JP6922306B2 - Audio playback device and audio playback program - Google Patents

Description

The present invention relates to audio reproduction technology.

Examples of applications of the voice reproduction technology include voice interaction between a person and a machine or voice interaction between a machine and a machine. An example of voice interaction between a person and a machine is a voice dialogue system that synthesizes and reproduces the voice of the answer to a user's question by voice. An example of machine-to-machine voice interaction is that the voice playback device B recognizes the voice of the question played by the voice playback device A according to a predetermined scenario and reproduces the answer voice. Specific examples include plays and comics in which a machine (voice playback device) plays all of the characters. When synthesizing the voice of the answer to the question by voice, it is preferable to reproduce the voice of the answer with intention in order to realize a human-like answer to the question by the user's voice without any unnaturalness. For example, in the technique disclosed in Patent Document 1, in order to express the intention contained in the answer, the pitch of the ending is different between the positive answer and the negative answer.

Japanese Unexamined Patent Publication No. 2015-064480

However, it is not possible to express various intentions only by adjusting the pitch of the ending of the answer as in the technique disclosed in Patent Document 1. If the voice data of the answer is prepared for each of the various intentions, a rich intention expression becomes possible, but there is a problem that the storage capacity of the storage device for storing the voice data increases.

The present invention has been made in view of the above-described problems, and an object of the present invention is to provide a technique capable of expressing a rich intention while suppressing an increase in the storage capacity required for storing the voice data of the answer. And.

In order to achieve the above object, the voice reproduction device according to one aspect of the present invention specifies an answer acquisition unit that acquires voice data of an answer to a question represented by an input voice signal, and an intention to be given to the answer. The intent designation unit, the rhyme control data acquisition unit that acquires the rhyme control data representing the time change of the rhyme according to the intention specified by the intention designation unit, and the rhyme control data that obtains the time change of the rhyme based on the voice data. It is characterized by including an answer reproduction unit that reproduces the voice of the answer controlled according to the above.

The time change of prosody refers to the time change of each prosodic component such as pitch, speaking speed, volume and utterance timing. The speaking speed is the number of phonemes that are pronounced per unit time. When a person utters an answer to a question, by adjusting the time change of prosody according to the intention contained in the answer, a variety of things such as "easy" and "cautious", or "anger" and "awareness" are available. Express your intentions. According to this aspect, it is possible to reproduce the voice of the answer in which the time change of the prosody is finely controlled according to the intention to be included in the answer, and it is possible to express a rich intention. Here, the prosody control data is an array of changes in the prosody components at each time (time starting from the beginning of the answer) such as pitch, speech speed, volume, and speech timing, that is, sequence data. It suffices, and the amount of data is smaller than that of voice waveform data. Therefore, even if the prosody control data corresponding to each intention is stored in the storage device in advance for each answer, it is compared with the mode in which the voice data of each answer pronounced with different intentions is stored in the storage device. It is possible to handle with a small storage capacity. That is, according to this aspect, it is possible to express a rich intention while suppressing an increase in the storage capacity required for storing the voice data of the answer. It should be noted that the answers are not limited to specific answers to the questions, but also include aizuchi (interjections). In addition, the answers include answers to questions and aizuchi, as well as dialogues in theater and comics. In addition to human voices, animals such as "one" (bowwow) and "meow" (meow) The bark is also included. That is, the answer and the voice here are concepts including not only the voice uttered by a person but also the cry of an animal.

In a more preferred embodiment, the rhythm control data has a database in which the rhythm control data is stored for each intention, and the rhythm control data acquisition unit acquires the rhythm control data corresponding to the intention specified by the intention designation unit from the database. do. In another preferred embodiment, the rhythm control data acquisition unit uses the rhythm control data when the rhythm control data corresponding to the intention specified by the intention designation unit is not stored in the database. It is characterized in that it is acquired by interpolation using a plurality of rhythm control data stored in a database. According to such an aspect, it is possible to express a richer intention while suppressing an increase in the storage capacity of the database. Further, in another preferred embodiment, the input voice signal is analyzed to have an intention specifying unit for specifying the intention given to the question, and the intention specifying unit is specified by the intention specifying unit. It is characterized in that the intention to be given to the answer is specified according to the intention. According to such an aspect, it is possible to reproduce the voice of the answer with the intention according to the intention contained in the question.

The aspect of the present invention can be conceived as a program that causes a computer to function as the audio reproduction device as well as the audio reproduction device.

