JP2002269004A - Electronic mail reader - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an electronic mail reader where the listening sense of a user on the reading output of an electronic mail can be improved. SOLUTION: In the electronic mail reader, a plurality of pieces of sound quality information of senders (or the other persons such as entertainers), which are made to correspond to the respective senders of the electronic mails, are registered in a sound quality data base 9b. The sender of the electronic mail is automatically discriminated by header information of the electronic mail being the object of reading and by information in a sender data base 9a. Sound for reading electronic mail is synthesized based on sound quality information in the sound quality data base 9b, which is made to correspond to the sender.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an electronic mail reading device.

[0002]

2. Description of the Related Art A conventional text-to-speech apparatus that reads out and outputs an e-mail by voice uses a text-to-speech synthesis technique. In this general text-to-speech synthesis technology, two types of male and female voice quality information are provided as standard voice quality information. Then, based on the e-mail address of the sender of the e-mail or the like, one of the two types of voice quality information is selected, and the e-mail is read aloud by a voice created based on the voice quality information. Has become.

[0003] Such an e-mail reading device is used, for example, for in-vehicle use, and is suitably used for checking the contents of received e-mails during driving.

[0004]

However, in the configuration in which the voice quality is switched by gender according to the sender's mail address or the like and the e-mail is read out as in the above-mentioned prior art, there are few types of voice quality prepared. In many cases, the voice quality at the time of reading out is significantly different from that of the sender, giving a sense of incongruity to the user of the reading out output, and causing a problem that the listening feeling is impaired.

SUMMARY OF THE INVENTION In view of the above problems, an object of the present invention is to provide an e-mail reading apparatus which can improve the user's perceived readability of e-mail reading output.

[0006]

A technical means for achieving the above object is an electronic mail reading device for reading out an electronic mail by voice and outputting the read out electronic mail.
A first storage unit for storing a sender database in which the sender of the electronic mail is registered, and a second storage unit for storing a voice quality database in which voice quality information associated with each of the senders included in the sender database is registered. Storage means, a sender discriminating means for analyzing header information of the given e-mail, and discriminating which of the senders included in the sender database the e-mail is from; Based on the voice quality information corresponding to the sender determined by the sender determination means among the voice quality information included in the voice quality database of the storage means, the given electronic mail is converted into a voice quality corresponding to the voice quality information. Voice synthesizing means for synthesizing voice to be read out and outputting the voice via the speaker. And wherein the Rukoto.

[0007] Preferably, the electronic mail reading device is fetched by first information fetching means for fetching voice information and sender information indicating the sender corresponding to the voice information, and the first information fetching means. Based on the voice information, a voice quality feature of the voice information is extracted, and the voice quality information created by the voice feature information is generated by the voice quality information indicating the voice quality feature. First database creation means for registering the sender specified by the sender information and corresponding to the voice information used to create the voice quality information in the voice quality database and the sender database in association with each other. Is good.

[0008] Preferably, the first information fetching means converts the voice information from a portable recording medium.
Alternatively, it is good to take in via a communication line.

[0009] Further, preferably, the feature extracting means, when creating the voice quality information, stores the voice quality information in the existing information corresponding to the sender corresponding to the voice information captured by the first information capturing means. When the voice quality information already exists in the voice quality database, the voice quality characteristics indicated by the existing voice quality information and the characteristics newly extracted from the voice information captured by the first information capturing unit are added. Based on this, it is preferable to create new voice quality information that further reflects the characteristics of voice quality corresponding to the sender.

[0010] Preferably, the electronic mail reading device further comprises: second information capturing means for capturing the voice quality information and sender information indicating the sender corresponding to the voice quality information; and second information capturing means. The voice quality information captured, and, the second database creation means, which is specified by the sender information, and registers the sender corresponding to the voice quality information in the voice quality database and the sender database in association with each other. It is better to have more.

Further, preferably, the second information capturing means captures the voice quality information from a portable recording medium or via a communication line.

