JP2005031150A - Apparatus and method for speech processing - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide speech recognition and speech synthesis when the speech recognition or speech synthesis are not corresponding to the mother tongue of a user, a language that the user is not reluctant to use is used and a consideration is given to the language which is not the mother tongue. <P>SOLUTION: Information regarding the language skill of the user is acquired (step S201), a language as an object of speech recognition is selected out of a plurality of languages according to the acquired information regarding the language skill (step S202), and operation conditions of speech recognition are set according to the information regarding the language skill and the language as the object of recognition (step S203). <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a speech processing apparatus and method capable of recognizing multilingual speech, and a speech processing apparatus and method capable of outputting multilingual speech.
[0002]
[Prior art]
In recent years, speech recognition devices that can recognize a plurality of languages and a speech synthesizer that can output a plurality of languages can be developed. However, in the current speech recognition apparatus and speech synthesis apparatus that support multiple languages, it is necessary to specify the language used by the user in advance. Here, if the user's native language is not included in the languages that can be processed by these devices, a language that is as resistant as possible to the user is selected, and these devices are configured so that the user can easily operate them. It is desirable to operate.
[0003]
Also, when considering a multilingual speech dialogue system using a multilingual speech recognition device and a multilingual speech synthesizer, ideally the language type handled by the speech recognition device and the language type handled by the speech synthesizer are the same. It is desirable, but not necessarily in reality. For example, in a multilingual speech dialogue system, speech recognition is available in five languages: English, Japanese, German, French, and Italian, but speech synthesis is available in three languages: English, Japanese, and Chinese. In some cases, the language is only supported. In this case, there is no problem for a user whose native language is English or Japanese because both speech recognition and speech synthesis are compatible with these languages. However, for German, for example, speech recognition is possible but speech synthesis is not possible. Conversely, for Chinese, speech synthesis is possible but speech recognition is not possible.
[0004]
Therefore, as a workaround in such a case, it is desirable to set an appropriate language other than German as a speech synthesis language for users whose native language is German. Similarly, for a user whose native language is Chinese, it is desirable that an appropriate language other than Chinese is set as the speech recognition language. Also, for example, when a user whose native language is Dutch, which does not support either speech recognition or speech synthesis, uses this multilingual speech dialogue system, an appropriate language out of the usable languages is spoken. It is desirable to set the language for recognition and speech synthesis.
[0005]
In response to such a request, for example, Patent Document 1 discloses a language-dependent object designated in a speech translation system in which a language-dependent object is configured by speech recognition, language analysis, language generation, and speech synthesis objects for each language. It is described that the language closest to the specified language (there is no description of the definition of proximity) is selected when it does not exist.
[0006]
[Patent Document 1]
JP 2001-222530 A
[Problems to be solved by the invention]
However, if an object is operated by simply selecting a close language by the method disclosed in Patent Document 1, a non-native language object is used as a native language object for a user. There is a possibility that the user cannot understand because the output utterance speed is too fast, or that speech recognition fails because the native pronunciation is not possible.
[0008]
Therefore, the present invention uses a language that is as resistant as possible to the user when speech recognition or speech synthesis does not correspond to the user's native language in the speech processing system, and the language is a non-native language. It is an object of the present invention to provide speech recognition and speech synthesis in consideration of the above.
[0009]
[Means for Solving the Problems]
According to one aspect of the present invention, there is provided a speech processing apparatus that performs speech recognition of a language selected from a plurality of languages, an acquisition unit that acquires information related to a user's language ability, and an information related to the acquired language ability. And a selection unit that selects a language to be a speech recognition target from the plurality of languages, and a setting unit that sets a speech recognition operation condition based on the information on the language ability and the language to be recognized. An audio processing device is provided.
[0010]
According to another aspect of the present invention, a speech processing apparatus that performs speech synthesis of a language selected from a plurality of languages, an acquisition unit that acquires information about a user's language ability, and information about the acquired language ability And selecting means for selecting a language for speech synthesis from the plurality of languages, and setting means for setting operating conditions for speech synthesis based on the language capability information and the language for speech synthesis. An audio processing device is provided.
[0011]
DETAILED DESCRIPTION OF THE INVENTION
DESCRIPTION OF EMBODIMENTS Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings.
[0012]
(First embodiment)
FIG. 1 is a block diagram showing the configuration of the speech processing apparatus according to the first embodiment of the present invention. This audio processing apparatus can be typically realized by a computer system using a CPU. Of course, you may implement | achieve with a dedicated hardware logic, without using CPU.
