Voice broadcasting method and system, navigation system and vehicle

Abstract

The invention relates to the technical field of voice dialogue, and provides a voice broadcast method and system, a navigation system and a vehicle. When navigation is started, the following steps are executed: acquiring passenger voice information and navigation information and / or road condition information; passenger semantics are obtained according to the passenger voice information; acquiring a passenger emotion label according to the passenger voice information; obtaining a sound effect model and a broadcast text according to the passenger emotion label, the passenger semantics, the navigation information and / or the road condition information; and obtaining broadcast voice according to the sound effect model and the broadcast text. According to the scheme, after navigation is started, the emotion of the user can be obtained through the voice instruction sent by the user, then the voice model is adjusted in real time according to the emotion of the user, and the broadcasting tone is changed by changing the tone, intonation, speed and the like of the broadcasting voice; therefore, the purpose of emotional dredging is achieved through mood effects such as warning, reminding, pacifying and direct description according to the emotion of the user, and the voice dialogue experience can be improved.

Description

Technical Field

[0001] The present invention relates to the field of voice dialogue technology, and in particular to a voice broadcasting method, system, navigation system and vehicle. Background Art

[0002] Voice dialogue is a popular technology and has been widely used in smart cockpits, especially in navigation. In existing navigation systems and electronic map products, navigation can be started after the starting point and end point are specified on the electronic map and the planned route is determined. However, there is a problem of single voice material. The existing typical technology uses a preset character voice package to broadcast the voice of a fixed character, which is monotonous and lacks changes in the specific content of the navigation information. Summary of the invention

[0003] In view of the above-mentioned shortcomings of the prior art, an object of the present invention is to provide a voice broadcast method, system, navigation system and vehicle to enhance the voice dialogue experience.

[0004] To achieve the above purpose and other related purposes, the present invention provides a voice broadcast method, which performs the following steps when navigation is started:

[0005] Obtaining passenger voice information, navigation information and / or road condition information;

[0006] Obtaining occupant semantics based on occupant voice information;

[0007] Obtaining occupant emotion labels based on occupant voice information;

[0008] Obtaining a sound effect model and a broadcast text according to the occupant emotion label and the occupant semantics and the navigation information and / or the road condition information;

[0009] The announcement voice is obtained according to the sound effect model and the announcement text.

[0010] In a specific embodiment of the present invention, the step of obtaining the occupant's semantics according to the occupant's voice information includes:

[0011] Obtaining voice text according to the passenger's voice information;

[0012] The speech text is subjected to natural language processing to obtain the occupant semantics.

[0013] In a specific embodiment of the present invention, the step of obtaining the occupant's emotion label according to the occupant's voice information includes:

[0014] Acquire a speech waveform according to the occupant's speech information;

[0015] Obtaining an acoustic feature sequence through an acoustic feature extractor according to the speech waveform;

[0016] The occupant emotion label is obtained according to the acoustic feature sequence.

[0017] In a specific embodiment of the present invention, the step of obtaining a sound effect model according to the occupant emotion label and the occupant semantics and the current navigation information and / or road condition information includes:

[0018] The sound effect model is obtained through the HIFI-Gan method according to the occupant emotion label and occupant semantics and the current navigation information and / or road condition information.

[0019] In a specific embodiment of the present invention, the step of obtaining the sound effect model by the HIFI-Gan method according to the occupant emotion label and the occupant semantics and the current navigation information and / or road condition information includes:

[0020] Obtaining an intermediate model according to the occupant emotion label and the occupant semantics and current navigation information and / or road condition information;

[0021] Obtaining a false speech waveform according to the intermediate model;

[0022] The occupant's voice waveform and the false voice waveform are discriminated by a discriminator;

[0023] When the discriminator cannot distinguish the occupant's speech waveform from the false speech waveform, the intermediate model corresponding to the current false speech waveform is used as the sound effect model;

[0024] When the discriminator is able to distinguish between the occupant's speech waveform and the false speech waveform, the intermediate model is adjusted until the discriminator is unable to distinguish between the occupant's speech waveform and the false speech waveform.

[0025] In a specific embodiment of the present invention, the step of obtaining a sound effect model and a broadcast text according to the occupant's emotion tag and the occupant's semantics and the current navigation information and / or road condition information includes:

[0026] Obtaining an initial text according to occupant semantics and navigation information and / or road condition information;

[0027] Conduct content analysis on the initial text to obtain the initial text meaning;

[0028] The announcement text is obtained according to the occupant emotion label and the initial text meaning.