[0012] Preferably, the second database creation means, when registering the fetched voice quality information in the voice quality database, stores the existing voice quality information corresponding to the sender corresponding to the voice quality information. If already present in the voice quality database, based on the voice quality characteristics indicated by the voice quality information indicated by the existing voice quality information and the voice quality characteristics indicated by the voice quality information newly captured by the second information capturing means, It is preferable to further include a function of newly creating voice quality information that further reflects characteristics of voice quality corresponding to the sender and registering the information in the voice quality database.

[0013]

DESCRIPTION OF THE PREFERRED EMBODIMENTS <Schematic Description of Overall Configuration> FIG.
1 is a block diagram of an electronic mail reading device according to an embodiment of the present invention. This e-mail reading device is shown in FIG.
As shown in FIG. 1, an operation unit 1 as an input unit and a communication unit 3
A reading unit 5, an information processing unit 7, a storage unit (first and second storage means) 9, a display unit 10, an amplifier 11,
It is configured to include a speaker 13, and here, it is assumed to be mounted on a vehicle. The information processing unit 7 includes a central control unit 7
a, a mail transmission / reception unit 7b, an information acquisition unit 7c, a mail management unit 7d, a voice quality information management unit (feature extraction unit, first and second database creation units) 7e, a display processing unit 7f, and a speech synthesis unit (sender identification) 7g). In this configuration, the first and second information capturing means according to the present invention include an operation unit 1, a communication unit 3, a reading unit 5, a central control unit 7a, and an information capturing unit 7c.
Is supported.

Here, this reading device generally includes
It has a function of sending and receiving e-mails and a function of reading out received e-mails by voice. The text-to-speech output is performed by a voice synthesized according to the characteristics of the voice quality of the sender of the electronic mail, or a voice of voice quality previously associated with the sender.

As for the reading of the e-mail, a reading mode in which reading is performed and a non-reading mode in which reading is not performed are prepared so as to be switchable. In the reading mode, first and second sub-modes are provided. In the first sub-mode, an e-mail is read out with a voice having a voice quality associated with the sender of the e-mail. In the second sub-mode, the e-mail is read aloud in a manually set male or female standard voice.

The mode switching between the reading mode and the non-reading mode and between the first sub-mode and the second sub-mode is based on an operation input from the operating unit 1 for mode switching. This is performed by the central control unit 7a described later.

In the first sub-mode of the reading mode, if the voice quality of the sender or another voice quality such as a particular entertainer is associated with each sender, the associated voice quality is automatically set. , And the e-mail from each sender is read out using the voice of the selected voice quality.

<Principle explanation of voice quality feature extraction and reproduction>
Next, before describing a specific configuration of the reading device, a brief description will be given of a method of extracting voice characteristic features and a method of reproducing the extracted voice characteristic employed in the present embodiment. For each of the above-mentioned techniques, a known technique (a technique using a vector envelope) shown in a document such as “Digital signal processing of voice / sound information” (co-authored by Shokodo, 1997, Kiyohiro Kano, etc.) is used. Can be

The features of voice quality when reading a document by voice include acoustic features (tone colors) and prosodic features (sound duration,
The feature extraction of voice quality is performed by digitally extracting feature parameters (pitch component (height), intensity, and spectral envelope component) indicating these two features for each phoneme used for speech synthesis. This is performed by extracting from audio information. This feature extraction process is performed by a voice quality information management unit 7e described later.

The extraction step will be described in more detail.
It looks like this: First, by performing speech recognition on the digital voice information of the sender or the like (tentatively, Mr. A) to be associated with the sender, meaningful text information (= phoneme sequence) and start / end time of each phoneme Get. In the speech recognition, only the spectral envelope among the feature parameters is used. However, when analyzing the speech information, the pitch component (height) and the strength indicating the characteristics of the voice quality of each phoneme can be obtained at the same time. Therefore, by associating the phoneme sequence obtained as a result of the speech recognition with the result of analyzing the speech information, voice quality information indicating the characteristics of the voice quality of each phoneme can be obtained. The extraction of text information is performed using a phoneme model stored in a phoneme database 9d described later and a vocabulary dictionary stored in a vocabulary database 9e described later.

