Audio data processing method, device and equipment, readable storage medium and product

Abstract

The invention provides an audio data processing method, device and equipment, a readable storage medium and a product, and relates to the field of artificial intelligence, in particular to the field of voice technologies and large models. According to the specific implementation scheme, a keyword input by a user is obtained, data processing is conducted on the keyword based on a preset large language model, and a target text related to the keyword is obtained; obtaining an audio synthesis parameter determined by the user for the target text; sending the target text and the audio synthesis parameters to a preset speech synthesis algorithm, and obtaining audio data fed back by the speech synthesis algorithm; and storing the audio data to a preset storage path, wherein the audio data is used for performing a performance test on the voice recognition technology. Therefore, diversified and high-quality audio data can be quickly generated.

Description

Technical Field

[0001] The present disclosure relates to speech technology and large models in artificial intelligence, and in particular to an audio data processing method, device, equipment, readable storage medium and product. Background Art

[0002] Speech recognition technology is widely used in many different scenarios, such as digital human scenarios and metaverse scenarios. In order to test the performance of speech recognition technology, a large amount of audio data is often required.

[0003] In the related art, audio data is generally obtained by manual recording by testers or by web crawling.

[0004] However, the audio data obtained by the above method often has a high acquisition cost and a relatively single data type. Summary of the invention

[0005] The present disclosure provides an audio data processing method, apparatus, device, readable storage medium and product for rapidly generating diverse and high-quality audio data.

[0006] According to a first aspect of the present disclosure, there is provided an audio data processing method, comprising:

[0007] Obtain keywords input by the user, perform data processing on the keywords based on a preset large language model, and obtain target text related to the keywords;

[0008] Acquiring audio synthesis parameters determined by the user for the target text;

[0009] Sending the target text and the audio synthesis parameters to a preset speech synthesis algorithm, and obtaining audio data fed back by the speech synthesis algorithm;

[0010] The audio data is stored in a preset storage path, and the audio data is used to perform a performance test on the speech recognition technology.

[0011] According to a second aspect of the present disclosure, there is provided an audio data processing device, comprising:

[0012] An acquisition module is used to acquire keywords input by a user, perform data processing on the keywords based on a preset large language model, and obtain target text related to the keywords;

[0013] A determination module, used to obtain audio synthesis parameters determined by the user for the target text;

[0014] A sending module, used to send the target text and the audio synthesis parameters to a preset speech synthesis algorithm, and obtain audio data fed back by the speech synthesis algorithm;

[0015] The storage module is used to store the audio data in a preset storage path, and the audio data is used to perform performance testing on the speech recognition technology.

[0016] According to a third aspect of the present disclosure, there is provided an electronic device, including:

[0017] at least one processor; and

[0018] a memory communicatively connected to the at least one processor; wherein,

[0019] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method described in the first aspect.

[0020] According to a fourth aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause the computer to execute the method described in the first aspect.

[0021] According to a fifth aspect of the present disclosure, a computer program product is provided, comprising: a computer program, wherein the computer program is stored in a readable storage medium, at least one processor of an electronic device can read the computer program from the readable storage medium, and the at least one processor executes the computer program so that the electronic device executes the method described in the first aspect.

[0022] The technology disclosed in the present invention solves the problems of high acquisition cost and relatively single data type in existing audio data acquisition methods.

[0023] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0025] Figure 1 A diagram of the system architecture on which the embodiments of the present disclosure are based;

[0026] Figure 2 A flowchart of an audio data processing method provided by an embodiment of the present disclosure;

[0027] Figure 3A flowchart of an audio data processing method provided by another embodiment of the present disclosure;

[0028] Figure 4 A schematic diagram of interface interaction provided by an embodiment of the present disclosure;

[0029] Figure 5 A flowchart of an audio data processing method provided by another embodiment of the present disclosure;

[0030] Figure 6 A schematic diagram of an application scenario provided for an embodiment of the present disclosure;

[0031] Figure 7 A schematic diagram of the structure of an audio data processing device provided in an embodiment of the present disclosure;

[0032] Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0033] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0034] The present disclosure provides an audio data processing method, device, equipment, readable storage medium and product, which are applied to speech technology and large models in the field of artificial intelligence to achieve the technical effect of quickly generating a large amount of high-quality audio data.

[0035] In the technical solution of the present disclosure, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0036] In order to enable readers to more deeply understand the implementation principle of the present disclosure, the following Figure 1-Figure 8 The embodiments of the present disclosure are further refined.

[0037] Currently, speech recognition technology has been widely used in various software, such as digital human and metaverse products. In order to test the speech recognition performance of speech recognition technology, a large amount of audio data needs to be obtained.

[0038] The current speech evaluation set mainly relies on manual recording by testers or online crawling solutions. However, the acquisition of speech evaluation sets using the above methods often requires professional recording equipment and technology as well as a recording environment, which is costly. In addition, it is difficult to obtain all possible languages, accents, intonations, and speech speeds, and the data type is relatively single, which cannot meet actual testing needs.