[0013]
A CPU 101 controls the entire voice processing apparatus according to a program loaded from the ROM 102 or the external storage device 104 to the RAM 103. The ROM 102 stores a boot program and various parameters. The RAM 103 functions as a main storage device that provides a work area when the CPU 101 executes various controls.
[0014]
Reference numeral 104 denotes a hard disk device serving as an external storage device. As shown in the figure, a voice processing program is installed in addition to the OS. This speech processing program includes a multi-language speech recognition program. The voice processing program is provided by being stored in, for example, the CD-ROM 110 a and installed in the external storage device 104 via the CD-ROM drive 110. Alternatively, it is possible to receive a voice processing program via a network (not shown).
[0015]
Reference numeral 105 denotes an audio input unit such as a microphone. Reference numeral 106 denotes an operation display unit such as a liquid crystal touch panel, which performs display / output of processing content setting / input, notification by characters, images, and the like. Reference numeral 107 denotes an auxiliary input / output unit, which can be composed of, for example, optical information such as buttons, numeric keys, a keyboard, a mouse, a pen, a switch, and an LED, Braille, and an actuator. An audio output unit 108 such as a speaker notifies a user of a message. Reference numeral 109 denotes a bus for connecting the above-described units.
[0016]
FIG. 2 is a diagram showing the configuration of the speech recognition module of the speech processing program in the present embodiment.
[0017]
The speech recognition module is roughly divided into an acoustic analysis module 1 and a search module 2. The acoustic analysis module 1 converts the voice input via the voice input unit 105 into an acoustic feature vector (for example, composed of Δ power, MFCC, etc.) at a fixed frame interval. The search module 2 uses the acoustic model 3 and the word dictionary 5 to perform a search while applying linguistic restrictions with the language model (or grammar) 4.
[0018]
The speech recognition in the embodiment is multilingual, and for example, corresponds to five languages: Japanese, English, German, French, and Italian. Here, as shown in the figure, the acoustic model 3, the language model 4, and the word dictionary 5 are respectively for native speakers and non-native speakers for each language described above. Includes one for each. Then, an acoustic model, a language model, and a word dictionary of a language selected by a method described later are selected, and search processing is executed.
[0019]
FIG. 3 is a flowchart showing processing for setting the operation conditions for speech recognition in the present embodiment.
[0020]
First, in step S201, utterance language information relating to the user's native language, the language that the user can utter, and its level, etc. is acquired as information about language ability. The user's native language is acquired by, for example, displaying a native language selection screen as shown in FIG. 4 on the operation display unit 106 and allowing the user to select it. Here, for example, when the user selects “Dutch”, the native language of the user is Dutch.
[0021]
In addition, instead of acquiring the utterance language information via the native language selection screen as described above, the utterance language information for each user may be stored in a file in advance and acquired based on this information. . In this case, it is preferable that the user notifies information related to the user in response to an operation such as input of the user ID.
[0022]
Alternatively, the utterance language information may be obtained by using a method for identifying the language based on the utterance content of the user.
[0023]
A language that the user can speak and information related to the level are acquired by displaying the speech level selection screen as shown in FIG. 5 on the operation display unit 106 and allowing the user to select it. For example, the speech level of each language is given in five levels (1 is low and 5 is high), and the user can select any appropriate one. The example in the figure shows a state in which English utterance level 4 is selected.
[0024]
It should be noted that the language that can be spoken by the user and the information on the level thereof are not acquired via the speech level selection screen as described above, but the user speaks a predetermined voice, and the voice content is subjected to voice recognition. Further, it may be acquired based on information such as a speech recognition rate, likelihood, and score at that time. In addition, as a method not based on speech recognition, the speech level may be acquired using a test for measuring a general language level.
[0025]
Next, in step S202, a language for speech recognition is selected based on the utterance language information. If the user's native language is English, Japanese, German, French, or Italian, speech recognition in the native language may be performed. However, if the user's native language is other than these, it is necessary to select a language that has the least burden on the user. That is, it is necessary to acquire information about which languages a user can speak among a plurality of languages that can be a target of speech recognition. If this information can be acquired via the utterance level selection screen shown in FIG. 5, the language set as the highest utterance level is set as the speech recognition target language. At this time, if the highest utterance level exists for a plurality of languages, the language is presented to the user and selected, or the language with a higher recognition rate for unspecified speakers in the speech recognition module Can be determined by automatically selecting.