By the way, although the basic phonemes of Japanese are composed of 43, the characteristics of the phonemes change depending on the following phonemes, so that in speech synthesis, for example, phonemes expanded to about 1000 are used. Has become. For example,
The phoneme of "k" in the pronunciation of "red (aka)" and the phoneme of "k" in the pronunciation of "autumn (aki)" have different characteristics due to the difference in the following phonemes, and are different in speech synthesis. Is treated as

Subsequently, a mapping relation between the extracted voice quality pattern of each phoneme and a standard pattern indicating a standard voice quality feature prepared for each phoneme is calculated, and based on the calculation result, Then, a conversion codebook necessary for reproducing the voice pattern of Mr. A by converting a standard pattern of voice quality of all phonemes required for speech synthesis is created. In the present embodiment, this conversion codebook is registered in a voice quality database 9b described later as voice quality information for reproducing the voice quality of Mr. A. In general, the number of phonemes for which a voice quality pattern can be actually obtained from speech information is the total number of phonemes required for speech synthesis (here, about 1000).
For other phonemes that are only a part of the phonemes of which the voice pattern could not be actually obtained, based on the sampling results of the phonemes for which the voice quality pattern was actually obtained,
By analogy, a conversion code for the standard pattern is created.

In the case of speech synthesis for reading aloud,
By converting the standard pattern of each phoneme using the conversion codebook stored in the voice quality database 9b, the voice quality pattern of each phoneme associated with each sender is reproduced, and based on the voice quality pattern. Speech for reading is synthesized. This speech synthesis processing is performed by a speech synthesis unit 7g described later. As a modification, the voice quality pattern of each phoneme may be registered as voice quality information in the voice quality database 9b, and the voice quality for reading aloud may be reproduced based on the voice quality pattern.

<Detailed Description of Apparatus Configuration> Next, the configuration of the reading apparatus will be described. The operation unit 1 is for receiving an operation input from a user.

The communication unit 3 is provided with a portable telephone or the like, and is controlled by a mail transmission / reception unit 7b and an information acquisition unit 7c of the information processing unit 7 to use a communication line (such as the Internet) using a wireless telephone line. And a function of transmitting and receiving information to and from a predetermined server or the like via a communication line. The communication unit 3 has a function of making a call by connecting to a telephone line, and under the control of the information acquisition unit 7c, a telephone that performs a call process with a predetermined communication partner via the telephone line. It also has the ability.

The reading section 5 fetches information such as voice information or voice quality information from a portable recording medium 5a such as a CD-ROM under the control of the information fetching section 7c.

The storage unit 9 stores a sender database 9a,
A voice quality database 9b, a mail database 9c, a phoneme database 9d, and a vocabulary database 9e are provided. The sender database 9a includes address information of a plurality of senders (name, telephone number, mail address, gender, voice quality information identification ID, etc.). The received e-mail is stored in the mail database 9c.

The voice quality database 9b has standard pattern information indicating the characteristics of the standard voice quality of each phoneme used for speech synthesis, and is associated with each sender registered in the sender database 9a. , And a plurality of conversion codebooks for reproducing voice quality associated with each sender by using, and two standard male and female voice quality patterns.

The phoneme database 9d and the vocabulary database 9e store phoneme models and vocabulary dictionaries used for speech recognition.

The central control unit 7a of the information processing unit 7 controls the functional elements (7b to 7g) of the information processing unit 7 based on an operation input input via the operation unit 1. And a function of creating an e-mail based on an input from the operation unit 1 or the like.

The mail transmission / reception unit 7b has a function of receiving an electronic mail from a predetermined server via the communication unit 3 and the communication network in response to a mail reception command given from the central control unit 7a, and a function provided by the central control unit 7a. And a function of transmitting an e-mail provided from the central control unit 7a to a predetermined destination via the communication unit 3 and the communication network in response to a given mail transmission command. The e-mail received by the mail transmitting / receiving unit 7b is once sent to the mail management unit 7d.

