CN112786007A

CN112786007A - Speech synthesis method, device, readable medium and electronic equipment

Info

Publication number: CN112786007A
Application number: CN202110075973.1A
Authority: CN
Inventors: 吴鹏飞; 潘俊杰
Original assignee: Beijing Youzhuju Network Technology Co Ltd
Current assignee: Beijing Youzhuju Network Technology Co Ltd
Priority date: 2021-01-20
Filing date: 2021-01-20
Publication date: 2021-05-11
Anticipated expiration: 2041-01-20
Also published as: WO2022156544A1; CN112786007B

Abstract

The present disclosure relates to a speech synthesis method, apparatus, readable medium and electronic device, and relates to the technical field of electronic information processing, wherein the method comprises: the method comprises the steps of obtaining a text to be synthesized, specifying acoustic features and an appointed emotion type, wherein the specified acoustic features are used for indicating prosodic features of audio, extracting a phoneme sequence corresponding to the text to be synthesized, extending the specified acoustic features according to the phoneme sequence to obtain an acoustic feature sequence, inputting the phoneme sequence, the acoustic feature sequence and the appointed emotion type into a pre-trained voice synthesis model to obtain target audio which is output by the voice synthesis model, corresponds to the text to be synthesized and has the appointed emotion type, and the acoustic features of the target audio are matched with the specified acoustic features. According to the method and the device, the speech synthesis of the text is controlled by specifying the acoustic characteristics and the emotion types, so that the explicit control of the emotion types and the acoustic characteristics in the speech synthesis process can be realized, and the expressive force of the target audio is improved.

Description

Speech synthesis method, device, readable medium and electronic equipment

Technical Field

The present disclosure relates to the field of electronic information processing technologies, and in particular, to a speech synthesis method, apparatus, readable medium, and electronic device.

Background

With the continuous development of electronic information processing technology, voice is widely used in daily life and work as an important carrier for people to obtain information. In an application scenario involving speech, processing of speech synthesis is usually included, and speech synthesis refers to synthesizing text designated by a user into audio. In the speech synthesis process, speech with corresponding emotion can be synthesized through the designated emotion tag. However, the types of emotion tags are limited, and it is difficult to meet the diversified demands of users.

Disclosure of Invention

This summary is provided to introduce a selection of concepts in a simplified form that are further described below in the detailed description. This summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.

In a first aspect, the present disclosure provides a method of speech synthesis, the method comprising:

acquiring a text to be synthesized, specified acoustic features and specified emotion types, wherein the specified acoustic features are used for indicating rhythm features of audio;

extracting a phoneme sequence corresponding to the text to be synthesized;

expanding the specified acoustic features according to the phoneme sequence to obtain an acoustic feature sequence;

inputting the phoneme sequence, the acoustic feature sequence and the specified emotion type into a pre-trained speech synthesis model to obtain a target audio which is output by the speech synthesis model and corresponds to the text to be synthesized and has the specified emotion type, wherein the acoustic feature of the target audio is matched with the specified acoustic feature.

In a second aspect, the present disclosure provides a speech synthesis apparatus, the apparatus comprising:

the system comprises an acquisition module, a synthesis module and a feedback module, wherein the acquisition module is used for acquiring a text to be synthesized, specified acoustic features and specified emotion types, and the specified acoustic features are used for indicating rhythm features of audio;

the extraction module is used for extracting a phoneme sequence corresponding to the text to be synthesized;

the extension module is used for extending the specified acoustic features according to the phoneme sequence to obtain an acoustic feature sequence;

and the synthesis module is used for inputting the phoneme sequence, the acoustic feature sequence and the specified emotion type into a pre-trained voice synthesis model so as to obtain a target audio which is output by the voice synthesis model and corresponds to the text to be synthesized and has the specified emotion type, and the acoustic feature of the target audio is matched with the specified acoustic feature.

In a third aspect, the present disclosure provides a computer readable medium having stored thereon a computer program which, when executed by a processing apparatus, performs the steps of the method of the first aspect of the present disclosure.

In a fourth aspect, the present disclosure provides an electronic device comprising:

a storage device having a computer program stored thereon;

processing means for executing the computer program in the storage means to implement the steps of the method of the first aspect of the present disclosure.

According to the technical scheme, the method comprises the steps of firstly obtaining a text to be synthesized, an appointed emotion type and appointed acoustic features used for indicating prosodic features of audio, then extracting a corresponding phoneme sequence from the text to be synthesized, then expanding the appointed acoustic features according to the phoneme sequence to obtain an acoustic feature sequence, and finally inputting the phoneme sequence, the acoustic feature sequence and the appointed emotion type into a pre-trained voice synthesis model, so that target audio which is output by the voice synthesis model, corresponds to the text to be synthesized, has the appointed emotion type and is matched with the appointed acoustic features is obtained. According to the method and the device, the voice synthesis of the text is controlled by specifying the acoustic characteristics and the emotion types, so that the target audio output by the voice synthesis model can accord with the specified acoustic characteristics on the basis of the specified emotion types, the explicit control of the emotion types and the acoustic characteristics in the voice synthesis process can be realized, and the expressive force of the target audio is improved.

Additional features and advantages of the disclosure will be set forth in the detailed description which follows.

Drawings

The above and other features, advantages and aspects of various embodiments of the present disclosure will become more apparent by referring to the following detailed description when taken in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numbers refer to the same or similar elements. It should be understood that the drawings are schematic and that elements and features are not necessarily drawn to scale. In the drawings:

FIG. 1 is a flow diagram illustrating a method of speech synthesis according to an exemplary embodiment;

FIG. 2 is a flow diagram illustrating another method of speech synthesis according to an example embodiment;

FIG. 3 is a process flow diagram illustrating a speech synthesis model in accordance with an exemplary embodiment;

FIG. 4 is a block diagram illustrating a speech synthesis model in accordance with an exemplary embodiment;

FIG. 5 is a flow diagram illustrating training a speech synthesis model according to an exemplary embodiment;

FIG. 6 is a flow diagram illustrating another method of training a speech synthesis model in accordance with an illustrative embodiment;

FIG. 7 is a flow diagram illustrating another method of speech synthesis according to an example embodiment;

FIG. 8 is a block diagram illustrating a speech synthesis apparatus according to an exemplary embodiment;

FIG. 9 is a block diagram illustrating another speech synthesis apparatus in accordance with an illustrative embodiment;

FIG. 10 is a block diagram illustrating another speech synthesis apparatus in accordance with an illustrative embodiment;

FIG. 11 is a block diagram illustrating an electronic device in accordance with an example embodiment.

Detailed Description

Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While certain embodiments of the present disclosure are shown in the drawings, it is to be understood that the present disclosure may be embodied in various forms and should not be construed as limited to the embodiments set forth herein, but rather are provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the disclosure are for illustration purposes only and are not intended to limit the scope of the disclosure.