[0026]
In the previous step S201, information on the user's native language and the user's utterance are sent via the native language selection screen as shown in FIG. 4 and the speech level selection screen as shown in FIG. Both the language that can be used and the information related to the level have been acquired as the utterance language information, but it is sufficient that at least one of them can be acquired. For example, if information about the native language cannot be obtained via the native language selection screen of FIG. 4, it is not possible to directly select the corresponding native language acoustic model or language model, but at least the utterance of FIG. Based on the information obtained via the level selection screen, it is possible to set the speech recognition operating conditions and language. On the other hand, when the language that can be spoken by the user and the information about the level cannot be obtained via the utterance level selection screen of FIG. 5, at least the native language obtained through the native language selection screen of FIG. Based on the information, information on the language and the level at which a general user of the person who speaks his / her native language can speak is shown in a form shown in FIG. By obtaining in advance that the density is high), it is possible to set the speech recognition operating conditions and language.
[0027]
Next, in step S203, operation conditions for speech recognition are set. At this time, the utterance level for the language used by the user obtained through the utterance level selection screen shown in FIG. 5 is regarded as “intimacy” for the language, and the operation condition of voice recognition is determined according to the intimacy value. change. For example, the familiarity can be defined by doubling the utterance level from 1 to 5 in FIG. The operation condition for speech recognition according to the familiarity is determined by referring to an operation condition table describing a correspondence relationship between the familiarity and the operation condition. This operating condition table has a structure as shown in FIG. 7, for example, and is stored in the hard disk 104 along with the voice processing program. Speech recognition operating conditions include search processing such as search conditions for speech recognition, results output such as the number of recognition candidates, acoustic models such as the type of acoustic model, linguistics such as vocabulary and grammar for speech recognition, etc. There are things related to various models.
[0028]
In the operation condition table shown in FIG. 7, as search conditions, the beam width value in the beam search, the N number of N best outputs (sentence hypotheses up to the top N) are generated, and native / non− in the acoustic model. Selection of native, selection of native / non-native in the language model, and the like are defined. Referring to FIG. 7, for example, when the familiarity is 6, the speech width of 200, the N best output number, 5, the non-native acoustic model, and the non-native recognition grammar are used. Operating conditions are set. Therefore, in this step S203, for each of the acoustic model, language model, and word dictionary of the language selected in step S202, whether for native or non-native is selected according to the familiarity.
[0029]
In FIG. 2, there are non-native models for all of the acoustic model, language model, and word dictionary. However, only some of these models may include a non-native model, or a non-native model may be included. It is good also as a structure which does not include this model. Further, the number of non-native models is not limited to one, and a configuration using a plurality of models corresponding to the degree of non-native, the type and category of the native language may be used. Here, as a native model, an acoustic model, a language model, and a word dictionary may be created by a commonly used method. Next, as a non-native acoustic model, an acoustic model can be created using a speech database uttered by non-native. The non-native language model includes a method for creating a simple grammar composed of plain words and short sentences, and a method for creating a statistical language model such as N-gram using plain words and short sentences. Can be used. In addition, the non-native word dictionary can create a pronunciation dictionary that is easy to utter non-native.
[0030]
In this way, the speech recognition operating conditions are set according to the user's utterance language and utterance level. For this reason, when a non-native language is recognized by speech, it is set so that speech recognition is executed using a non-native acoustic model, a language model, and a word dictionary. It is less likely that voice recognition will fail because it is not possible.
[0031]
(Second Embodiment)
The second embodiment relates to speech synthesis processing. The configuration of the speech processing apparatus according to this embodiment is the same as that shown in FIG. 1, but the speech processing program in this embodiment is as shown in FIG. 8 instead of the speech recognition module in the first embodiment. The speech synthesis module 10 having the above configuration is included.
