JPH10260815A - Voice synthesizing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a voice synthesizing method that makes it possible to grasp the attribute and display form of text from the speech mode of synthetic voice and the difference of a voice quality. SOLUTION: In this method, a voice synthesizing part 12 generates synthetic voice from synthetic parameters 102 to 105 that are generated by a synthetic parameter generating part 11 by using a characteristic parameter 200 by using the analysis result 101 of a text 10 of a text analyzing part 10. In such cases, the speech form and voice quality of a synthetic voice 106 are controlled through a speech mode parameter generating part 42 and a voice guality conversion parameter generating part 43 according to a text attribute analysis data 401, which represents the kind, content and generating source of a text that is acquired by a text attribute analyzing part 40 and a display form analysis data 402, which represents the size, kind, position, change, etc., of a window that shows the text, which is acquired by a display form analyzing part 41.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a text-to-speech synthesis method for generating text-to-speech from text.

[0002]

2. Description of the Related Art Text-to-speech synthesis is a technique for automatically generating synthesized speech from an input text.
As shown in FIG. 3, the text analysis unit 10 analyzes the text 100, a synthesis parameter generation unit 11 that generates synthesis parameters, and a speech synthesis unit 12 that generates a synthesized speech 106. Each component basically performs the following processing.

The input text 100 is first subjected to morphological analysis and syntax analysis in the text analysis unit 10. Next, the synthesis parameter generation unit 11 uses the text analysis data 101 to generate a phoneme symbol string 102, a phoneme duration 103, a pitch pattern 104, and a power 1.
A synthesis parameter such as 05 is generated. Then, in the speech synthesizing unit 12, the characteristic parameters 200 of the basic small units (referred to as speech synthesizing units) such as syllables, phonemes, and one-pitch sections stored in the synthesis unit set storage unit 20 are stored in the phoneme symbol string 102 The synthesized speech 106 is generated by selecting according to the information such as the duration 103 and the pitch pattern 104 and connecting them after controlling the pitch and the duration.

[0004] The synthesized speech 106 thus generated
Is determined mainly by the feature parameter 200 which is a unit of speech synthesis. Therefore, when it is necessary to synthesize a plurality of voices of different voice qualities in voice synthesis, as shown in FIG. 23, a plurality of synthesis unit set storage units 20 to 2n corresponding to the respective voice qualities are prepared. Is switched by using the voice quality conversion parameter from the voice quality conversion parameter input unit 31 and the feature parameter 200 is switched to generate the synthesized voice 106 having a different voice quality.

Further, as shown in FIG. 23, by controlling the synthesis parameter generation unit 11 with the utterance mode conversion parameter 300 from the utterance mode conversion parameter input unit 30, it is possible to generate the synthesized speech 106 having a different utterance mode. Can also. For example, by using pitch average value information, inflection magnitude information, and voice loudness information as speech style conversion parameters, changing the average value of the pitch pattern, inflection strength, and voice loudness in the synthesized speech. Can be.

On the other hand, as a method of generating synthesized voices of various voice qualities without preparing a plurality of synthesis unit set storage units, for example, there is a method of filtering the synthesized voice.
According to this method, for example, as shown in FIG. 24, the filter unit 13 is arranged on the output side of the speech synthesis unit 12, and the low frequency range and the high frequency range of the synthesized voice are emphasized by the filter unit 13. Can be changed.

In such a conventional speech synthesis method, in order to determine the speech style and voice quality of the synthesized speech, the user himself / herself receives the speech style parameters and speech quality parameters via the speech style parameter input unit 30 and the voice quality conversion parameter input unit 31. An operation for giving a voice quality conversion parameter is required.

[0008]

As described above, in the conventional speech synthesis method, the user himself / herself determines the speech style and voice quality of the synthesized speech by giving the speech style conversion parameter and the voice quality conversion parameter. Therefore, since the speech style and voice quality of the synthesized speech simply reflect the user's own sensibility and taste, it is necessary to grasp items such as differences in text attributes and text display forms from the speech style and voice quality. Can not.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a speech synthesizing method capable of recognizing differences in text attributes and display forms from differences in speech style and voice quality of synthesized speech.

[0010]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, the present invention relates to a speech synthesis method for generating a synthesized speech from a text. And controlling at least one of voice quality.

Here, as the attribute of the text, for example, at least one of the font, type, content, importance, and source of the text is used.