[0029] In a specific embodiment of the present invention, the step of obtaining an acoustic feature sequence according to the speech waveform through an acoustic feature extractor includes:

[0030] According to the speech waveform, an acoustic feature sequence Mel spectrum or Mel cepstral coefficient is obtained by an acoustic feature extractor.

[0031] The acoustic feature sequence is obtained according to the Mel spectrum or Mel cepstral coefficients.

[0032] The present invention also provides a voice broadcast system, characterized in that it includes:

[0033] An information acquisition module, which acquires passenger voice information, navigation information and / or road condition information;

[0034] Semantic recognition module, which obtains the occupant's semantics based on the occupant's voice information;

[0035] Emotion recognition module, which obtains the occupant's emotion label based on the occupant's voice information;

[0036] A generation module, which obtains a sound effect model and a broadcast text according to the occupant emotion label and the occupant semantics and the navigation information and / or the road condition information;

[0037] The announcement module obtains the announcement voice according to the sound effect model and the announcement text.

[0038] The present invention also provides a navigation system, comprising the above-mentioned voice broadcast system.

[0039] The present invention also provides a vehicle, comprising the above navigation system.

[0040] The present invention proposes a voice broadcast method, system, navigation system and vehicle. In the above scheme, after the navigation is started, the user's emotions can be obtained through the voice commands issued by the user, and then the voice model is adjusted in real time according to the user's emotions, and the timbre, intonation, speaking speed, etc. of the broadcast voice are changed to change the broadcast tone, thereby achieving the purpose of venting emotions and safe driving through tone effects such as warnings, reminders, soothing, and direct narration according to the user's emotions, and the voice dialogue experience can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0042] Figure 1 is a flow chart of a voice broadcasting method in one embodiment of the present invention;

[0043] Figure 2 is a flowchart of a navigation system in one embodiment of the present invention;

[0044] Figure 3 is a timing diagram of a voice broadcasting method in one embodiment of the present invention;

[0045] Figure 4 The present invention is a flowchart of the sound effect modeling optimization of the voice broadcasting method in one embodiment of the present invention. DETAILED DESCRIPTION

[0046] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0047] It should be noted that the illustrations provided in the following embodiments are only used to schematically illustrate the basic concept of the present invention, and thus the illustrations only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0048] Common voice conversations consist of the following core links.

[0049] ASR, or Automatic Speech Recognition, is the process of converting sound into text, which is equivalent to the ear.

[0050] NLP, or Natural Language Processing, is the process of understanding and processing text. It is equivalent to the brain. After understanding the voice input by the user, it is logically processed and converted into text before being given to TTS.

[0051] TTS, or text-to-speech, is the process of converting text into speech, which is equivalent to the mouth, and broadcasting through a preset voice package model. However, the existing typical technology uses a preset character voice package to broadcast a fixed character's timbre.

[0052] like Figure 1 As shown, a voice broadcast method is provided, and the following steps are performed when the navigation is started:

[0053] S1. Obtaining passenger voice information, navigation information and / or road condition information;

[0054] S2, obtaining the occupant's semantics based on the occupant's voice information;

[0055] S3, obtaining the occupant's emotion label according to the occupant's voice information;

[0056] S4, obtaining a sound effect model and a broadcast text according to the occupant emotion label and the occupant semantics and the navigation information and / or the road condition information;

[0057] S5. Obtaining the announcement voice according to the sound effect model and the announcement text.

[0058] In the above scheme, the passenger voice information is obtained through the microphone in the car. When the passenger issues a voice reception command by pressing a button or waking up the voice (for example, "Hello Gaode"), the microphone in the car records the passenger voice information. Navigation information and road condition information are provided by the navigation itself. The passenger emotion labels mainly include anger, happiness, surprise, disgust, sadness, fear, etc. Navigation information and road condition information are mainly obtained from the uploaded navigation SDK content. Navigation information includes driving time, driving distance, current speed, traffic information ahead and current time, etc., among which the traffic information ahead mainly includes information such as traffic lights, forks, and service area distances. Road condition information includes congestion distance, congestion time, traffic flow information, and traffic accident information. The sound effect model is mainly a model for adjusting the broadcast speed, broadcast tone, broadcast volume, and broadcast timbre.

[0059] This application also includes a speech analysis service AutoSpeechService (ASS) responsible for the state management and functional interface of the sound effect model.