The information fetching unit 7c responds to an information fetching instruction given from the central control unit 7a, from a predetermined server via the communication unit 3 and the communication network, or from the recording medium 5a via the reading unit 5. Has a function of taking in voice information or voice quality information (here, voice quality parameters) from the. Whether the information is obtained from the server or the recording medium 5a via the communication network, and whether the voice information or the voice quality information is obtained as the obtained information is determined according to the contents indicated by the information obtaining command. The voice information and voice quality information include personal information (sender information) indicating the identity of the voice information and voice quality information.
May be given, and in that case, the person information is also taken in together with the voice information and the voice quality information.

Further, the information fetching unit 7c includes a central control unit 7
Under the control of a, the communication unit 3 also has a function of enabling communication with a communication partner via a telephone line, and taking in voice information of the communication partner via the communication unit 3.

The mail management section 7d is provided with a mail transmission / reception section 7b.
In response to a read-out command given from the central control unit 7a, the electronic mail specified by the read-out command is read from the mail database 9c, and the voice synthesizing unit 7g is read. And a function of reading an e-mail specified by the display command from the mail database 9c in response to a display command given from the central control unit 7a and sending it to the display processing unit 7f.

The voice quality information management section 7e generally performs conversion codebook creation processing, management of the sender database 9a and voice quality database 9b, etc. in accordance with an instruction given from the central control section 7a. More detailed processing content of the voice quality information management unit 7e will be described later based on the flowchart of FIG.

The display processing unit 7f controls the display contents of the display unit 10 under the control of the central control unit 7a, and outputs and outputs, for example, operation information and the contents of an e-mail via the display unit 10.

In general, when the operation mode of the reading device is the reading mode, the voice synthesizing unit 7g responds to the instruction given from the central control unit 7a to send each of the sub-modes described above for each e-mail. And synthesizes a read-out voice having a voice quality corresponding to the above, and reads out and outputs an e-mail provided from the mail management unit 7d via the amplifier 11 and the speaker 13. More detailed processing content of the voice synthesizing unit 7g will be described later based on the flowchart of FIG.

<Description of Operation> FIG. 2 is a flow chart showing the processing contents for creating a voice quality database and the like.
In step S1, when the central control unit 7a receives an information fetch instruction input via the operation unit 1, information fetch processing is performed in step S2. In this information fetching process, the information fetching section 7 is instructed by the central control section 7a.
c captures voice information or voice quality information (here, voice quality parameters) via the communication unit 3 or the reading unit 5. The captured information is provided to the voice quality information management unit 7e.

Here, the manner in which voice information is taken in through the communication unit 3 includes a mode in which voice information and the like are acquired from a specific server via a communication network, and a speech content of a communication partner through a telephone line. , The voice information of the other party is obtained, and in which mode the voice information is obtained is specified by the user. Further, when the voice information or voice quality information to be acquired is given the personal information of the voice information or voice quality information, the personal information is also taken in.

In the following step S3, when the voice information is obtained in step S2, the voice quality information management unit 7e extracts the text information included in the voice information by voice recognition as described above, and Based on the information and the voice information, voice quality parameters indicating characteristics of voice quality of each phoneme included in the voice information are extracted, and the process proceeds to step S4. When the voice quality parameter is read instead of the voice information in step S2, the process proceeds to the next step S4 without performing the feature extraction process.

In step S4, person information is included in the acquired information obtained in step S2, and the sender in the sender database 9a corresponding to the person information is determined based on the person information and the information in the sender database 9a. Is determined by the voice quality information management unit 7e. Then, the sender in the sender database 9a corresponding to the person information is specified based on the information such as the name, mail address, telephone number, etc. of the person included in the person information and the address information of each sender in the sender database 9a. If it is possible, the process proceeds to step S5. On the other hand, if the acquired information does not include the person information, or if the person information or the information in the sender database 9a is insufficient, the corresponding sender cannot be specified. Proceed to.

In step S5, the operation mode of the voice quality information management unit 7e is set to an automatic codebook creation mode for automatically creating and registering a conversion codebook, or the necessity of creation by the user. Is confirmed, the voice quality information management unit 7e determines whether or not the mode is set to the manual mode for creating and registering the codebook for conversion. If the mode is set to the automatic codebook creation mode, the process proceeds to step S6. On the other hand, if not set, the process proceeds to step S8. Here, switching between the automatic codebook creation mode and the manual mode can be performed by a user input via the operation unit 1 and the central control unit 7a.