It should be understood that the various steps recited in the method embodiments of the present disclosure may be performed in a different order, and/or performed in parallel. Moreover, method embodiments may include additional steps and/or omit performing the illustrated steps. The scope of the present disclosure is not limited in this respect.

The term "include" and variations thereof as used herein are open-ended, i.e., "including but not limited to". The term "based on" is "based, at least in part, on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Relevant definitions for other terms will be given in the following description.

It should be noted that the terms "first", "second", and the like in the present disclosure are only used for distinguishing different devices, modules or units, and are not used for limiting the order or interdependence relationship of the functions performed by the devices, modules or units.

It is noted that references to "a", "an", and "the" modifications in this disclosure are intended to be illustrative rather than limiting, and that those skilled in the art will recognize that "one or more" may be used unless the context clearly dictates otherwise.

The names of messages or information exchanged between devices in the embodiments of the present disclosure are for illustrative purposes only, and are not intended to limit the scope of the messages or information.

FIG. 1 is a flow diagram illustrating a method of speech synthesis, as shown in FIG. 1, according to an exemplary embodiment, the method comprising:

step 101, acquiring a text to be synthesized, specifying acoustic features and a specified emotion type, wherein the specified acoustic features are used for indicating prosodic features of audio.

For example, a text to be synthesized that needs to be synthesized is first obtained. The text to be synthesized may be, for example, one or more sentences in a text file specified by a user, one or more paragraphs, one or more chapters in the text file, or one or more words in one text file. The text file may be, for example, an electronic book, or may be other types of files, such as news, public articles, blogs, and the like. At the same time, specified acoustic features and specified emotion types may also be obtained. Specifying acoustic features and specified emotion types may be understood as meaning that the user specifies that it is desirable to synthesize the text to be synthesized into audio having the specified emotion type and conforming to the specified acoustic features (i.e., target audio mentioned later). That is, the user can control the target audio by specifying both the acoustic features and the emotion types. The specified acoustic features may include multiple dimensions, for example, one or more of fundamental frequency (Pitch), volume (Energy), and speech rate (Duration), and may further include: noise level, pitch, timbre, loudness, etc. Wherein the noise level may be understood as a characteristic that reflects the magnitude of the noise in the audio. The specified emotion type may be in the form of a tag, for example, happy for 0001, surprised for 0011, hated for 1010, angry for 1011, shy for 0101, frightened for 0100, sad for 1000, and shivering for 1001.

Step 102, extracting a phoneme sequence corresponding to the text to be synthesized.

For example, the text to be synthesized may be input into a pre-trained recognition model to obtain a phoneme sequence corresponding to the text to be synthesized, which is output by the recognition model. Or searching the phoneme corresponding to each word in the text to be synthesized in a pre-established dictionary, and then forming the phoneme corresponding to each word into a phoneme sequence corresponding to the text to be synthesized. The phoneme can be understood as a phonetic unit divided according to the pronunciation of each word, and can also be understood as a vowel and a consonant in the pinyin corresponding to each word. The phoneme sequence includes a phoneme corresponding to each word in the text to be synthesized (one word may correspond to one or more phonemes). For example, the text to be synthesized is "this is a sample case", and phonemes corresponding to each word may be sequentially searched in the dictionary, thereby determining that the phoneme sequence is "zheshiyigeyangli".

And 103, expanding the specified acoustic features according to the phoneme sequence to obtain an acoustic feature sequence.

And 104, inputting the phoneme sequence, the acoustic feature sequence and the specified emotion type into a pre-trained voice synthesis model to obtain a target audio which is output by the voice synthesis model and corresponds to the text to be synthesized and has the specified emotion type, wherein the acoustic feature of the target audio is matched with the specified acoustic feature.

For example, after obtaining the phoneme sequence, the specified acoustic features may be extended according to the phoneme sequence to obtain an acoustic feature sequence, where the acoustic feature sequence includes the acoustic features corresponding to each phoneme in the phoneme sequence. In one implementation, the acoustic feature sequence may be generated according to the length of the phoneme sequence (i.e., the number of phonemes included in the phoneme sequence), where the acoustic feature corresponding to each phoneme is a specific acoustic feature. In another implementation, the acoustic feature corresponding to each phoneme may also be generated according to a preset distribution (for example, a gaussian distribution or a uniform distribution) by using the specified acoustic feature as an average (or standard deviation).

And then, the phoneme sequence, the acoustic feature sequence and the specified emotion type can be used as the input of a pre-trained speech synthesis model, and the output of the speech synthesis model is the target audio which corresponds to the text to be synthesized, has the specified emotion type and is matched with the specified acoustic feature. The voice synthesis model can be pre-trained, can be understood as a TTS (Text To Speech, Chinese from Text To voice) model, and can generate a target audio which corresponds To the Text To be synthesized, has a specified emotion type and is matched with the specified acoustic characteristic according To the Text To be synthesized, the specified acoustic characteristic and the specified emotion type. Specifically, the speech synthesis model may be obtained based on training of a Tacotron model, a Deepvoice 3 model, a Tacotron 2 model, a Wavenet model, and the like, which is not specifically limited by the present disclosure. Therefore, in the process of carrying out voice synthesis in the text to be synthesized, in addition to the semantics included in the text to be synthesized, the appointed acoustic features and the appointed emotion types are also considered, the target audio can have the appointed emotion types and can be matched with the appointed acoustic features, so that a user can control the voice synthesis from two dimensions of the emotion types and the acoustic features according to specific requirements without being limited to the number of the emotion types, the expressive force of the target audio is improved, and the auditory experience of the user is improved.

In summary, according to the present disclosure, a text to be synthesized, an appointed emotion type, and an appointed acoustic feature for indicating a prosodic feature of an audio are first obtained, then a corresponding phoneme sequence is extracted from the text to be synthesized, the appointed acoustic feature is extended according to the phoneme sequence to obtain an acoustic feature sequence, and finally the phoneme sequence, the acoustic feature sequence, and the appointed emotion type are input into a pre-trained speech synthesis model, so that a target audio which is output by the speech synthesis model, corresponds to the text to be synthesized, has the appointed emotion type, and is matched with the appointed acoustic feature is obtained. According to the method and the device, the voice synthesis of the text is controlled by specifying the acoustic characteristics and the emotion types, so that the target audio output by the voice synthesis model can accord with the specified acoustic characteristics on the basis of the specified emotion types, the explicit control of the emotion types and the acoustic characteristics in the voice synthesis process can be realized, and the expressive force of the target audio is improved.

Fig. 2 is a flow chart illustrating another speech synthesis method according to an exemplary embodiment, and as shown in fig. 2, step 103 may be implemented by:

step 1031, determining an acoustic feature corresponding to each phoneme in the phoneme sequence according to the specified acoustic features.

Step 1032, the acoustic features corresponding to each phoneme are combined into an acoustic feature sequence.