It is a block diagram which shows the structure of the audio reproduction apparatus which concerns on embodiment. It is a figure which shows an example of the time change of prosody in an intentional answer. It is a flowchart which shows the operation of the voice reproduction apparatus.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
(A: Composition)
FIG. 1 is a diagram showing a configuration of an audio reproduction device 10 according to an embodiment of the present invention.
The audio reproduction device 10 is, for example, a device incorporated in a stuffed animal. When the user asks a question to the stuffed animal, the voice reproduction device 10 synthesizes and reproduces the voice of the answer with the intention specified by the user. When a person utters an answer to a question, he or she adjusts the time change of prosody according to the intention of the answer, so that a variety of things such as "easy", "cautious", "anger", and "awareness" can be obtained. Express your intentions. For example, FIG. 2 shows the prosody of the time waveform TW of the reference voice "that" that was pronounced without a specific intention and the prosody of the reference voice "that" that included "comfort". The change pattern P1 and the prosodic change pattern P2 from the reference voice of the voice "that" with "cautiousness" are exemplified. In FIG. 2, the amount of change from the reference voice of each prosodic component of "pitch", "speaking speed", and "volume" is represented by the position on the coordinate axis from the center of gravity of the triangle to each apex. The farther away from the center of gravity, the higher the pitch, the faster the speaking speed, and the louder the volume than the reference voice.

The voice reproduction device 10 has a CPU (Central Processing Unit), a voice input unit 102, and a speaker 114, and the CPU executes a pre-installed application program to construct a plurality of functional blocks as follows. Will be done. Specifically, in the voice reproduction device 10, a language analysis unit 104, an answer acquisition unit 106, an intention designation unit 108, a prosody control data acquisition unit 110, and an answer reproduction unit 112 are constructed. Although not shown in particular, the audio reproduction device 10 also has a display unit, an operation input unit, and the like, so that the user can check the status of the device and input various operations to the device. , Various settings can be made. Further, the voice reproduction device 10 is not limited to a toy such as a stuffed animal, but may be a so-called pet robot, a terminal device such as a mobile phone, a tablet-type personal computer, or the like.

Although details are omitted, the voice input unit 102 includes a microphone that converts voice into an electric signal and an A / D converter that converts the converted voice signal into a digital signal. The language analysis unit 104 analyzes the meaning content of the question defined by the voice signal input from the voice input unit 102, and receives the meaning content data indicating the analysis result (that is, the meaning content of the above question) in the answer acquisition unit 106. Give to.

The answer library 124 is a database in which a plurality of sets of an identifier uniquely indicating an answer to a question by a user (hereinafter referred to as an answer identifier) and voice data of the answer are stored in advance. This voice data is a recording of the voice of a model person, such as answers to questions such as "yes", "no", "yes", "yes", "hmm", "I see". Aizuchi and so on. The voice data of the answer is in a format such as wav or mp3. Further, specific examples of the answer identifier include a character string representing the content of the answer such as "yes" and "no" and a serial number. In addition, the answer to the question is not limited to the answer and the aizuchi, for example, it is a sentence that presents the information requested by the question, such as the answer "It is sunny" to the question "What is the weather today?" You may.

The answer acquisition unit 106 selects one of the answer libraries 124 for the answer identifier and the voice data of the answer to the question whose meaning content is represented by the meaning content data given by the language analysis unit 104, and the selected answer identifier and the selected answer identifier and the voice data. The voice data is read from the answer library 124 and acquired. Then, the answer acquisition unit 106 outputs the answer identifier acquired from the answer library 124 to the prosody control data acquisition unit 110, and outputs the voice data acquired from the answer library 124 to the answer reproduction unit 112. The answer acquisition unit 106 of the present embodiment includes a table that stores answer identifiers of answers suitable for the meaning content question represented by the meaning content data in association with the meaning content data that can be output from the language analysis unit 104. ing. The answer acquisition unit 106 refers to the stored contents of the table and selects one answer suitable for the question of the meaning content represented by the meaning content data.

In the present embodiment, a mode of selecting an answer according to the meaning and content of the question will be described, but the answer may be randomly selected regardless of the meaning and content of the question. In this case, the language analysis unit 104 and the above table are unnecessary. Specifically, the output signal of the voice input unit 102 is given to the answer acquisition unit 106, and the answer acquisition unit 106 receives the voice signal from the voice input unit 102 as an opportunity to receive the answer identifier and the voice data from the answer library 124. It suffices to execute the process of reading out at random.

