JP2001324991A - Voice synthesizer and voice data storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a voice synthesizer in which connection between voices is smoothly conducted in reproduced voice, there exists no sense of incongruity and voice reproducing that is easily heard is conducted. SOLUTION: In the voice synthesizer, voice data are stored in a storage medium so that the voice data are read in terms of every unit voice data. Plural unit voice data are read from the storage medium and combined to generate sentence voice data and voice is reproduced. Priority data are added to each of the unit voice data to indicate the priority.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice synthesizer and a voice data storage medium, and more particularly, to a method for smoothly connecting each voice in a reproduced voice without causing a sense of incongruity.
The present invention relates to a voice synthesizing device capable of reproducing a voice that is easy to hear and a voice data storage medium used in the voice synthesizing device.

[0002]

2. Description of the Related Art In recent years, a voice synthesizing apparatus is used in many electronic devices, and guidance of operations and the like is often performed by voice. In particular, in a navigation system mounted on a car, it is difficult to see a screen while the vehicle is running, so that voice guidance is usually performed using a voice synthesizer.

[0003] There are various types of speech synthesis methods. In a navigation system, a CD-ROM, a DV
As storage media such as optical discs such as D increase in storage capacity, a large storage capacity is required, but a speech synthesis method using speech waveform data itself having good sound quality has been widely adopted. This is because a lot of unit audio data divided into units convenient for speech synthesis is created, and these unit audio data are written to a storage medium in a readable state for each unit audio data, and the audio playback is performed. In this case, a plurality of unit sound data necessary for a sentence to be read are read from the storage medium, and these are combined to generate a sentence sound data for reproduction, and the generated sentence sound data is subjected to digital-analog conversion. Thereafter, the signal is amplified and output from a speaker to reproduce the sound.

[0004]

However, in order to read the unit audio data from the storage medium, an operation such as moving a pickup for reading to a position on the storage medium where each unit audio data is stored is performed. Requires a certain amount of time. For this reason, an unnecessary silent part may be generated between the unit sound data to be reproduced, and the sound reproduction may be uncomfortable. In order to solve this problem, there is a method of once reading out all the unit audio data constituting the text to be reproduced and then starting the audio reproduction, but in this case, it takes time until the audio reproduction starts, and the audio notification However, there is a problem that the timing is deviated or the user feels uncomfortable.

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and it is possible to smoothly perform connection between respective sounds in a reproduced sound, and to perform sound reproduction that does not cause a sense of incongruity and is easy to hear. It is an object of the present invention to provide a voice synthesizing device that can be used and a voice data storage medium used in the voice synthesizing device.

[0006]

Means for Solving the Problems and Effects There is provided a voice synthesizing apparatus (1) according to the present invention, wherein a voice data is read from a storage medium in which voice data is readable for each unit voice data. In a voice synthesizing apparatus for generating text voice data by reading and combining a plurality of unit voice data and reproducing the voice, priority data indicating a priority is assigned to each of the unit voice data.
According to the voice synthesizing device (1), the priority data indicating the priority is assigned to each of the unit voice data. Therefore, the priority of data writing to the cache memory and the data retention are stored based on the priority data. It is possible to control the priority order, and it is possible to improve the efficiency of processing and the responsiveness of synthesized speech output.

[0007] Further, a speech synthesizer (2) according to the present invention.
A data reading means for reading out the unit voice data from a storage medium in which voice data is readable and stored for each unit voice data and writing the data into a cache memory in the voice synthesizing apparatus (1); The unit voice data read out, comprising a plurality of unit voice data read out to generate text voice data and voice conversion means for voice conversion, based on the priority data assigned to the unit voice data It is characterized by comprising a priority processing means for determining a priority order of the unit audio data to be left in the cache memory. According to the speech synthesizer (2), the unit speech data having a high priority, for example, the frequently used unit speech data is preferentially held in the cache memory, so that the necessary unit speech data is re-read from the storage medium. The required frequency is reduced, and processing efficiency and high-speed response of synthesized speech output can be realized.

[0008] A voice synthesizing device (3) according to the present invention.
In the above-mentioned voice synthesizing device (2), when the priority data is stored in the storage medium in advance for each of the unit voice data, the priority processing means may output the priority data from the storage medium. It is characterized by comprising priority reading means for reading data, and priority calculating means for calculating a priority based on the priority data read by the priority reading means. The speech synthesizer (3)
According to the method, the priority can be determined based on the priority data stored in the storage medium. Therefore, the priority can be quickly obtained without performing a complicated operation.

[0009] A voice synthesizing apparatus (4) according to the present invention.
In the voice synthesizer (2), the priority processing unit detects a frequency of reading the unit voice data stored in the cache memory, and a frequency detection unit detects a read frequency of the unit voice data based on the frequency detected by the frequency detection unit. And a priority calculating means for calculating the priority. According to the speech synthesizer (4), since the priority is obtained from the actual use frequency, a priority suitable for the actual use can be obtained, and it is necessary to prepare special priority data in advance. There is no.

[0010] A voice synthesizing device (5) according to the present invention.
In the above-mentioned voice synthesizing device (2), when the priority data is stored in the storage medium in advance for each of the unit voice data, the priority processing means may output the priority data from the storage medium. Priority reading means for reading data, frequency detecting means for detecting a reading frequency of the unit voice data stored in the cache memory, and priority data read by the priority reading means and the frequency detecting means. And a priority calculating means for calculating a priority based on the detected frequency. According to the speech synthesizer (5), the priority is calculated in consideration of both the priority data stored in the storage medium and the frequency of use based on actual use.
A substantially appropriate priority can be obtained even if the use period of the device is not long, and a priority that is more suitable for actual use can be obtained as the use period of the device becomes longer.