As a text display mode, for example, at least one of a type, a size, a position, and a change of a window for displaying the text is used.

The speech style of the synthesized speech refers to, for example, the volume, speech speed, average pitch, intonation and phoneme duration of the synthesized speech.

Further, the present invention is characterized in that the attribute of the text is obtained by analyzing the text.

As described above, according to the present invention, since the synthesized speech whose speech style and voice quality are controlled in accordance with the attributes and display forms of the text is generated, differences in the attributes and display forms of the text are determined from the speech style and voice properties of the synthesized speech. It can be grasped, and the added value of text-to-speech synthesis is improved.

In the method of attaching a parameter representing the speech style or voice quality of the synthesized speech to the text, or a synthetic parameter varied according to the speech style or voice quality,
There is a problem that the additional information attached to the text increases. However, if the text is analyzed at the time of speech synthesis to obtain the attribute of the text or the display form of the text is analyzed and obtained, such additional information is obtained. There is no need to attach information. Furthermore, even when additional information is added to text,
Since only the attribute information of the text and the information indicating the display mode are required as the additional information, the data amount of the additional information is small as compared with the method of adding the synthesis parameter and the like.

[0017]

Embodiments of the present invention will be described below with reference to the drawings. (First Embodiment) FIG. 1 shows the configuration of a speech synthesizer according to a first embodiment to which the speech synthesis method of the present invention is applied. This speech synthesis device includes a text analysis unit 10 for analyzing a text 100, a synthesis parameter generation unit 11 for generating synthesis parameters, and a speech synthesis unit 12 for synthesizing speech.
In the conventional speech synthesizer shown in FIG. 22 having a configuration having
0, display form analysis unit 4 for analyzing the display form of text
1. Speech style conversion parameter generation unit 42 for generating speech style conversion parameter 300 and voice quality conversion parameter 3
01 is added.

The basic operation of this speech synthesizer is the same as that of the conventional device. That is, the input text 100
First, the text analysis unit 10 performs morphological analysis and syntax analysis. Next, the synthesis parameter generation unit 11 generates synthesis parameters such as a phoneme symbol string 102, a phoneme duration time 102, a pitch pattern 104, and a power 105 using the text analysis data 101. Finally, the syllable, phoneme, and 1
After a characteristic parameter 200 of a basic small unit (speech synthesis unit) such as a pitch section is selected according to information such as a phoneme symbol string 102, a phoneme duration 103 and a pitch pattern 104, and the pitch and duration are controlled. By connecting these, a synthesized speech 106 is generated.

Next, a characteristic configuration of the present embodiment will be described. First, the text attribute analysis unit 40 analyzes the attributes of the text, such as the type, content, and source of the text, from the format of the text 100 and the information in the header portion of the text, and obtains information 401 relating to them.
(Hereinafter referred to as text attribute analysis data). The importance of the text can also be determined from information such as the type, content, and source of the text.

On the other hand, the display form analyzing section 41 analyzes the display form of the text, thereby obtaining information 402 (hereinafter referred to as display form analysis data) such as the size, type, position and change of the window for displaying the text. obtain.

Next, an utterance style conversion parameter generation unit 42
And the voice quality conversion parameter generation unit 43, the text attribute analysis data 401 and the display form analysis data 402 prepared in advance, and the speech style conversion parameter 300
Each conversion parameter is generated with reference to a correspondence table (referred to as a conversion parameter generation table) with the voice quality conversion parameter 301. After that, the synthesis parameter generation unit 11
In, a synthesis parameter is generated according to the utterance style conversion parameter 300. Then, a plurality of types of the synthesis unit set storage units 20 to 2n are switched, and voices having different utterance styles and voice qualities are synthesized by synthesizing voices according to synthesis parameters.

As described above, according to the present embodiment, it is possible to grasp the difference in the attribute and display form of the text from the difference in the speech style and voice quality of the synthesized speech.

When generating the voice conversion parameter 301, the text attribute is analyzed in advance, and the text attribute analysis data 401 as the analysis result is attached to the text, so that the text attribute analysis unit 40 can be omitted. It is possible. Next, a specific operation example of the speech synthesis device according to the present embodiment will be described.

[Operation Example 1] First, with reference to FIGS. 2 to 4, a case will be described in which a plurality of windows are opened on a computer and a message from the computer displayed in the window is synthesized by voice. Text attribute analysis unit 4 in 1
0, the operation of the display form analysis unit 41, the utterance style conversion parameter generation unit 42, and the voice quality conversion parameter generation unit 43 will be described.