[0039] In the process of solving the above technical problems, the inventors discovered through research that in order to quickly generate diversified audio data, a large language model can be combined with TTS to generate audio data.

[0040] Optionally, keywords may be input into the large language model so that the large language model generates text content associated with the keywords based on preset prompt rules, and generates audio data based on the text content through TTS.

[0041] Furthermore, the text content may consist of multiple interrogative sentences. When generating text content, the large language model may add delimiters between interrogative sentences or paragraphs. For example, the delimiter may be \n. Thus, the subsequent TTS can perform content segmentation based on the delimiter to generate multiple audio data.

[0042] In addition, users can also set audio conversion parameters according to actual needs, where the audio conversion parameters include but are not limited to timbre parameters, pitch parameters, volume parameters, audio data format parameters, etc. Therefore, TTS can generate more diverse audio data based on the audio conversion parameters and text content.

[0043] Figure 1 This is a system architecture diagram based on the embodiments of the present disclosure, such as Figure 1 As shown, the embodiment of the present disclosure includes at least a server 11 and a server 12, wherein a client is provided in the server 11 to provide local services for users. The server 12 may be a cloud server, which is provided with a preset large language model and a speech synthesis algorithm, and can provide cloud services for users.

[0044] Based on the above system architecture, the client can obtain the keywords input by the user, transmit the keywords to the cloud server, and generate the target text through the large language model. The user can edit the target text on the client, and can also determine the audio synthesis parameters in the client. The target text and audio synthesis parameters are sent to the speech synthesis algorithm, so that the speech synthesis algorithm can generate audio data. After obtaining the audio data, the client can store the audio data locally, or send it to the cloud server for backup.

[0045] Figure 2 A flowchart of an audio data processing method provided by an embodiment of the present disclosure is shown in FIG. Figure 2 As shown, the method includes:

[0046] Step 201: Obtain keywords input by a user, perform data processing on the keywords based on a preset large language model, and obtain target text related to the keywords.

[0047] The execution subject of this embodiment is an audio data processing device. The audio data processing device can be coupled to a client. The client can provide local services to users. The client can communicate with a cloud server. The cloud server is preset with a large language model and a speech synthesis algorithm. Therefore, cloud services can be provided to users.

[0048] In this embodiment, in order to obtain more text content for audio generation, the user can enter keywords. For example, the keywords can be words determined by the user according to actual needs, such as the intelligent metaverse.

[0049] After obtaining the keywords, the data of the keywords can be processed based on the preset large language model to obtain the target text related to the keywords. The large language model can generate the target text based on the preset prompt rules. Compared with the manual text generation method, the large language model can generate richer and more comprehensive text content based on keywords to adapt to more speech recognition technology test scenarios.

[0050] As an implementable manner, the user may also generate the target text according to actual needs and input the target text into the client, and the present disclosure does not impose any limitation on this.

[0051] Step 202: Obtain audio synthesis parameters determined by the user for the target text.

[0052] In this embodiment, in order to generate richer and more diverse audio data to adapt to more different test scenarios, the user can also set the audio synthesis parameters according to actual needs.

[0053] The audio synthesis parameter includes but is not limited to at least one of a pitch parameter, a timbre parameter, a volume parameter, an audio format parameter, and a sound speed parameter.

[0054] Step 203: Send the target text and the audio synthesis parameters to a preset speech synthesis algorithm, and obtain audio data fed back by the speech synthesis algorithm.

[0055] In this embodiment, a speech synthesis algorithm may be pre-set in the cloud server. For example, the speech synthesis algorithm may be a text to speech (TTS) algorithm, which can intelligently convert text into a natural speech stream through the design of a neural network.

[0056] Furthermore, after obtaining the target text and the audio synthesis parameters set by the user, the target text and the audio synthesis parameters can be sent to a preset speech synthesis algorithm, so that the speech synthesis algorithm can generate audio data based on the target text and the audio synthesis parameters. The client can obtain the audio data fed back by the speech synthesis algorithm.

[0057] Step 204: store the audio data in a preset storage path, and the audio data is used to perform a performance test on the speech recognition technology.

[0058] In this embodiment, after obtaining the audio data fed back by the speech synthesis algorithm, a preset storage path can be stored according to actual needs, so that subsequent users can obtain the pre-stored audio data from the storage path to perform performance testing on the speech recognition technology based on the audio data.

[0059] As an implementable method, after obtaining the audio data fed back by the speech synthesis algorithm, the audio data can be displayed on a preset display page, and the user can trigger a play operation according to actual needs to play the audio data.

[0060] Optionally, a preset storage control may be displayed in the display interface, and the user may determine a storage path based on the storage control. The storage path includes a local storage path and a cloud storage path. In response to a storage operation triggered by the user after selecting a storage path, the audio data may be stored.

