JP3838193B2 - Text-to-speech device, program for the device, and recording medium - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
  The present invention relates to a text reading device, a program for reading a text, and a recording medium on which the program is recorded.
[0002]
[Prior art]
  We have already known a text-to-speech device that synthesizes speech signals based on text data such as e-books, and that can adjust the volume and speed while reading aloud text.Ising.
[0003]
  For example, in Patent Document 1, a display input device in which a transparent touch panel is integrally formed on a display screen is used, and the operator performs a tracing operation of tracing the touch panel with a finger or the like, thereby expressing the sound production speed and volume. A text-to-speech device is described that can synthesize speech by reflecting parameters and thereby obtain a highly understandable synthesized speech in accordance with the operator's intention.
[0004]
  In addition to the above, changes in reading conditions during reading include, for example, tone changes, voice quality changes, and gender changes.
[0005]
[Patent Document 1]
  JP-A-9-265299 (abstract, paragraph (0009), paragraph (0010), FIGS. 1 and 2)
[0006]
  By the way, when changing the reading condition during reading aloud, it takes time to reflect the change operation on the reading in real time and even if it is changed, that is, there is a time difference between input and output. There are cases.
[0007]
  For example, in the case of volume control, it is only necessary to adjust the gain of the amplifier for audio output, so it can be adjusted after the audio conversion process, but when speed adjustment is performed, the change is made to reflect it in the output Since voice conversion processing based on speed is required, it cannot be reflected in the output voice unless it is performed before the voice conversion processing.
[0008]
  As described above, there are some changes in the reading conditions in the text-to-speech device that can be adjusted after the speech conversion process. For example, changes in volume and effects (tone (high / low), echo, frequency, etc.) can be adjusted even after the voice conversion process, but changes in speed, voice quality (gender (sex), speaker) Since the voice conversion processing based on this is required, it is not reflected in the output voice unless it is before the voice conversion processing.
[0009]
  FIG. 4 is a timing chart showing the voice conversion process and the actual voice output timing in the reading apparatus. As shown in the figure, there is a time difference between the voice conversion process and the actual voice output timing. This means that if, for example, the reading speed is changed during the text reading, a time difference occurs until the reading is output at the changed speed, so that the reading becomes unnatural.
[0010]
  This point will be further described with reference to FIGS.
[0011]
  FIG. 5 is a front view showing an example of a reading device which is not based on the prior art document but is a prerequisite technology of the text reading device of the present invention. On the screen of the text-to-speech reading apparatus 1, a slide bar 4 for speed, speed change and volume change, a speaker 5 and the like are displayed together with a button 3 for reading (start of reading) and a stop. A transparent touch panel 2 is disposed on the entire surface of the screen, and the touch panel 2 and the display content of the screen are associated with each other.
[0012]
  FIG. 6 is a block diagram schematically showing the configuration of the text-to-speech device 1 shown in FIG.
[0013]
  The CPU 10 of the text-to-speech reading apparatus 1 includes a display means 50 such as a touch panel 2, for example, a liquid crystal display 6, an electronic book data to be read out, a RAM 30 in which volume values, speed values, and the like are recorded, a speech synthesis program, A ROM 40 storing word dictionary data, synthesized phoneme data, and the like used for document analysis and speech synthesis for speech synthesis is connected. Further, the CPU 10 receives an input from the touch panel 2 and checks and identifies the operation content, and receives a signal from the tap unit 11 and reads and changes the speed value recorded in the RAM 30. The speed value is transferred to the voice conversion processing unit 14, and the speed control unit 12 that updates the speed value of the RAM 30 to the changed speed value, and similarly receives the signal from the tap unit 11 and reads the volume value recorded in the RAM 30. The volume control unit 13 that controls the amplifier 60 with the changed volume value and outputs it from the speaker 5 and updates the volume value of the RAM 30 with the changed volume value, and reads the electronic book data recorded in the RAM 30, Speech synthesis is performed using a speech synthesis program and dictionary data stored in the ROM 40, and an output signal from the speed control unit 12 The voice conversion processing unit 14 that outputs to the amplifier 60 at a voice speed based on it, and the display control unit 15 that controls to display the text data read from the book data stored in the RAM 30 on the display means 50 such as the liquid crystal display 6. It is made up of.