The prosody library 122 associates with each of the plurality of answer identifiers stored in the answer library 124, and sets an identifier (hereinafter, intention identifier) indicating the intention for each intention included in the answer indicated by the answer identifier and the intention. It is a database that stores the prosody control data that defines the time change of the prosody in the corresponding answer. Here, the prosody control data is an array of changes in the amount of change at each time (time starting from the beginning of the answer) of the prosody components such as pitch, speaking speed, and volume, that is, sequence data. Further, as a specific example of the intention identifier, a character string representing the content of the intention such as "anger" or "akire" can be mentioned.

Prosody control data can be created as follows. For example, in the case of the answer "that", the waveform data of the voice "that" pronounced on a flat plate without a specific intention is used as the reference data, and the specific intention such as "anger" or "akire" is used. Compare the waveform data of the voice "Ano" pronounced with It is conceivable to calculate and arrange the differences for each component in chronological order to obtain rhyme control data. In this case, the answer library 124 may store the voice data of the voice pronounced on the flat plate without a specific intention as the voice data of the answer. In addition, the waveform data of the voice sounded with a specific intention may be used as the reference data, and the rhyme control data for the voice sounded with another intention may be generated. In this case, the voice data of the answer. The voice data of the voice pronounced with the specific intention may be stored in the answer library 124.

The intention designation unit 108 is a device that allows the user to specify one of the intentions in which the intention identifier and the prosody control data are stored in the prosody library 122 as the intention to be included in the answer to the question. For example, the intention designation unit 108 displays a list of intention identifiers on a display unit (not shown in FIG. 1), and a prosody control data acquisition unit obtains the intention identifier designated by an operation on the operation input unit (not shown in FIG. 1). Give to 110.

The prosody control data acquisition unit 110 reads and acquires the prosody control data corresponding to the intention identifier given by the intention designation unit 108 and the answer identifier given by the answer acquisition unit 106 from the prosody library 122. The prosody control data acquisition unit 110 gives the prosody control data read from the prosody library 122 to the answer reproduction unit 112.

The answer reproduction unit 112 makes the voice represented by the voice data given by the answer acquisition unit 106 the time of each prosody component of pitch, speech speed, and volume according to the prosody control data given by the prosody control data acquisition unit 110. The change is controlled independently and reproduced (synthesized).
The above is the configuration of the audio reproduction device 10.

(B: Operation)
Next, the operation of the voice reproduction device 10 will be described.
FIG. 3 is a flowchart showing a processing operation in the audio reproduction device 10. In the present embodiment, when the user asks a question by voice to the stuffed animal to which the voice reproduction device 10 is applied, the process shown in this flowchart is activated.

When the user asks a question by voice, the voice of the question is converted into a voice signal by the voice input unit 102, and the voice signal is supplied from the voice input unit 102 to the language analysis unit 104. In step Sa11, the language analysis unit 104 stores the voice signal supplied from the voice input unit 102 in a memory or the like, and determines whether or not the question by voice is completed. Whether or not the question is completed is determined by whether or not the state in which the volume of the audio signal is less than a predetermined threshold value continues for a predetermined time. If the question is not completed (if the determination result of step Sa11 is "No"), the language analysis unit 104 executes the process of step Sa11 again and waits for the end of the utterance of the question.

In step Sa12, which is executed when the determination result of step Sa11 is "Yes", the language analysis unit 104 analyzes the meaning content of the question defined by the audio signal stored in the memory or the like, and indicates the analysis result. The data is given to the answer acquisition unit 106. In the present embodiment, the determination of whether or not the utterance of the question is completed and the semantic analysis of the question are sequentially executed, but both may be sequentially executed in parallel. By doing so, it is possible to reduce the delay from the end of the question utterance to the completion of the semantic analysis. In step Sa13 following step Sa12, the answer acquisition unit 106 acquires the answer identifier and the voice data of the answer to the question whose meaning content is represented by the above meaning content data from the answer library 124, and obtains the former as the rhyme control data acquisition unit. It is given to 110, and the latter is given to the answer reproduction unit 112.