[0061] The audio data processing method provided in this embodiment pre-sets a large language model and an audio synthesis algorithm, so after obtaining the keywords input by the user, it can generate a more diverse target text covering more thinking directions based on the keywords. The audio synthesis parameters determined by the user are obtained, and the target text and the audio synthesis parameters are sent to the preset speech synthesis algorithm, so that the speech synthesis algorithm can generate audio data that better meets the actual application requirements based on the target text and the audio synthesis parameters. The difficulty of audio data acquisition is relatively low, and the audio parameters and audio content of the acquired audio data are more diverse, which can meet different speech recognition technology test scenarios.

[0062] Optionally, based on any of the above embodiments, step 201 includes:

[0063] Preset guide description information is input into the large language model, where the guide description information is used to guide the large language model to generate a target text associated with the keyword based on the keyword, where the target text consists of a preset number of text segments.

[0064] The keyword is input into the large language model, and a target text output by the large language model is obtained.

[0065] In this embodiment, in order to enable the large language model to generate target text that is more in line with actual needs, preset guidance description information can be input into the large language model in advance. The guidance description information is used to guide the large language model to generate target text associated with the keywords based on the keywords. The target text consists of a preset number of text fragments.

[0066] For example, the guiding description information may be: I will enter a keyword, and please generate ten question sentences related to the keyword based on the keyword, which is required to cover multiple thinking dimensions.

[0067] Based on the above guidance operation, when the keyword is intelligent metaverse, the large language model can generate ten interrogative sentences around the intelligent metaverse. For example, interrogative sentences include but are not limited to what is the intelligent metaverse; what functions does the intelligent metaverse have; how to use the intelligent metaverse; how is the intelligent metaverse different from the traditional metaverse; what application scenarios does the intelligent metaverse have; how does the intelligent metaverse provide a better user experience; what are the advantages of the intelligent metaverse, etc.

[0068] After obtaining the keyword input by the user, the keyword can be input into the large language model to obtain the target text output by the large language model. Therefore, the large language model can generate target text covering more thinking dimensions around the keyword.

[0069] The audio data processing method provided in this embodiment inputs the guide description information into the large language model in advance, so that the large language model can accurately generate target text that is more in line with actual application needs based on keywords.

[0070] Figure 3 A flowchart of an audio data processing method provided by another embodiment of the present disclosure is provided. Based on any of the above embodiments, step 202 includes:

[0071] Step 301: Display the target text and a preset first trigger control in a preset content display interface.

[0072] Step 302: In response to the user's trigger operation on the first trigger control, a parameter setting interface is displayed, wherein the parameter setting interface includes preset parameter determination controls, and the parameter determination controls include at least one of a pitch determination control, a timbre determination control, a volume determination control, an audio format determination control, and a speed determination control.

[0073] Step 303: Obtain the audio synthesis parameters determined by the user based on the parameter determination control.

[0074] In this embodiment, after obtaining the target text generated by the large prediction model, the target text can be displayed in a preset content display interface. Thus, the user can view the target text in the content display interface. In addition, the user can also edit the target text according to actual needs, wherein the editing operation includes but is not limited to operations such as adding text, deleting text, and adding separators.

[0075] Furthermore, in order to facilitate the setting of audio synthesis parameters, a first trigger control associated with the target text may be displayed in the content display interface. In response to the user's trigger operation on the first trigger control, a parameter setting interface is displayed, and the parameter setting interface includes preset parameter determination controls, and the parameter determination controls include at least one of a tone determination control, a timbre determination control, a volume determination control, an audio format determination control, and a speed determination control. The user can determine the audio synthesis parameters by triggering one or more of the above parameter determination controls.

[0076] For example, the pitch determination control may be associated with a parameter slider, and the user may slide the slider to determine the pitch parameter. Alternatively, the timbre determination control may be associated with multiple timbres to be selected, and the user may determine the timbre parameter by triggering any of the timbres to be selected.

[0077] Figure 4 A schematic diagram of the interface interaction provided by the embodiment of the present disclosure, such as Figure 4 As shown, a content input control 41 and a second trigger control 42 can be displayed, so the user can input keywords based on the content input control 41, for example, the keyword can be: intelligent metaverse. In response to the user's trigger operation on the second trigger control 42, it can jump to the content display interface 43, and display the target text 44 generated by the large language model based on the keyword and the preset first trigger control 45 in the content display interface 43. In response to the user's trigger operation on the first trigger control 45, a parameter setting interface 46 is displayed, and the parameter setting interface 46 includes a preset parameter determination control 47, and the parameter determination control 47 includes a tone determination control, a timbre determination control, a volume determination control, an audio format determination control, and a speed determination control. Therefore, the user can accurately determine the audio synthesis parameters based on the parameter determination control.