For example, in one implementation, the length of the phoneme sequence, i.e., the number of phonemes included in the phoneme sequence, may be determined first. And then copying the specified acoustic features to obtain an acoustic feature sequence with the same length as the phoneme sequence, wherein each acoustic feature is the same as the specified acoustic feature, that is, the acoustic feature corresponding to each phoneme in the acoustic feature sequence is the specified acoustic feature. For example, if the length of the phoneme sequence is 100 (i.e., 100 phonemes are included), the acoustic feature corresponding to each phoneme may be determined as the specified acoustic feature, and then the acoustic features corresponding to 100 phonemes may be combined into the acoustic feature sequence. Taking the example of a vector with 1 × 5 dimensions as the acoustic features, the acoustic feature sequence includes 100 vectors with 1 × 5 dimensions, and may constitute a vector with 100 × 5 dimensions.

FIG. 3 is a process flow diagram illustrating a speech synthesis model according to an exemplary embodiment, as shown in FIG. 3, which may be used to perform the following steps:

and step A, determining a text characteristic sequence corresponding to the text to be synthesized according to the phoneme sequence, wherein the text characteristic sequence comprises text characteristics corresponding to each phoneme in the phoneme sequence.

And step B, determining the appointed emotional characteristics corresponding to the appointed emotional types, and expanding the appointed emotional characteristics according to the phoneme sequence to obtain an emotional characteristic sequence.

And C, generating a target audio according to the text characteristic sequence, the acoustic characteristic sequence and the emotion characteristic sequence.

For example, in a specific process of synthesizing a target audio by using a speech synthesis model, a Text feature sequence (Text Embedding) corresponding to a Text to be synthesized may be extracted according to a phoneme sequence, where the Text feature sequence includes Text features corresponding to each phoneme in the phoneme sequence, and the Text features may be understood as Text vectors capable of representing the phonemes. For example, the phoneme sequence includes 100 phonemes, and the text vector corresponding to each phoneme is a 1 × 80 dimensional vector, so the text feature sequence may be a 100 × 80 dimensional vector.

And then, determining corresponding specified emotional characteristics according to the specified emotional types, wherein the specified emotional characteristics can be understood as emotional vectors capable of representing the specified emotional types. Specifically, the speech synthesis model may include a Lookup Table (english: Lookup Table), and the Lookup Table may map a tag corresponding to the specified emotion type into a multidimensional emotion vector. For example, if the emotion type is designated as mimicry and the corresponding label is 0101, the look-up table can map 0101 to a 1 x 50-dimensional emotion vector as the designated emotion feature for characterizing happiness. In another application scenario, the speech synthesis model may include an encoder (i.e., the second encoder in fig. 4), and the encoder may extract a corresponding emotion vector according to a specified emotion type, that is, according to different emotion types, obtain an emotion vector capable of characterizing the emotion type. The encoder may be trained independently in advance based on a large number of training samples, or may be trained in combination with a speech synthesis model, which is not specifically limited by the present disclosure. Further, after obtaining the specified emotion feature corresponding to the specified emotion type, the specified emotion feature may be further extended according to the phoneme sequence to obtain an emotion feature sequence, where the emotion feature sequence includes the emotion feature corresponding to each phoneme in the phoneme sequence. For example, the length of the phoneme sequence, i.e. the number of phonemes comprised in the phoneme sequence, may be determined first. And then copying the appointed emotional features to obtain an emotional feature sequence with the same length as the phoneme sequence, wherein each emotional feature is the same as the appointed emotional feature, namely, the emotional features corresponding to each phoneme in the emotional feature sequence are the appointed emotional features. For example, if the length of the phoneme sequence is 100 (i.e. 100 phonemes are included), the emotion characteristics corresponding to each phoneme can be determined as the specified emotion characteristics, and then the emotion characteristics corresponding to 100 phonemes can be combined into the emotion characteristic sequence. Taking the example of a vector with 1 × 50 dimensions as the assigned emotional features, the sequence of emotional features includes 100 vectors with 1 × 50 dimensions, and may constitute a vector with 100 × 50 dimensions.

After the text feature sequence and the emotion feature sequence are obtained, the text feature sequence and the emotion feature sequence can be combined with the acoustic feature sequence to generate target audio which has a specified emotion type and is matched with specified acoustic features. For example, the text feature sequence, the emotion feature sequence, and the acoustic feature sequence may be spliced to obtain a combined sequence, and then the target audio may be generated according to the combined sequence. For example, if the phoneme sequence includes 100 phonemes, the text feature sequence may be a vector of 100 × 80 dimensions, the corresponding emotion feature sequence is a vector of 100 × 50 dimensions, the acoustic feature sequence is a vector of 100 × 5 dimensions, and the combined sequence may be a vector of 100 × 135 dimensions. The target audio may be generated from this 100 x 135 dimensional vector.

Taking the speech synthesis model shown in fig. 4 as an example, the speech synthesis model is a tacontron model, which includes: a look-up table (or second Encoder), a first Encoder (i.e., Encoder), an Attention network (i.e., Attention), a Decoder (i.e., Decoder), and a Post-processing network (i.e., Post-processing). The look-up table may be, for example, a preset matrix, and the tag corresponding to the specified emotion type may be multiplied by the look-up table, so that the tag corresponding to the specified emotion type is mapped to the specified emotion feature, and the specified emotion feature is expanded to obtain an emotion feature sequence. The first encoder may include an Embedding layer (i.e., a Character Embedding layer), a Pre-processing network (Pre-network) sub-model, and a CBHG (english: convergence Bank + high-way network + bidirectional Gated Recurrent Unit, chinese: convolutional layer + high-speed network + bidirectional Recurrent neural network) sub-model. The phoneme sequence can be input into a first encoder, firstly, the phoneme sequence is converted into a word vector through an embedding layer, then, the word vector is input into a Pre-net sub-model to carry out nonlinear transformation on the word vector, so that the convergence and generalization capability of a speech synthesis model is improved, and finally, a text feature sequence capable of representing a text to be synthesized is obtained through a CBHG sub-model according to the word vector after the nonlinear transformation.

And then splicing the acoustic characteristic sequence, the emotional characteristic sequence output by the lookup table and the text characteristic sequence output by the encoder to obtain a combined sequence, and inputting the combined sequence into an attention network, wherein the attention network can add an attention weight to each element in the combined sequence. Specifically, the Attention network may be a location Sensitive Attention (location Sensitive Attention) network, a GMM (Gaussian Mixture Model, abbreviated as GMM) authentication network, or a Multi-Head authentication network, which is not limited in this disclosure.

The output of the attention network is then used as the input of the decoder. The Decoder may include a pre-processing network sub-model (which may be the same as the pre-processing network sub-model included in the encoder), an Attention-RNN, a Decoder-RNN. The preprocessing network submodel is used for carrying out nonlinear transformation on input, the structure of the Attention-RNN is a layer of one-way zoneout-based LSTM (English Short-Term Memory, Chinese: Long Short-Term Memory network), and the output of the preprocessing network submodel can be used as input and is output to the Decoder-RNN after passing through the LSTM unit. The Decode-RNN is a two-layer one-way zoneout-based LSTM, and outputs Mel frequency spectrum information through an LSTM unit, wherein the Mel frequency spectrum information can comprise one or more Mel frequency spectrum characteristics. The mel-frequency spectrum information is finally input into a post-processing network, which may include a vocoder (e.g., a Wavenet vocoder, a Griffin-Lim vocoder, etc.) for converting the mel-frequency spectrum feature information to obtain the target audio.