[0032]
The speech synthesis in this embodiment is multilingual compatible, for example, in three languages, Japanese, English and Chinese. Accordingly, as shown in FIG. 8, the speech synthesis module 10 includes synthesis modules 11, 12, and 13 for national languages, and a synthesis module corresponding to a language selected by a method described later is executed. The synthesis module 11 for Japanese has a configuration as shown in the figure, for example. The text analysis module 14 performs syntax analysis (specifically, morphological analysis) of the input Japanese text using the language dictionary 14a. As shown in the figure, the language dictionary 14a preferably has a native one and a non-native one individually. The language processing module 15 includes a phoneme processing module 15a and a prosody processing module 15b. The phoneme processing module 15a outputs a phoneme symbol string based on the analysis result of the text, and the prosody processing module 15b performs pause, accent, intonation, and continuation. Prosodic information such as time length is output. The acoustic processing module 16 generates synthesized speech using the waveform dictionary 17 based on the input phoneme symbol string and prosodic information. In addition, the waveform dictionary is individually provided for native speakers and for non-native speakers. The English synthesizing module 12 and the Chinese synthesizing module 13 have the same configuration as the Japanese synthesizing module 11, and thus illustration and description thereof are omitted.
[0033]
FIG. 9 is a flowchart showing processing for setting the operating conditions for speech synthesis in the present embodiment.
[0034]
First, in step S301, listening language information about the user's native language, the language that the user can understand (listening) by listening to sound, and the level thereof is acquired. The user's native language is obtained by displaying the native language selection screen similar to that shown in FIG. 4 described in the first embodiment on the operation display unit 106 and allowing the user to select it. In addition, the language that can be understood by the user and the information related to the level are also acquired by displaying the utterance level selection screen similar to that in FIG. 5 described in the first embodiment on the operation display unit 106 and making the user select it. To do.
[0035]
Note that the listening language information may be obtained based on this information by previously storing listening language information for each user in a file or the like. In this case, it is preferable that the user notifies the information about the user by performing an operation such as inputting a user ID.
[0036]
Information about the language level that the user can listen to may be acquired based on a response result or a response time of the user with respect to speech synthesis of a predetermined content.
[0037]
Next, in step S302, the language of the audio output is selected based on the listening language information. If the user's native language is English, Japanese, or Chinese, speech synthesis using the native language may be performed. However, when the user's native language is other than these three languages, it is necessary to select a language with as little burden on the user as possible. That is, it is necessary to acquire information as to which language is easy for a user to understand among languages that can be output. If this information can be acquired via the utterance level selection screen, the language for voice output is set using the information.
[0038]
In step S301, both the user's native language and the language that the user can listen to and the level related to the level are acquired as listening language information via the native language selection screen and the speech level selection screen. However, it is sufficient that at least one of them can be acquired. For example, if information about the native language cannot be obtained via the native language selection screen of FIG. 4, it is not possible to specify the corresponding native language synthesis module, but at least the speech level selection screen of FIG. It is possible to set the speech synthesis operating conditions and language based on the information obtained through the menu. On the other hand, when the language that can be understood by the user via the utterance level selection screen in FIG. 5 and information on the level cannot be obtained, the information on the native language obtained through the native language selection screen in FIG. Based on the information on the language and the level that can be understood by a general user who speaks his / her native language as the familiarity in the form as shown in FIG. You can set the language.
[0039]
Next, in step S303, speech synthesis operating conditions are set. At this time, the listening level for the language selected by the user obtained through the utterance level selection screen similar to that in FIG. 5 is regarded as “intimacy” for the language, and the speech synthesis operation condition is determined according to the familiarity value. To change. The listening level and familiarity can be defined in the same manner as in the method described in the first embodiment.
[0040]
The speech synthesis operation condition corresponding to the familiarity is determined by referring to an operation condition table describing the correspondence between the familiarity and the operation condition. This operating condition table has a structure as shown in FIG. 11, for example, and is stored in the hard disk 104 along with the voice processing program. Speech synthesis operating conditions include prosody factors such as speech synthesis speech rate, output factors such as volume, phonological factors such as waveform dictionary types, and linguistic factors such as the contents of response sentences There is something about.
[0041]
In the operation condition table shown in FIG. 11, the utterance speed in prosodic control as a prosodic factor, the volume as a factor of output, the selection of a native or non-native waveform dictionary as a phonological factor, the linguistic factor The contents of the response sentence and the number of repetitions are specified. Referring to FIG. 11, when the intimacy is 6, for example, the speaking rate is slow, the volume is 7 (7 out of 10 steps where 0 is the minimum volume and 9 is the maximum volume), and non-native voice is used. The speech synthesis operating conditions such as using the waveform dictionary and the polite response sentence are set. Accordingly, in this step S303, the speech synthesis module of the language selected in step S302 is selected, and further, the native or non-native one of the waveform dictionaries is selected according to the familiarity.
