JPH04180000A - Speech synthesizer - Google Patents

Abstract

PURPOSE:To synthesize voice as a user intends by controlling the acoustic effect of the synthesized voice at an optional position of an input sentence. CONSTITUTION:A source sentence input means 21 inputs a source sentence given a synthesized sound control parameter and a decomposing means 22 decomposes the source sentence into the control parameter and a sentence which does not contain the control parameter. A voicing symbol generating means 23 processes the sentence containing no control parameters linguistically to generate a corresponding voice and a parameter adding means 24 adds an acoustic effect parameter and the control parameter to a voicing symbol corresponding to a synchronous point. Then a voice synthesizing parameter generating means is constituted. An acoustic effect parameter control member 26 controls the acoustic effect parameter with the key signal from an external input means 25 and the control parameter to generate the voice synthesis parameter. Consequently, the acoustic effect of the synthesized voice can be controlled at the optional position of the sentence.

Description

[Detailed description of the invention] Industrial applications The present invention relates to a speech synthesis device.

2. Description of the Related Art This type of speech synthesis device synthesizes speech by inputting sentences, and is realized, for example, by the processing configuration shown in FIG.

The original text to be voice synthesized is input through an input unit l such as a keyboard. Then, in the abbreviation/symbol/number processing section 2,
Abbreviations, symbols, and numbers in the original text are converted to characters.

Next, the morpheme decomposition processing unit 3 matches the morphemes, which are the smallest meaningful units that make up a word, with a dictionary.
The character string in the sentence is decomposed into morphemes.

Then, the sentence is decomposed into phrases and words by the syntactic analysis unit 4, and the phonetic symbol generation unit 5 generates the phonetic symbol of the word using the morphemes obtained by decomposing the character string and the phonetic symbol thereof in the dictionary. do. At this time, make adjustments so that the connection between morphemes does not look unnatural. Furthermore, for the words that cannot be broken down into morphemes, phonetic symbols are generated character by character.

Next, the sentence with this phonetic symbol is given a fundamental frequency that determines the pause, which is the pause time between words, the utterance duration of each phoneme, and the pitch of the uttered sound, to the sentence by the sound effect parameter adding unit 6. Ru.

Furthermore, a speech synthesis parameter generation section 7 generates speech synthesis parameters using the phonetic symbol, the pause, utterance duration time, and fundamental frequency, and a speech synthesis section 8 synthesizes speech using the speech synthesis parameters.

Problems to be solved by the inventionIn recent years, with the development of computer publishing technology,
Such speech synthesis devices are increasingly being used in many electronic devices.

However, there are cases where it is desired to synthesize and output audio in synchronization with the operations of other devices, such as synthesizing text on multiple screens displayed on a computer display in accordance with the timing of switching between display screens. In this type of usage, conventional speech synthesis devices cannot perform appropriate synthesis due to their principles.

In other words, conventional speech synthesis devices perform speech synthesis only based on the results of sentence analysis, so if you aim to synthesize speech in synchronization with the operations of other devices, There is a problem in that the input of a sentence must be input, and when that sentence is synthesized into speech, an analysis error occurs and the synthesis is incomplete.

In this way, the function that requires a speech synthesizer that operates in synchronization with the operations of various devices is to synthesize speech in synchronization with external input such as user input using a mouse or keyboard. This function requires a device that controls the sound effects of synthesized speech at any position in the sentence to be synthesized using external input.

In view of the above-mentioned circumstances, an object of the present invention is to provide a speech synthesis device that can control the sound effects of synthesized speech at any position of a sentence using external input.

Means for Solving the Problems The present invention synthesizes sentences to which control parameters for controlling synthesized speech are added in synchronization with external inputs such as mouse input. an original text input means for inputting the original text for the text; a decomposition means for disassembling the original text into control parameters and sentences that do not include the control parameters; a phonetic symbol generation means for generating;
a parameter assigning means for assigning a sound effect parameter to the generated phonetic symbol and a control parameter corresponding to the peer point; an external input means for inputting an external input; and an external input, a control parameter, and an acoustic effect parameter. According to the present invention, it is possible to generate sound effect parameters by external input. By using a sentence in which the control parameters to be controlled are written at any position as the original sentence, it is possible to control the acoustic parameters at any position in the original sentence.This allows us to control the acoustic parameters of the synthesized speech at any position in the sentence. It can be controlled and can synthesize speech in synchronization with the operation of various devices.

Example Embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 2 is a block diagram showing an embodiment of the speech synthesis device of the present invention.