First, in the text attribute analyzer 40,
The type and contents of the message generated in the window are identified by referring to a message table prepared in advance. on the other hand,
The display form analyzing unit 41 analyzes the form of the window displaying the message, and obtains information on the type and size of the window.

Next, the utterance style conversion parameter generation unit 42
The voice conversion parameter generation unit 43 refers to the conversion parameter generation table prepared in advance, and generates the speech style conversion parameter 300 and the voice conversion parameter 301 according to the information of the message content, the type of the window, and the size of the window. Later, a synthesized speech is generated according to each of the generated conversion parameters.

For example, consider a case where an error message "Error Message 0" is displayed on an application "kterm" started by a computer "hikari" with three windows opened on the computer as shown in FIG. . In this case, first, in the text attribute analysis unit 40,
The message displayed in the window is identified as an error message by referring to the message table. In the display form analysis unit 41, the window displaying the message is displayed in the window “k” generated by the application “kterm” of the computer “hikari”.
term @ hikari ”, and its size is“ 10
0 ".

Next, an utterance style conversion parameter generation unit 42
The voice quality conversion parameter generation unit 43 refers to the conversion parameter generation table shown in FIG. 4 prepared in advance, and converts the speech style conversion parameter 300 and the voice quality conversion parameter 301 according to the type of the message, the type of the window, and the size of the window. Generated.

Here, in the case of FIG. 2, the size is "100".
Message "Error in window" kterm @ hikari "
Since "Message 0" is displayed, the volume is "10" and the utterance speed is "2" according to the conversion parameter generation table shown in FIG.
After the conversion parameters of “0”, the average pitch “30”, and the synthesis unit set storage unit number “0” are generated, synthesized speech is generated according to these parameters.

Similarly, in the case of FIG. 3, although no error message is displayed in the application “kterm”, a window “xterm @ nozomi” having a size of “50” is displayed.
Is displayed, a warning message "Warning Message 1" is displayed in accordance with the conversion parameter generation table shown in FIG.
Volume "5", utterance speed "10", average pitch "20",
After the conversion parameters of the synthesis unit set storage unit number “8” are generated, a synthesized speech is generated according to these parameters.

As described above, in FIG. 2 and FIG. 3, the speech style and voice quality of the synthesized speech are different, and the difference in the type of message, the type of window, and the size of the window can be easily grasped from these differences. can do.

[Operation Example 2] Next, referring to FIGS. 5 to 8, a case will be described where a text transmitted from a transmission side of a communication system is received by a reception side, and the received text is subjected to speech synthesis on the reception side. For example, the operations of the text attribute analysis unit 40, the utterance style conversion parameter generation unit 42, and the voice quality conversion parameter generation unit 43 in FIG. 1 will be described.

First, in the text attribute analysis unit 40,
After the type of the text is identified from the format of the input text 100, the attached text information such as the sender and the subject is identified with reference to the header portion.

On the other hand, the display form analysis unit 41 analyzes the type and size of the window for displaying the text. Here, consider a case where a text as shown in FIG. 5 is received. FIG. 6 shows a display screen on the receiving side and a state of speech synthesis in this case.

The text attribute analyzer 40 refers to a correspondence table between the text format and the text type prepared as shown in FIG. 7 prepared in advance, whereby the text type in FIG. 5 is identified as mail. You. further,
By analyzing the information in the header portion, it is identified that the sender is A and the subject is S. On the other hand, the display mode analysis unit 41 analyzes the type and size of the window for displaying the text, and obtains display mode analysis data 402. 8 and 9 show specific examples of format 1 and format 2 in the correspondence table of FIG.

Next, the utterance style conversion parameter generation unit 42
The speech quality conversion parameter generation unit 43 refers to a conversion parameter generation table prepared in advance, and generates the speech style conversion parameter 300 and the voice quality conversion parameter 301 according to the information of the message content, the type of the window, and the size of the window. After that, a synthesized speech is generated according to each of these conversion parameters. Therefore, the difference in the type of text, the sender, the subject, the type of window, and the size of the window can be easily grasped from the difference in the speech style and voice quality of the synthesized speech.