Further, a speech synthesizer (6) according to the present invention.
Is characterized in that, in the voice synthesizing device (4) or (5), the frequency data detected by the frequency detecting means is stored in a backup memory whose stored content is retained even in a power-off state. According to the speech synthesizer (6), since the frequency data is held for a long period of time, a priority suitable for the user is always required.

Also, a speech synthesizer (7) according to the present invention.
A voice synthesizing apparatus for reading and combining a plurality of unit voice data from a storage medium in which voice data is readablely stored for each unit voice data to generate sentence voice data and reproduce voice; Is provided with delimiter type data indicating a first delimiter type in audio reproduction. According to the voice synthesizing device (7), since the delimiter type data having a high correlation with the continuous reproducibility (silence length) with respect to the preceding voice reproduction is assigned to each of the unit voice data, a cache is generated based on the delimiter type data. This makes it possible to control the priority of writing data to the memory and the priority of holding data, thereby improving processing efficiency and improving the response of synthesized speech output.

[0013] A speech synthesizer (8) according to the present invention.
A data reading means for reading the unit voice data from a storage medium in which the voice data is readable and stored for each unit voice data and writing the data into the cache memory; Read out the unit voice data, comprising a plurality of unit voice data read out to generate text voice data and voice conversion means for voice conversion, based on the delimiter type data given to the unit voice data It is characterized by comprising a priority processing means for determining a priority order of the unit audio data to be left in the cache memory. According to the voice synthesizing device (8), the unit voice data having high continuous reproducibility with respect to the previous voice reproduction, that is, for example, the unit voice data which needs to be reproduced continuously (without a silence part) after the previous voice reproduction is preferentially given to the above-mentioned voice data. Since the data is held in the cache memory, data that must be read quickly during reproduction can be quickly read, and continuous reproduction can be performed without any trouble.

Further, a speech synthesizer (9) according to the present invention.
A voice synthesizing apparatus for reading and combining a plurality of unit voice data from a storage medium in which voice data is readablely stored for each unit voice data to generate sentence voice data and reproduce voice; Is provided with connection strength data indicating the strength of connection with the immediately preceding unit sound in sound reproduction. According to the speech synthesizer (9), the connection strength data having a high correlation with the continuous reproducibility (silence length) with respect to the preceding voice reproduction is given. It is possible to control the order and the priority of data holding, and it is possible to improve the efficiency of processing, improve the response of synthesized speech output, and the like.

Further, a speech synthesizer (10) according to the present invention.
A data reading means for reading the unit voice data from a storage medium in which the voice data is readable and stored for each unit voice data and writing the data into the cache memory in the voice synthesizer (9); Read out the unit voice data, comprising a plurality of unit voice data read out to generate text voice data and voice conversion means for voice conversion, based on the connection strength data given to the unit voice data It is characterized by comprising a priority processing means for determining a priority order of the unit audio data to be left in the cache memory. According to the speech synthesizer (10),
The unit voice data which has a high continuous reproduction property with respect to the previous voice reproduction, that is, for example, which needs to be reproduced continuously (without a silence portion) after the previous voice reproduction is preferentially held in the cache memory. The data that must be read can be read quickly, and continuous reproduction can be performed without any trouble.

Also, a speech synthesizer (11) according to the present invention.
A voice synthesizing apparatus for reading and combining a plurality of unit voice data from a storage medium in which voice data is readablely stored for each unit voice data to generate sentence voice data and reproduce voice; Is provided with use position data indicating the use position in the text. According to the voice synthesizing device (11), highly correlated use position data is added to unit voice data that is easily affected by a delay time of reading data from an external storage medium. This makes it possible to control the priority of writing data to the memory and the priority of holding data, thereby improving processing efficiency and improving the response of synthesized speech output.

Further, a speech synthesizer (12) according to the present invention.
A data reading means for reading out the unit voice data from a storage medium in which voice data is readable and stored for each unit voice data and writing the data into a cache memory in the voice synthesizer (11); Read out the unit voice data, comprising a plurality of unit voice data read out to generate sentence voice data and voice conversion means for voice conversion, based on the use position data given to the unit voice data Priority processing means for determining the priority of the unit audio data to be left in the cache memory is provided. According to the speech synthesizer (12),
The unit voice data of the sentence head, which is most susceptible to the delay time of reading data from the external storage medium, can be preferentially held in the cache memory, and the data that must be read quickly at the time of reproduction can be read quickly. It is possible to provide a voice synthesizing device that can perform continuous reproduction without any delay until the start of reproduction.

Further, a speech synthesizer (13) according to the present invention.
A voice synthesizing apparatus for reading and combining a plurality of unit voice data from a storage medium in which voice data is readablely stored for each unit voice data to generate sentence voice data and reproduce voice; Is provided with playback time data indicating the audio playback time. According to the speech synthesizer (13), it is possible to control the priority of writing data to the cache memory and the priority of holding data based on the reproduction time data indicating the difference between the characteristics of the read time and the storage capacity. Thus, it is possible to improve the efficiency of processing and the response of synthesized speech output.

Also, a speech synthesizer (14) according to the present invention.
A data reading means for reading the unit voice data from a storage medium in which the voice data is readable for each unit voice data and writing the read unit voice data into a cache memory, and Voice conversion means for reading out the unit voice data, generating text voice data by combining the plurality of read unit voice data, and converting the voice into voice, and reproducing the voice data based on the playback time data given to the unit voice data. It is characterized in that a priority processing means is provided which performs processing for preferentially leaving the unit audio data in the cache memory as the time is shorter. The speech synthesizer (1)
According to 4), the short reproduction time (small capacity) unit audio data, which can be stored in the cache memory, is preferentially held in the cache memory, so that the use efficiency of the cache memory is improved and the text is composed. By storing in the cache memory the short playback time (small capacity) unit audio data that takes a long time to read in order to
It is possible to improve the efficiency of processing during voice reproduction and the responsiveness of synthesized voice output.