[0060] Speech Algorithm (SA) is responsible for text analysis, tone analysis, and speech speed analysis. SA is used to dynamically update the sound effect model.

[0061] The audio management module AudioPlayerManger (APM) is responsible for converting the algorithm-modeled audio into PCM (Pulse Code Modulation) metadata and transmitting it to the functional speaker;

[0062] The RecordManager (RM) module is responsible for receiving voice commands from the user (for example, navigation to xxx, how far is it from the destination, where is the next service area).

[0063] In a specific embodiment of the present invention, the step of obtaining the occupant's semantics according to the occupant's voice information includes:

[0064] Obtaining voice text according to the passenger's voice information;

[0065] The speech text is subjected to natural language processing to obtain the passenger semantics. The basic task of natural language processing is to segment the processed corpus based on ontology dictionary, word frequency statistics, context semantic analysis, etc., to form a semantically rich word unit with the smallest part of speech as the unit.

[0066] In a specific embodiment of the present invention, the step of obtaining the occupant's emotion label according to the occupant's voice information includes:

[0067] Acquire a speech waveform according to the occupant's speech information;

[0068] Obtaining an acoustic feature sequence through an acoustic feature extractor according to the speech waveform;

[0069] The occupant's emotion label is obtained according to the acoustic feature sequence. The acoustic feature sequence refers to the physical quantity that represents the acoustic characteristics of speech, and is also a general term for the acoustic performance of various sound elements. Such as the energy concentration area, resonance peak frequency, resonance peak intensity and bandwidth that represent the timbre, and the duration, fundamental frequency, average voice power, etc. that represent the rhythmic characteristics of speech. According to the acoustic feature sequence, the occupant's timbre, volume, speech speed and other information can be obtained.

[0070] Speech Emotion Recognition is a biometric attribute recognition method that can identify the speaker's emotional state through the acoustic features of a speech (regardless of the speech content and language). The speaker can express different emotions by adjusting the movements of the vocal organs and changing the acoustic features of the speech signal. For example, when a person is angry, the sympathetic nerves will affect the respiratory system and cause corresponding muscle tension, thereby affecting the vibration of the vocal cords and the shape of the vocal tract, resulting in changes in acoustic features.

[0071] Common speech emotion recognition algorithms include model-based algorithms (hidden Markov model, Gaussian mixture model, K nearest neighbor method), feature-based algorithms (spectral features, sound quality features, fusion features, deep learning features), and vector quantization (VQ) methods based on non-parametric models. In addition, there are algorithms based on artificial neural networks (ANN) and hybrid algorithms, such as DNN / HMM method. In this application, algorithms based on artificial neural networks (ANN) and hybrid algorithms are preferred. Feature-based algorithms usually perform emotion recognition based on features extracted from speech signals. The advantage of this type of algorithm is that it does not require too much manually labeled data and can resist a certain degree of speech signal interference, but the disadvantage is that it may not be able to handle complex speech information well. However, there is no complex speech environment in navigation speech, and its speech is relatively fixed and easy to calculate.

[0072] In a specific embodiment of the present invention, the step of obtaining a sound effect model according to the occupant emotion label and the occupant semantics and the current navigation information and / or road condition information includes:

[0073] The sound effect model is obtained through the HIFI-Gan method according to the occupant emotion label and occupant semantics and the current navigation information and / or road condition information. The HIFI-Gan method is a speech synthesis model based on the generative adversarial network (GAN). Its main function is to convert acoustic features, such as Mel spectrum or Mel cepstral coefficients, into high-quality speech waveforms. The core idea of ​​HIFIGAN is to learn how to convert Mel spectrum into realistic speech waveforms through the game process between the generator and the discriminator. The Mel spectrum is converted into a realistic speech waveform through a deep neural network.

[0074] like Figure 4 As shown, the sound effect model is obtained by the HIFI-Gan method according to the occupant emotion label and the occupant semantics and the current navigation information and / or road condition information, and the steps include:

[0075] Obtaining an intermediate model according to the occupant emotion label and the occupant semantics and current navigation information and / or road condition information;

[0076] Obtaining a false speech waveform according to the intermediate model;

[0077] The occupant's voice waveform and the false voice waveform are discriminated by a discriminator;

[0078] When the discriminator cannot distinguish the occupant's speech waveform from the false speech waveform, the intermediate model corresponding to the current false speech waveform is used as the sound effect model;

[0079] When the discriminator is able to distinguish between the occupant's speech waveform and the false speech waveform, the intermediate model is adjusted until the discriminator is unable to distinguish between the occupant's speech waveform and the false speech waveform.