[0014]
  Here, the speech synthesis processing procedure in the speech conversion processing unit 14 will be described with reference to the flowchart shown in FIG.
[0015]
  In step S <b> 101, the CPU 10 performs reverse kanji conversion on the extracted sentence extracted for reading from the electronic book data stored in the RAM 30. That is, reading is given to the extracted sentence. For example, if the extracted sentence is “Long Time,...”, “Kumashi Mashi,...” Can be obtained by reverse kanji conversion. Subsequently, in step S102, an accent type is assigned to the reversely kana-kanji converted character (accent processing). For example, “0” (zero) type is given as an accent type for “Mukashimakashi”.
[0016]
  In step S103, the duration time T of each syllable after the accent type is given in step S102 is calculated by multiplying the utterance speed coefficient α by the syllable duration length L stored in the ROM 40 ( T = α × L).
[0017]
  In step S104, the fundamental frequency of each syllable is calculated, and then the step.TheIn S105, the volume of each syllable is calculated. For example, for a syllable with higher accent, the fundamental frequency is controlled so that the fundamental frequency of the syllable becomes higher (step S104), and the volume is controlled so that the volume becomes larger (step S105). These processes are used to accentuate text reading in response to accents.DoIt is processing.
[0018]
  In step S106, the last character (modified) displayed on the liquid crystal display 6 from the first character of the extracted sentence extracted in step S102 (the last sentence in the sentence displayed on the display screen of the liquid crystal display 6). The expected time (time required for page switching) is calculated by adding the respective duration times calculated in step S103 (characters immediately before the page tag).
[0019]
  In step S107, the language processing dictionary and speech synthesis speech data stored in the ROM 40, the duration of each syllable calculated in step S103, the fundamental frequency of each syllable calculated in step S104, Using each volume calculated in step S105, voice synthesis data of a desired speed and volume of the extracted sentence extracted in step S102 is created, and a voice output device such as the speaker 5 is provided via the amplifier 60. Output to.
[0020]
  In this way, when the last sentence displayed on the display screen is extracted from one or a plurality of sentences displayed on the display screen of the liquid crystal display 6, the liquid crystal display is displayed from the beginning of the last sentence. Assuming the time required to read up to the last character displayed on the display screen 6, the display screen of the liquid crystal display 6 is displayed when the estimated time has elapsed since the start of reading the last sentence. The display content is switched, and the timing for switching the display content on the display screen of the liquid crystal display 6 is controlled to match the timing when the display content is read out.
[0021]
  In the text-to-speech reading device 1 shown above, when the volume adjustment is performed by operating the lower slide bar 4 in the example of FIG. 5, the operation of the touch panel 2 causes the tap unit 11 to adjust the volume. According to the instruction, the volume control unit 13 reads out and changes the volume value from the RAM 30 and adjusts the volume by directly controlling the amplifier 60 based on the changed volume value. Write the changed volume value in the value area. As described above, the sound volume changing operation can be performed without the voice conversion processing unit 14, and can be reflected in the output in real time.
[0022]
  On the other hand, in the case of changing the reading speed, the tap unit 11 specifies that the operation is speed adjustment from the operation of the slide bar 4 above the touch panel 2 and notifies the speed control unit 12, and the speed control unit 12 The speed value read from the RAM 30 is changed and sent to the voice conversion processing unit 14, and at the same time, the changed speed is recorded in the RAM 30. The voice conversion processing unit 14 performs voice conversion processing based on the changed signal, controls the amplifier 60, and outputs the voice from the speaker 5 at the changed speed.
[0023]
  In the above processing operation, when the reading speed is changed, as described above, there is a time difference between the voice conversion input and the output, and when the speed change is read out without being reflected in the output in real time, a sense of incongruity remains. Therefore, once the stop button 3 is operated to stop the voice conversion process, the change is performed.
[0024]
  Therefore, when the user wants to change the reading conditions during the text reading (including function setting), the user determines whether or not the change should be made by stopping the voice conversion process, and for the reading or according to the determination. The stop button 3 must be operated. Specifically, youTo theThe current reading speed is not appropriate (for example, it is difficult to hear).TheWhen it is desired to change the reading speed, the stop button 3 is operated to stop reading, then the speed change slide bar 4 is operated, and further, reading is started to start reading. The up button 3 is operated. Alternatively, when the user wants to change the volume when reading is not performed, the user operates the slide bar 4 for changing the volume and further operates the reading button 3 for starting reading. Many such operations are problematic because they are cumbersome and inconvenient.