The type of text need not be limited to mail, and various types such as news, magazines, and books can be considered. Further, the text attached information added to the header portion does not need to be limited to the sender and the subject, and may be other information. Further, by adding the type of text to the header section as text-attached information, a change can be made so as to identify the type of text from the text-attached information. Further, the elements associated with the voice quality conversion parameters are not limited to the type of text, sender, subject, and window type, but may be any elements as long as they are text attribute analysis data and display form analysis data.

Further, the text analysis section 10 in the present embodiment performs a process of extracting a reading area from the input text and analyzing text information of the extracted area.

[Operation Example 3] Next, referring to FIGS. 10 to 11, taking as an example a case where a text composed of a plurality of segments is subjected to speech synthesis, the text attribute analysis unit 4 in FIG.
0, the operation of the speech style conversion parameter generation unit 42 and the voice quality conversion parameter generation unit 43 will be described.

First, in the text attribute analysis unit 40,
It is determined which segment the input text belongs to. In addition, the display mode analysis unit 41 analyzes the type and size of the window for displaying the text. Here, consider the case where the text shown in FIG. 10 is synthesized by speech. FIG. 11 shows a display screen and speech synthesis in this case. In this case, first, the text attribute analysis unit 40 identifies which segment of the input text belongs to the name, the home telephone number, and the office telephone number. The display mode is analyzed by the display mode analysis unit 41.

Next, the utterance style conversion parameter generation unit 42
And voice conversion parameter generation unit 43, referring to the conversion parameter generation table, according to the type of segment, the type of window, and the size of window, utterance style conversion parameter 300 and voice conversion parameter 3
After the 01 is generated, a synthesized speech is generated according to each of these conversion parameters.

Accordingly, the window and segment (in the above example, any of the name, the home telephone number, and the office telephone number) in which the reading is performed can be easily grasped from the speech style and voice quality of the synthesized voice. Become.

(Second Embodiment) FIG. 12 shows the configuration of a speech synthesis apparatus according to a second embodiment to which the speech synthesis method of the present invention is applied. 12, elements denoted by the same reference numerals as those in FIG. 1 have the same functions as those in FIG.

In this embodiment, a synthesis unit conversion unit 32 for converting a synthesis unit is provided, and a synthesis unit is converted according to a voice quality conversion parameter 301. The speech synthesis unit 12 generates a synthesized speech 106 using the converted synthesis unit. The processing of the synthesis unit conversion unit 32 can be realized by, for example, mapping of a frequency and an amplitude to a spectrum parameter. In this case, the voice conversion parameter generation unit 43 outputs a mapping function of frequency and amplitude as the voice conversion parameter 301. If f is a frequency, the power spectrum s in the feature parameter 200
The conversion from (f) to the power spectrum s ′ (f) in the feature parameter 200 is represented by the following equation.

S ′ (f) = m (s (ω (f))) (1) where ω (f) is a function representing frequency mapping, and m () is a function representing amplitude mapping. .

For example, s (f) and s' (f) are shown in FIG.
It is assumed that each is represented in FIG. Where F ₁ ,
F ₂ , F ₃ ′, F ₄ ′ are formant frequencies,
P ₁ , P ₂ , P ₃ ′, P ₄ ′ are the corresponding formant powers. In this case, the mapping functions ω (f) of the frequencies corresponding to the conversion parameters are F ₁ and F ₁ ′, respectively.
And a function that maps F ₂ to F ₂ ′,
The amplitude mapping function m () is a function that maps P ₁ to P ₁ ′ and P ₂ to P ₂ , respectively. Examples of these functions ω (f) and m () are shown in FIG.
5 and FIG.

In this embodiment, the synthesis unit converter 32
In the conversion of the feature parameter 200, for example, it is also possible to obtain a conversion parameter corresponding to all the synthesis units and perform a different conversion for each feature parameter, and classify the synthesis unit into several classes. It is also possible to perform conversion by obtaining a common conversion parameter for one class. For example, a plurality of power spectra si (f) are converted to corresponding power spectra Si '.
(F), a common mapping function ω (), m () to be converted into (i = 1, 2,..., N) is a conversion error function E expressed by the following equation using a known optimization method. It is obtained by minimizing (ω, m).

[0048]

(Equation 1)

Here, fs is a sampling frequency. Since the voice quality information is mainly included in the vowel, the voice quality may be converted by performing the conversion of the characteristic parameter 200 only on the synthesis unit including the vowel and using the other synthesis units without conversion. It is possible. By doing so, it is possible to reduce the amount of processing in the synthesis unit conversion unit 32. Also, the synthesis unit conversion unit 32
It is also possible to perform the following conversion in

S ′ (f) = g (f) s (ω (f)) (3) This is an example in which the conversion of the amplitude is realized by the product of the function g (f).

