JP3862478B2 - Speech synthesis apparatus and speech synthesis method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide information playback equipment with which listening comprehension of synthetic voice is made possible to those who are in vehicles by controlling the parameter about a synthetic voice according to the situation of vehicles or the circumstance. SOLUTION: The text data which should be uttered is first chosen and is taken in from the utterance contents storing memory 101 by the utterance contents selection part 102 (S201). Based on the taken-in text, a default rhythm parameter is created by the rhythm parameter generation part 103 (S202). The present own vehicle speed (speed per hour) S is acquired from the own vehicle speed measurement part 109 (S203). Based on the value S, the default rhythm parameter is updated (S204-S209). By sticking the phoneme data gained according to the text sentence which the utterance contents selection part 102 chooses from the phoneme piece dictionary 106 on the updated rhythm parameter, a synthetic sound is generated (S210). A synthetic sound is outputted from the voice output unit 108 (S211).

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a speech synthesizer and a speech synthesis method that are mounted on a vehicle and perform speech synthesis based on text.
[0002]
[Prior art]
A conventional car navigation device displays the position of the vehicle on a map displayed on a display in the vehicle. Further, by setting the destination before or during traveling, the route to the destination can be searched, and the route to the destination can be guided to the driver by displaying the route on the display or by voice guidance.
[0003]
The voice guidance of the car navigation apparatus includes a method of reproducing a recording of actual human guidance voice and a method using a regular voice synthesis technique. In addition, as the vehicle approaches the intersection where the vehicle should bend, the route guidance timing is changed in consideration of the speed of the vehicle and the congestion of the road, thereby improving the reliability of route guidance to the driver.
[0004]
[Problems to be solved by the invention]
When regular voice synthesis is used for voice guidance, there is an advantage that it requires less storage space than recording voice, but currently it is inferior to human voices and the speed of the vehicle increases and the surroundings increase. When the noise increases, it becomes more difficult to hear. However, since the conventional synthesized speech generation unit does not create a synthesized waveform in consideration of the surrounding situation of the host vehicle, it cannot generate synthesized speech that matches the surrounding situation.
[0005]
The present invention has been made in view of the above-described problems, and by controlling parameters related to synthesized speech in accordance with the vehicle and surrounding conditions, the speech synthesis result can be heard by a person riding in the vehicle. For the purpose.
[0006]
[Means for Solving the Problems]
In order to achieve the object of the present invention, for example, a speech synthesizer of the present invention comprises the following arrangement. That is, speech synthesis means for generating synthesized speech corresponding to text,
An acquisition means for acquiring road congestion conditions;
When the road congestion situation is the first congestion situation, the speaker of the synthesized speech is changed, and when the road congestion situation is the second congestion situation, the prosody is a parameter related to the prosody of the synthesized speech. And a changing means for changing the parameter.
In order to achieve the object of the present invention, for example, the speech synthesis method of the present invention comprises the following arrangement. That is, a speech synthesis process for generating synthesized speech corresponding to text,
An acquisition process for acquiring road congestion;
A first changing step of changing a speaker of the synthesized speech when the congestion situation of the road is a first congestion situation;
And a second changing step of changing a prosodic parameter that is a parameter related to the prosody of the synthesized speech when the road congestion state is a second congestion state.
[0007]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, the present invention will be described in detail according to preferred embodiments with reference to the accompanying drawings. In the following embodiment, a case where a car navigation device is used as the information reproducing device will be described.
[0008]
[First Embodiment]
FIG. 1 shows a schematic configuration of a car navigation apparatus according to the present embodiment.
[0009]
An utterance content storage memory 101 stores an utterance content text which is a text sentence of the content uttered by the car navigation device. When the car navigation device makes an announcement, the utterance content selection unit 102 refers to the utterance content storage memory 101 and selects the utterance content text corresponding to the content to be uttered. Depending on circumstances, the content of the utterance may be an email to the driver downloaded from an information transmission server to which the car navigation device is connected through radio or a mobile phone, news, and the like.
[0010]
The utterance content text selected by the utterance content selection unit 102 is sent to the prosodic parameter generation unit 103 in the car navigation device.