[0025]
[Problems to be solved by the invention]
The above-described user operations are complicated and inconvenient.
[0026]
  The present invention solves the above problems (complex and inconvenient operations).InThe purpose is to allow the user to easily change the reading conditions.
[0027]
[Means for Solving the Problems]
  The invention of claim 1 is a sentence reading apparatus capable of changing a sentence reading condition, and means for inputting change information for changing the sentence reading condition;When the change in the text-to-speech reading condition due to the change information is reflected in the voice output with a delay, the change processing is performed by controlling the reading stop so that the change in the text-to-speech reading condition based on the change information is not delayed in the voice output. Change the text-to-speech condition to be changed without controlling the reading stop when it is reflectedA sentence reading apparatus comprising: means; and means for controlling reading based on the changed sentence reading condition. That is,Whether changes in text-to-speech reading conditions are delayed and reflected in the audio outputRead aloudStop systemAre associated with each other.
[0028]
  The invention of claim 2 is a text reading apparatus according to claim 1,The text-to-speech condition changing means reflects the change of the text-to-speech reading condition based on the change information without delay in the voice output, and when the text-to-speech reading is stopped, the reading process is controlled to perform a change process. It is characterized byTheChanges in text-to-speech conditions are delayed to voice output Whether or notWhether the text is being read out or not, and start or stop readingcontrolAnd associateEtIt is.
[0029]
  The invention of claim 3 is the text-to-speech device according to claim 1 or 2,The sentence reading condition changing means is:Change informationButRead aloudWhen it is volume or tone change information, it is determined whether or not the text-to-speech device is reading aloud, and if it is not being read out, it is controlled to start reading, and the volume or tone is changed. Specific changes such asStart readingcontrolAnd relatedWithTheEtIt is.
[0030]
  The invention of claim 4 is claimed in claimAny one of 1-3In the text-to-speech device described in 1.When stop control is performed by the text-to-speech condition changing means, control is performed so that the text is read from the head of the text to be read. If there is a change in the text-to-speech condition, the text is read from the head.
[0031]
  The invention of claim 5In order to read the text so that the text-to-speech reading condition can be changed, if the change in the text-to-speech condition due to the input change information is delayed and reflected in the voice output, it is controlled to stop reading and change processing, When the change of the text-to-speech reading condition by the change information is reflected without delay in the voice output, a procedure for performing the change processing without controlling the reading stop, and a procedure for controlling the reading based on the changed text-to-speech condition , Is a program characterized by being executed. That is, whether or not the change in the text-to-speech condition is delayed and reflected in the voice output is associated with the reading-out stop control.
[0032]
  The invention of claim 6A computer-readable recording medium on which the program according to claim 5 is recorded. According to the sixth aspect of the invention, the same effect as that of the fifth aspect is obtained.
[0033]
[0034]
DETAILED DESCRIPTION OF THE INVENTION
  Embodiments of the present invention will be described with reference to the accompanying drawings.
[0035]
  FIG. 1 shows an embodiment of a text-to-speech device 1 according to the present invention.
[0036]
  In this embodiment, a reading start control unit 16 and a reading stop control unit 17 are added to the CPU 10 of the text reading device 1 shown in FIG.
[0037]
  Both the reading start control unit 16 and the reading stop control unit 17 receive a signal from the tap unit 11, and the reading start control unit 16 starts and controls the speech conversion processing unit 14 when the text reading device 1 is not reading, The reading stop control unit 17 controls to stop the voice conversion processing unit 14 while reading a sentence. Further, the RAM 30 is provided with a storage area for a reading flag to indicate whether or not the text is being read out. Other structural functions are the same as those described with reference to FIG.
[0038]
  Next, the change (including the setting) of the text reading condition during reading by the text reading apparatus 1 described above will be described taking the reading volume and speed as an example.
[0039]
  FIG. 2 is a flowchart for explaining a processing procedure (first embodiment) in the text-to-speech reading apparatus.