FIG. 5 is a flow diagram illustrating training of a speech synthesis model, as shown in FIG. 5, trained in the following manner, according to an exemplary embodiment:

and D, extracting real acoustic features of the training audio corresponding to the training text, wherein the real acoustic features are used for indicating prosodic features of the audio.

And E, extending the real acoustic features according to the training phoneme sequence corresponding to the training text to obtain a training acoustic feature sequence.

And F, inputting the training phoneme sequence, the training acoustic feature sequence and the training emotion type corresponding to the training audio into the voice synthesis model, and training the voice synthesis model according to the output of the voice synthesis model and the training audio.

When a speech synthesis model is trained, training texts and training audios corresponding to the training texts need to be obtained first, wherein the number of the training texts can be multiple, and correspondingly, the number of the training audios is also multiple. For example, a large amount of text may be captured on the internet as training text, and then audio corresponding to the training text may be used as training audio. For the training audio, the real acoustic features corresponding to the training audio can be extracted. For example, the real acoustic features corresponding to the training audio may be obtained through signal processing, labeling, and the like, where the real acoustic features are used to indicate prosodic features of the training audio, and the method may include: at least one of fundamental frequency, volume and speech speed of the training audio may further include: noise level, pitch, timbre, loudness, etc. Meanwhile, a training phoneme sequence corresponding to the training text may be extracted, where the training phoneme sequence may include a training phoneme corresponding to each word in the training text (a word may correspond to one or more training phonemes).

And then, extending the real acoustic features according to a training phoneme sequence corresponding to the training text to obtain a training acoustic feature sequence. The training acoustic feature sequence includes training acoustic features corresponding to each training phoneme. For example, a training acoustic feature sequence may be generated according to the number of training phonemes included in the training phoneme sequence, where the training acoustic feature corresponding to each training phoneme is a real acoustic feature.

And finally, taking the training phoneme sequence, the training acoustic feature sequence and the training emotion type corresponding to the training audio as the input of the speech synthesis model, and training the speech synthesis model according to the output of the speech synthesis model and the training audio. For example, the parameters of neurons in the speech synthesis model, such as weights (in English: Weight) and offsets (in English: Bias) of the neurons, can be modified by a back propagation algorithm with the goal of reducing the loss function according to the output of the speech synthesis model and the difference (or mean square error) from the training audio as the loss function of the speech synthesis model. And repeating the steps until the loss function meets a preset condition, for example, the loss function is smaller than a preset loss threshold.

In one application scenario, a speech synthesis model includes: a first module, a second module and a synthesis module, the implementation of step F may comprise the steps of:

step F1, extracting, by the first module, a training text feature sequence corresponding to the training phoneme sequence, where the training text feature sequence includes a text feature corresponding to each training phoneme in the training phoneme sequence.

And step F2, extracting training emotional characteristics corresponding to the training emotional types through a second module, and expanding the training emotional characteristics according to the training phoneme sequence to obtain a training emotional characteristic sequence.

And F3, generating the output of the voice synthesis model according to the training text characteristic sequence, the training acoustic characteristic sequence and the training emotion characteristic sequence through a synthesis module.

Step F4, according to the output of the speech synthesis model and the training audio, determining the loss function of the speech synthesis model, and updating the first module, the second module and the synthesis module according to the loss function.

For example, the speech synthesis model may include a first module, a second module, and a synthesis module, where the first module is configured to extract a training text feature sequence including a text feature corresponding to each training phoneme in the training phoneme sequence, and the first module may be, for example, an encoder (e.g., a first encoder shown in fig. 4). The second module is configured to extract training emotional features and expand the training emotional features to obtain an emotional feature sequence including the emotional features corresponding to each training phoneme, and the second module may be, for example, a lookup table or an encoder (e.g., the second encoder shown in fig. 4). And the synthesis module is used for generating corresponding audio, namely the output of the voice synthesis model, according to the training text characteristic sequence, the training acoustic characteristic sequence and the training emotion characteristic sequence.

Further, during the training process of the speech synthesis model, a loss function of the speech synthesis model (for example, the difference between the output of the speech synthesis model and the training audio) may be determined according to the output of the speech synthesis model and the training audio, and the first module, the second module, and the synthesis module in the speech synthesis model may be updated by using a back propagation algorithm with the goal of reducing the loss function. It should be noted that, if the second module is a lookup table, the lookup table may be updated simultaneously in the process of training the speech synthesis model, and if the second module is an encoder, the parameters of the neurons in the encoder may be updated simultaneously in the process of training the speech synthesis model.

FIG. 6 is a flow diagram illustrating another method for training a speech synthesis model, according to an example embodiment, where the true acoustic features include, as shown in FIG. 6: at least one of fundamental frequency, volume and speech rate, and the step D may include the steps of:

step D1, if the real acoustic features include the speech rate, determining the duration corresponding to each training phoneme in the training phoneme sequence according to the training audio and the training phoneme sequence to determine the speech rate of the training audio.

The specific implementation mode can include:

first, according to the training audio and the training phoneme sequence, the duration corresponding to each training phoneme is determined. For example, training audio may be divided according to training phonemes included in the training phoneme sequence by using HTS (HMM-based Speech Synthesis System) to obtain a duration corresponding to each training phoneme, which may be expressed as a duration_iAnd the duration corresponding to the ith training phoneme is shown.

And then, carrying out logarithm operation on the duration corresponding to each training phoneme to obtain the logarithm duration corresponding to each training phoneme, and compressing the variation range of the duration through the logarithm operation so as to amplify the variation degree of the duration. For example, can be expressed as log _ duration_iAnd representing the log duration corresponding to the ith training phoneme.

And finally, taking the statistic value of the log duration corresponding to each training phoneme in the training phoneme sequence as the speech speed of the training audio. For example, the average (or standard deviation, extremum, etc.) of the log duration corresponding to each training phoneme can be used as the speech rate of the training audio, and can be expressed as log _ duration _ mean.

Step D2, if the real acoustic features include a fundamental frequency, extracting the fundamental frequency of each audio frame included in the training audio to determine the fundamental frequency of the training audio.