Further, a speech synthesizer (15) according to the present invention.
Means for reading out the unit voice data from a storage medium in which the voice data is readable for each unit voice data and writing the unit voice data to a cache memory in the voice synthesizing device (1); Voice conversion means for reading the unit voice data, generating text voice data by combining the read plurality of unit voice data, and converting the voice into voice, wherein the priority given to the unit voice data when the apparatus is started up It is characterized by comprising a start-up writing means for writing the unit audio data in the cache memory based on data. According to the speech synthesizer (15), data can be written in the cache memory in order from the highest priority when the apparatus is started, so that sound reproduction with good responsiveness can be realized immediately after the apparatus is started. .

Further, a speech synthesizer (16) according to the present invention.
Is characterized in that, in the speech synthesizer (15), the start-up writing means writes the unit sound data in descending order of priority until the cache memory becomes full at the start-up of the device. And According to the speech synthesizer (16), since the data having the highest priority can be written in the cache memory in order from the high-priority data at the time of the activation of the device, the state of the best responsiveness at the time immediately after the activation of the device is possible. Can be kept.

Also, a speech synthesizer (17) according to the present invention.
In the speech synthesizer (15), the cache memory is divided into a resident area in which data is not deleted after the apparatus is started and a temporary area in which data is also deleted after the apparatus is started. However, unit voice data having a higher priority than a predetermined priority is written in the resident area at the time of startup. According to the speech synthesizer (17), the unit voice data having a high priority is not deleted from the cache memory during use of the device, and the unit voice data having a high priority is unexpectedly deleted during use of the device. It is possible to prevent the occurrence of a situation such as deletion and deterioration of responsiveness.

Also, in the audio data storage medium (1) according to the present invention, in the audio data storage medium in which the audio data is stored in a readable manner for each unit audio data, the priority indicating the priority for each of the unit audio data is provided. It is characterized in that data is added and stored in advance. According to the audio data storage medium (1), the priority data indicating the priority is assigned to each of the unit audio data. Therefore, in the audio synthesizer or the like, the priority data is read and stored in the cache memory. It is possible to control the priority of writing and the priority of data holding, and it is possible to improve the efficiency of processing in a speech synthesizer or the like, improve the response of synthesized speech output, and the like.

The audio data storage medium (2) according to the present invention is an audio data storage medium in which audio data is stored in a readable manner for each unit of audio data. It is characterized in that delimiter type data indicating a type is stored. According to the audio data storage medium (2), the delimiter type data having a high correlation with the continuous reproducibility (silence length) with respect to the preceding voice reproduction is stored. It is possible to control the priority of writing data to the read cache memory and the priority of holding data, thereby improving the efficiency of processing in a speech synthesizer or the like and improving the response of synthesized speech output.

The audio data storage medium (3) according to the present invention is an audio data storage medium in which audio data is stored in a readable manner for each unit of audio data. It is characterized in that connection strength data indicating the strength of connection with the unit voice is stored. According to the audio data storage medium (3),
Since the connection strength data having a high correlation with the continuous reproducibility (silence length) with respect to the previous voice reproduction is stored, in the voice synthesizer or the like, the connection strength data is read out, and the priority order of the data writing to the cache memory or the data It is possible to control the priority of holding, and it is possible to improve the efficiency of processing in a speech synthesizer or the like, improve the responsiveness of synthesized speech output, and the like.

Further, in the audio data storage medium (4) according to the present invention, in the audio data storage medium in which the audio data is stored so as to be readable for each unit audio data, the use position in the text for each unit audio data is determined. The use position data shown is stored. According to the audio data storage medium (4), since the use position data having a high correlation is stored in the unit audio data which is easily affected by the delay time of the data reading from the external storage medium, the audio synthesis apparatus or the like uses Can read out the usage position data and control the priority of data writing to the cache memory and the priority of data retention, thereby improving the efficiency of processing in a voice synthesizer and the like and improving the responsiveness of synthesized voice output. Etc. can be achieved.

The audio data storage medium (5) according to the present invention is a sound data storage medium in which audio data is stored in a readable manner for each unit audio data, wherein the size data indicating the data length for each unit audio data is provided. Is stored.

According to the audio data storage medium (5),
Since size data indicating the difference in storage capacity characteristics is stored, in a voice synthesizer or the like, it is possible to read out the reproduction time data and control the priority of writing data to the cache memory and the priority of holding data. This makes it possible to improve the efficiency of processing in a speech synthesizer or the like, improve the responsiveness of synthesized speech output, and the like.

[0029]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A speech synthesizing apparatus and a speech data storage medium according to an embodiment of the present invention will be described below. FIG. 1 is a block diagram showing a configuration of a navigation / radio system equipped with a speech synthesizer according to an embodiment. The GPS sensor 1 is connected to a microcomputer (navigation microcomputer) 5, receives a signal from a GPS satellite, calculates a position from the signal, and outputs the calculation result to a navigation system control microcomputer (navigation microcomputer) 5. It is supposed to. The gyro sensor 2 is a sensor that detects a change in the direction of the vehicle, and is configured by a gyro. The gyro sensor 2 is also connected to the navigation microcomputer 5, and a detection signal of the gyro sensor 2 is output to the navigation microcomputer 5. . Navigation microcomputer 5
Then, the direction of the vehicle is calculated by integrating the detection signals of the gyro sensor 2. The vehicle speed pulse input unit 3 captures a vehicle speed pulse composed of a pulse (the number of pulses in a predetermined period is proportional to the vehicle speed) generated every predetermined traveling distance from a vehicle speed sensor (not shown) installed on the vehicle side, and generates noise. After performing removal, waveform shaping, and the like, a vehicle speed pulse is output to the navigation microcomputer 5. The signal from the vehicle speed sensor installed on the vehicle side is also used for control of the drive system of the vehicle, for example, fuel injection control and ignition timing control, and this vehicle speed sensor (not shown) is installed in the vehicle. Things. This vehicle speed sensor, for example, a magnet that rotates in synchronization with the axle,
A magnetic sensor consisting of a reed switch whose contacting state changes according to the state of the magnetic field that changes according to the rotational position of the magnet, a shielding plate that rotates in synchronization with the axle, and a blocking state of the optical path according to the rotational position of the shielding plate Light receiving element,
It is composed of an optical sensor or the like composed of a light emitting element.