[0040]
  When changing the volume and speed of the text-to-speech reading device 1, first, it is detected that a part of the touch panel 2 has been tapped by the user (S201, YES). When the tap unit 11 receives a signal from the touch panel 2 and checks its contents, and determines (specifies) that the operation is for starting reading (reproduction) by the reading button 3 (S202, (YES), reading (reproduction) is started from the head of the sentence stored in the ROM 30 (S203) and reading is performed as it is.
[0041]
  When it is determined (specified) that the tap unit 11 is not an operation for starting reading (reproduction) (S202, NO) but an operation for volume change control by the volume change slide bar 4 (S204, YES) The reading start control unit 16 checks the “reading flag” indicating whether or not reading is being performed (reading mode) recorded in the RAM 30 (S205), and if reading is stopped (flag is 0) ( (S206, YES) The reading starts from the beginning of the sentence stored in the RAM 30 (S207). If the reading is in progress (S206, NO), the volume control unit 13 reads the volume value from the RAM 30 in the reading state. The volume is adjusted with the changed volume value (S208), and the changed volume value is written in the RAM 30 (S209) and the amplifier 6 Voice output at the volume is changed from the speaker 5 via a. Here, the user adjusts the volume while listening to the reading volume by using the slide bar 4 for changing the volume, and repeats the processes of steps S208 and S209 until the necessary setting is completed, and if the setting is completed (S210, YES), the process is performed. finish.
[0042]
  In step S204, the tap unit 11 specifies (determines) that the tap operation on the touch panel 2 is not an operation for changing the volume (S204, NO) but an operation for changing the speed by the speed change slide bar 4. ) (S211, YES), the reading stop control unit 17 checks the “reading flag” recorded in the RAM 30 (S212), and if reading is in progress (reading mode) (S213, YES), stops reading. At the same time, the “reading flag” recorded in the RAM 30 is reset (for example, changed from “1” to “0”) (S214). If it is not being read out (reading mode) (S213, NO), the speed controller 12 reads the speed value from the RAM 30, adjusts the speed with the input speed value (S215), and combines with the changed speed value. Audio is output from the speaker 5 via the amplifier 60, and the changed (adjusted) speed value is written in the RAM 30 (S216).
[0043]
  Steps S215 to S216 are performed until the speed setting is completed, and when the setting is completed (S218, YES), the reading position is returned to the head of the sentence (that is, the head of the sentence if reading is in progress) ( In step S218, reading is started at the newly set reading speed, and at the same time, the “reading flag” stored in the RAM 30 is set (for example, changed from “0” to “1”) (S219). Thus, the speed control processing procedure is ended, and if there is a sentence to be read out, the reading is continued under the changed condition.
[0044]
  In addition, although explanation is omitted about the case where it is not speed control in Step S211 (S211, NO), for example, change of reading voice other than the slide bar 4 for speed change, change of male and female voices, etc. are changed after reading is stopped once. The processing procedure in the case where it is better to perform is similar to the speed control, and the tone change or the like that may not be adjusted properly unless it is being read out is executed in the same procedure as the volume change.
[0045]
  When an operation to change the reading condition is made during reading aloud, it is necessary to change or set the reading condition. If there is a need to stop it, it will automatically stop reading, and it can only be set during reading, like changing or setting the volume or tone. For those that cannot be read out, even when reading is stopped, reading can be started automatically, thereby reducing the burden on the user when changing the reading conditions.
[0046]
  FIG. 3 is a flowchart for explaining another processing procedure (second embodiment) in the text-to-speech reading apparatus 1.
[0047]
  When changing the volume and speed of the text-to-speech reading device 1, first, it is detected that a part of the touch panel 2 has been tapped by the user (S301, YES), and the tap unit 11 checks the content of the tap. When it is determined that the operation is to start reading (reproduction) by the button 3 for reading (S302, YES), reading is started (S303) and reading is performed from the head of the sentence stored in the RAM 30.