The specific implementation mode can include:

first, sox, librosa, strain can be utilizedAnd the audio processing tools such as the right and the like process the training audio to obtain the fundamental frequency of each audio in the training audio, and then perform logarithmic operation on the fundamental frequency corresponding to each audio frame to obtain the logarithmic fundamental frequency corresponding to each audio frame. Through logarithmic operation, the variation range of the fundamental frequency can be compressed, so that the variation degree of the fundamental frequency is amplified. For example, the corresponding fundamental frequency of each audio frame may be denoted as pitch_jRepresents the corresponding fundamental frequency of the jth audio frame, and correspondingly, the logarithmic fundamental frequency of the jth audio frame can be represented as log _ pitch_j。

And then, taking the statistic value of the logarithmic fundamental frequency corresponding to each audio frame in the training audio as the fundamental frequency of the training audio. For example, the mean value of the logarithmic fundamental frequency and the standard deviation of the logarithmic fundamental frequency corresponding to each audio frame can be expressed as log _ pitch _ mean and log _ pitch _ std as the fundamental frequency of the training audio, where log _ pitch _ mean can reflect the magnitude of the fundamental frequency of the training audio as a whole, and log _ pitch _ std can reflect the variation amplitude of the fundamental frequency of the training audio.

Step D3, if the real acoustic features include volume, extracting the volume of each audio frame included in the training audio to determine the volume of the training audio.

First, the training audio may be processed by using audio processing tools such as sox, library, and right to obtain the volume of each audio in the training audio, and then the logarithmic operation is performed on the volume corresponding to each audio frame to obtain the logarithmic volume corresponding to each audio frame. By the logarithm operation, the variation range of the volume can be compressed, and the variation degree of the volume can be amplified. For example, the volume corresponding to each audio frame can be expressed as energy_jAnd represents the volume corresponding to the jth audio frame. Accordingly, the logarithmic volume corresponding to the jth audio frame can be expressed as log _ energy_j。

Then, the statistic of the logarithmic volume corresponding to each audio frame is used as the volume of the training audio. For example, the average of the logarithmic volume corresponding to each audio frame may be used as the volume of the training audio, and may be represented as log _ energy _ mean.

Further, if the real acoustic features further include a noise level, step D may further include: and determining the noise level corresponding to the training audio according to the linear prediction coefficient corresponding to the training audio.

For example, an LPC Coefficient (chinese: Linear Prediction Coefficient) of the training audio may be determined, and then a 1 st dimension of the LPC Coefficient of the training audio may be subjected to a logarithm operation, and a result of the logarithm operation is used as a noise level corresponding to the training audio, and by the logarithm operation, a variation range of the noise level may be compressed, thereby amplifying a variation degree of the noise level. The noise level may be expressed as log _ spectral _ tilt, for example.

Under the condition that the real acoustic features comprise fundamental frequency, volume, speech rate and noise level, the fundamental frequency, the volume, the speech rate and the noise level of the training audio can be combined into the real acoustic features of the training audio. For example, the real acoustic feature may be a vector of 1 x 5 dimensions: { fundamental frequency: (log _ pitch _ mean, log _ pitch _ std), volume: log _ energy _ mean, speech rate: log _ duration _ mean, noise level: log _ spectral _ tilt }. Accordingly, in the case that the specified acoustic features include fundamental frequency, volume, speech rate and noise level, the specified acoustic features obtained in step 101 may also include the above-mentioned 5 dimensions.

In another application scenario, the speech synthesis model may also be obtained by training as follows:

and G, determining the statistical acoustic characteristics of the training set according to the real acoustic characteristics of a plurality of training audios in the preset training set.

And H, normalizing the real acoustic features of each training audio according to the statistical acoustic features.

Correspondingly, the implementation manner of step E may be:

and expanding the real acoustic features after the normalization processing according to the training phoneme sequence to obtain a training acoustic feature sequence.

For example, before the true acoustic features are expanded, normalization processing may be performed on the true acoustic features. For example, the training set includes a plurality of training texts, each corresponding to a training audio. The true acoustic features of each training audio may be determined and the statistical acoustic features of the training set determined in the manner of steps D1 through D4. The statistical acoustic feature may be, for example, a mean, a standard deviation, a variance, an extremum, or the like of the true acoustic feature. Then, the real acoustic features of each training audio are normalized according to the statistical acoustic features. For example, the mean μ and standard deviation σ of the real acoustic features can be used as statistical acoustic features, and then the real acoustic features between [ μ -3 σ, μ +3 σ ] are mapped into [ -1,1], and the real acoustic features outside [ μ -3 σ, μ +3 σ ] can be truncated to-1 or 1. And respectively obtaining the average value and the standard deviation of each dimension in the real acoustic features, and performing normalization processing on each dimension in the real acoustic features. For example, taking log _ pitch _ mean in real acoustic features as an example, the mean value of log _ pitch _ mean and the standard deviation of log _ pitch _ mean for each training audio can be found, then log _ pitch _ mean between [ pitch _ μ -3pitch _ σ, pitch _ μ +3pitch _ σ ] is mapped into [ -1,1], and log _ pitch _ mean outside [ pitch _ μ -3pitch _ σ, pitch _ μ +3pitch _ σ ] is truncated to-1 or 1 to normalize log _ pitch _ mean.

Further, the real acoustic features after normalization processing may be extended according to the training phoneme sequence to obtain a training acoustic feature sequence. For example, a training acoustic feature sequence may be generated according to the number of training phonemes included in the training phoneme sequence, where the training acoustic feature corresponding to each training phoneme is a real acoustic feature after normalization processing.

It should be noted that the specified acoustic features obtained in step 101 may also include the above 5 dimensions subjected to the normalization processing. The specified acoustic features after normalization processing are more explanatory. Taking the specified acoustic feature as { -1,1,0,1,0} for example, where the value of log _ pitch _ mean is-1, it means that the target audio generated by the speech synthesis model and conforming to the specified acoustic feature is characterized by heaviness. The value of log _ pitch _ std is 1, indicating that the fundamental frequency of the target audio varies greatly. The value of log _ energy _ mean is 0, indicating that the target audio is of normal volume. The log _ duration _ mean corresponds to a value of 1, which indicates that the speech rate of the target audio is slow (i.e., the average duration corresponding to the phoneme). The log _ spectral _ tilt corresponds to a value of 0, indicating that the noise level of the target audio is normal.

FIG. 7 is a flow diagram illustrating another method of speech synthesis according to an exemplary embodiment, as shown in FIG. 7, after step 104, the method further comprising:

and 105, updating the specified acoustic characteristics according to the preset step value.

And repeating the steps 103 to 105 until the target audio meets the preset condition.

For example, when a user is controlling speech synthesis, the desired emotion type (i.e., the specified emotion type) is often relatively definite, e.g., the target audio to be generated is expected to have a happy emotion. And the desired acoustic characteristics are often only a rough range. Therefore, after the target audio is output once by the speech synthesis model, the step value for updating the specified acoustic feature is set according to the specific requirements of the user, then the specified acoustic feature is updated by using the step value, and the steps 103 to 105 are repeatedly executed according to the updated specified acoustic feature until the target audio meets the preset condition. The preset condition may be that the user listens to the target audio, and a stop instruction is triggered when the requirement is satisfied, and the preset condition may be that the number of times of repeatedly executing steps 103 to 105 reaches a specified number of times (for example, 5 times). For example, the emotion type obtained in step 101 is designated as happy, and the acoustic features are designated as {0,0,0,0,0}, i.e., the fundamental frequency, the volume, the speech rate, and the noise level are average levels. Then the first time step 103 to step 104 are performed, the generated target audio has an emotional type with fun, and the fundamental frequency of the target audio is the average level, the volume is normal, the speech speed is normal, and the noise level is normal. After a user listens to the target audio, the speech speed is considered to be slow, the step value can be set to be {0,0,0, -0.5,0}, the step value and the specified acoustic feature are added to obtain the updated specified acoustic feature {0,0,0, -0.5,0}, and the process is repeated until a stop instruction triggered when the user considers that the target audio meets the requirement is received.