[0011]
In the prosody parameter generation unit 103, prosody parameters (synthesized speech parameters) such as sound power (voice volume), pitch frequency (voice pitch), and phoneme duration (voice speed) are written from the text in which the utterance content is written. Generate. Originally, these parameters can be generated from the utterance content text to which the position and strength of the accent are added, but in this embodiment, the information described later via the running state information acquisition unit 105 is also included. reflect. The driving state information acquisition unit 105 is connected to the own vehicle speed measurement unit 109, the own vehicle position measurement unit 110, and the congestion state acquisition unit 111 outside the speech synthesizer 107.
[0012]
The prosody parameters created in the prosody parameter generation unit 103 are sent to the phoneme segment connection unit 104, and the phoneme segment connection unit 104 stores the phoneme data as the basis of the regular synthesized sound according to the prosodic parameters. The synthesized speech is generated by connecting the acquired phoneme data.
[0013]
The driving state information acquisition unit 105 also controls the phoneme piece connection unit 104, and may change phoneme data to be used. As the phoneme data changes, the speaker of the synthesized sound changes.
[0014]
When the synthesized speech is created by the speech synthesizer 107, the data is sent to the speech output device 108 such as a speaker and outputted as speech.
[0015]
FIG. 2 is a flowchart when the car navigation apparatus according to this embodiment performs the above-described processing. The program code according to the flowchart shown in the figure is stored in a memory such as a ROM or a RAM (not shown) in the car navigation apparatus according to this embodiment, and is read and executed by the CPU. As a result, the car navigation device of the present embodiment can perform each process described below.
[0016]
Here, as an example, a process for changing a prosodic parameter according to the own vehicle speed and creating a synthesized sound that is uttered louder, clearly and slowly as the own vehicle speed increases will be described. Further, it is assumed that the content to be uttered by the car navigation device is determined before the processing according to this flowchart is executed.
[0017]
First, text data corresponding to the content to be uttered is selected and fetched by the utterance content selection unit 102 from the utterance content storage memory 101 (step S201). A default prosodic parameter is created by the prosodic parameter generating unit 103 based on the captured text (step S202). The processing here is the same as the general rule speech synthesis processing.
[0018]
Next, the current vehicle speed (speed) S is acquired from the vehicle speed measuring unit 109 (step S203). Based on the value of S, the previously determined default prosodic parameter is updated (steps S204 to S209). In the example shown in the figure, as the host vehicle speed S increases, a large and slow synthesized sound is output. In this example, the vehicle speed S is 80 km, 50 km, 20 km, and stepwise control is performed.
Prosodic parameter = α x S + β (α and β are constants)
As described above, the vehicle speed S may be calculated as a variable.
[0019]
Next, the phoneme segment connection unit 104 acquires necessary phoneme data from the phoneme dictionary 106 according to the text sentence selected by the utterance content selection unit 102, and acquires it on the above-mentioned prosodic parameters updated based on the vehicle speed S. The phoneme data thus pasted is pasted to generate synthesized speech (step S210). The generated synthesized speech is output to the speech output device 108, and the synthesized speech is output from the speech output device 108 (step S211).
[0020]
As described above, the car navigation apparatus according to the present embodiment can control the prosodic parameters as the synthesized speech parameters based on the own vehicle speed. As a result, even if the announcement is difficult to hear due to, for example, an increase in the volume of the engine sound due to an increase in the vehicle speed, the content of the announcement can be heard more easily by increasing the volume of the synthesized speech, for example.
[0021]
[Second Embodiment]
In the first embodiment, the own vehicle speed is used as a parameter for controlling the synthesized voice parameter. In this embodiment, the situation around the own vehicle (road congestion situation) is used as a parameter for controlling the prosodic parameter. Will be explained. The configuration of the car navigation device used in the present embodiment is the same as that used in the first embodiment.
[0022]
FIG. 3 is a flowchart of processing performed by the car navigation device according to the present embodiment. The program code according to the flowchart shown in the figure is stored in a memory such as a ROM or a RAM (not shown) in the car navigation apparatus according to this embodiment, and is read and executed by the CPU. As a result, the car navigation device of the present embodiment can perform each process described below.
[0023]
Here, as an example, an example in which the synthesized sound of the male voice becomes a synthesized sound (synthetic sound of the female voice) calmed by the congestion situation of the road by default is shown. Similarly to the first embodiment, it is assumed that the content to be uttered by the car navigation device is determined before the processing according to the flowchart is executed.
[0024]
First, text data corresponding to the content to be uttered is selected and taken in by the utterance content selection unit 102 from the utterance content storage memory 101 (step S301). The default phoneme dictionary is set to male (step S302), and the prosody parameter generation unit 103 creates default prosody parameters using the text captured by the utterance content selection unit 102 (step S303).