In step Sa14 following step Sa13, the answer reproduction unit 112 determines whether or not the answer is being reproduced. In step Sa14, the case where the answer reproduction unit 112 determines that the answer is being reproduced is a case where the next question is asked by the user while the answer is being reproduced in response to a certain question. If the answer is being reproduced (if the determination result in step Sa14 is "Yes"), the answer reproduction unit 112 re-executes the process of step Sa14 until the determination result in step Sa14 becomes "No". When the determination result of step Sa14 is "No", the intention designation unit 108 causes the user to specify the intention to be included in the answer, and outputs an intention identifier indicating the intention specified by the user to the prosody control data acquisition unit 110. (Step Sa15). In step Sa16 following step Sa15, the prosody control data acquisition unit 110 acquires the prosody control data corresponding to the intention identifier given by the intention designation unit 108 and the answer identifier given by the answer acquisition unit 106 from the prosody library 122. Then, the acquired prosody control data is notified to the answer reproduction unit 112, and the answer acquisition unit 106 instructs the answer acquisition unit 106 to reproduce the audio data of the answer. According to this instruction, the answer reproduction unit 112 controls the time change of each prosody component of pitch, speech speed, and volume according to the prosody control data given from the prosody control data acquisition unit 110, and controls the time change of each prosody component separately and independently. The voice represented by the voice data is reproduced (step Sa17).

For example, when the answer "that" is selected in step Sa13 and the intention of "easy" is specified in step Sa15, in step Sa17, the prosody changes with time in the time change pattern P1 of FIG. The voice is synthesized, and the user who listens to this voice feels "comfort" from the time change of the prosody in the voice. On the other hand, when the answer "that" is selected in step Sa13 and the intention of "cautious" is specified in step Sa15, in step Sa17, the prosody changes with time in the time change pattern P2 of FIG. The voice is synthesized, and the user who listens to this voice feels "cautiousness" from the time change of the prosody in the voice.

As described above, according to the present embodiment, it is possible to reproduce the answer voice in which the time change of the prosody is finely controlled according to the intention to be included in the answer, and it is possible to express a rich intention. In the voice reproduction device 10 of the present embodiment, it is necessary to store the prosody library 122 in a storage device or the like in addition to the answer library 124. However, since the prosody control data stored in the prosody library 122 is sequence data, the total amount of data is smaller than the sum of the voice data of the voice with various intentions for each answer. Therefore, it is possible to suppress an increase in the storage capacity of the storage device as compared with the mode in which all of these voice data are stored. That is, according to the present embodiment, it is possible to express a rich intention while suppressing an increase in the storage capacity of the storage device that stores the data required for synthesizing the voice of the answer. In the above embodiment, the case of controlling the time change of the pitch, the speaking speed, and the volume has been described, but the utterance timing may also be controlled.

(C: Deformation and application example)
The present invention is not limited to the above-described embodiment, and various applications and modifications as described below are possible, for example. In addition, one or a plurality of arbitrarily selected modes of application / modification described below can be appropriately combined.

<Voice input section>
In the above embodiment, the voice input unit 102 is configured to input the user's voice (speech) with a microphone and convert it into a voice signal, but the present invention is not limited to this configuration. That is, the voice input unit 102 may be configured to input or input a speech by a voice signal in some form. Specifically, the audio input unit 102 has a configuration for inputting an audio signal processed by another processing unit or an audio signal supplied (or transferred) from another device, and further, it is built in the LSI and is simply audio. It is a concept that includes an input interface circuit that receives a signal and transfers it to the subsequent stage. Further, the question to the voice reproduction device 10 is not limited to the question by voice, and may be the input of text data representing a sentence in which the question is written down. In this case, a text data input unit may be provided in the voice reproduction device 10 instead of the voice input unit 102.

<Prosody control unit>
In the above embodiment, prosody control data is prepared for each voice data of the answer to the question, and the time change of the prosody of the answer voice is controlled for each voice data. However, the prosody control data is used in phoneme units that compose the answer voice, in waveform sample units when the answer voice is sampled to generate voice data, or in frame units when the answer voice is divided into frames of a predetermined time. Prosody control data may be prepared and the time change of prosody may be controlled in units of these data.

