optimal recording volume training is performed. Because the microphone is known, the optimal recording volume value for the microphone is obtained from the Microphone Information - Optimal Volume Value Table and set as the initial training value. Training is then performed, and the trained value is updated to the User Account - Microphone Information - Optimal Volume Value Table.

[0074] If the microphone is unknown and the user is known, register the microphone in the Microphone Information - Volume Optimization Value table and update the User Account - Microphone Information - Volume Optimization Value table. Set the initial training value to 50% and perform optimal recording volume training. The trained value is updated in the Microphone Information - Volume Optimization Value table and the User Account - Microphone Information - Volume Optimization Value table. Because this microphone is not pre-set in the system, its optimal value in the Microphone Information - Volume Optimization Value table needs to be calculated and updated using a mathematical average method, that is, the new value is mathematically averaged with the old value.

[0075] If the microphone and user are unknown, the user account-voiceprint information table, the microphone information-volume optimal value table, and the user account-microphone information-volume optimal value table are updated simultaneously. The initial training value is set to 50%, and optimal recording volume training is performed. The trained value is updated to the microphone information-volume optimal value table and the user account-microphone information-volume optimal value table. In the microphone information-volume optimal value table, because the microphone is not pre-set by the system, its optimal value needs to be calculated and updated using a mathematical average method, that is, the new value is mathematically averaged with the old value.

[0076] Furthermore, by regularly maintaining the microphone information-volume optimal value table, the non-system preset microphone information, optimal values, etc. that have accumulated a large amount of data are sorted out to form a microphone knowledge base.

[0077] In this embodiment, microphone recording volume training is performed on a given user to obtain an optimal microphone recording volume. The microphone recording volume training steps include:

[0078] Preset training text, which is part of the speech recognition model training set;

[0079] The training user orally recites the training text at a preset recording distance to obtain the recognition result, and compares it with the standard text to obtain the recognition rate;

[0080] When the recognition rate reaches a preset value, the corresponding microphone recording volume is taken as the optimal value to complete the microphone recording volume training.

[0081] See also Figure 2Specifically, microphone recording volume training is used to set the microphone recording volume to an appropriate value, or optimal value, tailored to the specific user. This is achieved as follows: First, a training text is pre-set. This training text is included in the speech recognition model training set. This text, when spoken under normal conditions, achieves a 100% recognition rate, such as "On the first day after surgery, the patient is in good general condition, with tolerable incision pain, no fever, no nausea or vomiting, no abdominal pain or bloating, and no flatulence or bowel movements." The user is instructed to recite this training text at their usual recording distance, speed, and volume. Recognition results are obtained and compared with standard text to determine the recognition rate. The optimal value is found by adjusting the recording volume. If the recognition rate reaches 100%, training is completed and recorded. If the recognition rate consistently fails to reach 100%, training is discontinued after five attempts. The recording volume value with the highest recognition rate is used as the optimal value. Different initial training values ​​are used for pre-set and non-pre-set microphones: the pre-set microphone is set to the empirically optimal value, while the non-pre-set microphone is set to 50%.

[0082] In summary, the present invention extracts voiceprint features from user speech and distinguishes user objects based on the voiceprint features; obtains attribute information of the microphone and saves the attribute information of the microphone in a first database; a user account-microphone information-volume optimal value table is constructed in the first database, and the relationship between the user account, microphone information and volume optimal value is recorded through the user account-microphone information-volume optimal value table; the recording volume of the microphone is adapted according to the user's voiceprint features, user account, microphone information and volume optimal value. At the same time, microphone recording volume training is performed on a given user to obtain the optimal value of microphone recording volume. The distinguished user objects are marked, user accounts are assigned to the marked user objects, and the voiceprint features and user accounts are saved through a second database; a user account-voiceprint information table is constructed in the second database, and the user account and the corresponding voiceprint feature matrix are saved through the user account-voiceprint information table. A microphone information-volume optimal value table is constructed in the third database, and the microphone information and the corresponding volume optimal value are recorded through the microphone information-volume optimal value table. The present invention uses voiceprint technology and microphone recording volume training to form relational data of users, microphones, and optimal values. This allows users to automatically set the optimal microphone volume value without having to understand professional knowledge of speech recognition, thereby improving the accuracy of speech recognition; saving a lot of repeated setting work for technicians; and simultaneously completing the automatic update and accumulation of a large amount of microphone data.