The original text input means 21 is a means for inputting an original text in which a control parameter for controlling the synthesized sound is added to the sentence to be synthesized in accordance with synchronization with an external input such as a mouse.

Next, the decomposition means 22 is a means for decomposing the original sentence into the control parameter and a sentence that does not include the control parameter.

The phonetic symbol generating means 23 is a means for linguistically processing a sentence that does not include the control parameters to generate a phonetic symbol corresponding to the sentence.

The parameter providing means 24 is a parameter providing means that provides the sound effect parameter and the control parameter to the pronunciation symbol in correspondence with the synchronization point.

The external input means 25 is means for inputting key signals from a keyboard.

Based on the phonetic symbols, sound effect parameters, control parameters, and external input signals obtained in this way, the sound effect parameter control unit 26 controls the sound effect parameters using the external inputs and the control parameters, and synthesizes speech. The parameter generation unit 27 generates speech synthesis parameters using the phonetic symbols and their controlled sound effect parameters. Note that the sound effect parameter control section 26 and the speech synthesis parameter generation section 27 constitute speech synthesis parameter generation means.

The speech synthesis means 28 is a means for synthesizing speech using the speech synthesis parameters.

Next, as a specific example of the speech synthesis device shown in FIG. 1, the operation of the above embodiment will be explained by taking as an example a case where speech synthesis is performed in synchronization with mouse and keyboard input by a computer user. .

First, write the sentence shown at 31 in Figure 2 ([Mouse C1ic
k] Th1s is a pen, [niey
Press Tt+is is a de
sk) into the original text input means 21. here,
[Mouse C11ck] is a control parameter for mouse input, and [Keypress (
Key Press] represents a control parameter for key input.

Next, the decomposition means 22 decomposes these 31 sentences into 32 sentences and control parameters.

Then, the phonetic symbol generation means 23 converts these 32 sentences into 33
Convert to phonetic symbols.

Then, the parameter assigning means 24 generates 34 sentences in which the 33 phonetic symbols are assigned sound effect parameters such as pause, utterance duration, and fundamental frequency, and the control parameters.

Therefore, when there is an input signal from the outside, if the external input is a mouse input, for example, voice synthesis is started from the position of the control parameter [mouse click (MouseCl 1ck)] for the mouse input, and the next control parameter is written. Synthesizes the sentence up to the point where it is.

In addition, if the external input is a key input, for example, the control parameter for the key input [Key Pre
Speech synthesis starts from the point where "sS)" is written and continues to synthesize speech until the end. In this way, audio can be synthesized in synchronization with external input.

Although the external input of the present invention is a signal from a mouse or a keyboard in the above embodiment, it goes without saying that it may be any external input signal such as a screen switching signal of a display device.

Furthermore, although each means of the present invention is usually realized in software using a computer, it may also be realized using a dedicated hardware circuit having each function.

DETAILED DESCRIPTION OF THE INVENTION According to the present invention, acoustic parameters can be controlled at any position by control parameters included at any position in an input sentence. That is, by controlling the acoustic effects of synthesized speech at any position of the input sentence, it is possible to synthesize speech that matches the user's intention, and this has great practical effects.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an outline of the speech synthesis device of the present invention, FIG. 2 is a flowchart showing the operation of an embodiment of the same speech synthesis device, and FIG. 3 is a block diagram showing an outline of a conventional speech synthesis device. It is. 21... Original text input means, 22... Decomposition means, 23.
... Phonetic symbol generation means, 24 ... Parameter provision means, 25 ... External input means, 26 ... Sound effect parameter control section, 27 ... Speech synthesis parameter generation section, 2
8...Speech synthesis means. Agent Patent Attorney Matsu 1) Tadashi Michi Figure 1

Claims (1)

[Claims] an original text input means for inputting a sentence to be synthesized as an original text for synthesis, in which a control parameter for controlling the synthesized sound is added to the sentence to be synthesized in synchronization with an external input such as a mouse input; , a decomposition unit that decomposes the original text into the control parameter and a sentence that does not include the control parameter, and a phonetic symbol generation unit that performs language processing on the sentence that does not include the control parameter to generate a phonetic symbol corresponding to the sentence. , a parameter assigning means for assigning a sound effect parameter to the generated phonetic symbol and assigning the control parameter corresponding to the synchronization point; an external input means for inputting the external input; A speech synthesis device comprising: a speech synthesis parameter generating means for producing a speech synthesis parameter from a control parameter, the sound effect parameter and the pronunciation symbol; and a speech synthesis means for synthesizing speech using the speech synthesis parameter.