[0078] The audio data processing method provided in this embodiment displays a parameter setting interface based on a user's trigger operation, so that the user can determine the audio synthesis parameters according to actual application requirements in the parameter setting interface. Therefore, it is possible to generate more diversified audio data that better meets actual application requirements based on the audio synthesis parameters, thereby improving the content quality of the audio data.

[0079] Optionally, based on any of the above embodiments, the target text is composed of a preset number of text segments, and a separator is included between two adjacent text segments. Step 203 includes:

[0080] The target text and the audio synthesis parameters are sent to a preset speech synthesis algorithm through a preset calling interface, so that the speech synthesis algorithm segments the target text based on the separator, and asynchronously generates a preset number of audio data based on a preset number of text segments obtained by segmentation.

[0081] In this embodiment, the target text is composed of a preset number of text segments, and a separator is included between two adjacent text segments. The separator may be \n, or any symbol set by the user, which is not limited in the present disclosure.

[0082] Optionally, the speech synthesis algorithm may be called through a preset calling interface to generate audio data. Therefore, after obtaining the target text and audio synthesis parameters, the target text and audio synthesis parameters may be sent to the preset speech synthesis algorithm through the preset calling interface.

[0083] Furthermore, after the speech synthesis algorithm obtains the target text and the audio synthesis parameters, since the target text includes a separator, the speech synthesis algorithm can segment the target text based on the separator to obtain multiple text segments. Thus, the speech synthesis algorithm can generate a piece of audio data based on each text segment to obtain multiple audio data.

[0084] As an implementable approach, in order to improve the efficiency of audio data generation, after completing the segmentation operation on the target text, the speech synthesis algorithm can asynchronously perform the synthesis operation of the audio data based on the text segments.

[0085] The audio data processing method provided in this embodiment can generate multiple audio data based on a target text by adding separators to the target text. In addition, the asynchronous generation operation of the speech synthesis algorithm can effectively improve the generation efficiency of the audio data.

[0086] Further, based on any of the above embodiments, the separator is added between two adjacent text segments when the large language model generates the target text.

[0087] Alternatively, the separator is manually added by the user after determining the target position.

[0088] Step 201 also includes:

[0089] The target text is displayed in a preset content display interface.

[0090] In response to an adding operation triggered by a user on the target text, a target position currently selected by the user is determined, and the separator is added at the target position.

[0091] In this embodiment, the large language model can generate a preset number of text segments based on keywords. For example, when the keyword is intelligent metaverse, the large language model can generate ten interrogative sentences around the intelligent metaverse. For example, interrogative sentences include but are not limited to what is the intelligent metaverse; what functions does the intelligent metaverse have; how to use the intelligent metaverse; how is the intelligent metaverse different from the traditional metaverse; what application scenarios does the intelligent metaverse have; how does the intelligent metaverse provide a better user experience; what are the advantages of the intelligent metaverse, etc.

[0092] In order to segment different text segments to generate more audio data, the large language model can add a separator between two adjacent text segments. The separator can be \n, or any symbol set by the user, which is not limited in the present disclosure.

[0093] As an implementable method, the user can also add separators in the target text according to actual needs.

[0094] Optionally, the target text may be displayed in a preset content display interface. The user may determine the target position in the target text according to actual needs, and in response to the adding operation triggered by the user on the target text, determine the target position currently selected by the user, and add a separator at the target position. For example, the user may determine the target position by clicking the mouse, and add a separator at the target position by triggering the Enter key on the keyboard.

[0095] The audio data processing method provided in this embodiment adds delimiters according to the adding operation triggered by the user, so that the segmentation method of the target text can be more in line with the actual needs of the user, and can generate more audio data based on a piece of target text based on the adding operation, thereby improving the efficiency of audio data generation.

[0096] Figure 5 A flowchart of an audio data processing method provided by another embodiment of the present disclosure is provided. Based on any of the above embodiments, Figure 5 As shown, step 204 includes:

[0097] Step 501: Identify the content quality of the audio data.

[0098] Step 502: Store the audio data whose content quality meets the preset conditions in a preset storage path.

[0099] In this embodiment, after the audio data is generated, the content quality of the audio data can also be identified, and the audio data can be marked based on the identification result. Therefore, the audio data whose content quality meets the preset conditions can be stored in the preset storage path. In this way, the content quality of the currently stored audio data can be guaranteed to improve the test effect of the subsequent speech recognition technology test.

[0100] The audio data processing method provided in this embodiment can realize storage operation of better quality audio data and improve the content quality of the stored audio data by identifying the content quality of the audio data before storing the audio data.

[0101] Further, based on any of the above embodiments, step 501 includes:

[0102] The audio data is converted into verification text by a preset conversion algorithm.

[0103] The text similarity between the verification text and the target text is calculated.

[0104] The content quality corresponding to the audio data is determined based on the text similarity.