[0048]
  When it is determined (specified) that the operation is not the operation for starting the reading (S302, NO) but the speed changing by the slide bar 4 for changing the volume (S304, YES), the reading start control unit 16 Checks the “reading flag” recorded in the RAM 30 (S305), and as a result, if reading is stopped (S306, YES), there is a possibility that the volume cannot be adjusted or cannot be appropriately adjusted without reading.ThereTherefore, reading is started from the head of the text stored in the RAM 30 (S307), and the “reading flag” in the RAM 30 is set (S308). In step S306, if it is being read out (S306, NO), the volume control unit 13 performs volume adjustment by reading the volume value from the RAM 30 and changing it to the input volume value in the state of being read out as it is (S309). The sound is output from the speaker 5 through the amplifier 60, and the changed (adjusted) volume value is written in the RAM 30 (S310).
[0049]
  Here, the user adjusts the volume with the sound-changing slide bar 4 while listening to the reading volume, repeats the processing of steps S309 and S310 until the necessary setting is completed, and if the adjustment (setting) is completed (S311, YES) ) End the process.
[0050]
  In step S304, the tap unit 11 determines that the tap operation on the touch panel 2 is not an operation for changing the volume (NO in S304) but is for changing the speed by operating the slide bar 4 for changing the speed. When the determination is made (YES at S312), when the user performs a speed adjustment operation (S313), the input of the touch panel 2 is specified by the tap unit 11, and the speed control unit 12 is read and changed from the RAM 30. The speed is transferred to the voice conversion processing unit 14 and the changed speed value is written in the RAM 30 (S314). At this stage, the reading stop control unit 17 checks the reading flag recorded in the RAM 30 (S315). As a result of the check, the “reading flag” is being read in the set state (for example, the flag is 1) (reading mode). When it is determined that there is (S316, YES), reading is temporarily stopped (S317) and the “reading flag” is reset. Thereafter, reading is started, and the synthesized voice of the changed speed is output from the speaker 5 through the amplifier 60 and the “reading flag” is set (S318). Thereafter, the processing of steps S313 to S318 is executed until the setting is completed, and if the setting is completed (S319, YES), the processing is terminated, and if there is a sentence to be read out, the reading is continued under the changed condition.
[0051]
  In step S316, if reading is not in progress (S316, NO), the process is terminated after the setting is completed (S319, YES).
[0052]
  Note that the processing in step 312 when it is not the speed-changing slide bar 4 (S312: NO) is the same as in the first embodiment. That is, for example, regarding the change of the reading voice other than the slide bar 4 for changing the speed, the processing procedure in the case where it is better to stop the reading once, such as the change of the voice quality of men and women, the other party, etc. is the same as the speed control. In addition, changes in effects such as tone, echo, frequency, etc. that cannot be adjusted without reading are performed in the same procedure as the change in volume.
[0053]
  In the above-described embodiments, all the processes shown in FIGS. 2 and 3 are performed in one apparatus including the speaker 5 for generating synthesized speech. However, each process and a part of the processes are performed separately. Then, the synthesized speech may be finally generated from the speaker 5 in accordance with the changed condition, and the text may be read out. For example, the first computer only receives user input, and other substantial processing is performed by another second computer. Furthermore, the device is not limited to a dedicated device for reading a text, but may be a PDA, a personal computer, a mobile phone, a car navigation terminal, a TV, or the like having a function other than reading.
[0054]
  The text to be read out is not limited to books, but may be letters (including e-mails), directions, advertisements, and lyrics. The data stored in the RAM 30 is lost when the apparatus is turned off, but may be stored continuously even when the apparatus is turned off.
[0055]
  In the above-described embodiments, both of the automatic start and stop of the reading are performed in accordance with the change of the reading conditions. However, either of the automatic starting or the automatic stopping is performed. May be. The starting position of reading is at the beginning of the sentence, but it may be in the middle of the sentence. Furthermore, the object read out until the condition change is completed may be a dedicated sentence used when the condition is changed.
[0056]
  The processing described above can be executed by the CPU 10 of the text reading device 1 by a program describing the procedure of the processing. In addition, this program is FD (flexible disk), CD-ROM, MO, DVD-In addition to being provided by being recorded on a known recording medium for recording a program such as a ROM, the program can be provided via a network such as the Internet.