In this embodiment, one composite unit set 2
Since the voice quality of the synthesized voice 106 is changed by processing 0, it is not necessary to store a plurality of synthesis unit sets, so that the synthesized voice 106 of a plurality of types of voice quality can be obtained with a smaller memory capacity. Can be

(Third Embodiment) FIG. 17 shows the configuration of a speech synthesis apparatus according to a third embodiment to which the speech synthesis method of the present invention is applied. 17, elements denoted by the same reference numerals as those in FIG. 1 have the same functions as those in FIG.

In this embodiment, a filter unit 13 for filtering the synthesized speech 106 is provided,
The first embodiment is different from the first embodiment in that the filter unit 13 is controlled by the control unit 01. By controlling the filter unit 13 to emphasize the low and high frequencies of the synthesized speech 106, the voice quality of the synthesized speech 106 can be changed. Further, in the present embodiment, since the voice quality of the synthesized voice is changed by using a filter, it is not necessary to store a plurality of synthesis unit sets. Voice 106 is obtained.

(Fourth Embodiment) FIG. 18 shows the configuration of a speech synthesis apparatus according to a fourth embodiment to which the speech synthesis method of the present invention is applied. 18, elements denoted by the same reference numerals as those in FIG. 1 have the same functions as those in FIG.

This embodiment is an example in which the utterance style conversion parameter generation unit 42 in FIG. 1 in the first embodiment is omitted and only voice conversion is performed. As in the first embodiment, voice conversion is performed using voice conversion parameters according to text attribute and display mode information.

(Fifth Embodiment) FIG. 19 shows the configuration of a speech synthesis apparatus according to a fifth embodiment to which the speech synthesis method of the present invention is applied. 19, elements denoted by the same reference numerals as those in FIG. 12 have the same functions as those in FIG.

In the present embodiment, the utterance style conversion parameter generation unit 42 in FIG. 12 in the second embodiment is omitted,
This is an example in which only voice conversion is performed. The point that voice conversion is performed using voice conversion parameters in accordance with text attribute and display mode information is the same as in the second embodiment.

(Sixth Embodiment) FIG. 20 shows the configuration of a speech synthesis apparatus according to a sixth embodiment to which the speech synthesis method of the present invention is applied. 20, elements denoted by the same reference numerals as those in FIG. 17 have the same functions as those in FIG.

In the present embodiment, the utterance style conversion parameter generation unit 42 in FIG. 17 in the third embodiment is omitted.
This is an example in which only voice conversion is performed. As in the third embodiment, voice conversion is performed using voice conversion parameters according to text attribute and display mode information.

(Seventh Embodiment) FIG. 21 shows the configuration of a speech synthesis apparatus according to a seventh embodiment to which the speech synthesis method of the present invention is applied. In FIG. 21, elements denoted by the same reference numerals as those in FIG. 17 have the same functions as those in FIG.

This embodiment is an example in which the voice quality conversion parameter generation unit 43 in FIG. 17 in the third embodiment is omitted, and only the speech style conversion is performed. The conversion of the utterance style by the utterance style conversion parameter according to the information on the attribute of the text and the display form is the same as in the third embodiment.

Although some embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments, and can be implemented with various modifications.

For example, in each of the above embodiments, the data for controlling the utterance style conversion parameter generation section 42 and the voice quality conversion parameter generation section 43 is one of the text attribute analysis data and the text display form analysis data. May be used.

Further, in each of the above embodiments, the text attribute analysis data and the text display form analysis data are not limited to the elements described in the embodiments.
Other elements may be used.

[0065]

As described above in detail, according to the present invention, by generating a synthesized speech in which the speech style and voice quality are controlled according to the attributes and display forms of the text, the text style is converted from the speech style and voice quality of the synthesized speech. It is possible to grasp the attributes, that is, the font, type, content, importance, source, and the like of the text, and the difference in the display form of the text, for example, the type, size, position, and change of the window that displays the text.

Therefore, according to the present invention, in addition to the conventional simple text-to-speech function, a new function of knowing the attributes of the text read out from the synthesized speech and the display form on the computer is added. There is an advantage that the added value of the synthesis is improved.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a speech synthesizer according to a first embodiment of the present invention.