Fig. 8 is a block diagram illustrating a speech synthesis apparatus according to an exemplary embodiment, and as shown in fig. 8, the apparatus 200 includes:

the obtaining module 201 is configured to obtain a text to be synthesized, specify an acoustic feature, and specify an emotion type, where the specified acoustic feature is used to indicate a prosodic feature of an audio.

The extracting module 202 is configured to extract a phoneme sequence corresponding to the text to be synthesized.

And the extension module 203 is configured to extend the specified acoustic features according to the phoneme sequence to obtain an acoustic feature sequence.

And the synthesis module 204 is configured to input the phoneme sequence, the acoustic feature sequence, and the specified emotion type into a pre-trained speech synthesis model to obtain a target audio with the specified emotion type corresponding to the text to be synthesized, which is output by the speech synthesis model, where the acoustic feature of the target audio matches the specified acoustic feature.

In one application scenario, specifying acoustic features includes: at least one of fundamental frequency, volume and speech rate.

Fig. 9 is a block diagram illustrating another speech synthesis apparatus according to an exemplary embodiment, and as shown in fig. 9, the extension module 203 may include:

the determining sub-module 2031 is configured to determine, according to the specified acoustic features, the acoustic features corresponding to each phoneme in the phoneme sequence.

The expansion submodule 2032 is configured to combine the acoustic features corresponding to each phoneme into an acoustic feature sequence.

In one application scenario, the speech synthesis model may be used to perform the following steps:

In another application scenario, the speech synthesis model is obtained by training as follows:

and D, extracting real acoustic features of the training audio corresponding to the training text, wherein the real acoustic features are used for indicating prosodic features of the training audio.

In another application scenario, the speech synthesis model includes: a first module, a second module and a synthesis module, the implementation of step F may comprise the steps of:

In yet another application scenario, the real acoustic features include: at least one of fundamental frequency, volume and speech rate, and the step D may include the steps of:

First, according to the training audio and the training phoneme sequence, the duration corresponding to each training phoneme is determined.

And then, carrying out logarithm operation on the duration corresponding to each training phoneme to obtain the logarithm duration corresponding to each training phoneme.

And finally, taking the statistic value of the log duration corresponding to each training phoneme in the training phoneme sequence as the speech speed of the training audio.

Firstly, logarithmic operation is carried out on the fundamental frequency corresponding to each audio frame to obtain the logarithmic fundamental frequency corresponding to each audio frame.

And then, taking the statistic value of the logarithmic fundamental frequency corresponding to each audio frame in the training audio as the fundamental frequency of the training audio.

Firstly, carrying out logarithmic operation on the volume corresponding to each audio frame to obtain the logarithmic volume corresponding to each audio frame.

And then, taking the statistic value of the logarithmic volume corresponding to each audio frame as the volume of the training audio.

Fig. 10 is a block diagram illustrating another speech synthesis apparatus according to an exemplary embodiment, and as shown in fig. 10, the apparatus 200 may further include:

and the updating module 205 is configured to update the specified acoustic feature according to a preset step value after inputting the phoneme sequence, the acoustic feature sequence, and the specified emotion type into the pre-trained speech synthesis model to obtain a target audio with the specified emotion type corresponding to the text to be synthesized, which is output by the speech synthesis model.

And repeatedly executing the step of expanding the specified acoustic features according to the phoneme sequence to obtain an acoustic feature sequence, and updating the specified acoustic features according to the preset step value until the target audio meets the preset condition.

With regard to the apparatus in the above-described embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.

Referring now to fig. 11, a schematic diagram of an electronic device (i.e., the execution body of the speech synthesis method described above)) 300 suitable for implementing embodiments of the present disclosure is shown. The terminal device in the embodiments of the present disclosure may include, but is not limited to, a mobile terminal such as a mobile phone, a notebook computer, a digital broadcast receiver, a PDA (personal digital assistant), a PAD (tablet computer), a PMP (portable multimedia player), a vehicle terminal (e.g., a car navigation terminal), and the like, and a stationary terminal such as a digital TV, a desktop computer, and the like. The electronic device shown in fig. 11 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present disclosure.

As shown in fig. 11, the electronic device 300 may include a processing means (e.g., a central processing unit, a graphics processor, etc.) 301 that may perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM)302 or a program loaded from a storage means 308 into a Random Access Memory (RAM) 303. In the RAM 303, various programs and data necessary for the operation of the electronic apparatus 300 are also stored. The processing device 301, the ROM 302, and the RAM 303 are connected to each other via a bus 304. An input/output (I/O) interface 305 is also connected to bus 304.

Generally, the following devices may be connected to the I/O interface 305: input devices 306 including, for example, a touch screen, touch pad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; an output device 307 including, for example, a Liquid Crystal Display (LCD), a speaker, a vibrator, and the like; storage devices 308 including, for example, magnetic tape, hard disk, etc.; and a communication device 309. The communication means 309 may allow the electronic device 300 to communicate wirelessly or by wire with other devices to exchange data. While fig. 11 illustrates an electronic device 300 having various means, it is to be understood that not all illustrated means are required to be implemented or provided. More or fewer devices may alternatively be implemented or provided.

In particular, according to an embodiment of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program carried on a non-transitory computer readable medium, the computer program containing program code for performing the method illustrated by the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network through the communication means 309, or installed from the storage means 308, or installed from the ROM 302. The computer program, when executed by the processing device 301, performs the above-described functions defined in the methods of the embodiments of the present disclosure.

It should be noted that the computer readable medium in the present disclosure can be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, 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 disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present disclosure, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In contrast, in the present disclosure, a computer readable signal medium may comprise a propagated data signal with computer readable program code embodied therein, either in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, optical cables, RF (radio frequency), etc., or any suitable combination of the foregoing.

In some embodiments, the terminal devices, servers may communicate using any currently known or future developed network Protocol, such as HTTP (HyperText Transfer Protocol), and may interconnect with any form or medium of digital data communication (e.g., a communications network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), the Internet (e.g., the Internet), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future developed network.

The computer readable medium may be embodied in the electronic device; or may exist separately without being assembled into the electronic device.

The computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to: acquiring a text to be synthesized, specified acoustic features and specified emotion types, wherein the specified acoustic features are used for indicating rhythm features of audio; extracting a phoneme sequence corresponding to the text to be synthesized; expanding the specified acoustic features according to the phoneme sequence to obtain an acoustic feature sequence; inputting the phoneme sequence, the acoustic feature sequence and the specified emotion type into a pre-trained speech synthesis model to obtain a target audio which is output by the speech synthesis model and corresponds to the text to be synthesized and has the specified emotion type, wherein the acoustic feature of the target audio is matched with the specified acoustic feature.

Computer program code for carrying out operations for the present disclosure may be written in any combination of one or more programming languages, including but not limited to an object oriented programming language such as Java, Smalltalk, C + +, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).

The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The modules described in the embodiments of the present disclosure may be implemented by software or hardware. Where the name of a module does not in some cases constitute a limitation of the module itself, for example, the retrieving module may also be described as a "module that retrieves text to be synthesized, specifies acoustic features, and specifies emotion types".

The functions described herein above may be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field Programmable Gate Arrays (FPGAs), Application Specific Integrated Circuits (ASICs), Application Specific Standard Products (ASSPs), systems on a chip (SOCs), Complex Programmable Logic Devices (CPLDs), and the like.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The 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, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, 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 disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

Example 1 provides a speech synthesis method, according to one or more embodiments of the present disclosure, including: acquiring a text to be synthesized, specified acoustic features and specified emotion types, wherein the specified acoustic features are used for indicating rhythm features of audio; extracting a phoneme sequence corresponding to the text to be synthesized; expanding the specified acoustic features according to the phoneme sequence to obtain an acoustic feature sequence; inputting the phoneme sequence, the acoustic feature sequence and the specified emotion type into a pre-trained speech synthesis model to obtain a target audio which is output by the speech synthesis model and corresponds to the text to be synthesized and has the specified emotion type, wherein the acoustic feature of the target audio is matched with the specified acoustic feature.

Example 2 provides the method of example 1, wherein the extending the specified acoustic feature according to the phoneme sequence to obtain an acoustic feature sequence includes: determining the acoustic feature corresponding to each phoneme in the phoneme sequence according to the specified acoustic feature; and combining the acoustic features corresponding to each phoneme into the acoustic feature sequence.

Example 3 provides the method of example 1, the speech synthesis model to: determining a text feature sequence corresponding to the text to be synthesized according to the phoneme sequence, wherein the text feature sequence comprises text features corresponding to each phoneme in the phoneme sequence; determining appointed emotional characteristics corresponding to the appointed emotional types, and expanding the appointed emotional characteristics according to the phoneme sequence to obtain an emotional characteristic sequence; and generating the target audio according to the text feature sequence, the acoustic feature sequence and the emotion feature sequence.

Example 4 provides the method of example 1, the specifying acoustic characteristics including: at least one of fundamental frequency, volume and speech rate.

Example 5 provides the method of example 1, the speech synthesis model being obtained by training in the following manner: extracting real acoustic features of training audio corresponding to a training text, wherein the real acoustic features are used for indicating prosodic features of the training audio; extending the real acoustic features according to a training phoneme sequence corresponding to the training text to obtain a training acoustic feature sequence; inputting the training phoneme sequence, the training acoustic feature sequence and the training emotion type corresponding to the training audio into the speech synthesis model, and training the speech synthesis model according to the output of the speech synthesis model and the training audio.

Example 6 provides the method of example 5, the true acoustic features comprising: at least one of fundamental frequency, volume and speech rate; the extracting of the real acoustic features of the training audio corresponding to the training text includes: if the real acoustic features comprise the speech rate, determining the duration corresponding to each training phoneme in the training phoneme sequence according to the training audio and the training phoneme sequence so as to determine the speech rate of the training audio; if the real acoustic features comprise fundamental frequencies, extracting the fundamental frequency of each audio frame included in the training audio to determine the fundamental frequency of the training audio; if the real acoustic features comprise volume, extracting the volume of each audio frame included in the training audio to determine the volume of the training audio.

Example 7 provides the method of example 6, wherein determining a duration corresponding to each training phoneme in the training phoneme sequence to determine a speech rate of the training audio according to the training audio and the training phoneme sequence includes: determining the duration corresponding to each training phoneme according to the training audio and the training phoneme sequence; carrying out logarithm operation on the duration corresponding to each training phoneme to obtain the logarithm duration corresponding to each training phoneme; taking the statistic value of the log duration corresponding to each training phoneme in the training phoneme sequence as the speech speed of the training audio; the extracting a fundamental frequency of each audio frame comprised by the training audio to determine a fundamental frequency of the training audio comprises: carrying out logarithmic operation on the fundamental frequency corresponding to each audio frame to obtain the logarithmic fundamental frequency corresponding to each audio frame; taking the statistic value of the logarithmic fundamental frequency corresponding to each audio frame in the training audio as the fundamental frequency of the training audio; the extracting a volume of each audio frame included in the training audio to determine the volume of the training audio includes: carrying out logarithmic operation on the volume corresponding to each audio frame to obtain the logarithmic volume corresponding to each audio frame; and taking the statistic value of the logarithmic volume corresponding to each audio frame in the training audio as the volume of the training audio.

Example 8 provides the method of example 5, the speech synthesis model comprising: the method comprises a first module, a second module and a synthesis module, wherein the inputting of the training phoneme sequence, the training acoustic feature sequence and the training emotion type corresponding to the training audio into the speech synthesis model comprises: extracting a training text feature sequence corresponding to the training phoneme sequence through the first module, wherein the training text feature sequence comprises a text feature corresponding to each training phoneme in the training phoneme sequence; extracting training emotional characteristics corresponding to the training emotional types through the second module, and expanding the training emotional characteristics according to the training phoneme sequence to obtain a training emotional characteristic sequence; generating, by the synthesis module, an output of the speech synthesis model according to the training text feature sequence, the training acoustic feature sequence, and the training emotional feature sequence; the training the speech synthesis model based on the output of the speech synthesis model and the training audio includes: and determining a loss function of the speech synthesis model according to the output of the speech synthesis model and the training audio, and updating the first module, the second module and the synthesis module according to the loss function.

Example 9 provides the method of examples 1 to 8, where after the inputting the phoneme sequence, the acoustic feature sequence, and the specified emotion type into a pre-trained speech synthesis model to obtain a target audio output by the speech synthesis model and corresponding to the text to be synthesized and having the specified emotion type, the method further includes: updating the specified acoustic features according to a preset stepping value; and repeatedly executing the step of expanding the specified acoustic features according to the phoneme sequence to obtain an acoustic feature sequence, and updating the specified acoustic features according to a preset step value until the target audio meets a preset condition.