[0042]
In FIG. 8, there is a non-native waveform dictionary in the waveform dictionary, but the non-native waveform dictionary may not be included. Further, the number of non-native waveform dictionaries is not limited to one, and a plurality of waveform dictionaries according to the degree of non-native, the type and category of the native language may be used. Here, the native waveform dictionary may be created by a commonly used method. Next, a non-native waveform dictionary can be created by using voice data obtained by uttering a language in which non-native is native. In addition, it is created by associating speech data in which non-native speaks non-native language, phoneme system (phonetic system) of non-native language and phoneme system (phonetic system) of native language. You can also.
[0043]
In this way, the operation conditions for speech synthesis are set according to the listening level of the user's native language and audible language. For this reason, when synthesizing a non-native language, it is set so that speech synthesis is performed in accordance with the phoneme and prosody information adjusted for non-native, which helps to improve the understanding of the contents of the speech output. .
[0044]
(Third embodiment)
In the first embodiment, a speech processing apparatus having a speech recognition function has been described. In the second embodiment, a speech processing apparatus having a speech synthesis function has been described. However, the first embodiment and the second embodiment are described. Can be combined to realize a speech processing apparatus having both a speech recognition function and a speech synthesis function.
[0045]
The configuration of the speech processing apparatus according to the third embodiment is the same as that shown in FIG. 1, but the speech processing program in this embodiment is the speech recognition module in the first embodiment and the second embodiment. And a speech synthesis module.
[0046]
FIG. 12 is a flowchart showing a process for setting operation conditions for speech recognition and speech synthesis in the present embodiment.
[0047]
The procedure of the operation is the same as the processing in the first and second embodiments except that in step S401, information related to the utterance and the listening language is acquired at the same time. Here, the language used in the speech recognition language and the speech synthesis language can be obtained independently in steps S402 and S404 according to the methods described in the first and second embodiments, respectively. Some users may find it strange to use different languages for speech recognition and speech synthesis, so it is possible to apply the language that maximizes the sum of the intimacy of speech recognition and speech synthesis to speech recognition and speech synthesis. It is.
[0048]
(Other embodiments)
The embodiment of the present invention has been described in detail above. However, the present invention can take an embodiment as a system, apparatus, method, program, storage medium, or the like. In addition, the present invention may be applied to a system composed of a plurality of devices, or may be applied to an apparatus composed of a single device.
[0049]
In the present invention, a software program that realizes the functions of the above-described embodiments is directly or remotely supplied to a system or apparatus, and the computer of the system or apparatus reads and executes the supplied program code. Including the case where it is also achieved by. In that case, as long as it has the function of a program, the form does not need to be a program.
[0050]
Accordingly, since the functions of the present invention are implemented by computer, the program code installed in the computer also implements the present invention. That is, the scope of the claims of the present invention includes the computer program itself for realizing the functional processing of the present invention.
[0051]
In this case, the program may be in any form as long as it has a program function, such as an object code, a program executed by an interpreter, or script data supplied to the OS.
[0052]
As a recording medium for supplying the program, for example, flexible disk, hard disk, optical disk, magneto-optical disk, MO, CD-ROM, CD-R, CD-RW, magnetic tape, nonvolatile memory card, ROM, DVD (DVD-ROM, DVD-R).
[0053]
As another program supply method, a client computer browser is used to connect to an Internet homepage, and the computer program itself of the present invention or a compressed file including an automatic installation function is downloaded from the homepage to a recording medium such as a hard disk. Can also be supplied. It can also be realized by dividing the program code constituting the program of the present invention into a plurality of files and downloading each file from a different homepage. That is, a WWW server that allows a plurality of users to download a program file for realizing the functional processing of the present invention on a computer is also included in the claims of the present invention.
[0054]
In addition, the program of the present invention is encrypted, stored in a storage medium such as a CD-ROM, distributed to users, and key information for decryption is downloaded from a homepage via the Internet to users who have cleared predetermined conditions. It is also possible to execute the encrypted program by using the key information and install the program on a computer.
[0055]
In addition to the functions of the above-described embodiments being realized by the computer executing the read program, the OS running on the computer based on the instruction of the program is a part of the actual processing. Alternatively, the functions of the above-described embodiment can be realized by performing all of them and performing the processing.
[0056]
Furthermore, after the program read from the recording medium is written in a memory provided in a function expansion board inserted into the computer or a function expansion unit connected to the computer, the function expansion board or The CPU or the like provided in the function expansion unit performs part or all of the actual processing, and the functions of the above-described embodiments are realized by the processing.