[0105] In this embodiment, in order to verify the content quality of the audio data, the audio data can be converted into verification text by a preset conversion algorithm. For example, the audio data can be played according to preset playback parameters, and a voice recognition operation can be performed based on the playback content to obtain the verification text.

[0106] Further, after obtaining the verification text associated with the audio data, the text similarity between the verification text and the target text generating the audio data can be calculated. Among them, any similarity calculation method can be used to calculate the text similarity, and the present disclosure does not limit this.

[0107] After determining the text similarity associated with the audio data, the content quality corresponding to the audio data can be determined based on the text similarity. For example, a similarity threshold can be pre-set. If the text similarity associated with the audio data is higher than the similarity threshold, it can be determined that the content quality of the audio data is high. Conversely, if the text similarity associated with the audio data is lower than the similarity threshold, it can be determined that the content quality of the audio data is poor.

[0108] The audio data processing method provided in this embodiment converts audio data into verification text, thereby calculating the text similarity between the verification text and the target text, and then accurately determining the content quality of the currently generated audio data based on the text similarity.

[0109] Further, based on any of the above embodiments, step 501 includes:

[0110] The audio data is played according to a plurality of preset different playing parameters, wherein the playing parameters include at least one of a playing volume and a playing speed.

[0111] The playback contents corresponding to the multiple different playback parameters are converted using a preset conversion algorithm to obtain multiple verification texts.

[0112] The text similarities between each verification text and the target text are calculated respectively to obtain text similarities corresponding to a plurality of different playback parameters.

[0113] The content quality corresponding to the audio data is determined based on the weight information corresponding to each playback parameter and the text similarity corresponding to each playback parameter.

[0114] In this embodiment, in order to further determine the comprehensive content quality of the audio data under different playback parameters, multiple different playback parameters can be preset. The audio data is played according to the preset multiple different playback parameters, wherein the playback parameters include at least one of the playback volume and the playback speed. The playback contents corresponding to the multiple different playback parameters are converted by a preset conversion algorithm to obtain multiple verification texts. For example, the audio data can be played according to the preset playback parameters, and a voice recognition operation can be performed based on the playback content to obtain a verification text.

[0115] Furthermore, the text similarity between each verification text and the target text can be calculated respectively to obtain the text similarity corresponding to multiple different playback parameters. Among them, different weight information can be set for different playback parameters. The weight information can be set by the user according to actual needs, or it can be set based on experience values, and the present disclosure does not limit this. The content quality corresponding to the audio data is determined based on the weight information corresponding to each playback parameter and the text similarity corresponding to each playback parameter.

[0116] For example, a similarity threshold may be pre-set. If the text similarity is higher than the similarity threshold, it may be determined that the content quality of the audio data is high. Conversely, if the text similarity is lower than the similarity threshold, it may be determined that the content quality of the audio data is poor.

[0117] The audio data processing method provided in this embodiment determines the text similarity between the verification text and the target text under different playback parameters, and determines the content quality corresponding to the audio data based on the weight information corresponding to each playback parameter and the text similarity corresponding to each playback parameter, thereby being able to more accurately calculate the content quality of the audio data.

[0118] Optionally, based on any of the above embodiments, step 204 includes:

[0119] The audio data is stored in a preset local storage path.

[0120] and / or,

[0121] The audio data is sent to a preset cloud storage path for data backup operation.

[0122] In this embodiment, the local service may be provided with a local data service. The local data service may be a MySQL database. Therefore, after generating the audio data, the user may store the audio data in the local data service according to actual needs.

[0123] However, in order to avoid the loss and damage of audio data caused by failure of local data service, the audio data can also be sent to a preset cloud storage path for data backup operation. Among them, the cloud service can use BOS storage service for data storage operation.

[0124] As an implementable manner, after the audio data is generated, a storage path selection control may be displayed, and the user may choose to store the audio data in a local storage path and / or a cloud storage path according to actual needs.

[0125] The audio data processing method provided in this embodiment can avoid the loss of audio data when a local storage path fails by storing the audio data in a local storage path and a cloud storage path respectively, thereby improving the data security of the audio data.

[0126] Further, based on any of the above embodiments, the method further includes:

[0127] A data sharing operation is performed based on all audio data in the local storage path and / or the cloud storage path.

[0128] In this embodiment, after the audio data is stored in the local storage path and / or the cloud storage path, a data sharing operation can be performed based on all the audio data in the local storage path and / or the cloud storage path.

[0129] Optionally, for each user of the local service, the generated audio data can be stored in a local storage path and / or a cloud storage path to achieve data sharing. When the audio data needs to be called for testing the speech recognition technology, the user can call all the audio data stored in the local storage path and / or the cloud storage path.

[0130] The audio data processing method provided in this embodiment performs data sharing operations based on all audio data in the local storage path and / or cloud storage path, thereby being able to aggregate and subsequently apply audio data generated by different users, thereby further expanding the amount of available audio data.