In step S6, it is determined whether or not the conversion codebook associated with the sender in the sender database 9a specified in step S4 exists in the voice quality database 9b by checking the address of the corresponding sender in the sender database 9a. The voice quality information management unit 7e determines based on the presence / absence of the voice quality information identification ID in the information, and if there is a corresponding conversion codebook, the process proceeds to step S7, while the corresponding conversion codebook does not exist. In this case, the process proceeds to step S11.

In step S7, the voice quality information management unit 7e uses the voice quality parameter extracted in step S3 or the voice quality parameter obtained in step S2, and uses the voice quality parameter identified by the voice quality information identification ID to store the corresponding existing conversion information. The contents of the codebook are optimized.

By the way, in general, voice parameters of each phoneme are extracted from voice information as described above, a conversion codebook is created based on the voice quality parameters, and an original code of each phoneme is generated based on the conversion codebook. When reproducing voice quality, it is often difficult to reproduce voice quality close to true voice quality by one sampling. The symbol A in FIG.
FIG. 4 schematically shows a spectrogram showing characteristics of the true voice quality of a specific phoneme, and reference numeral B1 schematically shows a spectrogram showing characteristics of the voice quality of the phoneme reproduced based on one sampling result. From this, it can be seen that it is difficult to reproduce the characteristics of the true voice quality by one sampling.

Therefore, in the present embodiment, when a voice quality parameter indicating a voice quality characteristic used for voice quality reproduction of each phoneme is additionally obtained, in step S7, the voice quality parameter is obtained from the corresponding existing conversion codebook. Based on the existing voice quality parameters obtained and the voice quality parameters additionally obtained in step S3, the conversion codebook is reconfigured and optimized, and the optimized conversion codebook is stored in the voice quality database 9b. It is to be registered. The symbol B2 in FIG. 4 schematically shows a spectrogram showing characteristics of the voice quality of the phoneme reproduced based on the first and second sampling results, and a voice quality closer to the true voice quality is reproduced. You can see that it is done.

In step S11, voice quality information management section 7e
Based on the voice quality parameters extracted in step S3 or the voice quality parameters acquired in step S2, a conversion codebook is newly created and registered in the voice quality database 9b as described above,
A voice quality information identification ID for identifying the conversion codebook is created and registered in the address information of the corresponding sender in the sender database 9a.

In this step S11, if the information to be registered as address information in the sender database 9a has been obtained in step S2 or step S8 described later, the information is stored in the sender database 9a.
Registered within.

In step S8, the voice quality information management unit 7e instructs the central control unit 7a to perform a process of confirming the necessity of registration of voice quality information for the user and a process of accepting input of necessary information from the user. A necessity confirmation process and a necessary information input reception process are performed.

In the necessity confirmation processing, a confirmation output for prompting the user for the necessity confirmation input is performed through the display unit 10 or the voice synthesis unit 7g and the speaker 13 under the control of the central control unit 7a. In the input receiving process of the necessary information, when an input indicating the necessity of registration is performed by the user via the operation unit 1 in response to the confirmation output, the input content is provided to the voice quality information management unit 7e, Acceptance of sender information for specifying a sender to be associated with the voice quality information is performed through the operation unit 1. The input sender information is sent from the central control unit 7a to the voice quality information management unit 7e.
Given to.

In step S9, the voice quality information management section 7e determines whether or not a registration command for instructing registration has been input in step S8. If the registration command has been input, the process proceeds to step S10, while the registration command has been input. If is not input, the process ends without performing registration.

In step S10, the voice quality information management section 7e determines whether or not the voice quality information is newly registered. If the voice quality information is newly registered, the process proceeds to step S11, where the conversion codebook is newly created and registered. On the other hand, if it is not a new registration, the process proceeds to step S7, where the existing codebook for conversion is optimized.