Example 10 provides, in accordance with one or more embodiments of the present disclosure, a speech synthesis apparatus comprising: the system comprises an acquisition module, a synthesis module and a feedback module, wherein the acquisition module is used for acquiring a text to be synthesized, specified acoustic features and specified emotion types, and the specified acoustic features are used for indicating rhythm features of audio; the extraction module is used for extracting a phoneme sequence corresponding to the text to be synthesized; the extension module is used for extending the specified acoustic features according to the phoneme sequence to obtain an acoustic feature sequence; and the synthesis module is used for inputting the phoneme sequence, the acoustic feature sequence and the specified emotion type into a pre-trained voice synthesis model so as to obtain a target audio which is output by the voice synthesis model and corresponds to the text to be synthesized and has the specified emotion type, and the acoustic feature of the target audio is matched with the specified acoustic feature.

Example 11 provides a computer-readable medium having stored thereon a computer program that, when executed by a processing apparatus, implements the steps of the methods of examples 1-9, in accordance with one or more embodiments of the present disclosure.

Example 12 provides, in accordance with one or more embodiments of the present disclosure, an electronic device, comprising: a storage device having a computer program stored thereon; processing means for executing the computer program in the storage means to implement the steps of the methods of examples 1 to 9.

The foregoing description is only exemplary of the preferred embodiments of the disclosure and is illustrative of the principles of the technology employed. It will be appreciated by those skilled in the art that the scope of the disclosure herein is not limited to the particular combination of features described above, but also encompasses other embodiments in which any combination of the features described above or their equivalents does not depart from the spirit of the disclosure. For example, the above features and (but not limited to) the features disclosed in this disclosure having similar functions are replaced with each other to form the technical solution.

Further, while operations are depicted in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Likewise, while several specific implementation details are included in the above discussion, these should not be construed as limitations on the scope of the disclosure. Certain features that are described in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable subcombination.

Although the subject matter has been described in language specific to structural features and/or methodological acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are disclosed as example forms of implementing the claims. With regard to the apparatus in the above-described embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.

Claims

1. A method of speech synthesis, the method comprising:

extracting a phoneme sequence corresponding to the text to be synthesized;

2. The method of claim 1, wherein the extending the specified acoustic features according to the phoneme sequence to obtain an acoustic feature sequence comprises:

determining the acoustic feature corresponding to each phoneme in the phoneme sequence according to the specified acoustic feature;

and combining the acoustic features corresponding to each phoneme into the acoustic feature sequence.

3. The method of claim 1, wherein the speech synthesis model is used to:

determining a text feature sequence corresponding to the text to be synthesized according to the phoneme sequence, wherein the text feature sequence comprises text features corresponding to each phoneme in the phoneme sequence;

determining appointed emotional characteristics corresponding to the appointed emotional types, and expanding the appointed emotional characteristics according to the phoneme sequence to obtain an emotional characteristic sequence;

and generating the target audio according to the text feature sequence, the acoustic feature sequence and the emotion feature sequence.

4. The method of claim 1, wherein the specifying acoustic features comprises: at least one of fundamental frequency, volume and speech rate.

5. The method of claim 1, wherein the speech synthesis model is obtained by training as follows:

extracting real acoustic features of training audio corresponding to a training text, wherein the real acoustic features are used for indicating prosodic features of the training audio;

extending the real acoustic features according to a training phoneme sequence corresponding to the training text to obtain a training acoustic feature sequence;

inputting the training phoneme sequence, the training acoustic feature sequence and the training emotion type corresponding to the training audio into the speech synthesis model, and training the speech synthesis model according to the output of the speech synthesis model and the training audio.

6. The method of claim 5, wherein the true acoustic features comprise: at least one of fundamental frequency, volume and speech rate; the extracting of the real acoustic features of the training audio corresponding to the training text includes:

if the real acoustic features comprise the speech rate, determining the duration corresponding to each training phoneme in the training phoneme sequence according to the training audio and the training phoneme sequence so as to determine the speech rate of the training audio;

if the real acoustic features comprise fundamental frequencies, extracting the fundamental frequency of each audio frame included in the training audio to determine the fundamental frequency of the training audio;

if the real acoustic features comprise volume, extracting the volume of each audio frame included in the training audio to determine the volume of the training audio.

7. The method of claim 6, wherein determining a duration corresponding to each training phoneme in the training phoneme sequence based on the training audio and the training phoneme sequence to determine the speech rate of the training audio comprises:

determining the duration corresponding to each training phoneme according to the training audio and the training phoneme sequence;

carrying out logarithm operation on the duration corresponding to each training phoneme to obtain the logarithm duration corresponding to each training phoneme;

taking the statistic value of the log duration corresponding to each training phoneme in the training phoneme sequence as the speech speed of the training audio;

the extracting a fundamental frequency of each audio frame comprised by the training audio to determine a fundamental frequency of the training audio comprises:

carrying out logarithmic operation on the fundamental frequency corresponding to each audio frame to obtain the logarithmic fundamental frequency corresponding to each audio frame;

taking the statistic value of the logarithmic fundamental frequency corresponding to each audio frame in the training audio as the fundamental frequency of the training audio;

the extracting a volume of each audio frame included in the training audio to determine the volume of the training audio includes:

carrying out logarithmic operation on the volume corresponding to each audio frame to obtain the logarithmic volume corresponding to each audio frame;

and taking the statistic value of the logarithmic volume corresponding to each audio frame in the training audio as the volume of the training audio.

8. The method of claim 5, wherein the speech synthesis model comprises: the method comprises a first module, a second module and a synthesis module, wherein the inputting of the training phoneme sequence, the training acoustic feature sequence and the training emotion type corresponding to the training audio into the speech synthesis model comprises:

extracting a training text feature sequence corresponding to the training phoneme sequence through the first module, wherein the training text feature sequence comprises a text feature corresponding to each training phoneme in the training phoneme sequence;

extracting training emotional characteristics corresponding to the training emotional types through the second module, and expanding the training emotional characteristics according to the training phoneme sequence to obtain a training emotional characteristic sequence;

generating, by the synthesis module, an output of the speech synthesis model according to the training text feature sequence, the training acoustic feature sequence, and the training emotional feature sequence;

the training the speech synthesis model based on the output of the speech synthesis model and the training audio includes:

and determining a loss function of the speech synthesis model according to the output of the speech synthesis model and the training audio, and updating the first module, the second module and the synthesis module according to the loss function.

9. The method according to any one of claims 1-8, wherein after the inputting the phoneme sequence, the acoustic feature sequence and the specified emotion type into a pre-trained speech synthesis model to obtain the target audio output by the speech synthesis model, corresponding to the text to be synthesized and having the specified emotion type, the method further comprises:

updating the specified acoustic features according to a preset stepping value;

and repeatedly executing the step of expanding the specified acoustic features according to the phoneme sequence to obtain an acoustic feature sequence, and updating the specified acoustic features according to a preset step value until the target audio meets a preset condition.

10. A speech synthesis apparatus, characterized in that the apparatus comprises:

11. A computer-readable medium, on which a computer program is stored, characterized in that the program, when being executed by processing means, carries out the steps of the method of any one of claims 1-9.

12. An electronic device, comprising:

a storage device having a computer program stored thereon;

processing means for executing the computer program in the storage means to carry out the steps of the method according to any one of claims 1 to 9.