[0080] Wherein obtaining the false speech waveform according to the intermediate model is realized by a generator;

[0081] In the above scheme, the timbre, pitch, volume, and speech speed required for the corresponding emotion are obtained according to the occupant's emotional label and occupant's semantics, and then an intermediate model is generated according to the required timbre, pitch, volume, and speech speed. The intermediate model is then adversarially optimized to make it closer to human natural speech to improve the quality of the broadcast. When the discriminator distinguishes the occupant's speech waveform and the false speech waveform, it does not need to distinguish its semantics, but only distinguishes whether it is a synthetic sound.

[0082] In a specific embodiment of the present invention, the step of obtaining a sound effect model and a broadcast text according to the occupant's emotion tag and the occupant's semantics and the current navigation information and / or road condition information includes:

[0083] Obtaining an initial text according to occupant semantics and navigation information and / or road condition information;

[0084] Conduct content analysis on the initial text to obtain the initial text meaning;

[0085] The announcement text is obtained according to the passenger's emotional label and the meaning of the initial text. Different announcement texts can also express different emotions and can also be used to comfort. For example, when a passenger makes a hasty inquiry "How far is the next server?", the announcement text can be adjusted to "Please don't worry, the next service area is only ten kilometers away, and it is expected to take only ten minutes." If the passenger makes a gentle inquiry "How far is the next server?", the announcement text can be adjusted to "The next service area is ten kilometers away, and it is expected to take another ten minutes."

[0086] In a specific embodiment of the present invention, the step of obtaining an acoustic feature sequence according to the speech waveform through an acoustic feature extractor includes:

[0087] Obtaining an acoustic feature sequence Mel spectrum or Mel cepstral coefficients through an acoustic feature extractor according to the speech waveform;

[0088] The acoustic feature sequence is obtained according to the Mel spectrum or Mel cepstral coefficients.

[0089] Mel-Frequency Cepstrum (MFCC) is a feature extraction method for sound signals. Its purpose is to convert audio signals from time domain to frequency domain to facilitate recognition and classification analysis.

[0090] Mel-Frequency Cipstal Coefficients (MFCC) are a set of key coefficients used to create a Mel-Frequency Cepstrum. From a fragment of a music signal, we can get a set of cepstrum that is sufficient to represent the music signal, and the Mel-Frequency Cepstrum Coefficients are the cepstrum (i.e. the spectrum of the spectrum) derived from this cepstrum.

[0091] The present invention also provides a voice broadcast system, characterized in that it includes:

[0092] An information acquisition module, which acquires passenger voice information, navigation information and / or road condition information;

[0093] Semantic recognition module, which obtains the occupant's semantics based on the occupant's voice information;

[0094] Emotion recognition module, which obtains the occupant's emotion label based on the occupant's voice information;

[0095] A generation module, which obtains a sound effect model and a broadcast text according to the occupant emotion label and the occupant semantics and the navigation information and / or the road condition information;

[0096] The announcement module obtains the announcement voice according to the sound effect model and the announcement text.

[0097] The present invention also provides a navigation system, comprising the above-mentioned voice broadcast system.

[0098] The present invention also provides a vehicle, comprising the above navigation system.

[0099] like Figure 2 Shown is a workflow diagram of the navigation system, which is:

[0100] Navigation starts;

[0101] After initializing the voice service (voice service module), initialize the sound effect model, that is, initialize the background voice core service and load the general role sound effect;

[0102] If the user issues a voice command, text analysis - tone, speech speed and emotion analysis - model correction based on the above analysis results - navigation SDK report event, otherwise the navigation SDK directly reports the event;

[0103] Generate initial broadcast text

[0104] Conduct content analysis on the initial broadcast text;

[0105] Determine whether model correction is needed based on the content analysis results. If model correction is needed, generate a new report text and then make a voice broadcast. Otherwise, make a voice broadcast directly based on the initial report text. The model correction process is that the navigation SDK reports the report text, and the background voice service replaces the model according to the navigation SDK content for different road conditions.

[0106] Adopt new model for voice broadcast;

[0107] Repeat the above process until the navigation is completed.

[0108] like Figure 3 Shown is a timing diagram of the voice broadcast method in this application, which is:

[0109] When navigation starts: the navigation service NarigatorSerice, the speech analysis service AutoSpeechService (ASS), the speech algorithm SpeechAlgorithm (SA), the broadcast management module AudioPlayerManger (APM) and the recording management module RecordManager (RM) are initialized in sequence.

