JPH08146987A - Speech input device and its control method - Google Patents

Abstract

PURPOSE: To change a method for speech detection according to the state of speech input so that effective speech information is passed to a lower order processing. CONSTITUTION: As a mode wherein the start point and end point of a speech part inputted from a speech input device 4 are specified, there are three modes, i.e., a mode wherein they are specified automatically, a mode wherein only the start position is specified manually and the end point is specified automatically, and a mode wherein the start point and end point are both specified manually; and one of them is selected through an input device 3.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice input device and a control method therefor, and more particularly to a voice input device for specifying a voice section in input information and passing it to a subordinate process and a control method therefor.

[0002]

2. Description of the Related Art Generally, a voice input interface performs a process of passing a user's uttered content to another application such as voice recognition. Therefore, it is necessary to detect the voice section uttered by the user from the input voice data.

There are various methods for detecting the voice section, for example, the following method. Detection method 1: The user determines the start point and end point of utterance using an input device such as a mouse or a key. Detection method 2: The user determines the start point of utterance and the computer determines the end point. Detection method 3: The computer determines the start point of utterance and the end point is determined by the user. Detection method 4: The computer determines both the start point and the end point of utterance.

Further, an example of an operation required by the user at the time of voice input corresponding to each of the voice section detecting methods described above is shown below. However, in this case, a mouse is used as the input device. Operation method 1: Press the mouse button before speaking, hold it while speaking, and release the mouse button after speaking. Operation method 2: Press the mouse button once before speaking. Operation method 3: Press the mouse button once after utterance. Operation method 4: No user operation is required.

The conventional voice input interface has only one of the above-mentioned detection methods.

[0006]

Each of the above-mentioned detection methods and operation methods has advantages and disadvantages.

[0007] For example, in the detection method 1,
Since the user gives the end, there are few mistakes in cutting out the voice section. However, the operation is troublesome as shown in the operation method 1.

The detection methods 2 and 3 are slightly less burdensome on the user than the detection method 1, but on the other hand,
A computer may erroneously detect noise as speech, and noise may be included in the detected speech section.

In detection method 4, since the user is not required to operate at all, the usability is improved, but again the uttered voice is not detected or noise is detected as voice due to changes in the voice input environment or noise. It may get chilly.

As described above, in the conventional voice input interface, since only one of the detection methods is used, the disadvantage of the detection method remains a problem of the voice interface.

[0011]

[Means for Solving the Problems] and

The present invention has been made in view of the above problems, and makes it possible to change the detection method according to the situation of voice input, and a voice input device for passing effective voice information to lower processing and a control method thereof. Is to provide.

To solve this problem, for example, the voice input device of the present invention has the following configuration. That is, a voice input device that passes the voice information input from the voice input means to a lower process, and a plurality of modes for specifying the period of significant voice information input from the voice input means,
And a selecting means for selecting one from the plurality of modes.

According to a preferred embodiment of the present invention, the modes include a first mode for automatically identifying a start point and an end point of a voice, a second mode for manually instructing a start point position, a start point and an end point. A third mode for manually indicating position is preferably included. As a result, it becomes possible to cover all modes depending on the situation in which the user is placed.

Further, it is preferable that the selecting means selects the first mode at an initial stage, and selects the second mode and the third mode in this order when a predetermined change instruction is given. As a result, in a normal use environment, a mode that does not burden the user at all is selected. Further, when the usage environment changes, the mode is appropriately changed according to the change in the environment.

The selecting means may be manually selected. According to this, it becomes possible to immediately adapt to the environment in which the user is placed.

[0016]

Embodiments of the present invention will be described in detail below with reference to the accompanying drawings. FIG. 1 is a block diagram showing a schematic configuration of a first embodiment of a computer which is an information processing apparatus according to the present invention.