It should be noted that whether or not a new registration is made in step S10 is determined by determining that the sender indicated by the sender information input in step S8 is registered in the sender database 9a and that the conversion corresponding to the sender is to be performed. It is determined based on whether or not the sender code book is already registered in the voice quality database 9b. If the sender indicated by the sender information does not exist in the sender database 9a, or if the sender exists, the conversion code book corresponding to the sender is not determined. If it does not exist in the voice quality database 9b, it is determined to be a new registration.

As a result, the voice information (or voice quality information) of each sender of the electronic mail and the voice information (or voice quality information) of another specific person such as an entertainer are obtained.
Voice quality information indicating the voice characteristics of the voice information extracted from the obtained voice information can be easily stored and used as voice quality information for reading out the e-mail sent from each sender. Has become.

FIG. 5 is a flowchart showing the contents of processing for creating a voice quality for reading out an e-mail. The operation mode of this reading out device is a first sub-mode of a reading out mode for reading out an e-mail. Shows the processing contents when it is set to. In the first sub-mode, the electronic mail is read aloud in a voice having a voice quality associated with the sender of the electronic mail.

In step S21, when a reading command and mail specifying information for specifying an electronic mail to be read are input via the operation unit 1, in response to the input,
The central control unit 7a issues a mail reading instruction to the mail management unit 7d and the voice synthesis unit 7g. Here, the read-out command given to the mail management unit 7d is also provided with mail specifying information for specifying an electronic mail to be read out. The reading instruction given to the voice synthesizing unit 7g is given an instruction indicating that reading should be performed in the first sub-mode. This step S21
Then, in response to the reading instruction, the e-mail specified by the mail specifying information given to the reading instruction is sent to the mail database 9c by the mail managing unit 7d.
And sent out to the speech synthesizer 7g.

The following steps S22 to S30 are executed mainly by the voice synthesizer 7g.

In the following step S22, the mail management unit 7
header information of the e-mail given from d (for example,
The “from:” part is analyzed, and information (sender name, mail address, etc.) for identifying the sender of the electronic mail included in the header information is extracted.

In the following step S23, step S22
It is determined whether the address information of the sender corresponding to the information extracted from the header information exists in the sender database 9a, and if so, the sender database 9a
The corresponding sender is specified and the process proceeds to step S24. If the sender does not exist, the process proceeds to step S30.

In step S24, it is determined whether or not voice quality information (here, a conversion codebook) associated with the sender specified in step S23 exists in voice quality database 9b, and the address information of the specified sender is determined. The determination is made based on the presence or absence of the voice quality information identification ID, and if the corresponding voice quality information exists, the process proceeds to step S25. If the corresponding voice quality information does not exist, the process proceeds to step S27.

In step S25, a conversion codebook of voice quality associated with the sender of the electronic mail to be read out is read out from the voice quality database 9b based on the voice quality information identification ID in the sender database 9a. Based on the conversion codebook and the standard parameters of each phoneme stored in the voice quality database 9b, voice quality parameters indicating the voice quality characteristics of each phoneme associated with the sender are created as described above. ,
Proceed to step S26.

On the other hand, in step S30, standard voice parameters (standard male or female voice parameters) specified in advance by the user via the operation unit 1 and the central control unit 7a are read from the voice database 99b. Is issued and the process proceeds to step S26.

In step S27, it is determined whether or not the operation mode of the voice synthesizing unit 27 is set to a gender automatic discrimination mode for automatically discriminating the gender of the sender. While the process proceeds to S28, if not set, the process proceeds to step S30.
The setting of the operation mode of the voice synthesis unit 27 to the automatic gender discrimination mode and the release thereof are performed by a user input via the operation unit 1 and the central control unit 7a.

In step S28, the information extracted from the header information in step S22 and the sender database 9
Based on the information in a, it is determined whether or not the gender of the sender of the electronic mail to be read can be automatically determined, and if so, the gender is automatically determined, and step S2 is performed.
On the other hand, if it is impossible, the process proceeds to step S30.

In step S29, standard male or female voice parameters corresponding to the result of the automatic gender discrimination in step S28 are read from the voice database 9b, and the flow advances to step S26.