The CD-ROM player 4 reads necessary data from a CD-ROM (optical disk) storing map data and the like in accordance with a control signal from the navigation microcomputer 5, and transmits this data signal to the navigation microcomputer 5 Output. The CD-ROM stores audio data (waveform data) in a unit suitable for audio guidance for navigation (unit audio data) in a readable manner. For example, "approximately", "300", "m ahead",
They are stored in units such as "right direction" and "is." These are read out and combined to provide voice guidance such as "approximately 300 m ahead right direction." Note that the CD-ROM of the CD-ROM player 4 is replaceable, and is capable of updating maps and audio data (upgrading the map CD-ROM).

The operation switch 6 is a switch for operating the navigation system, and is constituted by a push button switch installed in the navigation system body, an infrared remote controller, and the like, and includes an ON-OFF switch, a direction specifying switch such as a joystick, and the like. ing. The navigation system also has a voice synthesis operation control function, and a switch for operating the control function is included in the operation switch 6.

The navigation microcomputer 5 calculates the position of the own vehicle in an independent manner based on the detection signal from the gyro sensor 2 and the vehicle speed pulse signal from the vehicle speed pulse input unit 3.
Further, the calculation result and the position signal from the GPS sensor 1 are complemented to determine the own vehicle position. The navigation microcomputer 5 controls the CD-ROM player 4 according to the determined vehicle position and the operation state of the operation switch 6 to fetch necessary map data and the like from the CD-ROM, and to reach the destination. And the like to calculate the route of the map, a corresponding map is displayed on the display 7 constituted by a liquid crystal display device.
Routes, various types of guidance, and operation guidance are displayed. The navigation microcomputer 5 has a built-in memory 9 used for storing various data and programs and for performing arithmetic processing. The memory 9 retains its stored contents even when the power is turned off.
It is configured to include an OM, a writable EEPROM, a RAM backed up by a power supply, and the like.

The navigation microcomputer 5 outputs audio data composed of unit audio data to a digital-to-analog converter (D / A) 8, and the D / A 8 outputs an analog signal of synthesized audio to the amplifier 12. The amplifier 12 amplifies the synthesized voice and outputs it as voice from a speaker 13 provided in the vehicle interior.

The operation performed by the navigation microcomputer 5 at the time of voice output will be briefly described as follows. Navigation microcomputer 5
When it is necessary to provide route guidance by voice when approaching an intersection or the like, unit voice data necessary for voice synthesis is obtained from the guidance content (reproduced text) at that time, and the CD-ROM
An instruction signal is output to the player 4 to read the unit audio data from the CD-ROM. Then, the navigation microcomputer 5 assembles the plurality of unit voice data read from the CD-ROM so as to form a sentence, and outputs the data to the D / A 8. Thereafter, the audio signal (text) converted into an analog signal by the D / A 8 is amplified by the amplifier 12 and
Will be output as audio.

Next, data processing in speech synthesis processing,
And a processing program structure will be described. FIG.
FIG. 4 is an explanatory diagram illustrating an outline of a speech synthesis process. The speech synthesis processing program includes a speech management task T1, a speech conversion task T
2. It is composed of a plurality of tasks called a voice output task T3 and a CD reading task T4 (a program for realizing a certain operation, and a plurality of necessary tasks are collected to constitute an entire program), and each task is a navigation microcomputer. 5 alternately occupy the processing of the five CPUs as needed, and constitute a so-called multitask.
Data exchange and operation instructions (control) are performed.

Temporary buffer TB for storing read data
1 is assigned a part of the RAM area of the memory 9, and the unit audio data read from the CD-ROM D1 is stored as cache data. When the same unit audio data becomes necessary thereafter, this readout is performed. The data is read from the data storage temporary buffer TB1. Therefore, since the unit audio data stored in the read data storage temporary buffer TB1 may be read from the read data storage temporary buffer TB1, there is no need to read from the CD-ROM D1, and high-speed reading becomes possible.

The voice management task T1 includes a CD-ROMD when a voice guidance is required.
The notification of the identification code of the unit audio data to be read out from No. 1 and a reading instruction are performed. Upon receiving the instruction, the CD read task T4 reads the corresponding unit audio data from the CD-ROM D1, and writes the read unit audio data to the read data storage temporary buffer TB1. The CD reading task T4 is
It notifies the voice management task T1 of the data read status (read completion, etc.). The voice management task T1
Is designed to give a read instruction only for unit audio data that is not stored in the read data storage temporary buffer TB1.

The audio management task T1 sends the audio data conversion task T2 to the audio conversion task T2 at an appropriate timing such as receiving a notification of completion of reading the unit audio data from the CD read task T4. The voice data is read, and an instruction is made to generate text voice data. The voice conversion task T2 having received the instruction includes a temporary buffer TB1 for storing read data.
, A predetermined unit of audio data is read out in a predetermined order, and the read data is converted into an output format from the D / A 8 and written to the audio output temporary buffer TB2. The audio conversion task T2 also decodes the unit audio data as needed (read from the CD-ROM D1 and stores the read data temporary buffer TB1).
Is encoded in the unit audio data stored in). The temporary buffer for audio output TB2 is usually composed of a ring buffer, and outputs unit audio data in the order of writing according to an output (read) instruction. Further, the audio conversion task T2 is provided with a temporary audio output buffer TB2 for the audio management task T1.
Notify the status of writing data to the memory (writing completed, etc.).