[0083] Example 2

[0084] See also Figure 3 Embodiment 2 of the present invention further provides a volume adaptation device in a speech recognition scenario, comprising:

[0085] Voiceprint recognition module 1, used to extract voiceprint features from user speech and distinguish user objects based on the voiceprint features;

[0086] Microphone information acquisition module 2, used to obtain attribute information of the microphone and save the attribute information of the microphone to the first database;

[0087] An information management module 3 is configured to construct a user account-microphone information-volume optimal value table in the first database, and record the relationship between the user account, microphone information, and volume optimal value through the user account-microphone information-volume optimal value table;

[0088] The volume adaptation module 4 is used to adapt the recording volume of the microphone according to the user's voiceprint characteristics, user account, microphone information and optimal volume value.

[0089] This embodiment further includes an account allocation module 5 for marking the differentiated user objects and allocating user accounts to the marked user objects;

[0090] The information management module 3 saves the voiceprint features and user accounts through the second database; a user account-voiceprint information table is constructed in the second database, and the user account and the corresponding voiceprint feature matrix are saved through the user account-voiceprint information table;

[0091] In the microphone information acquisition module 2, the attribute information includes the microphone name, model, and device manufacturer, and the microphone information in the user account-microphone information-volume optimal value table is the attribute information of the microphone.

[0092] In this embodiment, the information management module 3 further constructs a microphone information-volume optimal value table through a third database, and records the microphone information and the corresponding volume optimal value through the microphone information-volume optimal value table;

[0093] When the microphone and user are known, the optimal microphone recording volume is adapted based on the user account-microphone information-volume optimal value table;

[0094] When the microphone is known and the user is unknown, perform voiceprint feature extraction and user account creation for the unknown user;

[0095] When it is an unknown microphone + a known user, register the microphone into the microphone information-volume optimal value table and update the user account-microphone information-volume optimal value table;

[0096] When it is an unknown microphone + an unknown user, the user account-voiceprint information table, the microphone information-volume optimal value table and the user account-microphone information-volume optimal value table are updated.

[0097] In this embodiment, a recording volume training module 6 is further included, which is used to perform microphone recording volume training on a given user to obtain an optimal microphone recording volume. The recording volume training module 6 includes:

[0098] The training text presetting submodule 61 is used to presetting training text, which belongs to the content of the speech recognition model training set;

[0099] The recognition rate acquisition submodule 62 is used to enable the training user to orally recite the training text at a preset recording distance to obtain the recognition result, and compare it with the standard text to obtain the recognition rate;

[0100] In the optimal value acquisition submodule 63 , when the recognition rate reaches a preset value, the user takes the corresponding microphone recording volume as the optimal value to complete the microphone recording volume training.

[0101] It should be noted that the information interaction, execution process and other contents between the modules of the above-mentioned system are based on the same concept as the method embodiment in Example 1 of this application, and the technical effects they bring are the same as those of the method embodiment of this application. For specific contents, please refer to the description in the method embodiment shown above in this application, and no further details will be given here.

[0102] Example 3

[0103] Embodiment 3 of the present invention provides a non-transitory computer-readable storage medium, in which a program code for a volume adaptation method in a speech recognition scenario is stored. The program code includes instructions for executing the volume adaptation method in a speech recognition scenario of embodiment 1 or any possible implementation thereof.

[0104] Computer-readable storage media can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state drives (SSDs)).

[0105] Example 4

[0106] Embodiment 4 of the present invention provides an electronic device, including: a memory and a processor;

[0107] The processor and the memory communicate with each other through a bus; the memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to execute the volume adaptation method in the speech recognition scenario of Example 1 or any possible implementation thereof.

[0108] Specifically, the processor can be implemented by hardware or by software. When implemented by hardware, the processor can be a logic circuit, an integrated circuit, etc.; when implemented by software, the processor can be a general-purpose processor, which is implemented by reading software code stored in a memory. The memory can be integrated into the processor or located outside the processor and exist independently.

[0109] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website, computer, server or data center to another website, computer, server or data center via a wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode.

[0110] Obviously, those skilled in the art will appreciate that the various modules or steps of the present invention described above can be implemented using a general-purpose computing device, centralized on a single computing device, or distributed across a network of multiple computing devices. Alternatively, they can be implemented using program code executable by a computing device, which can then be stored in a storage device and executed by the computing device. In some cases, the steps shown or described can be performed in a different order than that shown, or can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0111] Although the present invention has been described in detail above using general descriptions and specific embodiments, it will be apparent to those skilled in the art that modifications and improvements may be made thereto. Therefore, such modifications and improvements, without departing from the spirit of the present invention, are intended to be within the scope of protection claimed herein.