[0131] Figure 6 A schematic diagram of an application scenario provided by an embodiment of the present disclosure, such as Figure 6 As shown, the user can call the local service 61, and the local service 61 can communicate with the cloud service 62. The local service 61 includes a text module 63, and the user can input keywords by interacting with the text module 63. After obtaining the keywords, the text module 63 can transmit the keywords to the large language model 64 in the cloud service 62, so that the large language model 64 can generate target text based on the keywords and feed the target text back to the local service 61.

[0132] Furthermore, the target text can be transmitted to the speech conversion module 65 in the local service 61, and the speech conversion module 65 transmits the target text to the TTS speech service 66 in the cloud service 62, so the TTS speech service 66 can generate audio data based on the target text. The audio data is then sent to the storage module 67 in the local service 61, so the storage module 67 can store the audio data in the MySQL data service 68 in the local service 61, or the storage module 67 can store the audio data in the BOS storage service 69 in the cloud service 62.

[0133] Figure 7 A schematic diagram of the structure of an audio data processing device provided in an embodiment of the present disclosure, such as Figure 7 As shown, the device includes: an acquisition module 71, a determination module 72, a sending module 73 and a storage module 74. Among them, the acquisition module 71 is used to obtain the keywords input by the user, perform data processing on the keywords based on the preset large language model, and obtain the target text related to the keywords. The determination module 72 is used to obtain the audio synthesis parameters determined by the user for the target text. The sending module 73 is used to send the target text and the audio synthesis parameters to the preset speech synthesis algorithm, and obtain the audio data fed back by the speech synthesis algorithm. The storage module 74 is used to store the audio data in a preset storage path, and the audio data is used to perform performance testing on the speech recognition technology.

[0134] Further, based on any of the above embodiments, the acquisition module includes: an input unit, used to input preset guide description information to the large language model, the guide description information is used to guide the large language model to generate a target text associated with the keyword based on the keyword, and the target text is composed of a preset number of text segments. An acquisition unit, used to input the keyword into the large language model and acquire the target text output by the large language model.

[0135] Further, based on any of the above embodiments, the determination module includes: a display unit, which is used to display the target text and a preset first trigger control in a preset content display interface. A processing unit, which is used to display a parameter setting interface in response to the user's trigger operation on the first trigger control, wherein the parameter setting interface includes a preset parameter determination control, and the parameter determination control includes at least one of a tone determination control, a timbre determination control, a volume determination control, an audio format determination control, and a speed determination control. An acquisition unit, which is used to acquire the audio synthesis parameters determined by the user based on the parameter determination control.

[0136] Further, based on any of the above embodiments, wherein the target text is composed of a preset number of text segments, and a separator is included between two adjacent text segments. The sending module includes: a sending unit, which is used to send the target text and the audio synthesis parameter to a preset speech synthesis algorithm through a preset calling interface, so that the speech synthesis algorithm performs a segmentation operation on the target text based on the separator, and asynchronously generates a preset number of audio data based on the preset number of text segments obtained by segmentation.

[0137] Further, based on any of the above embodiments, the separator is added between two adjacent text segments when the large language model generates the target text. Alternatively, the separator is manually added by the user after determining the target position. The device also includes: a display module, which is used to display the target text in a preset content display interface. An adding module is used to determine the target position currently selected by the user in response to an adding operation triggered by the user for the target text, and add the separator at the target position.

[0138] Further, based on any of the above embodiments, the storage module includes: an identification unit for identifying the content quality of the audio data; and a storage unit for storing the audio data whose content quality meets a preset condition in a preset storage path.

[0139] Further, based on any of the above embodiments, the recognition unit includes: a conversion subunit, configured to convert the audio data into a verification text by a preset conversion algorithm; a calculation subunit, configured to calculate the text similarity between the verification text and the target text; and a determination subunit, configured to determine the content quality corresponding to the audio data based on the text similarity.

[0140] Further, on the basis of any of the above embodiments, the recognition unit also includes: a playback subunit, which is used to play the audio data according to a plurality of preset different playback parameters, wherein the playback parameters include at least one of the playback volume and the playback speed. A conversion subunit, which is used to convert the playback content corresponding to a plurality of different playback parameters through a preset conversion algorithm to obtain a plurality of verification texts. A calculation subunit, which is used to respectively calculate the text similarity between each verification text and the target text to obtain the text similarity corresponding to a plurality of different playback parameters. A determination subunit, which is used to determine the content quality corresponding to the audio data based on the weight information corresponding to each playback parameter and the text similarity corresponding to each playback parameter.

[0141] Further, based on any of the above embodiments, the storage module includes: a first storage unit, used to store the audio data in a preset local storage path.

[0142] And / or, a second storage unit is used to send the audio data to a preset cloud storage path for data backup operation.