In the figure, reference numeral 1 is a system bus, on which a display device 2 such as a CRT display,
An input device 3 such as a keyboard or a mouse, a voice input device 4 such as a microphone, and an I / O for converting a voice signal supplied from the voice input device 4 into data that can be processed by a computer.
A device 5 and a CPU 6 for controlling the operation of the entire system are connected. In the CPU 6, a ROM storing a program for realizing the processing according to a flowchart described later and a RAM used as a work area
It is assumed that it has a built-in main memory.

In the system having the above configuration, for example, a mouse is used as the input device 3 in the embodiment. Considering the case where there are three types of voice input modes, the operation required by the user in each mode and the method for detecting the voice section corresponding to each mode are as follows.

Voice Input Mode Mode 1: No user operation is required. Mode 2: The user presses the mouse once before speaking. Mode 3: The user presses the mouse before speaking, keeps pressing the mouse during speaking, and releases the mouse after speaking.

Therefore, the voice section detection method in each mode is as follows. Mode 1: The computer detects the beginning and end of the voice section. Mode 2: The user determines the beginning of the voice section and the computer detects the end. Mode 3: The user determines the start and end of the voice section.

The operation of the system will be described with reference to the flowchart of FIG.

First, in step S1, a voice input mode is selected. FIG. 3 shows an example of this voice input mode selection processing.

First, in step S11, the selectable voice input mode is displayed, and in step S12, the user waits for selection. The user selects a desired input mode using the mouse (input device 3). When the user selects one of the voice input modes, the user is notified to confirm the selection input mode in step S13, and the selected content is stored in a predetermined storage area (a predetermined address of the main storage memory in the CPU 6 in the embodiment). Keep it in memory.

As described above, when the voice input mode is selected, the process proceeds to step S2 of FIG. 2, the time t is set to 0, and the variable St indicating the detection status of the start / end of the voice section is set. , Is set to NOTYET indicating that neither the start edge nor the end edge has been detected. The time t is set to "0" to reset a timer (not shown), and the variable St is secured at a predetermined address in the main memory of the CPU 6. After this initial setting process, the process proceeds to the voice capturing process of step S3 and the voice data analyzing process of step S5.

In the voice capturing process (step S3),
The voice uttered by the user is converted into data that can be processed by the computer by using the voice input device 4 such as a microphone and the I / O device 5, and is taken into the computer. Next, in the voice data holding process (step S4), the voice data fetched in step S3 is held in the storage device 7. After holding the voice data, the process returns to the voice capturing process again, and the same operation is repeated to continue the voice capturing and the voice data holding.

In parallel with the steps S3 and S4, the voice data analyzing process of the step S5 is performed, and the voice data stored in the storage device 7 in the step S4 is converted into data of a predetermined time width Δt. To analyze the voice data. An example of the voice data analysis process is shown in the flowchart of FIG.

In FIG. 4, in step S51, the voice data from time t to time t + Δt is read from the storage device 7. In a succeeding step S52, a root mean square value of the read voice data is calculated and obtained as P (t).

Returning to FIG. 2, in the following step S6, the voice section corresponding to the voice input mode set in step S1 is detected. Here, when the user selects mode 1 as the voice input mode in step S1, an example of voice section detection is shown in the flowchart of FIG. 5, and an example of voice section detection when mode 2 is selected is shown in the flowchart of FIG. An example of voice section detection when mode 3 is selected is shown in the flowchart of FIG.

First, the voice section detection processing when mode 1 is selected will be described.

In FIG. 5, a voice non-input judgment step S
In step 61, when St indicating the state of voice section detection is NOTYET (when the start end of the voice section is not determined), the process proceeds to step S621, and it is determined whether or not it is the start point of the voice section. If the result is not NOTYET, the process proceeds to step S64.