[0025]
Next, the driving state information acquisition unit 105 acquires the congestion state of the road on which the current vehicle is measured, which is measured by the congestion state acquisition unit 111 (step S304). If it is determined that the road is congested (step S305), the prosodic parameter generation unit 103 uses the above-mentioned default prosodic parameters so that the default speaker becomes a female voice in order to unravel the driver's feelings. Is updated (step S307). Specifically, the voice is made more feminine by increasing the pitch frequency.
[0026]
In general, the pitch frequency band of male voice is roughly 80 to 160 Hz, and the pitch frequency band of female voice is roughly 120 to 250 Hz. On average, the pitch frequency of female voice is Higher than that of men. Therefore, since the voice data of a male speaker is the default, if the default pitch frequency is set to 120 Hz, for example, in step S307, processing for increasing the pitch frequency to 200 Hz is performed in order to obtain female voice.
[0027]
If it is determined that the road is congested even if it is not traffic jam (step S306), the default prosodic parameters are manipulated (step S308) and updated. In the example of this flowchart, the voice is lowered by lowering the pitch frequency, the announcement is made slowly by lengthening the phoneme length, and the volume is lowered by lowering the sound power. As a result, it has been updated to a relaxed and calm voice.
[0028]
Next, the phoneme segment connection unit 104 acquires the necessary phoneme data from the phoneme dictionary 106 according to the text sentence selected by the utterance content selection unit 102, and pastes the acquired phoneme data on the prosodic parameters updated as described above. Then, synthesized speech is generated (step S309). The generated synthesized speech is output to the speech output device 108, and the synthesized speech is output from the speech output device 108 (step S310).
[0029]
As described above, the car navigation device according to the present embodiment can change the prosodic parameters according to the situation around the host vehicle. As a result, it is possible to hear a calm voice when driving on a crowded road and a female voice on a congested road.
[0030]
[Third Embodiment]
In the present embodiment, a case will be described in which the prosodic parameters are updated so that the arrival time at the point where the next route guidance should be performed is taken into account and the text is read out within that time. The car navigation apparatus used in this embodiment is the same as that used in the first embodiment.
[0031]
FIG. 5 shows a situation in which the car navigation device of this embodiment operates. Reference numeral 501 denotes the own vehicle, which is assumed to be running at the speed v at the time shown in FIG. Reference numeral 502 denotes a point where the next route guidance must be performed. The host vehicle 501 must finish reading a text sentence to be described later before reaching the point 502. It is assumed that there is an L distance from the position of the own vehicle 501 to the point 502 at the time shown in FIG.
[0032]
FIG. 4 is a flowchart of processing performed by the car navigation device according to this embodiment. The program code according to the flowchart shown in the figure is stored in a memory such as a ROM or a RAM (not shown) in the car navigation apparatus according to this embodiment, and is read and executed by the CPU. As a result, the car navigation device of the present embodiment can perform each process described below.
[0033]
The process according to this flowchart is called when a certain route guidance sentence utterance (for example, “turn right at the intersection”) is completed. First, the own vehicle speed and the own vehicle position are measured by the own vehicle speed measuring unit 109 and the own vehicle position measuring unit 110, respectively, and the own vehicle speed and the own vehicle position are obtained by the traveling state information obtaining unit 105 (step S401). . As a result, the traveling state information acquisition unit 105 estimates a time T until the host vehicle arrives at a point where next route guidance is performed (for example, a point 502 in FIG. 5) (step S402). The time T can be estimated by the following equation.
[0034]
T = Distance to the point where the next route guidance is to be provided / Own vehicle speed Furthermore, by using the average value of the own vehicle speed for several minutes up to the present time instead of the current own vehicle speed, Estimated value.
[0035]
If the value of the time T is not sufficiently large (in the example, the time T is 10 minutes or less), the process is terminated (step S403). On the other hand, if the value of the time T is 10 minutes or more, the utterance content selection unit 102 selects an e-mail or news (text text) to an unread driver that can be read out within the time T (step S404). That is, the above-mentioned mail or news text stored in the information transmission server connected to the car navigation device (speech content storage memory 101) via the wireless network may be referred to and selected, or the information transmission server You may access and select the above-mentioned mail or news text sentence downloaded to the utterance content storage memory 101 in the car navigation device. Whether or not it can be read out within the time T is obtained by acquiring the number of syllables constituting the text sentence and the time required for uttering each syllable, and the time to read out the text sentence (the number of syllables × speaking each syllable). It is necessary to determine whether or not the result is within time T.