<Variation of intention specification>
In the above embodiment, only one intention to be included in the answer to the question is specified by the user, but a plurality of intentions are specified by the user, for example, "Akire" and "easily" are specified. You may. In this case, the intention designating unit 108 causes the prosody control data acquisition unit 110 to output an intention identifier indicating each of the plurality of intentions designated by the user. One prosody control data that directly corresponds to the plurality of intentions specified by the user is not stored in the prosody library 122. Therefore, the time signature control data acquisition unit 110 acquires the time signature control data corresponding to each of the plurality of intention identifiers, and generates one time signature control data by interpolation using the plurality of time signature control data. Just do it. Specifically, the prosody control data acquisition unit 110 generates new prosody control data by weighting and adding a plurality of prosody control data acquired in the above manner, for example, 1: 1 and the new prosody control. The data is output to the answer reproduction unit 112 to synthesize the answer sound.

According to such an aspect, it is considered that the voice of the answer corresponding to the intermediate intention of the plurality of intentions specified by the user can be synthesized. In general, even if the waveform data of the voice "Ane" pronounced with a feeling of "disappointment" and the waveform data of the voice of "Ane" pronounced with a feeling of "easy" are weighted and added, it is accompanied by "comfort". This is because it is not possible to obtain waveform data of a voice that feels like "a slapstick", but since the rhyme control data is sequence data, it is possible to generate rhyme control data that expresses an intermediate intention by weighted addition. Similarly, it is considered that pronouncing on a flat plate without a specific intention is one aspect of the intention, and by specifying "flat plate" and "careful", "flat plate" and "careful" such as "somewhat careful" can be used. It is also possible to generate prosodic control data corresponding to intermediate intentions.

In the embodiment in which the user specifies a plurality of intentions, the user may also specify the weight of each intention in the weighting addition, and the user may further specify the weight for each prosodic component. This is because the importance of each prosodic component in producing the intention can be different between "anger" and "akire". In addition, the user is made to specify an intermediate intention of a plurality of intentions such as a feeling of "being bored with comfort", and such an intermediate intention is decomposed into a plurality of intentions, and metric control data corresponding to each intention is generated. The rhythm control data acquisition unit 110 may be made to acquire the data. In short, when the rhythm control data corresponding to the intention specified by the intention designation unit 108 is not stored in the rhythm library 122 (database), interpolation using a plurality of rhythm control data stored in the database is performed. Any mode may be used as long as the rhythm control data acquisition unit 110 is made to acquire the rhythm control data corresponding to the above-specified intention.

In the above embodiment, the user is made to specify the intention to be included in the answer to the question for each question to the voice reproduction device 10, but the intention to be included in the answer is to always synthesize the voice of the answer with the same intention. May be specified by the user in advance. Further, when the answers are synthesized according to a predetermined scenario, the intention may be specified according to the progress of the scenario. Specifically, scenario data in which the intention identifiers of the intentions to be included in the answer are arranged in the order of reproduction of the answer is input to the intention designation unit 108, and each time the voice of the question is input from the voice input unit 102, the intention designation unit 108 The intention identifiers included in the scenario data may be acquired in the order of description and output to the rhyme control data acquisition unit 110.

In addition, the personality, atmosphere, and speaking style characteristics of the voice playback device 10 or the answer character that anthropomorphizes the embedded stuffed character of the voice playback device 10 such as "always feel comfortable" or "always remain cautious" (hereinafter, , Character characteristics) are prepared in advance, an intention identifier corresponding to the character characteristics is associated with each character characteristic, and the user is asked to specify the character characteristics of the answer character to specify the intention to be included in the answer. You may let it. Further, a plurality of answer characters having different character characteristics may be prepared, and the user may specify the answer character to specify the intention to be included in the answer. When synthesizing the answers to the questions according to the scenario, the character characteristics may be defined for each character in the scenario. Furthermore, a plurality of character characteristics are associated with one answer character, and which character characteristic is used to synthesize the answer voice is determined in the scenario. You may produce the proper use of.

In addition, an intention specifying unit that analyzes the audio signal of the question and specifies the intention given to the question is provided, and the intention specifying unit 108 responds to the above question according to the intention specified by the intention specifying unit. You may let you specify the intention to give. Specifically, for each question in which the answer voice data is stored in the answer library 124, the waveform data of the voice in which the question is sounded on a flat plate without a specific intention is stored in advance in the intention identification unit as reference data. In the intention specifying unit, the voice signal input from the voice input unit 102 is converted into waveform data and compared with the corresponding reference data, so that the rhyme regulation data representing the rhyme in the voice represented by the waveform data is specified. (Data in the same format as the above rhyme control data) is generated. Then, the intention specifying unit reads out the intention identifier corresponding to the rhythm control data that is the same as or close to the rhythm regulation data from the rhythm library 122 and gives it to the intention designation unit 108. The intention designation unit 108 may cause the prosody control data acquisition unit 110 to output the intention identifier given by the intention identification unit and the intention identifier having a predetermined specific relationship. Here, the above-mentioned specific relationship may be a relationship of opposite intentions such as "carefulness" for "comfort" or a relationship of the same intention. When the "relationship of the same intention" is adopted as the specific relationship, the intention specification unit 108 is made to execute the process of outputting the intention identifier given by the intention identification unit to the prosody control data acquisition unit 110 as it is. Just do it.