The voice management task T1 includes a voice conversion task T
When receiving the write completion notification from the second device 2, the CPU outputs a sound output instruction to the sound output task T3 at an appropriate timing for the sound output. Upon receiving the voice output instruction, the voice output task T3 reads the text voice data to be output from the temporary buffer for voice output TB2 and outputs it to the D / A8. By outputting the sentence voice data to the D / A 8, the desired sentence voice data is converted into an analog signal by the D / A 8, amplified, and output from the speaker 13 as voice. The voice output task T3 notifies the voice management task T1 of the data output status (output completion, etc.).

Next, audio data stored in the CD-ROM D1 will be described. FIG. 3 is a configuration diagram showing a storage configuration of audio data in the CD-ROM D1. As shown in FIG. 3, the storage area for the audio data includes an audio data header section storing data indicating the storage state of the entire audio data, and an audio data storage section storing data indicating the storage state of each unit audio data. It is configured to include an address information section and an actual audio data section in which actual data (digital audio data) of each unit audio data is stored. In the audio data header section, the size (storage capacity) of the entire audio data area, the head address of the audio data storage address information section,
The head address and the like of the actual audio data section are stored.

The audio data storage address information section contains, for each unit audio data, a code number for identification, an address (a head address of an area where each unit audio data is stored), a size, and a priority to be left in the cache memory. Priority data indicating the order is stored. Regarding the priority data, the priority data may have different numerical values as shown by numerical values in the figure, or may be divided into several levels as shown by alphabets in the figure. good.

Next, a description will be given of a voice management task T1 and a CD reading task T4 performed by the navigation microcomputer 5 for voice synthesis processing. FIG. 4 is a flowchart showing a first processing example performed by the voice management task T1, which is executed when a voice output request is issued in a navigation operation task or the like.

In step S1, processing for taking in an audio output request from a navigation operation task or the like, for example, analysis of input data, selection of unit audio data to be read from the CD-ROM D1 (unit audio read to the temporary buffer TB1 for storing read data, etc.) (No data is required), and the routine goes to Step S2. At this time, the tasks of the audio reproduction system, that is, the audio conversion task T2 and the audio output task T
The processing of Step 3 is stopped (prohibited), and the processing of other tasks is set to the priority state (CPU power is concentrated to improve the processing speed).

In step S2, the CD reading task T
4, a read request for unit audio data is made, that is, a read instruction and a code number of the unit audio data to be read are transmitted to the CD read task T4, and the process proceeds to step S3. In step S3, it is determined whether or not the read data storage temporary buffer TB1 has an empty area for newly writing unit audio data. If it is determined that there is an empty area, the process proceeds to step S4. If it is determined that there is no empty area, Step S
Move to 6. In step S6, of the unit audio data stored in the read data storage temporary buffer TB1,
The low priority data is released to create a free area (in a state where deletion or overwriting is possible), and the process returns to step S3.

In step S4, the unit audio data read from the CD-ROM D1 is newly written in the temporary buffer TB1 for storing read data, and the flow advances to step S5. In step S5, a sound output process is performed and the process ends. In the sound output processing, the suspension (prohibition) of the processing of the sound reproduction system task, that is, the sound conversion task T2 and the sound output task T3 is released, and a sound conversion start request is issued to the sound conversion task T2. The instruction and the code number of the unit voice data to be voice-converted and the data indicating the conversion order are transmitted to the voice conversion task T2. With this processing, the sentence audio data is written into the audio output temporary buffer TB2. After that, a voice output start request is made to the voice output task T3, that is, data indicating the voice output start instruction and the text voice identification code number to be voice-outputted are transmitted to the voice output task T3, and the process is completed. With this processing, the audio output task T3 reads the corresponding sentence audio from the audio output temporary buffer TB2,
/ A8.

As described above, according to the present processing example, the read data storage temporary buffer TB1 is used up to its full capacity, and the storage state is maintained so that the unit audio data with high priority remains. The storage temporary buffer TB1 is used efficiently, the unit audio data can be read out quickly, and audio can be reproduced smoothly.

Next, a second processing example of the voice management task T1 and the CD reading task T4 performed by the navigation microcomputer 5 during the voice synthesis processing will be described. FIG. 5 is a flowchart showing a second processing example performed by the voice management task T1 and the CD reading task T4, which is executed when a voice output request is made in the navigation operation task or the like. Note that the same processes as those in the first processing example are denoted by the same reference numerals, and description thereof will be omitted.

The second processing example corresponds to step S of the first processing example.
Step S7 and Step S8 between Step 4 and Step S5
Is inserted, and step S6 is changed to step S9. At step S4, the CD-R is stored in the temporary buffer TB1 for reading data storage.
When the unit audio data read from the OMD 1 is newly written, the process proceeds to step S7. In step S7, the usage count data for the unit audio data written in the read data storage temporary buffer TB1 is counted up, and the process proceeds to step S8. In step S8, the usage count data counted up in step S7 is stored in the backed-up memory, and the process proceeds to step S5.

If it is determined in step S3 that there is no free space in the temporary buffer TB1 for storing read data,
Move to step S9. In step S9, using the usage count data stored in the backup memory, unit audio data with a low usage rate is searched, and the corresponding unit audio data is deleted from the read data storage temporary buffer TB1, and the process returns to step S3. . The usage rate is based on past performance,
For example, a value such as the number of times the unit audio data has been used in the last 10,000 uses in the past can be used.