Claims

1. A volume adaptation method in a speech recognition scenario, characterized in that: The following steps are involved: Extracting voiceprint features from user speech, and distinguishing user objects based on the voiceprint features; Acquire attribute information of the microphone, and save the attribute information of the microphone to a first database; A user account-microphone information-volume optimal value table is constructed in the first database, and the relationship between the user account, microphone information and volume optimal value is recorded through the user account-microphone information-volume optimal value table; Adapt the microphone recording volume based on the user's voiceprint features, user account, microphone information and optimal volume value; Marking the distinguished user objects, assigning user accounts to the marked user objects, and saving the voiceprint features and the user accounts through a second database; A user account-voiceprint information table is constructed in the second database, and the user account and the corresponding voiceprint feature matrix are stored in the user account-voiceprint information table; The attribute information includes the microphone name, model, and equipment manufacturer. The microphone information in the user account-microphone information-volume optimal value table is the attribute information of the microphone; The method further includes constructing a microphone information-volume optimal value table in a third database, and recording the microphone information and the corresponding volume optimal value through the microphone information-volume optimal value table; When it is a known microphone + a known user, the optimal value of the microphone recording volume is adapted through the user account-microphone information-volume optimal value table; When it is a known microphone + unknown user, perform voiceprint feature extraction and user account creation for the unknown user; When it is an unknown microphone + a known user, register the microphone into the microphone information-volume optimal value table, and update the user account-microphone information-volume optimal value table; When it is an unknown microphone + an unknown user, the user account-voiceprint information table, the microphone information-volume optimal value table and the user account-microphone information-volume optimal value table are updated.

2. The volume adaptation method in a speech recognition scenario according to claim 1, characterized in that: Perform microphone recording volume training for a given user to obtain the optimal microphone recording volume. The microphone recording volume training steps include: Preset training text, which belongs to the content of the speech recognition model training set; The training user orally recites the training text at a preset recording distance to obtain the recognition result, and compares it with the standard text to obtain the recognition rate; When the recognition rate reaches a preset value, the corresponding microphone recording volume is taken as the optimal value to complete the microphone recording volume training.

3. A volume adaptive device in a speech recognition scenario, characterized in that: include: A voiceprint recognition module is used to extract voiceprint features from user voices and distinguish user objects based on the voiceprint features; A microphone information acquisition module, used to acquire attribute information of the microphone and save the attribute information of the microphone to a first database; An information management module, configured to construct a user account-microphone information-volume optimal value table in the first database, and record the relationship between the user account, microphone information and volume optimal value through the user account-microphone information-volume optimal value table; The volume adaptation module is used to adapt the recording volume of the microphone according to the user's voiceprint characteristics, user account, microphone information and optimal volume value; It also includes an account allocation module, which is used to mark the differentiated user objects and allocate user accounts to the marked user objects; The information management module saves the voiceprint features and the user account through the second database; a user account-voiceprint information table is constructed in the second database, and the user account and the corresponding voiceprint feature matrix are saved through the user account-voiceprint information table; In the microphone information acquisition module, the attribute information includes the microphone name, model, and equipment manufacturer, and the microphone information in the user account-microphone information-volume optimal value table is the attribute information of the microphone; The information management module also constructs a microphone information-volume optimal value table through a third database, and records the microphone information and the corresponding volume optimal value through the microphone information-volume optimal value table; When it is a known microphone + a known user, the optimal value of the microphone recording volume is adapted through the user account-microphone information-volume optimal value table; When it is a known microphone + unknown user, perform voiceprint feature extraction and user account creation for the unknown user; When it is an unknown microphone + a known user, register the microphone into the microphone information-volume optimal value table, and update the user account-microphone information-volume optimal value table; When it is an unknown microphone + an unknown user, the user account-voiceprint information table, the microphone information-volume optimal value table and the user account-microphone information-volume optimal value table are updated.

4. The volume adaptive device in a speech recognition scenario according to claim 3, characterized in that: It also includes a recording volume training module, which is used to train the microphone recording volume of a given user to obtain the optimal value of the microphone recording volume. The recording volume training module includes: A training text presetting submodule is used to preset training text, which belongs to the content of the speech recognition model training set; The recognition rate acquisition submodule is used to enable the training user to orally recite the training text at a preset recording distance to obtain the recognition result, and compare it with the standard text to obtain the recognition rate; In the optimal value acquisition submodule, when the user's recognition rate reaches a preset value, the corresponding microphone recording volume is taken as the optimal value to complete the microphone recording volume training.

Images