[0110] During navigation: the navigation service NarigatorSerice sends the broadcast text to the speech analysis service AutoSpeechService (ASS). The speech analysis service AutoSpeechService (ASS) performs content analysis on the broadcast text and sends the result to the speech algorithm SpeechAlgorithm (SA) for model conversion on the sound effect model and returns it to the speech analysis service AutoSpeechService (ASS). Then the speech analysis service AutoSpeechService (ASS) generates the broadcast voice according to the converted sound effect model and the broadcast text and sends it to the broadcast management module AudioPlayerManger (APM) for broadcast.

[0111] When receiving voice commands: the sound recording management module RecordManager (RM) receives the voice navigation commands issued by the user, sends the voice navigation commands to the speech algorithm SpeechAlgorithm (SA) for text analysis and sentiment analysis, replaces the sound effect model based on the text analysis and sentiment analysis and returns it to the speech analysis service AutoSpeechService (ASS), and then the speech analysis service AutoSpeechService (ASS) generates a broadcast voice based on the converted sound effect model and the broadcast text, sends it to the broadcast management module AudioPlayerManger (APM) for broadcast, and then returns to the navigation service NarigatorSerice.

[0112] In summary, the present invention proposes a voice broadcast method, system, navigation system and vehicle. In the above scheme, after the navigation is started, the user's emotions can be obtained through the voice commands issued by the user, and then the voice model can be adjusted in real time according to the user's emotions, and the timbre, intonation, speaking speed, etc. of the broadcast voice can be changed to change the broadcast tone, thereby achieving the purpose of venting emotions and safe driving through tone effects such as warnings, reminders, soothing, and direct narration according to the user's emotions, and the voice dialogue experience can be improved.

[0113] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.

[0114] In the description herein, many specific details, such as examples of components and / or methods, are provided to provide a complete understanding of embodiments of the present invention. However, those skilled in the art will recognize that embodiments of the present invention may be practiced without one or more of the specific details or with other devices, systems, components, methods, components, materials, parts, etc. In other cases, well-known structures, materials, or operations are not specifically shown or described in detail to avoid obscuring aspects of embodiments of the present invention.

[0115] References throughout this specification to "one embodiment," "an embodiment," or "a specific embodiment" mean that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the invention, and not necessarily in all embodiments. Thus, various appearances of the phrases "in one embodiment," "in an embodiment," or "in a specific embodiment" in different places throughout this specification do not necessarily refer to the same embodiment. In addition, the particular features, structures, or characteristics of any specific embodiment of the invention may be combined with one or more other embodiments in any suitable manner. It should be understood that other variations and modifications of the embodiments of the invention described and illustrated herein may be possible in light of the teachings herein and are to be considered part of the spirit and scope of the invention.

[0116] It should also be understood that one or more of the elements shown in the figures may also be implemented in a more separate or more integrated manner, or even removed because they are inoperable in certain circumstances or provided because they may be useful depending on the application.

[0117] In addition, unless otherwise explicitly indicated, any marking arrows in the drawings should be regarded as exemplary only and not limiting. In addition, unless otherwise indicated, the term "or" used herein is generally intended to mean "and / or". In the case where the term is not clear because it is anticipated that the ability to separate or combine is provided, the combination of components or steps will also be regarded as indicated.

[0118] As used in the description herein and throughout the claims that follow, "a", "an", and "the" include plural references unless otherwise indicated. Likewise, as used in the description herein and throughout the claims that follow, the meaning of "in" includes "in" and "on" unless otherwise indicated.

[0119] The above description of the illustrated embodiments of the present invention (including the contents described in the Abstract) is not intended to be an exhaustive list or to limit the present invention to the precise form disclosed herein. Although specific embodiments of the present invention and examples of the present invention are described herein for illustrative purposes only, various equivalent modifications are possible within the spirit and scope of the present invention as will be recognized and appreciated by those skilled in the art. As noted, these modifications may be made to the present invention in accordance with the above description of the embodiments described in the present invention, and these modifications will be within the spirit and scope of the present invention.

[0120] Systems and methods have been generally described herein as details that aid in understanding the present invention. In addition, various specific details have been given to provide an overall understanding of embodiments of the present invention. However, those skilled in the relevant art will recognize that embodiments of the present invention may be practiced without one or more of the specific details, or may be practiced using other devices, systems, accessories, methods, components, materials, parts, etc. In other cases, well-known structures, materials, and / or operations are not specifically shown or described in detail to avoid confusion with various aspects of embodiments of the present invention.

