WO2017088136A1 - Translation method and terminal - Google Patents

Abstract

Disclosed are a translation method and a terminal, relating to an intelligent voice translation technique, and used to solve the problem that an existing translation method is relatively low in translation accuracy. The method provided in the embodiments of the present invention comprises: acquiring an audio signal sent by a user, wherein the audio signal contains a voice segment signal; performing semantic analysis on the voice segment signal in the audio signal, and if a feature point exists in the voice segment signal, dividing the voice segment signal into at least one voice sub-segment signal by taking the feature point as a division point; and translating the at least one voice sub-segment signal into a voice signal conforming to a target user language, and sending the translated voice signal to a target terminal.

Description

Translation method and terminal Technical field

The invention relates to a voice intelligent translation technology, in particular to a translation method and a terminal.

Background technique

With the continuous development and progress of the country's trade and exchange with the country, the trend of users' internationalization will inevitably lead to problems in people of different languages when using mobile phones for communication. For example, Chinese and English native speakers are used as examples. Proficiency in English can communicate with English users, and few English users understand Chinese. Therefore, language has become the biggest obstacle to international communication, making the demand for instant language translation during the call more and more important.

The current translation technology is mainly based on the speech endpoint detection (English full name: voice activity detection, English abbreviation: VAD) technology, detecting the silent segment in the continuous sentence, using the silent segment as the segmentation point, dividing the continuous sentence into multiple short sentences. To achieve real-time translation during the call. However, this method of deciphering and translating from the physical layer to judge the user's pause time is completely out of the translation scenario, because the user may have ambient noise, background sound, and a mantra during the conversation. This "such voice signal that does not provide the silent time required for VAD detection but does not have obvious semantics causes the sentence to fail or the sentence to be unreasonable, resulting in translation distortion and reduced translation accuracy.

Summary of the invention

Embodiments of the present invention provide a translation method and a terminal to solve the problem that the translation accuracy of the existing translation method is low.

In order to achieve the above object, embodiments of the present invention adopt the following technical solutions:

In a first aspect, the embodiment of the present invention provides a translation method, which is applied to a terminal that is performing a voice call, and the terminal may be a sender that is transmitting the local voice to the target terminal. The method may include:

Obtaining an audio signal input by the user and including a voice segment signal;

Performing semantic analysis on the voice segment signal, if there is a feature point in the voice segment signal, dividing the voice segment signal into at least one sub-segment segment signal by using the feature point as a segmentation point;

Translating the at least one sub-segment segment signal into a voice conforming to a target user language, and transmitting the translated voice signal to the target terminal.

In this way, the speech segment without the complete semantics in the speech segment signal can be eliminated, and at the same time, the sub-speech segment segmentation is guaranteed to be a speech segment with complete semantics, which improves the translation accuracy compared with the existing translation method.

At the same time, in order to achieve simultaneous translation and improve translation efficiency, the present invention performs a pipeline mode of playing after completion of a complete semantic statement translation, and a speech of the next complete semantics while the speech is played after the translation of the sentence. The speech synthesis technology superimposed in the speech for playing, superimposes the original speech and the translated speech synthesis, the original sound volume is reduced as the background sound, and the translated speech is sent to the target terminal as the master volume, and the specific implementation is as follows:

Transmitting, to the target terminal, a first sub-voice signal;

After the first sub-speech signal is played to the target user, the translated sub-speech signal and the second sub-speech signal are synthesized;

The synthesized voice signal is sent to the target terminal.

In this way, it is not necessary to wait until all the statements are played, and then the translated sentences are played sentence by sentence. Compared with the existing play mode, the translation play time is advanced, which reduces the translation waiting delay, improves the translation efficiency, and enhances the user experience.

Since the transmitting end and the receiving end are relative concepts during the dual-talking process, the calling party is determined to be the transmitting end and the listening party is determined as the receiving end according to the situation in which the double call is being called; therefore, at a certain end At the moment, the above transmitting end can serve as a receiving end. When the sending end is used as the receiving end and the translation function is performed, the method may further include:

Receiving an audio signal sent by the source terminal; the audio signal includes a voice segment signal; and the voice segment signal is a segment of the voice signal whose power value is greater than a preset threshold;

Performing semantic analysis on the speech segment signal in the audio signal;

If there is a feature point in the voice segment signal, the feature point is used as a segmentation point. Segmenting the speech segment signal into at least one sub-segment segment signal; the feature point is a time point at which the speech signal without complete semantics is located;

Translating the at least one sub-segment segment signal into a speech signal of a preset language;

Play the translated voice signal.

Similarly, in order to achieve simultaneous translation and improve translation efficiency, the method may further include:

Playing the first sub-speech signal;

Synthesizing the translated speech signal and the second sub-speech signal of the first sub-speech signal;

Play the synthesized speech signal.

In a second aspect, the embodiment of the present invention further provides a terminal, where the terminal may be a sender that is integrated with a translation function, and is configured to perform the foregoing translation method, where the terminal may include:

An audio processing module, configured to acquire an audio signal input by a user; the audio signal includes a voice segment signal; and the voice segment signal is a segment of a voice signal whose power value is greater than a preset threshold;

a voice endpoint detection module, configured to perform semantic analysis on a voice segment signal in an audio signal acquired by the audio processing module; if a feature point exists in the voice segment signal, the feature point is used as a segmentation point, and the The speech segment signal is segmented into at least one sub-segment segment signal; the feature point is a time point at which the speech signal without complete semantics is located;

a translation module, configured to translate at least one sub-segment segment signal detected by the speech endpoint detection module into a speech signal conforming to a target user language;

a voice synthesis module, configured to send the translated voice signal of the translation module to the target terminal.

In this way, the speech segment without the complete semantics in the speech segment signal can be eliminated, and at the same time, the sub-speech segment segmentation is guaranteed to be a speech segment with complete semantics, which improves the translation accuracy compared with the existing translation method.

At the same time, in order to achieve simultaneous translation and improve translation efficiency, the present invention implements a pipeline mode of playing after completion of a complete semantic statement translation, and At the same time as the speech playback after the translation of the sentence, the speech synthesis technology in which the next complete semantic speech is superimposed in the speech, the original sound volume is reduced as the background sound, and the translated speech is played as the master volume, specifically, The speech synthesis module is used to:

After the first sub-speech signal is played to the target user, the speech signal translated by the first sub-speech signal and the second sub-speech signal are synthesized by speech;

The synthesized voice is played to the target user.

In this way, it is not necessary to wait until all the statements are played, and then the translated sentences are played sentence by sentence. Compared with the existing play mode, the translation play time is advanced, which reduces the translation waiting delay, improves the translation efficiency, and enhances the user experience.