FIG. 2 is a diagram showing an embodiment in which a message generated in a window is synthesized by voice;

FIG. 3 is a diagram showing another mode in which a message generated in a window is synthesized by voice;

FIG. 4 is a diagram showing a conversion parameter generation table;

FIG. 5 is a diagram showing an example of an input text.

FIG. 6 is a diagram showing an embodiment in which a received text is subjected to speech synthesis;

FIG. 7 is a diagram showing a correspondence table between formats and text types;

FIG. 8 is a diagram showing an example of format 1 in FIG. 7;

9 is a diagram showing an example of format 2 in FIG.

FIG. 10 is a diagram showing another example of the input text.

FIG. 11 is a diagram showing an embodiment in which text composed of a plurality of segments is subjected to speech synthesis.

FIG. 12 is a block diagram showing a configuration of a speech synthesizer according to a second embodiment of the present invention.

FIG. 13 is a view showing an example of a power spectrum in a feature parameter according to the embodiment.

FIG. 14 is a diagram showing an example of a power spectrum obtained by converting a power spectrum in a feature parameter according to the embodiment;

FIG. 15 is a view showing an example of a frequency mapping function corresponding to a conversion parameter according to the embodiment;

FIG. 16 is a diagram showing an example of a function of mapping an amplitude corresponding to a conversion parameter in the embodiment.

FIG. 17 is a block diagram showing a configuration of a speech synthesizer according to a third embodiment of the present invention.

FIG. 18 is a block diagram showing a configuration of a speech synthesizer according to a fourth embodiment of the present invention.

FIG. 19 is a block diagram showing a configuration of a speech synthesizer according to a fifth embodiment of the present invention.

FIG. 20 is a block diagram showing a configuration of a speech synthesizer according to a sixth embodiment of the present invention.

FIG. 21 is a block diagram showing a configuration of a speech synthesizer according to a seventh embodiment of the present invention.

FIG. 22 is a block diagram showing a basic configuration of a speech synthesizer for text speech synthesis.

FIG. 23 is a block diagram showing the configuration of a conventional speech synthesizer in which the voice quality and speech style of synthesized speech are variable.

FIG. 24 is a block diagram showing the configuration of another conventional speech synthesizer in which the voice quality of the synthesized speech is made variable by applying a filter to the synthesized speech.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 ... Text analysis part 11 ... Synthesis parameter generation part 12 ... Speech synthesis part 13 ... Filter part 20-2n ... Synthesis unit set storage part 30 ... Speech style conversion parameter input part 31 ... Voice quality conversion parameter input part 32 ... Synthesis unit conversion Unit 40: Text attribute analysis unit 41: Display form analysis unit 42: Utterance style conversion parameter generation unit 43: Voice quality conversion parameter generation unit 100: Text 101: Text analysis data 102: Phoneme symbol string 103: Phoneme duration 104: Pitch Pattern 105 Power 106 Synthesized speech 200 Feature parameter 300 Speech style conversion parameter 301 Voice conversion parameter 400 Display mode 401 Text attribute analysis data 402 Display mode analysis data

 ──────────────────────────────────────────────────続 き Continuing from the front page (72) Inventor Masahiro Oshikiri 1 Toshiba R & D Center, Komukai-shi, Kawasaki-shi, Kanagawa Prefecture (72) Inventor Emperor Emperor Emperor Toshiba Komukai, Sai-ku, Kawasaki-shi, Kanagawa No. 1 Town, Toshiba R & D Center (72) Inventor Takehiko Kagoshima No. 1, Komukai Toshiba-cho, Saiwai-ku, Kawasaki-shi, Kanagawa, Japan Toshiba R & D Center

Claims (5)

[Claims] 1. A speech synthesis method for generating a synthesized speech from a text, wherein at least one of a speech style and a voice quality of the synthesized speech is controlled according to at least one of information on an attribute and a display form of the text. Speech synthesis method. 2. The speech synthesis method according to claim 1, wherein at least one of font, type, content, importance, and source of the text is used as the attribute of the text. 3. The speech synthesis method according to claim 1, wherein at least one of a type, a size, a position, and a change of a window for displaying the text is used as the display mode of the text. 4. The speech synthesis according to claim 1, wherein the speech style of the synthesized speech is at least one of a volume, a speech speed, an average pitch, intonation, and a phoneme duration of the synthesized speech. Method. 5. The speech synthesis method according to claim 1, wherein the attribute of the text is obtained by analyzing the text.