[0121] Thus, although the invention has been described herein with reference to specific embodiments thereof, freedom of modification, various changes and substitutions are also within the foregoing disclosure, and it should be understood that in some cases, some features of the invention will be employed without the corresponding use of other features without departing from the scope and spirit of the proposed invention. Thus, many modifications may be made to adapt a particular environment or material to the essential scope and spirit of the invention. The invention is not intended to be limited to the specific terms used in the claims below and / or the specific embodiments disclosed as the best mode contemplated for carrying out the invention, but the invention will include any and all embodiments and equivalents falling within the scope of the appended claims. Thus, the scope of the invention will be determined solely by the appended claims.

Claims

1. A voice broadcast method, characterized in that: The following steps are performed when the navigation starts: Obtaining passenger voice information, navigation information and / or road condition information; Obtaining occupant semantics based on occupant voice information; Obtaining occupant emotion labels based on occupant voice information; Obtaining a sound effect model and a broadcast text according to the occupant emotion label and the occupant semantics and the navigation information and / or the road condition information; The announcement voice is obtained according to the sound effect model and the announcement text.

2. The voice broadcasting method according to claim 1, characterized in that: The step of obtaining the occupant's semantics according to the occupant's voice information includes: Obtaining voice text according to the passenger's voice information; The speech text is subjected to natural language processing to obtain the occupant semantics.

3. The voice broadcasting method according to claim 1, characterized in that: The step of obtaining the occupant's emotion label according to the occupant's voice information includes: Acquire a speech waveform according to the occupant's speech information; Obtaining an acoustic feature sequence through an acoustic feature extractor according to the speech waveform; The occupant emotion label is obtained according to the acoustic feature sequence.

4. The voice broadcasting method according to claim 1, characterized in that: The step of obtaining a sound effect model according to the occupant emotion label and the occupant semantics and the current navigation information and / or road condition information includes: The sound effect model is obtained through the HIFI-Gan method according to the occupant emotion label and occupant semantics and the current navigation information and / or road condition information.

5. The voice broadcasting method according to claim 4, characterized in that: The step of obtaining the sound effect model by the HIFI-Gan method according to the occupant emotion label and the occupant semantics and the current navigation information and / or road condition information includes: Obtaining an intermediate model according to the occupant emotion label and the occupant semantics and current navigation information and / or road condition information; Obtaining a false speech waveform according to the intermediate model; The occupant's voice waveform and the false voice waveform are discriminated by a discriminator; When the discriminator cannot distinguish the occupant's speech waveform from the false speech waveform, the intermediate model corresponding to the current false speech waveform is used as the sound effect model; When the discriminator is able to distinguish between the occupant's speech waveform and the false speech waveform, the intermediate model is adjusted until the discriminator is unable to distinguish between the occupant's speech waveform and the false speech waveform.

6. The voice broadcasting method according to claim 1, characterized in that: The step of obtaining a sound effect model and a broadcast text according to the occupant emotion label and the occupant semantics and the current navigation information and / or road condition information includes: Obtaining an initial text according to occupant semantics and navigation information and / or road condition information; Conduct content analysis on the initial text to obtain the initial text meaning; The announcement text is obtained according to the occupant emotion label and the initial text meaning.

7. The voice broadcasting method according to claim 3, characterized in that: The step of obtaining an acoustic feature sequence according to the speech waveform through an acoustic feature extractor comprises: According to the speech waveform, an acoustic feature sequence Mel spectrum or Mel cepstral coefficient is obtained by an acoustic feature extractor. The acoustic feature sequence is obtained according to the Mel spectrum or Mel cepstral coefficients.

8. A voice broadcast system, characterized in that: include: An information acquisition module, which acquires passenger voice information, navigation information and / or road condition information; Semantic recognition module, which obtains the occupant's semantics based on the occupant's voice information; Emotion recognition module, which obtains the occupant's emotion label based on the occupant's voice information; A generation module, which obtains a sound effect model and a broadcast text according to the occupant emotion label and the occupant semantics and the navigation information and / or the road condition information; The announcement module obtains the announcement voice according to the sound effect model and the announcement text.

9. A navigation system, characterized in that: Comprising the voice broadcast system as claimed in claim 8.

10. A vehicle, characterized in that: Comprising a navigation system as claimed in claim 9.

Images