Since the transmitting end and the receiving end are relative concepts, usually according to the situation in which the double call is being called, the speaking party is determined as the transmitting end, and the listening party is determined as the receiving end. Therefore, at a certain moment, the terminal can be received as a receiving end. The audio processing module is further configured to: when the terminal is used as the receiving end, the audio processing module is further configured to:

Receiving an audio signal sent by the source terminal; the audio signal includes a voice segment signal; and the voice segment signal is a segment of the voice signal whose power value is greater than a preset threshold;

The voice endpoint detection module may be further configured to perform semantic analysis on a voice segment signal in the audio signal acquired by the audio processing module; if a feature point exists in the voice segment signal, the feature point is segmented Pointing, dividing the speech segment signal into at least one sub-speech segment signal; the feature point is a time point at which the speech signal without complete semantics is located;

The translation module may be further configured to translate the at least one sub-segment segment signal detected by the speech endpoint detection module into a speech signal of a preset language;

The speech synthesis module may be further configured to play the translated speech signal of the translation module.

Similarly, in order to achieve simultaneous translation and improve translation efficiency, the speech synthesis module can also be used to:

Playing the first sub-speech signal;

After playing the first sub-speech signal, synthesizing the translated sub-speech signal and the second sub-speech signal;

Play the synthesized speech signal.

In a third aspect, the embodiment of the present invention further provides a terminal, where the terminal may be a sender that is integrated with a translation function, and is configured to perform the foregoing translation method, where the terminal may include:

An input device, configured to acquire an audio signal input by a user; the audio signal includes a voice segment signal; and the voice segment signal is a segment of a voice signal whose power value is greater than a preset threshold;

a processor, configured to perform semantic analysis on a voice segment signal in the audio signal acquired by the input device; if a feature point exists in the voice segment signal, the feature segment is used as a segmentation point, and the voice segment is Separating the signal into at least one sub-segment segment signal; the feature point is a time point at which the speech signal without complete semantics is located;

Translating the at least one sub-segment segment signal into a speech signal conforming to a target user language;

And a transmitter, configured to send the translated voice signal of the processor to the target terminal.

In this way, the speech segment without the complete semantics in the speech segment signal can be eliminated, and at the same time, the sub-speech segment segmentation is guaranteed to be a speech segment with complete semantics, which improves the translation accuracy compared with the existing translation method.

At the same time, in order to achieve simultaneous translation and improve translation efficiency, the present invention performs a pipeline mode of playing after completion of a complete semantic statement translation, and a speech of the next complete semantics while the speech is played after the translation of the sentence. The speech synthesis technology superimposed on the speech for playing, the original sound volume is reduced as the background sound, and the translated speech is played as the master volume. Specifically, the processor is further used for:

Before the transmitter sends the translated voice signal of the processor to the target terminal, synthesizing the translated voice signal and the second sub-voice signal of the first sub-voice signal;

The transmitter is specifically configured to:

Transmitting the first sub-voice signal to the target terminal;

And transmitting the synthesized voice signal to the target terminal.

In this way, it is not necessary to wait until all the statements have been played, and then the translated sentences are played sentence by sentence. Compared with the existing play mode, the translation play time is advanced, and the translation is reduced. Time delays improve translation efficiency and enhance user experience.

Since the transmitting end and the receiving end are relative concepts, usually according to the situation in which the double call is being called, the speaking party is determined as the transmitting end, and the listening party is determined as the receiving end. Therefore, at a certain moment, the terminal can also be used as The receiving end, when the terminal is used as the receiving end, the terminal may further include:

a receiver, configured to receive an audio signal sent by the source terminal; the audio signal includes a voice segment signal; and the voice segment signal is a segment of the voice signal whose power value is greater than a preset threshold;

The processor may be further configured to perform semantic analysis on a voice segment signal in the audio signal acquired by the receiver; if a feature point exists in the voice segment signal, the feature point is used as a segmentation point, and The speech segment signal is segmented into at least one sub-segment segment signal; the feature point is a time point at which the speech signal without complete semantics is located;

Translating the at least one sub-segment segment signal into a speech signal of a preset language;

And an output device, configured to play the translated voice signal of the processor.

Similarly, in order to achieve simultaneous translation and improve translation efficiency, the processor is also used to:

Before the output device plays the translated speech signal of the processor, synthesizing the translated speech signal and the second sub-speech signal of the first sub-speech signal;

The output device is specifically configured to:

Playing the first sub-speech signal;

Playing the synthesized speech signal.

As can be seen from the above, an embodiment of the present invention provides a translation method and a terminal, which acquires a frame of an audio signal sent by a source user; the audio signal includes a voice segment signal; and performs semantic analysis on the voice segment signal in the audio signal, and detects Whether there is a feature point in the speech segment signal; the feature point is: a time point at which the speech signal without complete semantics is located; if there is a feature point in the speech segment signal, the feature point is used as a segmentation point And dividing the voice segment signal into at least one sub-segment segment signal; translating the at least one sub-segment segment signal into a voice conforming to a target user language, and playing the translated voice to the target user. So, based on semantic analysis, the statement detected by the VAD endpoint is not Speech culling with complete semantics is divided into shorter and complete semantic statements, which fully expresses the speaker's sentence meaning, avoids the occurrence of sentence or half sentence, and effectively improves the accuracy of instant translation during the call. At the same time, the speech synthesis technology through pipeline mode + two-channel audio superposition does not need to wait until all the statements are played, and then the translated sentences are played sentence by sentence. Compared with the existing play mode, the translation play time is advanced, which is reduced. Translation latency is delayed, which improves translation efficiency and enhances user experience.

DRAWINGS

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments or the description of the prior art will be briefly described below. Obviously, the drawings in the following description are only It is a certain embodiment of the present invention, and other drawings can be obtained from those skilled in the art without any creative work.

FIG. 1 is a schematic structural diagram of a terminal according to an embodiment of the present disclosure;

2 is a schematic flowchart of a translation method according to an embodiment of the present invention;

FIG. 3 is a schematic flowchart diagram of a translation method according to an embodiment of the present disclosure;

4 is a sequence diagram of real-time translation according to an embodiment of the present invention;

FIG. 5 is a schematic structural diagram of a terminal according to an embodiment of the present invention.

detailed description

The core idea of the invention is to integrate real-time speech translation function between different languages in an existing mobile phone, and to segment the voice information input by the calling user or the voice information sent by the called user based on semantic analysis, and eliminate the voice information. A speech that does not have any semantics, divides the speech information into shorter short sentences with complete semantics, and at the same time, adopts a pipeline mode in which a complete semantic statement is translated and then played, and after the statement is translated. At the same time of voice playback, the speech synthesis technology in which the next complete semantic voice is superimposed in the voice plays the translated voice to the target user, so as to realize the support for the mobile phone users of different languages.

It can be understood that the calling party and the called party are relative concepts according to the embodiment of the present invention, and the calling initiator is usually referred to as a “calling party” according to the order in which the call is sent, and correspondingly, the call is received. Called "called"; the voice information can be included A segment of speech segment signals with multiple semantic but no mute segment signals, each statement can express a meaning, and the time interval between the sentences is relatively short (basically using existing endpoint detection techniques is indistinguishable); for example, according to People's habits of communication usually end up with a language that expresses different meanings, such as: "What is good to eat ~ ah ~ amount ~ which ~ eat noodles" although an uninterrupted piece of voice signal, but "~ "Hmm ~ amount ~ which ~" does not have any special meaning, so you can divide the voice information into two meanings according to semantic analysis: First, what to eat; Second, eat noodles. It should be noted that the mute segment signal is a segment of the signal that the power value of the voice signal is lower than the preset threshold and the duration is greater than the preset time value, such as a segment of the complete sentence to be sent by the user to the peer end, such as The voice signal has a power value lower than 0 dB, and a voice signal having a duration greater than 500 ms can be used as the silent segment signal; correspondingly, the data with the power value of the voice signal greater than the preset threshold is a voice segment signal; wherein, the preset The threshold value and the preset time value can be set as needed, and the comparison of the embodiments of the present invention is not limited.