At step S26, steps S25, S
29, based on the voice quality parameters of each phoneme created or read out in S30, a voice for voice-to-voice output of voice quality corresponding to the voice quality parameter is synthesized as described above. The text of the e-mail and the name of the sender are read out and output via the amplifier 11 and the speaker 13.

The operation mode of the reading device is as follows.
When the second sub-mode of the reading mode is set, in response to the input of the reading command, the voice parameters of the standard male or female phonemes specified by the user are stored in the voice quality database. 9b is read by the voice synthesis unit 7g, and the voice for reading out is synthesized by the voice synthesis unit 7g based on the voice quality parameters.
The e-mail to be read aloud by the voice is amplifier 1
1 and a speaker 13.

As described above, according to the present embodiment, arbitrary voice quality information for reading out and outputting is prepared for each sender of the electronic mail, and the voice is associated with the voice quality associated with the sender of the electronic mail. And can automatically read out and output, improving the user's listening feeling and eliminating the need for cumbersome operation for voice quality selection.

Also, by preparing voice quality information indicating characteristics of the voice quality of each sender as voice quality information, an e-mail can be read out and output in the voice quality of the sender, and the listening feeling can be further improved. it can.

Further, since a function of taking in voice information and creating voice quality information based on the voice information and a function of taking in voice quality information are provided, the voice quality of the sender can be changed without changing hardware or software. Information and voice quality information associated with the sender can be obtained.

Further, by recording human voice and voice quality information on a recording medium 5a such as a CD-ROM, or by transmitting voice information or voice quality information through a communication network such as the Internet or a communication line such as a telephone line. Incorporation can be easily performed.

Further, every time voice information or voice quality information corresponding to the sender is obtained, the voice quality information corresponding to the sender can be further optimized. For example, a voice closer to the voice quality of the sender is synthesized and read out and output. It can be performed.

[0073]

According to the first aspect of the present invention, the second storage means stores a voice quality database of voice quality information indicating voice characteristics of each sender registered in advance. Based on the voice quality information corresponding to the sender determined by the sender determination means in the voice quality information included in the voice quality database of the storage means, a voice for reading out the given e-mail with the voice quality corresponding to the voice quality information is synthesized. Is output through the speaker, and any voice quality information for reading out and outputting is prepared for each sender of the e-mail, and the e-mail is automatically converted to the voice quality corresponding to the sender. It is possible to read out aloud and to improve the listening feeling of the user, and it is not necessary to perform a troublesome operation for voice quality selection.

Also, by preparing voice quality information indicating the characteristics of the voice quality of each sender as voice quality information, an e-mail can be read out and output in the voice quality of the sender, thereby further improving the listening feeling. Can be.

According to the second aspect of the present invention, since voice information is taken in and voice quality information is created based on the voice information, the sender's identity can be changed without changing hardware or software. Voice quality information and voice quality information associated with the sender can be acquired.

According to the third aspect of the present invention, audio information for feature extraction is captured from a portable recording medium or via a communication line.
-Voice information can be easily captured by recording human voice in a recording medium such as a ROM or through a communication network such as the Internet or a telephone line.

According to the fourth aspect of the present invention, each time voice information corresponding to a sender is acquired, voice quality information corresponding to the sender can be further optimized.
A voice closer to the voice quality of the sender can be synthesized and read out.

According to the fifth aspect of the present invention, since voice synthesis is performed by taking in voice quality information, the voice quality information of the sender is associated with the sender without changing hardware or software. Voice quality information can be obtained.

According to the sixth aspect of the present invention, since the voice quality information is taken in from a portable recording medium or through communication, the voice quality information of the person is recorded on a recording medium such as a CD-ROM. The recording of the voice quality information can be easily performed by recording or through a communication network such as the Internet or a communication line such as a telephone line.

According to the seventh aspect of the present invention, each time voice quality information corresponding to a sender is obtained, the voice quality information corresponding to the sender can be further optimized.
A voice closer to the voice quality of the sender can be synthesized and read out.

[Brief description of the drawings]

FIG. 1 is a block diagram of an electronic mail reading device according to an embodiment of the present invention.

FIG. 2 is a flowchart showing processing contents for creating a voice quality database and the like.