[0036]
As for the number of syllables in the entire text, measure the number of syllables that make up the text sentence of the mail or news stored in the above information transmission server in advance, and attach the measured number of syllables to this text. The syllable number is acquired when the text is referred to or downloaded by the car navigation device. In addition, the utterance content selection unit 102 may determine how to read a text sentence using morphological analysis and count the number of syllables based on this reading. The time required to utter each syllable is constant.
[0037]
FIG. 6 shows a method for calculating the time required to utter the content to be uttered.
[0038]
Reference numeral 601 denotes content to be uttered, and here, the word “akasaka” is used as an example. Reference numerals 601a to 601d denote syllables constituting the content 701 to be uttered. The time required to utter each syllable is t1, t2, t3, and t4, and each time (t1, t2, t3, and t4) may be a strict value measured in advance or all of them. The same average value may be used. As a result, the (estimated) time required for uttering the content 601 to be uttered is (t1 + t2 + t3 + t4).
[0039]
In addition, in order to determine whether or not the above-described mail or news is unread, it can be determined by attaching an unread flag to each mail or news. FIG. 7 shows the structure of a table in which this unread flag is attached to each mail, taking mail as an example. Reference numeral 701 denotes an unread flag item attached to each mail. The mail with the unread flag set to 0 indicates that the mail has been read, and the mail with the unread flag set to 1 indicates that the mail has not been read. Reference numeral 702 denotes an item in which the contents of each mail are described. This table is stored in the information transmission server described above, and this table is referred to when the car navigation apparatus refers to the mail in the information transmission server. This unread flag is set to 1 when mail arrives at the information transmission server, and is added to this table together with the contents of the mail.
[0040]
Although the structure of the above table has been described by taking e-mail as an example, the structure of the table is the same even for news, and the table exemplarily includes news by setting the item 702 as an item in which the content of the news is described. It can be. In step S404, the unread flag corresponding to the selected text sentence (the contents of mail or the contents of news) is set to zero.
[0041]
Next, it is determined whether or not there is still a text sentence (sentence) to be read out (step S405). If there is no more text sentence to be read out, the process according to this flowchart is terminated. On the other hand, if there is still a text sentence to be read, the prosodic parameter generation unit 103 takes in one sentence from the text sentence selected by the utterance content selection unit 101 as described above (step S406), and again the own vehicle speed measurement unit 109 and the own vehicle. The own vehicle speed and the own vehicle position are measured from the position measuring unit 110 and acquired by the traveling state information acquiring unit 105 (step S407), and the traveling state information acquiring unit 105 re-estimates the time T until the next route guidance ( Step S408). Here, instead of the vehicle speed, the accuracy of T is improved by using the average value of the vehicle speed for several minutes until now. In addition, by repeating this re-estimation every time one sentence is read out, it is possible to cope with a case where the vehicle speed suddenly changes after starting to read the sentence.
[0042]
The prosodic parameter generation unit 103 sets a default prosodic parameter in the captured sentence text (step S409), and updates the set prosodic parameter according to the value of T (step S410 to step S413). In the example shown in the figure, the phoneme duration is manipulated, which corresponds to the quickest possible opening when the time until the next route guidance is short.
[0043]
The phoneme segment connection unit 104 uses the prosodic parameters obtained by the prosody parameter generation unit 103 to generate synthesized speech in the same manner as in the first and second embodiments, and the synthesized speech is output by the speech output device 108. (Step S414).
[0044]
As described above, the car navigation apparatus according to the present embodiment considers the arrival time to the point where the next route guidance should be taken into consideration, and uses the prosodic parameters as synthesized speech parameters so that the text is read out within that time. Update. As a result, the speed at which the text is read out can be changed according to the above arrival time, and the text can be read out within the above arrival time.
[0045]
[Other Embodiments]
Note that the present invention can be applied to a system including a plurality of devices (for example, a host computer, an interface device, a reader, and a printer), and a device (for example, a copying machine and a facsimile device) including a single device. You may apply to.