In step S621, the root mean square value P (t) of the voice data from time t to time t + Δt, which is the previously calculated root mean square value calculation result, and a threshold value for determining a predetermined start point of the voice. Compare Tps, P
If (t) is larger than the threshold Tps, step S63.
Go to and decide the starting point. Here, St is changed to “IN” indicating that the beginning of the voice section is detected (that is, the voice is input), the time t is set to the start time STime of the voice section, and the voice section detection process ends. . If it is determined in step S621 that P (t) does not exceed the threshold Tps, the voice section detection process ends.

On the other hand, when the process proceeds to step S64 and it is determined that the voice input period is in progress, when St is IN (when the start end of the voice section is determined), the end of the voice section is set. The process proceeds to determination end detection step S651, and if St is other than IN, the voice section detection process ends. In the end detection process of step S651, P (t) calculated in step S52 is compared with a predetermined threshold Tpe for determining the end of voice. P
If (t) is smaller than Tpe, the process proceeds to step S66, and if P (t) is greater than or equal to the threshold Tpe, the voice section detection process ends.

In step S66, St is changed to END indicating that the end of the voice section has been detected (that is, the voice input has ended), and the time t is set as the end time ETime of the voice section, and the voice section detection processing is performed. To finish.

Next, the voice section detection process when the mode 2 is selected in step S1 in FIG. 2 will be described with reference to the flowchart of FIG.

The difference from the case of mode 1 is the processing when St indicating the state of voice section detection in the voice non-input judgment step S61 is NOTYET. The same reference numerals as those in FIG.

In mode 2, if St is NOTYET in the voice non-input determination process in step S61, the process proceeds to step S622 to determine whether or not the user has pressed the mouse to notify the start of the voice section, and the start input detection process is performed. . In this step, it is checked whether or not the user has pressed the mouse, which is a signal to start the voice section, and if the user has pressed the mouse, the process proceeds to step S63 to determine the start end, and if the mouse has not been pressed, , The processing of voice section detection ends. For the other steps, the same processing as in mode 1 is performed.

Next, the voice section detection process when the mode 3 is selected in step S1 of FIG. 2 will be described with reference to the flowchart of FIG.

The difference from the case of mode 2 is step S
This is a process in the case where St indicating the state of voice segment detection is IN in the voice input period determination process in 64.

In step S64, if St is IN, the process proceeds to step S652 to determine the end of the voice section, and if St is other than IN, the voice section detection process is ended.

In step S652, it is checked whether or not the mouse button has been released (from the pressed state),
If the mouse has been released, the process proceeds to step S66 to determine the end, and if the mouse has not been released (the user continues to press the mouse), the voice section detection process ends. The other steps are the same as those in the mode 2.

Returning to FIG. 2 again, after the voice section detection process is completed as described above, the process proceeds to step S7, and a voice input end determination is performed. Here, if St is END, the process proceeds to step S9, the voice section is displayed, and St is EN.
If it is other than D, the process proceeds to step S8, the analysis time is updated, and the time t is increased by Δt. And step S
Return to 5 to continue the voice data analysis.

In the voice section display process in step S9, the voice section detected in step S66 (time ST
(from time to time ETime) is displayed on the display device 2, and then voice data is extracted.

When the embodiment described above is applied to the case where the user inputs a voice in an environment where sudden noise is generated, the following is obtained.

First, at the same time when the program starts, step S
In the voice input mode selection process No. 1, selectable input modes are displayed on the display device 2 (step S11), and the user waits for selection (step S12). Here, it is assumed that the user selects the mode 1, for example. In the following step S13, the user is informed that the selected mode is the mode 1 by displaying it on the display device 2. FIG. 8 shows an example in which the selection of the voice input mode is realized on the display device 2.

In the figure, each voice input mode is named. Mode 1 is Keep Pressing
SpeechInput, Mode 2 is One Click SpeechInput, and Mode 3 is Hand Free SpeechInput. In this figure,
This indicates that Hand Free Speech Input corresponding to mode 3 is selected. 9 shows an example displayed on the display device 2 to inform the user of the selected mode in step S13. As shown in the figure, the current mode is notified to the user by clearly indicating the selection mode at the top of the screen.