FIG. 3 is a diagram comparing spectrograms of actually reproduced voice quality and true voice quality to be reproduced.

FIG. 4 is a diagram comparing spectrograms of actually reproduced voice quality and a true voice quality to be reproduced.

FIG. 5 is a flowchart illustrating processing for creating voice quality for reading out an e-mail.

[Explanation of symbols]

 Reference Signs List 1 operation unit 3 communication unit 5 reading unit 5a recording medium 7 information processing unit 7a central control unit 7b mail transmission / reception unit 7c information acquisition unit 7d mail management unit 7e voice quality information management unit 7f display processing unit 7g voice synthesis unit 9 storage unit 9a Sender database 9b Voice quality database 9c Email database 10 Display unit 13 Speaker

 ────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Katsumi Nishitani 1-7-10 Kikuzumi, Minami-ku, Nagoya-shi, Aichi F-term in Auto Network Engineering Laboratory Co., Ltd. 5D045 AA07 AB01 AB26 5K030 GA18 HA06 HB01 JT01 KA07 KA20 LB16

Claims (7)

[Claims] 1. An electronic mail reading device for reading an electronic mail by voice and outputting the read electronic mail, wherein: a speaker; a first storage unit for storing a sender database in which the sender of the electronic mail is registered; A second storage means for storing a voice quality database in which voice quality information associated with each of the senders included in the e-mail is analyzed, and the header information of the given e-mail is analyzed, and the e-mail is stored in the sender database. A sender discriminating unit for discriminating which of the senders is included, and the sender discriminated by the sender discriminating unit of the voice quality information included in the voice quality database included in the second storage unit. Based on the voice quality information corresponding to
A voice synthesizing means for synthesizing a voice for reading out the given e-mail with a voice quality corresponding to the voice quality information, and outputting the voice via the speaker. 2. The electronic mail reading device according to claim 1, wherein the first information capturing unit captures voice information and sender information indicating the sender corresponding to the voice information, and the first information capturing unit captures the voice information. Based on the voice information, extract voice quality features of the voice information,
A feature extraction unit that creates the voice quality information indicating the feature of the voice quality, the voice quality information created by the feature extraction unit, and the voice information specified by the sender information and used to create the voice quality information. The e-mail reading apparatus according to claim 1, further comprising: first database creating means for registering the corresponding senders in the voice quality database and the sender database in association with each other. 3. The apparatus according to claim 2, wherein the first information capturing unit captures the audio information from a portable recording medium or via a communication line.
The e-mail reading device according to the item. 4. The method according to claim 1, wherein, when the voice quality information is created, the existing voice quality information corresponding to the sender corresponding to the voice information captured by the first information capturing means is generated. If it already exists in the voice quality database, based on the voice quality characteristics indicated by the existing voice quality information and the characteristics newly extracted from the voice information captured by the first information capturing unit, 3. A new voice quality information which further reflects voice characteristics corresponding to the sender.
Or the electronic mail reading device according to 3. 5. The electronic mail reading device according to claim 2, wherein the voice quality information and second information capturing means for capturing sender information indicating the sender corresponding to the voice quality information, and wherein the second information capturing means captures the voice information. The voice quality information, and a second database creation unit that is specified by the sender information and registers the sender corresponding to the voice quality information in the voice quality database and the sender database in association with each other. The electronic mail reading device according to any one of claims 1 to 4, wherein: 6. The apparatus according to claim 5, wherein the second information capturing means captures the voice quality information from a portable recording medium or via a communication line.
An e-mail reading device according to the item. 7. The second database creating means, when registering the fetched voice quality information in the voice quality database, the existing voice quality information corresponding to the sender corresponding to the voice quality information already has the voice quality information. If it exists in the database, it corresponds to the sender based on the voice quality characteristics indicated by the existing voice quality information and the voice quality characteristics indicated by the voice quality information newly captured by the second information capturing means. 7. The electronic mail reading device according to claim 5, further comprising a function of newly creating voice quality information that further reflects characteristics of the voice quality to be performed and registering the voice quality information in the voice quality database.