[0057]
【The invention's effect】
  According to the invention described in claim 1 of the present application, when changing the reading condition while reading a sentence, the user needs to automatically stop reading without performing the operation for starting or stopping the reading as usual. If you do not want to stop reading, you can stop reading what you need to stop, and if the change in text-to-speech conditions should be done while reading, when the device is not reading Can automatically start reading. Therefore, the user does not have the trouble of operating the touch panel and the like, and the reading conditions can be easily changed.Further, even when the user changes the reading condition with a time difference from the input to the output of the change information such as the reading speed, the user can naturally hear the reading by the synthesized voice without being aware of the time difference.
[0058]
  According to the invention described in claim 2 of the present application, the effect of the invention described in claim 1 is achieved, whether or not the text is being read out and whether or not reading is started or stopped.controlCan be read out satisfactorily.
[0059]
  According to invention of Claim 3 of this application, there exists an effect of the invention of Claim 1 or 2,A specific change such as a volume or a tone is associated with a reading start control, so that quick reading can be performed.
[0060]
  According to the invention of claim 4 of the present application,Any one of 1-3The effects of the invention described inIf the text-to-speech conditions change, the text will be read from the headline, so it ’s goodCan be read aloud.
[0061]
  According to the invention described in claim 5 of the present application,There is no need for the user to operate the touch panel or the like as in the prior art, and the reading conditions can be easily changed. Also, reading speed Even if the reading condition is changed with a time difference from the input to the output of the change information such as the change information, it is possible to naturally hear the reading by the synthesized speech without being aware of the time difference.
[0062]
  According to the invention described in claim 6 of the present application,The same effect as that of the invention described in claim 5 can be obtained, and the above effect can be realized in any information device by causing the computer of the portable device or other information device to read the program.
[0063]
[Brief description of the drawings]
FIG. 1 is a block diagram showing a configuration according to a first embodiment of a text-to-speech device of the present invention.
FIG. 2 is a flowchart for explaining a reading condition changing procedure;
FIG. 3 is a flowchart for explaining another reading condition changing procedure;
FIG. 4 is a time chart showing a time difference between audio conversion processing and audio output.
FIG. 5 is a front view showing an example of a conventional text-to-speech device.
6 is a block diagram showing a schematic configuration of the text-to-speech device shown in FIG. 5. FIG.
FIG. 7 is a flowchart for explaining speech conversion processing of the text-to-speech device shown in FIG. 5;
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... Text-to-speech device, 2 ... Touch panel, 3 ... Button for reading and stopping, 4 ... Slide bar for speed and volume change, 5 ... Speaker, 6 ... Liquid crystal display DESCRIPTION OF SYMBOLS 10 ... CPU, 30 ... RAM, 40 ... ROM, 50 ... Display means, 60 ... Amplifier.

Claims (6)

In a text-to-speech device that can change text-to-speech conditions,
Means for inputting change information for changing the text-to-speech condition;
When the change in the text-to-speech reading condition due to the change information is reflected in the voice output with a delay, the change processing is performed by controlling the reading stop so that the change in the text-to-speech reading condition based on the change information is not delayed in the voice output. A text-to-speech reading condition changing means for performing a change process without reading stop control when it is reflected ,
Means for controlling reading based on the changed text reading conditions;
A text-to-speech device characterized by comprising: In the text-to-speech device according to claim 1,
The text-to-speech condition changing means reflects the change of the text-to-speech reading condition based on the change information without delay in the voice output, and when the text-to-speech reading is stopped, the reading process is controlled to perform a change process. A text-to-speech device characterized by: In the text-to-speech device according to claim 1 or 2,
The text-to-speech reading condition changing means determines whether or not the text-to-speech device is reading aloud when the change information is a reading volume or tone change information. A text-to-speech device characterized by: In the text-to-speech device according to any one of claims 1 to 3 ,
A text-to-speech device that controls to read out from the head of the text to be read when the text-to-speech condition changing means controls the stop . In order to read out the text, the text-to-speech conditions can be changed.
If the change in the text-to-speech reading condition due to the input change information is reflected in the voice output with a delay, it is controlled to stop reading, and the change in the text-to-speech reading condition based on the change information is delayed in the voice output. If it is reflected without change, the procedure for change processing without reading stop control ,
A procedure to control reading based on the changed text-to-speech conditions;
A program characterized by having executed . A computer-readable recording medium having recorded thereon the program according to claim 5.

Images