The technical solutions in the embodiments of the present invention are clearly and completely described in the following with reference to the drawings in the embodiments of the present invention. It is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts are within the scope of the present invention.

1 is a schematic structural diagram of a terminal 10 integrated with a translation function according to an embodiment of the present invention, which is used to implement real-time translation during a user call. As shown in FIG. 1 , the terminal 10 may be composed of the following modules: The processing module 101, the voice endpoint detection module 102, the voice recognition module 103, the translation module 104, and the voice synthesis module 105;

The audio processing module 101: may include an audio driver, a digital signal processing (DSP), a modem (Modem), a codec, a microphone (MIC), a speaker (Specker, SPK for short), and the like. The recording and playback function is mainly provided for receiving the audio signal sent by the calling user, and the audio signal is sent to the voice endpoint detection module 102 for subsequent translation work, and the voice translated into the called language is subjected to analog-to-digital conversion, modulation, and encoding. After being processed, it is played to the called party, or receives the audio signal sent by the called user, and the audio signal is subjected to digital-to-analog conversion, demodulation, After decoding and the like, the processing is sent to the voice endpoint detection module 102 for subsequent translation work, and the called voice translated into the calling language is played to the calling user; among them, the audio driver, the DSP, the Modem, the Codec, the MIC, the SPK, and the like are sub-modules. It is a common module in the existing audio processing, and will not be described in detail here.

The speech endpoint detection module 102: based on the semantic database in the speech recognition module 103, detects a semantically independent statement in the speech segment signal, and provides the detected statement to the speech recognition module 103 for text conversion.

The speech recognition module 103 can include a semantic database, and provides a judgment basis for the sentence detection of the speech endpoint detection module 102, and converts the statement detected by the speech endpoint detection module 102 into text information.

The translation module 104 is mainly used to translate the text information converted by the speech recognition module 103 into text information conforming to the target (calling or called) language.

The speech synthesis module 105 is mainly configured to convert the translated text information of the translation module 104 into voice information and send it to the audio processing module 101, and the audio processing module 101 plays the target information to the target user.

The voice synthesizing module 105 can also be used to superimpose the sounds translated by the original sound and the speech synthesis module 105, and the original sound volume is reduced as the background sound, and the translated speech is used in order to prevent the translated speech from affecting the original speech. As the master volume, it achieves the effect of similar on-site simultaneous translation.

It can be understood that, for any two mobile phone users who implement instant voice translation during a mobile phone call, the above-mentioned module for performing the translation function can be integrated into one terminal, or can be integrated into two terminals of the call, that is, In the present invention, any one of the call terminals can adopt the structure shown in FIG. 2 to implement the instant translation function during the call. Specifically, the terminal application of the present invention can adopt the following basic architectures in practical applications: 1 the calling terminal adopts the structure shown in FIG. 2, the called terminal remains unchanged; 2 the calling terminal remains unchanged, and the called terminal The structure shown in FIG. 2 is adopted; 3 the calling terminal and the called terminal adopt the structure shown in FIG. 2, that is, each communication terminal can support the translation of the first language to the second language, and the second language to the first language. The embodiment of the present invention is not limited, and the present invention only uses the module that will perform the translation function. The block is concentrated in one terminal and is described as an example.

For convenience of description, the following embodiment 1 shows and describes in detail the process of the automatic translation of the terminal 10 of the present invention in the form of steps, wherein the steps shown may also be in addition to the terminal 10 such as a set of executable instructions. Executed in a computer system. Moreover, although logical sequences are shown in the figures, in some cases the steps shown or described may be performed in a different order than the ones described herein.

Embodiment 1

2 is a flowchart of a translation method according to an embodiment of the present disclosure, which is applied to a terminal as shown in FIG. 1 , and the terminal and the opposite terminal are performing a voice call, and the local voice is being sent to the target at the current moment. The terminal (ie, the peer terminal), as shown in FIG. 2, the method may include:

Step 101: Acquire an audio signal input by a user; the audio signal includes a voice segment signal; and the voice segment signal is a segment of a voice signal whose power value is greater than a preset threshold.

The user is a user who is talking during the current call, and is a local user who holds the terminal.

Wherein, the audio signal is a carrier of frequency and amplitude change information of regular sound waves with voice, music and sound effects, and according to the characteristics of sound waves, it may be a continuously changing analog signal, which is divided in time and can be divided. For multi-segment speech signals, usually, the audio signal has three important parameters: frequency, amplitude and phase to determine the characteristics of the audio signal; the square value of the signal amplitude of the audio signal can be determined as the power value of the speech signal (in dB) Unit), used to indicate the intensity of the voice signal, that is, the volume.

The mute segment signal is a segment of a signal in which a power value of the voice signal is lower than a preset threshold and the duration is greater than a preset time value, such as a power value of the voice signal, in a complete sentence to be sent by the user to the opposite end. A voice signal below 0 dB and having a duration greater than 500 ms may be referred to as a silent segment signal; correspondingly, the data whose voice signal has a power value greater than a preset threshold is a voice segment signal; wherein, the preset threshold value The preset time value can be set as needed, and the comparison of the embodiments of the present invention is not limited.

Optionally, in the embodiment of the present invention, the existing voice endpoint detection (voice) may be adopted. The activity detection (VAD) technology detects the audio signal by first determining that the voice signal has a power value less than a preset threshold, and the voice signal having a duration greater than the preset time value is determined as a silent segment signal, and then The mute segment signal is a segmentation point, and the audio signal is segmented to obtain at least one segment of the speech segment.

For example, the opposite user said: "The weather is very good today, let's go eat, what is good to eat ~ um ~ amount ~ which ~ eat noodles", among them, "today is good weather" and "we go to eat" If the power value of the voice signal is lower than the preset threshold and the duration is greater than the preset time value, it is determined that the voice signal sent between the two sentences is a silent segment signal. Similarly, if "we go to eat." "And what is good to eat ~ ah ~ amount ~ which ~ eat noodles" between the power value of the voice signal is also lower than the preset threshold, and the duration is also greater than the preset time value, you can determine "we go Let's eat and "What do you eat? Well, the amount ~ which ~ eat noodles." The voice signal is a silent segment signal. Therefore, the peer user can say the paragraph "Today's weather is very good, let's go Eat, what to eat? ~Hmm~ amount~ which eat noodles?" Divided into three voice segment signals "Today's weather is very good", "Let's go eat", "What are you eating? Well~~~ ~ Eat noodles."