[0057]
【The invention's effect】
According to the present invention, in a speech processing system, when speech recognition or speech synthesis does not correspond to a user's native language, a language that is as resistant as possible to the user is used, and the language is a non-native language. Therefore, speech recognition and speech synthesis are provided.
[Brief description of the drawings]
FIG. 1 is a block diagram illustrating a configuration of a sound processing apparatus according to an embodiment.
FIG. 2 is a diagram illustrating a configuration of a voice recognition module in the embodiment.
FIG. 3 is a flowchart illustrating processing for setting operation conditions for speech recognition in the embodiment.
FIG. 4 is a diagram illustrating an example of a native language selection screen in the embodiment.
FIG. 5 is a diagram showing an example of an utterance level selection screen in the embodiment.
FIG. 6 is a diagram illustrating an example of intimacy with respect to another language for each native language.
FIG. 7 is a diagram illustrating an example of an operation condition table for speech recognition in the embodiment.
FIG. 8 is a diagram illustrating a configuration of a speech synthesis module in the embodiment.
FIG. 9 is a flowchart illustrating processing for setting operation conditions for speech synthesis in the embodiment.
FIG. 10 is a diagram showing an example of familiarity with another language for each native language.
FIG. 11 is a diagram illustrating an example of an operation condition table for speech synthesis in the embodiment.
FIG. 12 is a flowchart illustrating processing for setting operation conditions for speech recognition and speech synthesis in the embodiment.

Claims (12)

A speech processing apparatus that performs speech recognition of a language selected from a plurality of languages,
An acquisition means for acquiring information about the language ability of the user;
Selection means for selecting a language to be subjected to speech recognition from the plurality of languages based on the acquired information on the language ability;
Setting means for setting operation conditions for speech recognition based on the information on the language ability and the language to be recognized;
A speech processing apparatus comprising: The speech processing apparatus according to claim 1, wherein the information regarding the language ability includes native language information and utterance level information of at least one of the plurality of languages. The said selection means selects languages other than the said native language based on the said speech level information, when a user's native language is not contained in these languages. Audio processing device. The speech according to claim 1, wherein the operation condition includes at least one of a search condition for a recognition candidate, an output condition for a recognition result, an acoustic model selection, a language model selection, and a word dictionary selection. Processing equipment. A method of performing speech recognition by selecting one of a plurality of languages,
Obtaining information about the user's language skills;
Selecting a language to be subjected to speech recognition from the plurality of languages based on the acquired information on the language ability;
Setting speech recognition operating conditions based on the information about the language ability and the language to be recognized;
A method characterized by comprising: In order to select one of several languages and perform speech recognition,
Obtaining information about the user's language skills;
Selecting a language to be subjected to speech recognition from the plurality of languages based on the acquired information about the language ability;
Setting operating conditions for speech recognition based on the information about the language ability and the language to be recognized;
A program for running A speech processing apparatus that performs speech synthesis of a language selected from a plurality of languages,
An acquisition means for acquiring information about the language ability of the user;
Selection means for selecting a language for speech synthesis from the plurality of languages based on the acquired information on language ability;
Setting means for setting operation conditions for speech synthesis based on the information on the language ability and the language for speech synthesis;
A speech processing apparatus comprising: 8. The speech processing apparatus according to claim 7, wherein the information on the language ability includes native language information and listening level information of at least one of the plurality of languages. 9. The selection unit according to claim 8, wherein when the user's native language is not included in the plurality of languages, the selection unit selects a language other than the native language based on the listening level information. Audio processing device. The speech processing apparatus according to claim 7, wherein the operation condition includes at least one of an utterance speed, a volume, a waveform dictionary selection, and a response sentence content. A method for performing speech synthesis by selecting one of a plurality of languages,
Obtaining information about the user's language skills;
Selecting a language for speech synthesis from the plurality of languages based on the acquired information about the language ability;
Setting operating conditions for speech synthesis based on the information about the language ability and the language for speech synthesis;
A method characterized by comprising: In order to select a language from multiple languages and perform speech synthesis,
Obtaining information about the user's language skills;
Selecting a language for speech synthesis from the plurality of languages based on the acquired information about the language ability;
Setting operating conditions for speech synthesis based on the information about the language ability and the language for speech synthesis;
A program for running

Images