As described above, according to the second processing example, the temporary buffer TB1 for storing read data is used up to its full capacity, and the storage state is maintained such that unit audio data having a high usage rate remains. The temporary buffer TB1 for storing read data is used efficiently, so that the unit audio data can be read quickly and the sound can be reproduced smoothly.

Next, a second example of audio data stored in the CD-ROM D1 will be described. FIG. 6 is a configuration diagram showing an audio data storage configuration of the CD-ROM D1 in the second example. As shown in FIG. 6, the storage area of the audio data includes an audio data header section storing data indicating the storage state of the entire audio data, and an audio data storage section storing data indicating the storage state of each unit audio data. It is configured to include an address information section and an actual audio data section in which actual data (digital audio data) of each unit audio data is stored. In the audio data header section, the size (storage capacity) of the entire audio data area, the start address of the audio data storage address information section, the start address of the actual audio data section, and the like are stored. In the audio data storage address information section, a code number for identification,
Address (head address of the area where each unit audio data is stored), size, and sentence separation information indicating the type of sentence separation between preceding unit sounds are stored.

The sentence delimiter information indicates the type of the delimiter of the unit audio data after the audio is reproduced, and includes a "punctuation" where a punctuation is inserted after the audio is reproduced, such as "is.""Phraseboundary" where breathlessness / silence part enters,
For example, there is "no break" in which the subsequent part is reproduced continuously. For example, "approximately" is "reading point", "300" is "no delimiter", "m ahead" is "phrase boundary", "rightward" is "no delimiter", "is.""about,
○○ 300m ahead ○ Right direction. ○○○ (next sentence) "
When sound is reproduced (by inserting a silent part) at intervals (delimiters) according to the sentence separation information as in (（indicates no sound volume), sound reproduction that is natural and easy to hear can be realized.
In other words, the delimiter information between the preceding unit voice (the delimiter of the preceding unit voice is roughly determined for each unit voice. For example, before “approximately”, the unit of “punctuation” such as “is.” Audio comes), the preceding audio breaks are “No break”, “Phrase boundary”,
Since it is necessary to read out the unit audio data at high speed in the order of "reading point" and "punctuation mark", if the unit audio data is preferentially stored in the temporary buffer for reading data TB1 in this order, the temporary buffer for storing reading data may be used. The use efficiency of the TB1 is improved, and smooth sound reproduction becomes possible.

The audio data storage address information section also includes connection strength data (for example, “m” and “head”, etc.) indicating the strength of the connection with the preceding unit sound. ”And“ 300 ”have weak connection strengths)
Use location data indicating the use location in the text of the unit voice data (for example, "approximately" etc. is "start of sentence", "300" and "before" are "in the sentence", "is." Is "end of sentence") You can keep it. When the connection strength is weak, the reproduced sound can be heard naturally even if there is a relatively long silent portion between the preceding unit sound. That is, since the unit audio data needs to be read at a higher speed as the connection strength is higher, it is desirable to store the unit audio data in the temporary buffer TB1 for storing read data preferentially. If stored, a temporary buffer T for storing read data
The use efficiency of B1 is improved, and smooth sound reproduction becomes possible.

In the case of the unit sound used at the beginning of the sentence, the CD-ROM D1
, And this tendency is stronger for unit voice data used at the beginning of a sentence. For this reason, the unit voice data is read out in the order of “beginning of sentence”, “in the middle of sentence”, and “end of sentence” in the order of use, and the temporary buffer
, The use efficiency of the read data storage temporary buffer TB1 is improved, and smooth audio reproduction is possible.

When reading the unit audio data from the CD-ROM D1, the movement time of the optical pickup to the data storage position occupies much of the reading time, and the data size does not significantly affect the reading time. For this reason, CD
-If the number of unit audio data read from the ROMD1 is reduced, smooth audio reproduction becomes possible. Therefore, if the unit audio data having a small size is preferentially stored in the read data storage temporary buffer TB1, the number of storage unit audio data stored in the read data storage temporary buffer TB1 increases, and the read data storage temporary buffer TB1 increases. The use efficiency of the temporary buffer TB1 is improved, and smooth audio reproduction can be performed.

Next, a description will be given of a third processing example of the audio management task T1 and the CD reading task T4 for realizing such a method of using the temporary buffer TB1 for storing read data. FIG. 7 is a flowchart showing a third processing example performed by the voice management task T1 and the CD reading task T4, which is executed when a voice output request is made in the navigation operation task or the like. Note that the same processes as those in the first processing example are denoted by the same reference numerals, and description thereof will be omitted.

The third processing example corresponds to step S of the first processing example.
6 to step S10, step S11, step S12
Alternatively, the configuration is changed to the process of step S13. If it is determined in step S3 that there is no free space in the read data storage temporary buffer TB1, the process proceeds to step S1.
0, the process proceeds to step S11, step S12 or step S13.

In step S10, the unit speech having a strong delimiter is searched using the delimiter type data of the preceding unit speech ("phrase", "reading point", "phrase boundary", "no delimiter"). Is deleted from the read data storage temporary buffer TB1 in step S3.
Return to

Alternatively, in step S11, a unit voice having a weak connection strength is searched for using the connection strength data with the preceding unit voice, and the corresponding unit voice data is deleted from the read data storage temporary buffer TB1. S
Return to 3.

Alternatively, in step S12, use location data indicating a location where a unit voice is used in a sentence is used,
The unit voice is searched for in the order of "end of sentence", "in the middle of sentence", and "head of sentence", and the corresponding unit voice data is deleted from the read data storage temporary buffer TB1, and the process returns to step S3.

Alternatively, in step S13, a unit voice having a large size is searched using the size data of the unit voice, and the corresponding unit voice data is deleted from the read data storage temporary buffer TB1, and the process returns to step S3.