Step 102: Perform semantic analysis on the speech segment signal in the audio signal. If there is a feature point in the speech segment signal, divide the speech segment signal into at least one sub-speech by using the feature point as a segmentation point. Segment signal, the feature point is: the time point at which the speech signal without complete semantics is located.

In the embodiment of the present invention, some feature values of words or words that are not commonly used in practice may be stored in the semantic database as semantic feature values, and then the semantic database may be queried for the audio signal. The speech segment signal is semantically analyzed; if the first speech signal is present in the speech segment signal, and the feature value of the first speech signal is included in the semantic database, determining the first speech signal is the And a feature point; if the feature values of all the voice signals in the voice segment signal are not included in the semantic database, determining that the feature segment does not include a feature point. The first voice signal may be any one of the voice segment signals.

For example, during an actual call, according to people’s habits, The statements before and after the transition words such as first, second, um, forehead, and which are used as two different meanings are spoken to each other at the same time. However, these transition words do not have complete semantics. In this case, the present invention is implemented. For example, according to the custom application, the feature values of the transition words may be pre-existed in the semantic feature database as semantic feature values, so as to perform semantic analysis on the speech segment signals.

It can be understood that, in the embodiment of the present invention, feature values of speech that are not muted but have no semantics such as “noise, background sound” may be stored as semantic feature values in the semantic database to match the voice segment signal. The part of the voice.

Optionally, the dividing the sub-speech segment signal into the at least one sub-segment segment signal by using the feature point as a segmentation point may specifically include: for any feature point, the feature point and the feature point The speech signal between the adjacent previous feature points is used as a sub-speech segment signal, and the speech signal between the feature point and the next feature point adjacent to the feature point is used as another speech segment signal.

In this way, the speech segment signal without the complete semantics can be culled, and at the same time, the divided sub-speech segment is guaranteed to be a speech segment with complete semantics.

Still taking the example in step 101 as an example, the semantic analysis of the three speech segment signals in the audio signal is performed separately, and it is found that the speech segment signal "what is good to eat ~ um ~ amount ~ which ~ eat noodles" in the "hmm ~ amount ~ Which "the eigenvalue of the speech signal is included in the semantic feature database, then the speech segment signal can be divided into two complete semantic sub-speech segments at intervals of "hmm ~ amount ~ which", "What are you eating?" and " Eat noodles."

Of course, the acquisition of the sub-segment segment signal may also be a speech segment signal between a plurality of feature points. For example, if a certain speech segment includes 10 feature points, the speech segment signal between the first feature point and the fourth feature point may be extracted as a sub-speech segment signal. The extraction rule of the sub-speech segment signal can be determined according to the processing capability of the terminal or the like.

It can be understood that if there is no feature point in the speech segment signal, it indicates that the speech segment signal cannot be segmented into at least two complete semantic sentences, and the speech segment signal needs to be directly translated.

Step 103: Translate the at least one sub-speech segment signal into a speech signal conforming to the target user language, and send the translated speech signal to the target terminal.

The target user is a user who is listening to the voice during the current call, and the calling language supported by the local user and the target user is different; for example, the local user can call in Chinese, and the target user can You can talk in English.

Optionally, the translating the at least one sub-segment segment signal into the voice that conforms to the target user language may include:

The voice recognition module in the terminal converts each sub-speech segment signal into text conforming to the source user language;

Translating the module in the terminal to translate the converted text into text conforming to the target user language;

The speech synthesis module in the terminal converts the translated text into a speech signal.

Correspondingly, the translated voice signals can be sequentially played to the target user in chronological order. Since the original voice is completely filtered out in the actual application, only playing the translated voice signal will make the user feel confused and unnatural. To solve the problem, at present, people basically will send the original voice signal and the user. The translated speech signal is played as the target user. At this time, in order to prevent the translated speech signal from affecting the playback of the original speech signal, it is conventional practice to place the audio signal in the segment after the speech signal is completely stopped. The translated voice signal of each sub-audio signal is sequentially played out, which causes the target user to wait for a long time. In order to avoid the target user waiting for the translation for too long, the present invention makes the original through the pipeline mode + two-way audio superposition speech synthesis technology. The speech and the translated speech synthesis are superimposed together, the original sound volume is reduced as the background sound, and the translated speech is played as the master volume, and the specific implementation is as follows:

Transmitting, to the target terminal, a first sub-voice signal;

Synthesizing the translated speech signal and the second sub-speech signal of the first sub-speech signal;

The synthesized speech signal is transmitted to the target terminal.

Similarly, after the second sub-speech signal is played to the target terminal, the second sub-voice signal translated speech signal and the third sub-speech signal may be synthesized and played to the target terminal, according to this Ways until the sub-speech signal and the translated The voice signal is completely played to the target terminal, thus realizing the effect of translation while playing, and reducing the waiting delay of translation.

It should be noted that the first sub-speech signal, the second sub-speech signal, and the third sub-speech signal may be any one of the at least one sub-speech signals obtained in step 102, but, in terms of chronological order, The second sub-speech signal is: a speech signal after the time of the first sub-speech signal and adjacent to the first sub-speech signal, and the third sub-speech signal is: after the time of the second sub-speech signal and with the second sub- A speech signal adjacent to a speech signal.

For example, as shown in FIG. 3, which is a timing diagram of real-time translation provided by an embodiment of the present invention, the local end divides the voice information into three complete semantic voice information based on semantic analysis, and after the sentence is translated, the prior art usually After the three original sentences are fully played, the translated speech is played back to the target user in chronological order, resulting in a delay. The application adopts a pipeline mode, and the translated voice is played after a statement is played. At the same time, in order to make the original playback of the next sentence does not affect the current translated voice play, the two are subjected to audio synthesis processing, so It is necessary to wait until all the statements have been played, and then the translated sentences are played one by one. As can be seen from FIG. 3, the pipeline mode of the present application is earlier than the existing play mode, and the translation waiting time is reduced. Improve translation efficiency and enhance user experience.

Optionally, in the embodiment of the present invention, the synthesizing the translated voice signal and the second sub-voice signal of the first sub-voice signal may specifically include:

Performing weighted summation on the voice signal translated by the first sub-speech signal and the second sub-speech signal;

In the weighted summation process, the weight of the voice signal after the translation of the first sub-speech signal and the weight of the second sub-speech signal may be set as needed, and the comparison in the embodiment of the present invention is not performed. By definition, the weight can be understood as the proportion of a plurality of speech signals occupying the total speech signal in the weighted summation process. However, in order to prevent the original voice from affecting the playback of the translated voice signal, in the embodiment of the present invention, the weight of the voice signal after the first sub-voice signal is translated and the weight of the second sub-voice signal During the setting of the value, the translated language of the first sub-speech signal needs to be translated The weight of the tone signal is greater than the weight of the second sub-speech signal.

For example, suppose A is the original speech of the second sub-speech signal, and B is the translated speech of the first sub-speech signal, so that the weight of A is 10% for the second, and the weight of B is 90%, then the synthesized The voice is: 10%*A+90%*B, and the A and B mix effects can be obtained. Since the weight of B is greater than A, A can be regarded as the background sound, and B is the sound that the user mainly hears. It should be noted that the embodiments of the present invention include, but are not limited to, speech synthesis of the above weighting method.