<Evaluation and scoring of how to put intentions into questions>
In the embodiment in which the intention specifying unit is provided, the intention specified by the intention specifying unit may be compared with the stored contents of the prosody library 122 to evaluate and score how the intention is included in the question. Specifically, regarding the question uttered by the user with the intention such as "anger" or "awareness", whether or not the intention that the user put in the question was specified by the intention identification unit, or the intention identification According to the comparison result between the prosody regulation data generated from the voice signal of the question by the department and the prosody control data stored in the prosody library 122 for the intention, the suitability of the intention is evaluated and scored. Just do it. According to this aspect, it becomes possible to support the practice of how to put the intention into the question.

<Others>
In the above embodiment, the language analysis unit 104, the prosody library, and the answer library, which are configured to acquire the answer to the remark, are provided on the side of the voice reproduction device 10, but the processing load is heavy in the terminal device and the like. In consideration of the fact that the storage capacity is limited and the storage capacity is limited, the configuration may be provided on the external server side. Further, in the above embodiment, an example of application of the present invention to a voice reproduction device that synthesizes and reproduces the voice of an answer to a question by a user's voice to realize a voice dialogue has been described. The present invention may be applied to an audio reproduction device that plays the above.

102 ... voice input unit, 104 ... language analysis unit, 106 ... answer acquisition unit, 108 ... intention specification unit, 110 ... prosody control data acquisition unit, 112 ... answer reproduction unit, 122 ... prosody library, 124 ... answer library.

Claims (7)