Then, the process proceeds to the initialization process of step S2, and the time t and the voice detection status St are initialized. Succeedingly, voice capturing and voice data holding in the storage device 2 are started by voice voice capturing in step S3 and voice data holding in step S4, and steps S3 and S4 are repeated repeatedly. At the same time, in the voice analysis process in step S5, the voice data held in the storage device 2 is analyzed with a predetermined time width Δt, and the process proceeds to the voice section detection process in step S6.

Since the example in which the mode 1 is selected as the voice input mode has been described, the voice section detection shown in the flowchart of FIG. 5 will be performed. When the voice section detection process of step S6 is completed, the voice section detection status St is checked, and voice data analysis processing is performed for each time width Δt until the end of the voice section is detected (until the voice section is considered to end). S5) and the voice section detection process (step S6) are repeated.

For the sake of explanation, after the user selects the voice input mode and before the user generates, a sudden noise (a loud sound whose root mean square value exceeds the threshold Tps instantaneously) occurs from time tn to time tn2 ( It is assumed that this occurs over tn2> tn). In this case, in the voice section detection process (step S6) for processing the voice data at the time tn, since the root mean square value of the sudden noise is larger than the threshold value Tps, the start point of the noise is detected as the start point of the voice section, and the step S63 is performed. Then, the process proceeds to the start edge detection process. In step S63, St is changed to IN, and the time STime at the start of the voice section is set to tn.

After that, in the end detection processing (step S651) of the voice section detection processing (step S6) at time tn2 when the sudden noise weakens, P (t) becomes smaller than the threshold value Tpe because the noise weakens. Therefore, the termination is determined (step S66). That is, st is END
And tn2 is the time ETime at the end of the voice section.

The computer mistakenly recognizes sudden noise as voice,
The time when noise is generated (between time tn and tn2) is detected as a voice section, and the voice section is displayed in step S9. From this display, the user can confirm that the wrong voice section has been detected. FIG. 10 shows an example in which the voice input mode is set to mode 1 and voice input is performed, and a voice section is erroneously detected due to sudden noise. FIG. 10 shows an example of the waveform display unit of the voice input interface. Among the two white display units in the center, the upper display unit shows the original waveform of the voice and the lower display unit displays the root mean square value. are doing. Further, in the upper display section, the area A is the detected voice section.

In the second execution (because of the previous execution, it is known that the detection of the start edge of the voice section in mode 1 is erroneously detected in the environment where sudden noise is generated), and the voice input in step S1 is performed. In the mode selection process, the user selects mode 2 or the like.

The subsequent processing is the same as when the mode 1 is selected, and the different point is the processing performed in the voice section detection.
Since the mode 2 is selected as the voice input mode in step S12, the voice section detection corresponding to the mode 2 shown in the flowchart of FIG. 6 is performed.

Similarly to the previous time, it is assumed that sudden noise is generated again from time tn to time tn2 (tn2> tn). Unlike the previous case, in the voice section detection process (step S6) for processing the voice data at time tn, step S62
From the detection of the start input of No. 2, no matter how loud the noise (or voice) is, it does not mean that the start of the voice section is detected until the user presses the mouse. That is, since the time point when the user presses the mouse is set as the start end of the voice section, there is no possibility of erroneously determining the start end of the voice section in response to noise as in the previous time using mode 1. The process after detecting the start of the voice section is the same as the previous process.

FIG. 11 shows an example in which the voice input mode 2 is set in an environment where sudden noise is generated, the user speaks after pressing the mouse, and the voice section is correctly detected. In FIG. 11, a sudden noise was input before the user generated it, but it was not detected as a voice section because the user did not press the mouse.

As described above, when sudden noise is input, it can be dealt with by selecting the voice input mode as mode 2.