Further, in order to identify the language supported by the peer end, before the at least one sub-segment segment signal is translated into a voice signal conforming to the target user language, the method further includes:

And receiving the indication information sent by the target terminal, where the indication information is used to: indicate a language supported by the target terminal.

For example, the two sides of the call are Chinese and American respectively, that is, the language used by both parties is Chinese and English. The instructions are sent by the two parties (the Chinese say "Chinese" and the Americans say "English"). The languages spoken by both parties are Chinese and English. Thus, in the subsequent processing, if the input speech signal is a Chinese speech signal, the source language is Chinese and the target language is English; conversely, if the input speech signal is an English speech signal, the source language is English, and the target language is English. Chinese.

Further, before acquiring the audio signal input by the user, the method further includes:

Establishing a voice call with the target terminal;

Receiving a translation request sent by the target terminal, the translation request being used to request that the translated voice signal be sent to the target terminal.

The voice call in the example of the present invention may include a normal call through the data network, and may also include a voice call through an APP or voice chat software.

Since the sender and the receiver are relative concepts during the conversation between the two parties, usually according to the situation that the call is being sent, the speaker is determined to be the sender, and the listener is determined to be the receiver. Therefore, at a certain moment The terminal that is transmitting the voice signal described above may also serve as the receiving end. Correspondingly, when the terminal is used as the receiving end, the terminal The terminal can also perform several method steps as shown in FIG. 4 to implement translation and playback of the received voice signal:

Step 201: Receive an audio signal sent by the source terminal, where the audio signal includes a voice segment signal, and the voice segment signal is a segment of the voice signal whose power value is greater than a preset threshold.

Step 202: Perform semantic analysis on the voice segment signal in the audio signal.

If a feature point exists in the voice segment signal, the voice segment signal is segmented into at least one sub-segment segment signal by using the feature point as a segmentation point; and the feature point is a voice signal without complete semantics Time point.

Step 203: Translate the at least one sub-speech segment signal into a voice signal of a preset language, and play the translated voice signal.

The default language is the language supported by the local user, and is not limited herein. The specific implementation process of step 202 and step 102 is the same. In step 203, the translation of the sub-speech signal into the speech signal of the preset language is the same as the specific implementation process of translating the sub-speech signal into the target language in step 103. A detailed description.

Optionally, the step 204 playing the translated voice signal specifically refers to:

The translated voice signal is played to the local user through the audio processing module of the terminal itself.

Similarly, in order to achieve simultaneous translation and improve translation efficiency, the method may further include:

Playing the first sub-speech signal;

After playing the first sub-speech signal, synthesizing the translated sub-speech signal and the second sub-speech signal;

Play the synthesized speech signal.

Specifically, the voice signal can be synthesized by the above method of weighted summation, and details are not described herein again.

Further, before step 202, the method may further include:

Establishing a voice call with the source terminal;

Displaying prompt information on a user interface of the terminal; the prompt information is used to raise Indicate whether the user initiates the translation function;

Receiving the confirmation information sent by the user, and starting the translation function.

As can be seen from the above, an embodiment of the present invention provides a translation method, which is applied to a terminal that is engaged in a voice call, including: acquiring an audio signal sent by a user; the audio signal includes a voice segment signal; and a voice segment in the audio signal The signal is semantically analyzed. If there is a feature point in the voice segment signal, the segment is segmented into at least one sub-segment segment signal by using the feature point as a segmentation point; and at least one sub-segment segment signal is translated into The voice conforming to the target user's language, and the translated voice is sent to the target terminal. In this way, based on semantic analysis, the speech without the complete semantics in the statement detected by the VAD endpoint is culled into shorter sentences with complete semantics, which fully express the speaker's sentence meaning and avoid the occurrence of sentence or sentence. The situation effectively improves the accuracy of instant translation during a call.

It should be noted that the foregoing process may be performed by each unit in the terminal shown in FIG. 1, and details are not described herein. In addition, the audio processing module in the terminal shown in FIG. 1 of the present invention may be an input device or a transmitter of the terminal; the voice endpoint detection module, the voice recognition module, the translation module, and the voice synthesis module may be separate processors or integrated. It is implemented in a certain processor of the terminal, and may also be stored in the memory of the terminal in the form of program code, and is called by one of the processors of the terminal and performs the above translation function. The processor described herein may be a Central Processing Unit (CPU), or an Application Specific Integrated Circuit (ASIC), or one or more integrated systems configured to implement embodiments of the present invention. Circuit. Specifically, as described in the second embodiment, the present invention further provides a terminal, which is preferably used to implement the method in the foregoing method embodiment.

Embodiment 2

FIG. 5 is a structural diagram of a terminal 20 according to an embodiment of the present invention. The terminal 20 provided by the embodiment of the present invention may be used to implement the method shown in the foregoing method embodiment. For ease of description, only the implementation of the present invention is shown. For the relevant parts of the examples, the specific technical details are not disclosed, please refer to the description in the above method embodiments.

The terminal can be a mobile phone, a tablet, a laptop, a UMPC (Ultra-mobile) An instant call tool such as a personal computer, a netbook, a PDA (Personal Digital Assistant), and the like. The embodiment of the present invention uses a terminal as a mobile phone as an example, and FIG. 5 shows the implementation of the present invention. A block diagram of a portion of the structure of the associated handset 20.

As shown in FIG. 5, the mobile phone 20 includes components such as an input device 201, a memory 202, a processor 203, a transmitter 204, an output device 205, a receiver 206, and the like. It will be understood by those skilled in the art that the structure of the handset shown in FIG. 5 does not constitute a limitation to the handset, and may include more components than those illustrated, or some components may be combined, or different component arrangements.

The components of the mobile phone 20 will be specifically described below with reference to FIG. 5:

The input device 201 may include a touch screen, and may also include a microphone in the audio circuit for implementing the input function of the mobile phone 20. The voice signal sent by the user on or near the user may be collected, and the corresponding connection device is driven according to a preset program to convert the collected sound signal into an electrical signal, which is received by the audio circuit, converted into an audio signal, and then the audio signal is transmitted. Send to another handset or output an audio signal to memory 202 for further processing.

The memory 202 can be used to store data, software programs, and modules; and mainly includes a storage program area and a storage data area, wherein the storage program area can store an operating system, an application required for at least one function (such as a sound playing function, an image playing function, etc.) The storage data area can store data (such as audio data, image data, phone book, etc.) created according to the use of the mobile phone 20. Moreover, memory 202 can include high speed random access memory, and can also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid state storage device.

The processor 203 is the control center of the handset 20, which connects various portions of the entire handset using various interfaces and lines, by running or executing software programs and/or modules stored in the memory 202, and recalling data stored in the memory 202, The various functions and processing data of the mobile phone 20 are performed to perform overall monitoring of the mobile phone. Optionally, the processor 203 may include one or more processing units; preferably, the processor 203 may integrate an application processor and a modem processor, where the application processor mainly processes an operating system, The user interface and applications, etc., the modem processor primarily handles wireless communications. It can be understood that the above modem processor may not be integrated into the processor 203.