An intention specifying unit that analyzes the input audio signal and identifies the psychological state of the questioning source contained in the question represented by the voice signal as the intention of the questioning source.
The answer acquisition unit that acquires the voice data of the answer to the above question,
An intention specifying unit that specifies the intention given to the answer according to the intention specified by the intention specifying unit, and
A prosody control data acquisition unit that acquires prosody control data representing a time change of the prosody according to the intention specified by the intention designation unit, and a prosody control data acquisition unit.
A response reproduction unit for reproducing the voice of the answer in which the time change of the prosody based on the voice data is controlled according to the prosody control data is provided.
The intention specifying unit uses the waveform data of the voice that sounds the question represented by the voice signal on a flat plate as the reference data, and the waveform data obtained by converting the voice signal and the reference data. By comparing, the rhyme regulation data representing the time change of the rhyme in the question represented by the voice signal is generated, and the intention of the question source is specified according to the rhyme regulation data, so that the voice of the question represented by the voice signal is specified. Identify the intent of the questioner based on the temporal change of the rhythm in
The intention designating unit designates an intention having a predetermined specific relationship with the intention specified by the intention specifying unit as an intention to be given to the answer.
Audio playback device comprising a call. It has a database in which the prosody control data is stored for each intention.
The prosody control data acquisition unit acquires prosody control data corresponding to the intention specified by the intention designation unit from the database.
The audio reproduction device according to claim 1. An intention specifying unit that analyzes the input audio signal and identifies the psychological state of the questioning source contained in the question represented by the voice signal as the intention of the questioning source.
The answer acquisition unit that acquires the voice data of the answer to the above question,
An intention specifying unit that specifies the intention given to the answer according to the intention specified by the intention specifying unit, and
A prosody control data acquisition unit that acquires prosody control data representing a time change of the prosody according to the intention specified by the intention designation unit, and a prosody control data acquisition unit.
An answer reproduction unit that reproduces the voice of the answer in which the time change of the prosody based on the voice data is controlled according to the prosody control data.
It has a database in which the prosody control data is stored for each intention.
The rhyme control data acquisition unit acquires the rhyme control data corresponding to the intention specified by the intention designation unit from the database, and stores the rhyme control data corresponding to the intention specified by the intention designation unit in the database. If not, the rhyme control data is acquired by interpolation using a plurality of rhyme control data stored in the database.
An audio playback device characterized by this. Computer,
It is an intention specifying unit that analyzes the input voice signal and specifies the psychological state of the questioning source contained in the question represented by the voice signal as the intention of the questioning source, and specifies the question represented by the voice signal. By using the waveform data of the voice sounded on the flat plate without intention as the reference data and comparing the waveform data obtained by converting the voice signal with the reference data, the temporal change of the rhyme in the question represented by the voice signal. By generating the rhyme regulation data representing the above and specifying the intention of the question source according to the rhyme regulation data, the intention of the question source is specified based on the time change of the rhyme in the voice of the question represented by the voice signal. Intention identification part to do,
The answer acquisition unit that acquires the voice data of the answer to the above question,
An intention designating unit that specifies the intention given to the answer according to the intention specified by the intention specifying unit, and an intention having a predetermined specific relationship with the intention specified by the intention specifying unit. , The intention designation part specified as the intention to be given to the answer,
A prosody control data acquisition unit that acquires prosody control data representing a time change of the prosody according to the intention specified by the intention designation unit, and a prosody control data acquisition unit.
An answer reproduction unit that reproduces the voice of the answer in which the time change of the prosody based on the voice data is controlled according to the prosody control data.
An audio playback program characterized by the ability to function. Computer,
An intention specifying unit that analyzes the input audio signal and identifies the psychological state of the questioning source contained in the question represented by the voice signal as the intention of the questioning source.
The answer acquisition unit that acquires the voice data of the answer to the above question,
An intention specifying unit that specifies the intention given to the answer according to the intention specified by the intention specifying unit, and
It is a rhyme control data acquisition unit that acquires rhyme control data representing a time change of the rhyme according to the intention specified by the intention designation unit, and is the intention designation unit from a database in which the rhyme control data is stored for each intention. If the rhythm control data corresponding to the intention specified by is acquired and the rhythm control data corresponding to the intention specified by the intention designation unit is not stored in the database, the rhythm control data is stored in the database. A rhyme control data acquisition unit that acquires by interpolation using a plurality of stored rhyme control data,
An answer reproduction unit that reproduces the voice of the answer in which the time change of the prosody based on the voice data is controlled according to the prosody control data.
An audio playback program characterized by the ability to function. It is an intention specifying step of analyzing the input voice signal and specifying the psychological state of the questioning source contained in the question represented by the voice signal as the intention of the questioning source, and the question represented by the voice signal is specified. By using the waveform data of the voice sounded on the flat plate without intention as the reference data and comparing the waveform data obtained by converting the voice signal with the reference data, the temporal change of the rhyme in the question represented by the voice signal. By generating the rhyme regulation data representing the above and specifying the intention of the question source according to the rhyme regulation data, the intention of the question source is specified based on the time change of the rhyme in the voice of the question represented by the voice signal. Intention-specific steps to be performed and
The answer acquisition step to acquire the voice data of the answer to the above question, and
It is an intention designating step in which the intention given to the answer is specified according to the intention specified in the intention specifying step, and has a predetermined specific relationship with the intention specified in the intention specifying step. An intention specification step that specifies the intention as the intention to be given to the answer, and
A prosody control data acquisition step for acquiring prosody control data representing a time change of prosody according to the intention specified in the intention designation step, and a prosody control data acquisition step.
An answer reproduction step of reproducing the voice of the answer in which the time change of the prosody based on the voice data is controlled according to the prosody control data, and
An audio reproduction method characterized by including. An intention specifying step that analyzes the input audio signal and identifies the psychological state of the questioning source contained in the question represented by the audio signal as the intention of the questioning source.
The answer acquisition step to acquire the voice data of the answer to the above question, and
An intention specification step that specifies the intention to be given to the answer according to the intention specified in the intention identification step, and
This is a rhyme control data acquisition step for acquiring rhyme control data representing a time change of the rhyme according to the intention specified in the intention designation step, and the intention designation is performed from a database in which the rhyme control data is stored for each intention. If the rhyme control data corresponding to the intention specified in the step is acquired and the rhyme control data corresponding to the intention specified in the intention designation step is not stored in the database, the rhyme control data is stored. A rhyme control data acquisition step acquired by interpolation using a plurality of rhyme control data stored in the database, and
An answer reproduction step of reproducing the voice of the answer in which the time change of the prosody based on the voice data is controlled according to the prosody control data, and
An audio reproduction method characterized by including.

Images