It is to be noted that a calculation is performed by combining these data to determine the priority to be left in the temporary buffer for reading data storage TB1. Is deleted from the read data storage temporary buffer TB1, and the
The process may return to.

As described above, according to the third processing example, the read data storage temporary buffer TB1 is used up to its full capacity, and the read data storage temporary buffer TB is used.
Since the storage state is maintained such that unit audio data that increases the usage efficiency of the unit 1 remains, the temporary buffer TB1 for storing read data is used efficiently, the unit audio data is read out quickly, and audio reproduction is performed. It can be done smoothly.

Next, a fourth processing example of the voice management task T1 and the CD reading task T4 will be described. FIG. 8 is a flowchart showing a fourth processing example performed by the voice management task T1 and the CD reading task T4, which is executed when the voice synthesizing apparatus is started. Note that the same processes as those in the first processing example are denoted by the same reference numerals, and description thereof will be omitted.

The fourth processing example is different from the first processing example in that a write processing (steps S21 to S23) to the temporary buffer TB1 for accumulating read data at the time of starting the apparatus is added. Instead of the first processing example, it is also possible to add the writing processing to the read data storage temporary buffer TB1 of the present processing example at the time of starting the apparatus in the second processing example or the third processing example (voice synthesis apparatus). After the activation, when the voice output request is made in the navigation operation task or the like, the first processing example, the second processing example,
A processing example or a third processing example is executed).

In step S21, the system startup processing
For example, a process of clearing a memory or the like, a process of reading initial data, a process of assigning each temporary buffer to a memory, and the like are performed, and the process proceeds to step S22. In step S22,
It is determined whether there is unit audio data having a very high priority, that is, unit audio data to be resident (unit audio data to be resident and not stored in the read data storage temporary buffer TB1), If it is determined that there is, the process proceeds to step S23.
Move to 1. In step S23, the unit audio data to be resident is written into the read data storage temporary buffer TB1, and the process returns to step S22.

As described above, according to the fourth processing example, since the high-priority unit audio data is always stored in the temporary storage buffer TB1 for reading data immediately after the start of the voice synthesizing apparatus, the reading is performed. The data storage temporary buffer TB1 is used efficiently, so that the unit audio data can be read out quickly and audio can be reproduced smoothly.

Next, a fifth processing example of the voice management task T1 and the CD reading task T4 will be described. FIG. 9 is a flowchart illustrating a fifth processing example performed by the voice management task T1 and the CD reading task T4, which is executed when the voice synthesizing apparatus is started. Note that the same processes as those in the first processing example are denoted by the same reference numerals, and description thereof will be omitted.

The fifth processing example is different from the first processing example in that the temporary buffer TB1 for storing read data when the apparatus is started up.
Process for writing to (Steps S24 to S26)
However, in another processing example, instead of the first processing example, when the device is activated in the second processing example or the third processing example, the read data storage temporary buffer TB1 of this processing example is stored in the temporary buffer TB1. It is also possible to add a writing process (after the voice synthesizing device is started, the first processing example, the second processing example, or the third processing example is executed when a voice output request is made in a navigation operation task or the like).

In step S24, the system startup processing
For example, a process of clearing a memory or the like, a process of reading initial data, a process of allocating each temporary buffer to a memory, and the like are performed, and the process proceeds to step S25. In step S25,
It is determined whether there is an empty area in the temporary buffer TB1 for storing read data.
The process proceeds to step S1 if it is determined that the process is not performed. In step S26, the higher priority unit audio data (unit audio data not written in the read data storage temporary buffer TB1) is written in the read data storage temporary buffer TB1, and the process returns to step S24.

As described above, according to the fifth processing example, immediately after the voice synthesizing apparatus is started, the read-out data storage temporary buffer TB1 is in a state in which the high-priority unit voice data is fully stored. The temporary buffer TB1 for storing read data is used efficiently, so that the unit audio data can be read quickly and the sound can be reproduced smoothly.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a navigation system equipped with a speech synthesizer according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing an outline of a speech synthesis process.

FIG. 3 is a configuration diagram showing an audio data storage configuration of a CD-ROM.

FIG. 4 is a flowchart illustrating a first processing example performed by a voice management task and a CD reading task.

FIG. 5 is a flowchart showing a second processing example performed by the voice management task and the CD reading task.

FIG. 6 is a configuration diagram showing an audio data storage configuration of a CD-ROM.

FIG. 7 is a flowchart illustrating a third processing example performed by the voice management task and the CD reading task.

FIG. 8 is a flowchart illustrating a fourth processing example performed by the voice management task and the CD reading task.

FIG. 9 is a flowchart illustrating a fifth processing example performed by the voice management task and the CD reading task.

[Explanation of symbols]

 5 Navigation microcomputer 6 Operation switch 7 Display 9 Memory T1 Voice management task T2 Voice conversion task T3 Voice output task T4 CD read task TB1 ... temporary buffer for storing read data TB2 ... temporary buffer for audio output

Claims (22)