[0143] Furthermore, based on any of the above embodiments, the device also includes: a sharing module, which is used to perform data sharing operations based on all audio data in the local storage path and / or the cloud storage path.

[0144] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0145] According to an embodiment of the present disclosure, the present disclosure further provides an electronic device, including:

[0146] at least one processor; and

[0147] a memory communicatively connected to the at least one processor; wherein,

[0148] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method described in any one of the above embodiments.

[0149] According to an embodiment of the present disclosure, the present disclosure further provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable the computer to execute the method described in any of the above embodiments.

[0150] According to an embodiment of the present disclosure, the present disclosure also provides a computer program product, which includes: a computer program, the computer program is stored in a readable storage medium, at least one processor of an electronic device can read the computer program from the readable storage medium, and at least one processor executes the computer program so that the electronic device executes the solution provided by any of the above embodiments.

[0151] Figure 8 A schematic diagram of the structure of an electronic device provided for an embodiment of the present disclosure. The electronic device 800 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0152] like Figure 8 As shown, the device 800 includes a computing unit 801, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 802 or a computer program loaded from a storage unit 808 into a random access memory (RAM) 803. In the RAM 803, various programs and data required for the operation of the device 800 can also be stored. The computing unit 801, the ROM 802, and the RAM 803 are connected to each other via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0153] A number of components in the device 800 are connected to the I / O interface 805, including: an input unit 806, such as a keyboard, a mouse, etc.; an output unit 807, such as various types of displays, speakers, etc.; a storage unit 808, such as a disk, an optical disk, etc.; and a communication unit 809, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 809 allows the device 800 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0154] The computing unit 801 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 801 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 801 performs the various methods and processes described above, such as the audio data processing method. For example, in some embodiments, the audio data processing method may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 808. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 800 via ROM 802 and / or communication unit 809. When the computer program is loaded into RAM 803 and executed by the computing unit 801, one or more steps of the audio data processing method described above may be performed. Alternatively, in other embodiments, the computing unit 801 may be configured to perform the audio data processing method in any other appropriate manner (e.g., by means of firmware).

[0155] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0156] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0157] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0158] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0159] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0160] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship between the client and the server is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services ("Virtual Private Server", or "VPS" for short). The server may also be a server of a distributed system, or a server combined with a blockchain.

[0161] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0162] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. An audio data processing method, comprising: Obtain keywords input by the user, perform data processing on the keywords based on a preset large language model, and obtain target text related to the keywords; Acquiring audio synthesis parameters determined by the user for the target text; Sending the target text and the audio synthesis parameters to a preset speech synthesis algorithm, and obtaining audio data fed back by the speech synthesis algorithm; The audio data is stored in a preset storage path, and the audio data is used to perform a performance test on the speech recognition technology.

2. The method according to claim 1, wherein: The step of performing data processing on the keyword based on a preset large language model to obtain a target text related to the keyword includes: Inputting preset guide description information into the large language model, the guide description information is used to guide the large language model to generate a target text associated with the keyword based on the keyword, the target text consisting of a preset number of text segments; The keyword is input into the large language model, and a target text output by the large language model is obtained.

3. The method according to claim 1, wherein: The obtaining of the audio synthesis parameters determined by the user for the target text includes: Displaying the target text and a preset first trigger control in a preset content display interface; In response to the user's triggering operation on the first trigger control, displaying a parameter setting interface, wherein the parameter setting interface includes a preset parameter determination control, and the parameter determination control includes at least one of a tone determination control, a timbre determination control, a volume determination control, an audio format determination control, and a speed determination control; Acquire the audio synthesis parameters determined by the user based on the parameter determination control.

4. The method according to claim 1, wherein: The target text is composed of a preset number of text segments, and a separator is included between two adjacent text segments; The step of sending the target text and the audio synthesis parameters to a preset speech synthesis algorithm and obtaining audio data fed back by the speech synthesis algorithm includes: The target text and the audio synthesis parameters are sent to a preset speech synthesis algorithm through a preset calling interface, so that the speech synthesis algorithm segments the target text based on the separator, and asynchronously generates a preset number of audio data based on a preset number of text segments obtained by segmentation.

5. According to the method of claim 4, the separator is added between two adjacent text segments when the large language model generates the target text; Alternatively, the separator is manually added by the user after determining the target position; After the keyword is processed based on the preset large language model to obtain the target text related to the keyword, the method further includes: Displaying the target text in a preset content display interface; In response to an adding operation triggered by the user on the target text, the target position currently selected by the user is determined, and the separator is added at the target position.

6. The method according to claim 1, wherein: The step of storing the audio data in a preset storage path includes: Performing an identification operation on the content quality of the audio data; The audio data whose content quality meets the preset conditions is stored in the preset storage path.