However, in an environment in which noise that constantly exceeds the predetermined threshold value is generated, in the voice section detection in Mode 1 and Mode 2, P is always set in the end detection processing (step S651).
(T)> = Tpe, and the end detection of the voice section is not detected. When inputting voice in such an environment, as shown in FIG.
This can be dealt with by switching to the method mode 3 for detecting the voice section shown in the flowchart of FIG.

In mode 3, the user indicates the start and end of the voice section with the mouse, so that the voice section can be correctly detected even if noise occurs. FIG. 12 is an example in which, when the voice input mode is set to mode 1, the voice section is not correctly detected due to constant noise. On the other hand, FIG. 13 shows an example in which the voice input mode is set to mode 3 and the voice section is detected in a constant noise environment.

As described above, according to this embodiment, since the user can select the mode for determining the voice input section, the detection method is changed according to the situation in which the user is placed, and the voice section is correctly detected. It becomes possible to do.

By the way, it is the mode 1 that the user's operation is simple, followed by the mode 2 and the mode 3, but the user can select the mode according to his / her own state or the environment. It becomes possible to optimize both the operability and the detection method for correctly detecting the voice section.

[Explanation of the Second Embodiment] The contents of the operation processing in the second embodiment will be explained according to the flow chart of FIG. The device configuration is the same as that shown in FIG.

In the second embodiment, with respect to the voice input interface having the same three types of voice input modes as in the first embodiment, the computer automatically sets the input mode to input the voice, The result of voice section detection corresponding to the set input mode is shown to the user. The user determines from the displayed result whether or not the voice section is correctly detected. If the section is not correctly detected, the computer changes the voice input mode. If the section is correctly detected, the input mode is not changed.

Details will be described with reference to the flowcharts of FIGS. 14 and 15.

The processing different from that of the first embodiment is the voice input mode selection processing in step S1 'and the cancel determination processing in step S10 performed after the voice section display processing in step S9' in FIG. The other steps are the same as those in FIG.

An example of the voice input mode selection process of step S1 'will be described with reference to the flowchart of FIG.

In FIG. 15, first, step S101.
In the immediately after startup determination process, it is determined whether or not it is immediately after the program startup. If it has just been started, step S1
In step 02, mode 1 is selected as the voice input mode. Then, the process proceeds to the selection input mode transmission step S13 ′ (similar to step S13 in FIG. 3).

If it has not been started immediately, step S1
In step 03, it is determined whether the current mode is mode 1. When the current setting of the voice input mode is the mode 1, the process proceeds to step S104, the mode 2 is set as the voice input mode, and the selection input mode transmission process of step S13 ′ is performed.

Further, when it is determined that the current mode is not the mode 1, the process proceeds from step S103 to step S105, and it is determined whether the current voice input mode is set to the mode 2. When it is determined that the mode is the mode 2, the process proceeds to step S106, the voice input mode is set to the mode 3, and the input mode transmission process of step S13 'is performed.

If the set mode is not the mode 2, that is, the mode 3 is selected, the process proceeds to step S11 ', the selectable input mode is displayed on the display device 2, and at step S12'. Wait for user selection. After the user sets the voice input mode, the step S1 is performed.
Move to 3 '.

When the mode is determined as described above, the process returns to FIG. 14 and the same processing as in the first embodiment is performed.
Then, the section detection result is displayed in step S9 ', and a cancellation decision is made in step S10. The user does not cancel the voiced voice if the voice segment is correctly detected. If noise or the like is erroneously detected as a voice section, it is canceled. In the cancellation determination process of step S10, it is determined whether or not the user cancels, and if canceled, the process proceeds to step S1 ′, and if not canceled, the process proceeds to step S5. Cancellation is performed using the input device 3, and the method is determined in advance. For example, as a canceling method, a mouse is double-clicked, or a specific key on the keyboard is pressed.

A case where the user actually uses the embodiment shown above will be described.