The transmitter 204 may include a radio frequency circuit (RF), which may be used for sending a voice signal during a call, in particular, transmitting a voice signal processed by the processor 203 to another mobile phone through a wireless channel; usually, the transmitter 204 includes, but is not limited to, an antenna, at least one amplifier, a transceiver, a coupler, an LNA (low noise amplifier), a duplexer, and the like.

The receiver 205 may include an RF circuit including, but not limited to, an antenna, at least one amplifier, a transceiver, a coupler, an LNA (low noise amplifier), a duplexer, etc., which can communicate wirelessly. Communicate with the network and other devices to receive voice signals sent by other devices; usually, since the antenna has a reciprocal function, the above-mentioned transmitter 204 and receiver 205 can be integrated as a transceiver.

The output device 206 may include a speaker in the audio circuit, and may also include a touch screen, and may provide an audio interface between the user and the mobile phone 20, and transmit the converted electrical signal to the speaker, and convert the speaker into The sound signal is played to the local user.

Although not shown, the mobile phone 20 may further include: a wireless (wireless fidelity) module, a Bluetooth module, a power source (such as a battery) powered by various components, and the like, and details are not described herein.

In the embodiment of the present invention, if the mobile phone 20 is currently transmitting the local voice to the target terminal (ie, the peer terminal), the input device 201 may also be configured to acquire an audio signal input by the user; the audio signal includes a voice. a segment signal; the speech segment signal is a segment of the speech signal whose power value is greater than a preset threshold.

The user is a user who is talking during the current call, and is a local user who holds the terminal.

The processor 203 is further configured to perform semantic analysis on the voice segment signal in the audio signal acquired by the input device 201. If a feature point exists in the voice segment signal, the feature point is used as a segmentation point, and Segmenting the speech segment signal into at least one sub-language a segment signal; the feature point is: a time point at which the speech signal without complete semantics is located;

And translating the at least one sub-segment segment signal into a speech that conforms to the target user language;

The transmitter 204 can also be configured to send the translated voice signal to the target terminal.

In the embodiment of the present invention, in order to implement semantic analysis, some feature values of words or words that are not commonly used in practice may be stored in the semantic database of the memory 202 as semantic feature values, and then the processor 203 Specifically used for:

Querying a semantic database in the memory 202; wherein the semantic database includes at least one semantic feature value, and the semantic feature value is: a feature value of a word or a word that does not have complete semantics;

If the first voice signal exists in the voice segment signal, and the feature value of the first voice signal is included in the semantic database, determining that the first voice signal is the feature point;

If the feature values of all the voice signals in the voice segment signal are not included in the semantic database, it is determined that the feature segment does not include a feature point.

Further, in order to prevent the target user from waiting for the translation for too long, the present invention superimposes the original speech and the translated speech synthesis by the pipeline synthesis mode and the two-channel audio superposition speech synthesis technology, and the original sound volume is reduced as the background sound, and the translation is performed. After the voice is played as the master volume, the processor 203 is further configured to:

Transmitting, by the transmitter 204, the translated voice signal to the target terminal to synthesize the translated voice signal and the second sub-voice signal of the first sub-voice signal;

The transmitter 204 is specifically configured to:

Transmitting the first sub-voice signal to the target terminal;

The synthesized speech signal is transmitted to the target terminal.

Similarly, after the second sub-speech signal is played to the target terminal, the second sub-voice signal translated speech signal and the third sub-speech signal may be synthesized and played to the target terminal, according to this In a manner, until the sub-speech signal and the translated speech signal are completely played to the target terminal, so that translation is performed while playing. The effect is to reduce the waiting delay of the translation.

Optionally, in the embodiment of the present invention, the processor 203 is specifically configured to:

Performing weighted summation on the voice signal translated by the first sub-speech signal and the second sub-speech signal;

In the weighted summation process, the weight of the voice signal after the translation of the first sub-speech signal and the weight of the second sub-speech signal may be set as needed, and the comparison in the embodiment of the present invention is not performed. By definition, the weight can be understood as the proportion of a plurality of speech signals occupying the total speech signal in the weighted summation process. However, in order to prevent the original voice from affecting the playback of the translated voice signal, in the embodiment of the present invention, the weight of the voice signal after the first sub-voice signal is translated and the weight of the second sub-voice signal During the setting of the value, the weight of the voice signal translated by the first sub-speech signal needs to be greater than the weight of the second sub-speech signal.

Further, in order to identify the language supported by the peer, the receiver 205 can also be used to:

Before the processor 203 translates the at least one sub-speech segment signal into a voice signal conforming to the target user language, receiving the indication information sent by the target terminal, where the indication information is used to: indicate that the target terminal supports Language.

Further, the receiver 205 can also be used to:

Before the input device 201 acquires the audio signal input by the user, after the terminal establishes a voice call with the target terminal, receiving a translation request sent by the target terminal, where the translation request is used to request to send a translation to the target terminal. Voice signal.

Since the sender and the receiver are relative concepts during the conversation between the two parties, usually according to the situation that the call is being sent, the speaker is determined to be the sender, and the listener is determined to be the receiver. Therefore, at a certain moment The mobile phone 20 that is transmitting the voice signal described above can also serve as the receiving end. Correspondingly, when the mobile phone 20 is used as the receiving end, the receiver 205 in the mobile phone 20 can also be used to:

Receiving an audio signal sent by the source terminal; the audio signal includes a voice segment signal; and the voice segment signal is a segment of the voice signal whose power value is greater than a preset threshold.

The processor 203 is further configured to use the audio signal received by the receiver 205. The speech segment signal in the semantic analysis;

If a feature point exists in the voice segment signal, the voice segment signal is segmented into at least one sub-segment segment signal by using the feature point as a segmentation point; and the feature point is a voice signal having no complete semantics Time point

Translating the at least one sub-segment segment signal into a speech signal of a preset language.

The output device 206 can also be used to play the translated voice signal of the processor 203.

The detailed steps of the processor 203 performing the translation function are as described above, and will not be described in detail herein.

Similarly, in order to achieve simultaneous translation and improve translation efficiency, the processor 203 is further configured to:

Before the output device 203 plays the translated speech signal of the processor 203, synthesizing the translated speech signal and the second sub-speech signal of the first sub-speech signal;

The output device 206 can be specifically configured to:

Playing the first sub-speech signal;

Playing the synthesized speech signal.

Specifically, the voice signal can be synthesized by the above method of weighted summation, and details are not described herein again.

Further, the output device 206 can also be used to:

After the terminal establishes a voice call with the source terminal, displaying prompt information on the user interface of the terminal; the prompt information is used to prompt the user whether to initiate a translation function;

The input device 201 may be further configured to receive the confirmation information sent by the user.

The processor 203 can also be used to initiate a translation function.