7. The method according to claim 6, wherein: The operation of identifying the content quality of the audio data includes: Convert the audio data into verification text using a preset conversion algorithm; Calculating text similarity between the verification text and the target text; The content quality corresponding to the audio data is determined based on the text similarity.

8. The method according to claim 6, wherein: The operation of identifying the content quality of the audio data includes: Performing a playback operation on the audio data according to a plurality of preset different playback parameters, wherein the playback parameters include at least one of a playback volume and a playback speed; The playback contents corresponding to the multiple different playback parameters are converted by a preset conversion algorithm to obtain multiple verification texts; Calculating the text similarity between each verification text and the target text respectively, and obtaining the text similarity corresponding to a plurality of different playback parameters; The content quality corresponding to the audio data is determined based on the weight information corresponding to each playback parameter and the text similarity corresponding to each playback parameter.

9. The method according to any one of claims 1 to 8, wherein: The step of storing the audio data in a preset storage path includes: Storing the audio data in a preset local storage path; and / or, The audio data is sent to a preset cloud storage path for data backup operation.

10. The method according to claim 9, further comprising: A data sharing operation is performed based on all audio data in the local storage path and / or the cloud storage path.

11. An audio data processing device, comprising: An acquisition module is used to acquire keywords input by a user, perform data processing on the keywords based on a preset large language model, and obtain target text related to the keywords; A determination module, used to obtain audio synthesis parameters determined by the user for the target text; A sending module, used to send the target text and the audio synthesis parameters to a preset speech synthesis algorithm, and obtain audio data fed back by the speech synthesis algorithm; The storage module is used to store the audio data in a preset storage path, and the audio data is used to perform performance testing on the speech recognition technology.

12. The device according to claim 11, wherein The acquisition module comprises: An input unit, used to input preset guide description information into the large language model, wherein the guide description information is used to guide the large language model to generate a target text associated with the keyword based on the keyword, wherein the target text is composed of a preset number of text segments; The acquisition unit is used to input the keyword into the large language model and acquire the target text output by the large language model.

13. The device according to claim 11, wherein: The determination module comprises: A display unit, used to display the target text and a preset first trigger control in a preset content display interface; a processing unit, configured to display a parameter setting interface in response to a trigger operation of the user on the first trigger control, wherein the parameter setting interface includes a preset parameter determination control, and the parameter determination control includes at least one of a tone determination control, a timbre determination control, a volume determination control, an audio format determination control, and a speed determination control; An acquisition unit is used to acquire the audio synthesis parameters determined by the user based on the parameter determination control.

14. The device according to claim 11, wherein: The target text is composed of a preset number of text segments, and a separator is included between two adjacent text segments; The sending module comprises: A sending unit is used to send the target text and the audio synthesis parameters to a preset speech synthesis algorithm through a preset calling interface, so that the speech synthesis algorithm segments the target text based on the delimiter, and asynchronously generates a preset number of audio data based on a preset number of text segments obtained by the segmentation.

15. The device according to claim 14, wherein the separator is added between two adjacent text segments when the large language model generates the target text; Alternatively, the separator is manually added by the user after determining the target position; The device also includes: A display module, used to display the target text in a preset content display interface; The adding module is used to determine the target position currently selected by the user in response to the adding operation triggered by the user on the target text, and add the separator at the target position.

16. The device according to claim 11, wherein The storage module comprises: An identification unit, used to identify the content quality of the audio data; The storage unit is used to store audio data whose content quality meets preset conditions in a preset storage path.

17. The device according to claim 16, wherein: The identification unit comprises: A conversion subunit, used to convert the audio data into a verification text by using a preset conversion algorithm; A calculation subunit, used for calculating the text similarity between the verification text and the target text; A determination subunit is used to determine the content quality corresponding to the audio data based on the text similarity.

18. The device according to claim 16, wherein: The identification unit comprises: A playing subunit, used to play the audio data according to a plurality of preset different playing parameters, wherein the playing parameters include at least one of a playing volume and a playing speed; The conversion subunit is used to convert the playback contents corresponding to the multiple different playback parameters through a preset conversion algorithm to obtain multiple verification texts; A calculation subunit, used to calculate the text similarity between each verification text and the target text, and obtain the text similarity corresponding to a plurality of different playback parameters; A determination subunit is used to determine the content quality corresponding to the audio data based on the weight information corresponding to each playback parameter and the text similarity corresponding to each playback parameter.

19. The device according to any one of claims 11 to 18, wherein: The storage module comprises: A first storage unit, used for storing the audio data in a preset local storage path; and / or, The second storage unit is used to send the audio data to a preset cloud storage path for data backup operation.

20. The apparatus according to claim 19, further comprising: A sharing module is used to perform data sharing operations based on all audio data in the local storage path and / or the cloud storage path.

21. An electronic device, comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 10.

22. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-10.

23. A computer program product, comprising a computer program, which, when executed by a processor, implements the steps of the method according to any one of claims 1 to 10.

Images