First, at the same time when the program is started, step S
In step 102, the voice input mode is set to the mode 1 (that is, the mode in which the operation is the simplest), and the user is notified that the mode 1 has been selected in step S13 ′. After initial setting in step S2 ', voice acquisition and voice data holding are performed in steps S3' and S4 ', and step S5'.
The voice data from time t to time t + Δt is analyzed,
In step S6 ', the voice section detection process is performed. While the voice detection status St is not END in step S7 ', the voice data at the next time is analyzed, and steps S5' to S8 'for determining voice segment detection and voice input end are repeated. If it is determined in step S7 that the voice input has ended, the process moves to step S9 for displaying the voice section, and the detected voice section is displayed. Then, the cancellation determination process of step S10 is performed.

If it is assumed that the voice section is correctly detected, the user does not cancel it in step S10, so the process proceeds to step S2 ', and voice input is continued in the same voice input mode (mode 1). Here, it is assumed that, when a voice is input, sudden noise occurs before the user's utterance, and a voice segment is erroneously detected. The user uses step S9 '
Since it can be recognized that noise has been erroneously detected, it is canceled in the subsequent step S10. If it is canceled in step S10, the process proceeds to step S1, the voice input mode selection process is performed, and the process shown in the flowchart of FIG. 15 is performed. In this case, since the current voice input mode is mode 1, mode 2 is set in step S104 after steps S101 and S103. Further, the user is informed of the change to the mode 2 in step S13 '. Subsequently, voice input is performed in the voice input mode of mode 2.

Incidentally, even if the voice input mode of mode 2 is used, if a large amount of noise is constantly generated and the end of the voice section is not correctly detected, the user performs step S.
Cancellation in 10 returns to the voice input mode selection process of step S1 ′, and steps S101 and S1
After 03 and S105, the mode is changed to the mode 3 in step S106. When inputting in mode 3, since the user has set the voice section, there is no erroneous detection due to noise or the like.

When the voice input environment changes during voice input in mode 3 to reduce noise and change to an input mode that is easy to operate, the user cancels so that the user can set an arbitrary input mode. Become. In this case, the process proceeds from step S10 to the voice input mode selection process (step S1 '), and steps S101, S103, S
After 105, the selectable voice input mode is displayed in step S11 '. The user can select an arbitrary input mode and continue voice input in the subsequent user selection process (step S12 ′).

In the above, one embodiment for automatically selecting the voice input mode has been described. As described above, a method of performing voice input in a voice input mode that does not require user operation immediately after starting the program, and gradually changing the section detection to a user-dependent input mode each time the user cancels Is.

The present invention is not limited to the illustrated embodiment, and various modifications can be made. For example, there are the following modifications. (1) Although the mouse is used as the input device 2 in the above embodiment, the present invention is not limited to this, and a key on a keyboard, a light pen, a touch panel, or the like may be used. (2) In the above embodiment, the root mean square value was used as a parameter when detecting the voice section by the computer, but the present invention is not limited to this, and the number of zero crossings, windowed power, etc. may be used. Also, a plurality of these parameters may be used. (3) In the above embodiment, three types of voice section detection methods are used, but the present invention is not limited to this. For example, (a) the beginning of the voice section when both the computer and the user set the start (end) of the voice section. (End).

(B) A case where the user performs an operation of notifying the input device 2 of the start end (end) around the time when the computer determines that the start end (end) is the start end (end). Etc. may be used. (4) In the above embodiment, the selected mode name is displayed on the screen as a method of transmitting the selected voice input mode to the user, but the present invention is not limited to this. You may inform the user.

As described above, according to this embodiment,
A voice input mode setting means for setting a voice input mode from among a plurality of voice input modes, a voice capture holding means for capturing and holding contents generated by the user, and a voice input mode setting means for setting a plurality of voice section detection methods. With the voice section detecting means for detecting the voice section corresponding to the selected voice input mode, the voice input mode can be selected from a plurality of voice input modes according to the change of the voice input environment, noise, etc. Remarkably improved in usability and usability.