As can be seen from the above, an embodiment of the present invention provides a terminal that acquires an audio signal sent by a user; the audio signal includes a voice segment signal; and performs semantic analysis on a voice segment signal in the audio signal, if the voice segment signal is in the voice segment signal There is a feature point, wherein the feature point is used as a segmentation point, and the voice segment signal is segmented into at least one sub-segment segment signal; and at least one sub-segment segment signal is translated into a voice conforming to a target user language, and the translation is performed. The subsequent voice is sent to the target terminal. In this way, based on semantic analysis, the speech without the complete semantics in the statement detected by the VAD endpoint is culled into shorter sentences with complete semantics, which fully express the speaker's sentence meaning and avoid the occurrence of sentence or sentence. The situation effectively improves the accuracy of instant translation during a call.

The above is only a specific embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily think of changes or substitutions within the technical scope of the present invention. It should be covered by the scope of the present invention. Therefore, the scope of the invention should be determined by the scope of the claims.

Claims (33)

A translation method is applied to a terminal that is engaged in a voice call, and includes: Acquiring an audio signal input by the user; the audio signal includes a voice segment signal; and the voice segment signal is a segment of the voice signal whose power value is greater than a preset threshold; Performing semantic analysis on the speech segment signal in the audio signal; If a feature point exists in the voice segment signal, the voice segment signal is segmented into at least one sub-segment segment signal by using the feature point as a segmentation point; and the feature point is a voice signal having no complete semantics Time point Translating the at least one sub-segment segment signal into a speech signal conforming to the target user language, and transmitting the translated speech signal to the target terminal. The method according to claim 1, wherein the performing semantic analysis on the voice segment signal in the audio signal comprises: Querying a semantic database; the semantic database includes at least one semantic feature value, the semantic feature value being: a feature value of a word or a word having no complete semantics; If the first voice signal exists in the voice segment signal, and the feature value of the first voice signal is included in the semantic database, determining that the first voice signal is the feature point; If the feature values of all the voice signals in the voice segment signal are not included in the semantic database, it is determined that the feature segment does not include a feature point. The method according to claim 1 or 2, wherein before the transmitting the translated voice signal to the target terminal, the method further comprises: Combining the translated speech signal and the second sub-speech signal of the first sub-speech signal; The sending the translated voice signal to the target terminal specifically includes: Transmitting the first sub-voice signal to the target terminal; Transmitting the synthesized speech signal to the target terminal; the second sub-speech signal is: a speech signal after the time of the first sub-speech signal and adjacent to the first sub-speech signal. The method of claim 3 wherein said synthesizing said The translated voice signal and the second sub-voice signal of a sub-voice signal specifically include: Performing weighted summation on the voice signal translated by the first sub-speech signal and the second sub-speech signal; The weight of the translated voice signal of the first sub-speech signal is greater than the weight of the second sub-speech signal. The method according to any one of claims 1 to 4, wherein before the translating the at least one sub-segment segment signal into a speech signal conforming to a target user language, the method further comprises: And receiving the indication information sent by the target terminal, where the indication information is used to: indicate a language supported by the target terminal. The method according to any one of claims 1-5, wherein before the obtaining the audio signal input by the user, the method further comprises: Establishing a voice call with the target terminal; Receiving a translation request sent by the target terminal, the translation request being used to request that the translated voice signal be sent to the target terminal. A translation method is applied to a terminal that is engaged in a voice call, and includes: Receiving an audio signal sent by the source terminal; the audio signal includes a voice segment signal; and the voice segment signal is a segment of the voice signal whose power value is greater than a preset threshold; Performing semantic analysis on the speech segment signal in the audio signal; If a feature point exists in the voice segment signal, the voice segment signal is segmented into at least one sub-segment segment signal by using the feature point as a segmentation point; and the feature point is a voice signal having no complete semantics Time point Translating the at least one sub-segment segment signal into a speech signal of a preset language; Play the translated voice signal. The method according to claim 7, wherein the performing semantic analysis on the voice segment signal in the audio signal comprises: Querying a semantic database; the semantic database includes at least one semantic feature value, the semantic feature value being: a feature value of a word or a word having no complete semantics; If the first voice signal exists in the voice segment signal, and the feature value of the first voice signal is included in the semantic database, determining that the first voice signal is the feature point; If the feature values of all the voice signals in the voice segment signal are not included in the semantic database, it is determined that the feature segment does not include a feature point. The method according to claim 7 or 8, wherein before the playing the translated voice signal, the method further comprises: Combining the translated speech signal and the second sub-speech signal of the first sub-speech signal; The playing the translated voice signal specifically includes: Playing the first sub-speech signal; Playing the synthesized speech signal; the second sub-speech signal is: a speech signal after the time of the first sub-speech signal and adjacent to the first sub-speech signal. The method according to claim 9, wherein the synthesizing the translated speech signal and the second sub-speech signal of the first sub-speech signal comprises: Performing weighted summation on the voice signal translated by the first sub-speech signal and the second sub-speech signal; The weight of the translated voice signal of the first sub-speech signal is greater than the weight of the second sub-speech signal. The method according to any one of claims 7 to 10, further comprising: before receiving the audio signal transmitted by the source terminal, the method further comprising: Establishing a voice call with the source terminal; Displaying prompt information on a user interface of the terminal; the prompt information is used to prompt the user whether to initiate a translation function; Receiving the confirmation information sent by the user, and starting the translation function. A terminal, comprising: An audio processing module, configured to acquire an audio signal input by a user; the audio signal includes a voice segment signal; and the voice segment signal is a segment of a voice signal whose power value is greater than a preset threshold; a voice endpoint detection module, configured to acquire an audio signal from the audio processing module The speech segment signal is semantically analyzed; if there is a feature point in the speech segment signal, the speech segment signal is segmented into at least one sub-speech segment signal by using the feature point as a segmentation point; The point in time at which the speech signal without complete semantics is located; a translation module, configured to translate at least one sub-segment segment signal detected by the speech endpoint detection module into a speech signal conforming to a target user language; a voice synthesis module, configured to send the translated voice signal of the translation module to the target terminal. The terminal according to claim 12, wherein the terminal further comprises: a voice recognition module, wherein the voice endpoint detection module is specifically configured to: Querying a semantic database; the semantic database includes at least one semantic feature value, the semantic feature value being: a feature value of a word or a word having no complete semantics; If the first voice signal exists in the voice segment signal, and the feature value of the first voice signal is included in the semantic database, determining that the first voice signal is the feature point; If the feature values of all the voice signals in the voice segment signal are not included in the semantic database, it is determined that the feature segment does not include a feature point. The terminal according to claim 12 or 13, wherein the speech synthesis module is specifically configured to: Transmitting the first sub-voice signal to the target terminal; Synthesizing the translated speech signal and the second sub-speech signal of the first sub-speech signal; Transmitting the synthesized speech signal to the target terminal; the second sub-speech signal is: a speech signal after the time of the first sub-speech signal and adjacent to the first sub-speech signal. The terminal according to claim 14, wherein the speech synthesis module is specifically configured to: Performing weighted summation on the voice signal translated by the first sub-speech signal and the second sub-speech signal; The weight of the translated