[Claims] 1. A voice synthesizing apparatus for reading and combining a plurality of unit voice data from a storage medium in which voice data is readably stored for each unit voice data to generate sentence voice data and play back the voice, wherein the unit voice A voice synthesizing apparatus, wherein priority data indicating a priority is assigned to each data. 2. A data reading means for reading the unit audio data from a storage medium in which the audio data is readable and stored for each unit audio data and writing the unit audio data into a cache memory, and the unit audio data stored in the cache memory And voice conversion means for generating text voice data by combining a plurality of the read unit voice data and converting the voice into voice data, and leaving the voice data in the cache memory based on the priority data assigned to the unit voice data. 2. The voice synthesizing apparatus according to claim 1, further comprising a priority processing unit for determining a priority of said unit voice data. 3. The priority reading means for reading the priority data from the storage medium, wherein the priority data is stored in the storage medium beforehand assigned to each of the unit audio data. 3. The speech synthesizer according to claim 2, further comprising: means; and priority calculating means for calculating a priority based on the priority data read by the priority reading means. 4. A frequency detecting means for detecting a reading frequency of the unit audio data stored in the cache memory, wherein the priority processing means calculates a priority based on the frequency detected by the frequency detecting means. 3. The speech synthesizer according to claim 2, further comprising priority calculation means. 5. In the case where the priority data is stored in the storage medium by being added in advance for each of the unit audio data, the priority processing unit reads the priority data from the storage medium Means, frequency detection means for detecting the frequency of reading the unit audio data stored in the cache memory, and the priority data read by the priority reading means and the frequency detected by the frequency detecting means 3. The speech synthesizer according to claim 2, further comprising a priority calculating means for calculating a priority based on the priority. 6. The speech synthesis device according to claim 4, wherein the frequency data detected by the frequency detection means is stored in a backup memory whose stored contents are retained even in a power-off state. apparatus. 7. A voice synthesizing apparatus which reads out a plurality of unit voice data from a storage medium in which voice data is readablely stored for each unit voice data, generates sentence voice data by combining the voice data, and reproduces the voice, wherein: A speech synthesizing apparatus, wherein delimiter type data indicating a first delimiter type in audio reproduction is assigned to each data. 8. A data reading means for reading the unit audio data from a storage medium in which the audio data is readable and stored for each unit audio data and writing the unit audio data to a cache memory, and the unit audio data stored in the cache memory And voice conversion means for generating text voice data by combining a plurality of read unit voice data and converting the voice into voice data, and leaving the cache memory based on the division type data added to the unit voice data. 8. The speech synthesizer according to claim 7, further comprising priority processing means for determining a priority order of said unit audio data. 9. A voice synthesizing apparatus which reads out a plurality of unit voice data from a storage medium in which voice data is readable and stored for each unit voice data, generates sentence voice data and reproduces voice, wherein A speech synthesizer characterized by adding connection strength data indicating the strength of connection with the immediately preceding unit sound in sound reproduction to each piece of data. 10. A data reading means for reading the unit audio data from a storage medium in which the audio data is readable for each unit audio data and writing the unit audio data to a cache memory, and the unit audio data stored in the cache memory And voice conversion means for generating text voice data by combining a plurality of read unit voice data and converting the voice into voice data, and leaving the cache memory based on the connection strength data given to the unit voice data. 10. The speech synthesizing apparatus according to claim 9, further comprising a priority processing means for determining a priority of said unit audio data. 11. A voice synthesizing apparatus which reads out a plurality of unit voice data from a storage medium in which voice data is readably stored for each unit voice data, generates sentence voice data and reproduces the voice, and wherein the unit voice includes: A voice synthesizing apparatus, wherein use position data indicating a use position in a sentence is assigned to each data. 12. A data reading means for reading the unit audio data from a storage medium in which the audio data is readable and stored for each unit audio data and writing the unit audio data to a cache memory; Voice conversion means for generating text voice data by combining the read unit voice data, and converting the voice data into voice data, and leaving the voice data in the cache memory based on the use position data given to the unit voice data. 12. The speech synthesizer according to claim 11, further comprising priority processing means for determining a priority of said unit voice data. 13. A voice synthesizing apparatus which reads out a plurality of unit voice data from a storage medium in which voice data is readable and stored for each unit voice data, generates sentence voice data and reproduces the voice, and wherein the unit voice is reproduced. A voice synthesizing apparatus, wherein reproduction time data indicating a voice reproduction time is assigned to each data. 14. A data reading means for reading the unit audio data from a storage medium in which the audio data is readable for each unit audio data and writing the unit audio data to a cache memory; A voice converting means for generating sentence voice data by combining the read and read plurality of unit voice data and converting the voice into voice, based on the playback time data given to the unit voice data, the shorter the playback time, the higher the priority. 14. The voice synthesizing apparatus according to claim 13, further comprising priority processing means for performing a process of keeping said unit voice data in said cache memory. 15. A data read means for reading the unit audio data from a storage medium in which the audio data is readable for each unit audio data and writing the unit audio data to a cache memory; Voice conversion means for generating text voice data by combining read out plurality of read unit voice data and converting the voice into voice, and at the time of starting the apparatus, the cache based on the priority data assigned to the unit voice data 2. The speech synthesizer according to claim 1, further comprising a start-up writing unit for writing the unit voice data into a memory. 16. The apparatus according to claim 15, wherein said startup writing means writes the unit audio data in descending order of priority until the cache memory becomes full when the apparatus is started. Voice synthesizer. 17. The cache memory is divided into a resident area in which data is not deleted after the apparatus is started and a temporary area in which data is also deleted. 16. The speech synthesizer according to claim 15, wherein high-priority unit speech data is written into the resident area at the time of startup. 18. An audio data storage medium in which audio data is stored so as to be readable for each unit audio data, wherein priority data indicating a priority is assigned to each unit audio data and stored in advance. Audio data storage medium. 19. An audio data storage medium in which audio data is stored so as to be readable for each unit audio data, wherein delimiter type data indicating a forward delimiter type in audio reproduction is stored for each unit audio data. An audio data storage medium characterized by the following. 20. In an audio data storage medium in which audio data is stored in a readable manner for each unit of audio data, connection strength data indicating the strength of connection with the immediately preceding unit audio for audio reproduction for each of the unit audio data. An audio data storage medium, characterized by having stored therein. 21. A sound data storage medium in which sound data is stored in a readable manner for each unit sound data, wherein use position data indicating a use position in a sentence is stored for each unit sound data. Audio data storage medium. 22. An audio data storage medium in which audio data is stored so as to be readable for each unit audio data, wherein size data indicating a data length is stored for each unit audio data. Medium.