[0046]
Another object of the present invention is to supply a storage medium (or recording medium) in which a program code of software that realizes the functions of the above-described embodiments is recorded to a system or apparatus, and the computer (or CPU or Needless to say, this can also be achieved by the MPU) reading and executing the program code stored in the storage medium. In this case, the program code itself read from the storage medium realizes the functions of the above-described embodiments, and the storage medium storing the program code constitutes the present invention. In addition, by executing the program code read by the computer, not only the functions of the above-described embodiments are realized, but also an operating system (OS) running on the computer based on the instruction of the program code. It goes without saying that a case where the function of the above-described embodiment is realized by performing part or all of the actual processing and the processing is included.
[0047]
Furthermore, after the program code read from the storage medium is written into a memory provided in a function expansion card inserted into the computer or a function expansion unit connected to the computer, the function is determined based on the instruction of the program code. It goes without saying that the CPU or the like provided in the expansion card or the function expansion unit performs part or all of the actual processing and the functions of the above-described embodiments are realized by the processing.
[0048]
When the present invention is applied to the above-described storage medium, the storage medium stores program codes corresponding to the above-described flowchart (shown in FIG. 2, 3 or 4).
[0049]
【The invention's effect】
By controlling the parameters related to the synthesized speech in accordance with the vehicle and surrounding conditions, there is an effect that the speech synthesis result can be heard by a person riding in the vehicle.
[Brief description of the drawings]
FIG. 1 is a diagram showing a schematic configuration of a car navigation device according to a first embodiment of the present invention.
FIG. 2 is a flowchart of processing performed by the car navigation device according to the first embodiment of the present invention.
FIG. 3 is a flowchart of processing performed by a car navigation device according to a second embodiment of the present invention.
FIG. 4 is a flowchart of processing performed by a car navigation device according to a third embodiment of the present invention.
FIG. 5 is a diagram illustrating a situation in which a car navigation device according to a third embodiment of the present invention operates.
FIG. 6 is a diagram illustrating a method of calculating a time required for uttering content to be uttered.
FIG. 7 is a diagram illustrating a configuration of a table in which an unread flag is attached to each mail, taking mail as an example.

Claims (15)

Speech synthesis means for generating synthesized speech corresponding to text;
An acquisition means for acquiring road congestion conditions;
When the road congestion situation is the first congestion situation, the speaker of the synthesized speech is changed, and when the road congestion situation is the second congestion situation, the prosody is a parameter related to the prosody of the synthesized speech. A speech synthesizer comprising: a changing unit that changes a parameter.   The speech synthesis apparatus according to claim 1, wherein the prosodic parameter includes a parameter related to voice pitch.   The speech synthesizer according to claim 1, wherein the prosody parameter includes a parameter relating to a voice speed.   The speech synthesizer according to any one of claims 1 to 3, wherein the prosodic parameter includes a parameter related to a voice volume.   The speech synthesis apparatus according to claim 1, wherein the text is downloaded from a predetermined server.   The said change means changes the speaker of the said synthetic voice from a man to a woman, when the congestion condition of the said road is the said 1st congestion condition, The any one of Claim 1 thru | or 5 characterized by the above-mentioned. The speech synthesizer described.   7. The change unit according to claim 1, wherein when the congestion situation of the road is the second congestion situation, parameters relating to a volume, a height, and a speed of the synthesized speech are changed. The speech synthesizer of any one of Claims. A speech synthesis step for generating synthesized speech corresponding to the text;
An acquisition process for acquiring road congestion;
A first changing step of changing a speaker of the synthesized speech when the congestion situation of the road is a first congestion situation;
And a second changing step of changing a prosodic parameter that is a parameter relating to the prosody of the synthesized speech when the road congestion state is a second congestion state.   The speech synthesis method according to claim 8, wherein the prosodic parameter includes a parameter related to voice pitch.   The speech synthesis method according to claim 8 or 9, wherein the prosodic parameter includes a parameter relating to a voice speed.   The speech synthesis method according to claim 8, wherein the prosodic parameter includes a parameter related to a voice volume.   The speech synthesis method according to any one of claims 8 to 11, wherein the text is downloaded from a predetermined server.   The said change process WHEREIN: When the congestion condition of the said road is a 1st congestion condition, the speaker of the said synthetic | combination voice is changed from a man to a woman. Voice synthesis method. Wherein in the changing step, when the congestion state of the road is the second congestion, the synthesized speech loudness of claims 8 to 13, characterized in that to change the parameters related to height and speed The speech synthesis method according to any one of the above.   A storage medium storing a program code for realizing the speech synthesis method according to claim 8.

Images