In particular, according to the second embodiment, a mode having a simple operability is selected in the initial stage, and therefore, the mode is set so that the user is least burdened in an environment such as a place where there is relatively little noise. Will be possible.

Further, in the above embodiment, an example in which one independent device is applied has been described. However, as can be easily inferred from the above description, the present invention is applied to a system composed of a plurality of devices. It doesn't matter. Further, it goes without saying that the present invention can be applied to the case where it is achieved by supplying a program to a system or an apparatus.

[0081]

As described above, according to the present invention, it is possible to change the voice detection method according to the situation of voice input and pass valid voice information to the lower processing.

[0082]

[Brief description of drawings]

FIG. 1 is a block diagram of a first embodiment of an information device according to the present invention.

FIG. 2 is a flowchart showing a main process of the first embodiment.

FIG. 3 is a flowchart showing a voice input mode selection process in the first embodiment.

FIG. 4 is a flowchart showing a voice data analysis process in the first embodiment.

FIG. 5 is a mode 1 of voice section detection in the first embodiment.
It is a flowchart which shows the process of.

FIG. 6 is a second mode of voice section detection in the first embodiment.
It is a flowchart which shows the process of.

FIG. 7 is a mode 3 of voice section detection in the first embodiment.
It is a flowchart which shows the process of.

FIG. 8 is an example of displaying a voice input mode selectable when the voice input mode is selected in the first embodiment.

FIG. 9 is an example in which the selected voice input mode is displayed on the screen in order to notify the user in the first embodiment.

FIG. 10 is an example of erroneously detecting a voice section due to the occurrence of sudden noise in the voice section detection processing of mode 1 in the first embodiment.

FIG. 11 is an example in which the voice section detection processing of mode 2 correctly detects the voice section in the first embodiment.

FIG. 12 is an example in which the voice section cannot be detected in the voice section detection process of the mode 1 in the first embodiment due to the occurrence of steady noise.

FIG. 13 is an example in which the voice section detection processing of mode 3 correctly detects the voice section due to the occurrence of sudden noise in the first embodiment.

FIG. 14 is a flowchart showing a main process of the second embodiment.

FIG. 15 is a flowchart showing a voice input mode selection process in the second embodiment.

[Explanation of symbols]

 1 System Bus 2 Display Device 3 Input Device 4 Voice Input Device 5 I / O Device 6 CPU 7 Storage Device

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor, Yasunori Otodo 3-30-2 Shimomaruko, Ota-ku, Tokyo Canon Inc.

Claims (8)

[Claims] 1. A voice input device for passing voice information input from a voice input means to a lower process, comprising a plurality of modes for specifying a period of significant voice information input from the voice input means. A voice input device, comprising: a selection unit that selects one of the plurality of modes. 2. The mode includes a first mode for automatically identifying a start point and an end point of a voice, a second mode for manually instructing a start point position, and a third mode for manually instructing a start point and an end point position. The voice input device according to claim 1, wherein: 3. The selecting means includes the first unit at an initial stage.
3. The voice input device according to claim 2, wherein the second mode and the third mode are selected in that order when a predetermined change instruction is issued. 4. The voice input device according to claim 1, wherein the selection means selects manually. 5. A control method for a voice input device, which transfers voice information input from voice input means to a lower process, comprising a plurality of methods for specifying a period of significant voice information input from said voice input means. And a selection step of selecting one from the plurality of modes, and a method for controlling a voice input device. 6. The mode includes a first mode for automatically identifying a start point and an end point of a voice, a second mode for manually instructing a start point position, and a third mode for manually instructing a start point and an end point position. The method for controlling a voice input device according to claim 5, wherein. 7. The selecting step includes the first step at an initial stage.
7. The control method for the voice input device according to claim 6, wherein the second mode and the third mode are selected in this order when a predetermined change instruction is issued. 8. The method for controlling a voice input device according to claim 5, wherein the selecting means selects manually.