voice signal of the first sub-speech signal is greater than the weight of the second sub-speech signal. The terminal according to any one of claims 12-15, wherein the audio processing module is further configured to: Before the voice synthesizing module translates the at least one sub-speech segment signal into a voice signal that conforms to the target user language, receiving the indication information sent by the target terminal, where the indication information is used to: indicate that the target terminal supports Language. The terminal according to any one of claims 12 to 16, wherein the audio processing module is further configured to: Before acquiring the audio signal input by the user, receiving a translation request sent by the target terminal, the translation request is used to request to send the translated voice signal to the target terminal. A terminal, comprising: An audio processing module, configured to receive an audio signal sent by the source terminal; the audio signal includes a voice segment signal; and the voice segment signal is a segment of the voice signal whose power value is greater than a preset threshold; a voice endpoint detection module, configured to perform semantic analysis on a voice segment signal in an audio signal acquired by the audio processing module; if a feature point exists in the voice segment signal, the feature point is used as a segmentation point, and the The speech segment signal is segmented into at least one sub-segment segment signal; the feature point is a time point at which the speech signal without complete semantics is located; a translation module, configured to translate at least one sub-segment segment signal detected by the speech endpoint detection module into a speech signal of a preset language; a speech synthesis module, configured to play the translated speech signal of the translation module. The terminal according to claim 18, wherein the terminal further comprises: a voice recognition module, wherein the voice endpoint detection module is specifically configured to: Querying a semantic database; the semantic database includes at least one semantic feature value, the semantic feature value being: a feature value of a word or a word having no complete semantics; If the first voice signal exists in the voice segment signal, and the feature value of the first voice signal is included in the semantic database, determining that the first voice signal is the feature point; If the feature values of all the voice signals in the voice segment signal are not included in the language In the semantic database, it is determined that the feature segment does not include a feature point. The terminal according to claim 18 or 19, wherein the speech synthesis module is specifically configured to: Playing the first sub-speech signal; Synthesizing the translated speech signal and the second sub-speech signal of the first sub-speech signal; Playing the synthesized speech signal; the second sub-speech signal is: a speech signal after the time of the first sub-speech signal and adjacent to the first sub-speech signal. The terminal according to claim 20, wherein the speech synthesis module is specifically configured to: Performing weighted summation on the voice signal translated by the first sub-speech signal and the second sub-speech signal; The weight of the translated voice signal of the first sub-speech signal is greater than the weight of the second sub-speech signal. The terminal according to any one of claims 18 to 21, wherein before receiving the audio signal transmitted by the source terminal, the audio processing module is further configured to: Before receiving the audio signal sent by the source terminal, the terminal and the source terminal establish a voice call, receive the confirmation information sent by the user, and start a translation function. A terminal, comprising: An input device, configured to acquire an audio signal input by a user; the audio signal includes a voice segment signal; and the voice segment signal is a segment of a voice signal whose power value is greater than a preset threshold; a processor, configured to perform semantic analysis on a voice segment signal in the audio signal acquired by the input device; if a feature point exists in the voice segment signal, the feature segment is used as a segmentation point, and the voice segment is Separating the signal into at least one sub-segment segment signal; the feature point is a time point at which the speech signal without complete semantics is located; Translating the at least one sub-segment segment signal into a speech signal conforming to a target user language; And a transmitter, configured to send the translated voice signal of the processor to the target terminal. The terminal according to claim 23, wherein said treatment tool Body for: Querying a semantic database; the semantic database includes at least one semantic feature value, the semantic feature value being: a feature value of a word or a word having no complete semantics; If the first voice signal exists in the voice segment signal, and the feature value of the first voice signal is included in the semantic database, determining that the first voice signal is the feature point; If the feature values of all the voice signals in the voice segment signal are not included in the semantic database, it is determined that the feature segment does not include a feature point. The terminal according to claim 23 or 24, wherein the transmitter is further configured to: Before the processor sends the translated voice signal to the target terminal, synthesizing the first sub-voice signal translated speech signal and the second sub-speech signal; The transmitter is specifically configured to: Transmitting the first sub-voice signal to the target terminal; Transmitting the synthesized speech signal to the target terminal; the second sub-speech signal is: a speech signal after the time of the first sub-speech signal and adjacent to the first sub-speech signal. The terminal according to claim 25, wherein the processor is specifically configured to: Performing weighted summation on the voice signal translated by the first sub-speech signal and the second sub-speech signal; The weight of the translated voice signal of the first sub-speech signal is greater than the weight of the second sub-speech signal. The terminal according to any one of claims 23 to 26, wherein the terminal further comprises a receiver, The receiver is configured to receive, after the transmitter translates the at least one sub-speech segment signal into a voice signal conforming to a target user language, the indication information sent by the target terminal, where the indication information is used to: indicate The language supported by the target terminal. The terminal according to claim 27, wherein said receiver, Also used for: Before the input device acquires the audio signal input by the user, the terminal and the target terminal establish a voice call, and receive a translation request sent by the target terminal, where the translation request is used to request to send a translation to the target terminal. After the voice signal. A terminal, comprising: a receiver, configured to receive an audio signal sent by the source terminal; the audio signal includes a voice segment signal; and the voice segment signal is a segment of the voice signal whose power value is greater than a preset threshold; a processor, configured to perform semantic analysis on a voice segment signal in the audio signal acquired by the receiver; if a feature point exists in the voice segment signal, use the feature point as a segmentation point, and the voice segment is Separating the signal into at least one sub-segment segment signal; the feature point is a time point at which the speech signal without complete semantics is located; Translating the at least one sub-segment segment signal into a speech signal of a preset language; And an output device, configured to play the translated voice signal of the processor. The terminal according to claim 29, wherein the processor is specifically configured to: Querying a semantic database; the semantic database includes at least one semantic feature value, the semantic feature value being: a feature value of a word or a word having no complete semantics; If the first voice signal exists in the voice segment signal, and the feature value of the first voice signal is included in the semantic database, determining that the first voice signal is the feature point; If the feature values of all the voice signals in the voice segment signal are not included in the semantic database, it is determined that the feature segment does not include a feature point. The terminal according to claim 29 or 30, wherein the processor is further configured to: Before the output device plays the translated speech signal of the processor, synthesizing the translated speech signal and the second sub-speech signal of the first sub-speech signal; The output device is specifically configured to: Playing the first sub-speech signal; Playing the synthesized speech signal; the second sub-speech signal is: a speech signal after the time of the first sub-speech signal and adjacent to the first sub-speech signal. The terminal according to claim 31, wherein the processor is specifically configured to: Performing weighted summation on the voice signal translated by the first sub-speech signal and the second sub-speech signal; The weight of the translated voice signal of the first sub-speech signal is greater than the weight of the second sub-speech signal. The terminal according to any one of claims 29 to 32, wherein the terminal further comprises an output device and an input device; The output device is configured to display prompt information on a user interface of the terminal after the terminal establishes a voice call with the source terminal; the prompt information is used to prompt the user whether to initiate a translation function; The input device is further configured to receive confirmation information sent by the user, The processor is further configured to initiate a translation function.

Images