JPH06161487A - Speech editor - Google Patents

Abstract

PURPOSE:To provide the speech editor which can display and edit a speech waveform for a long period of time. CONSTITUTION:When an input signal is compared with a threshold value to recognize a voiceless sound signal (110) and the number of successive voiceless sound signals reaches to make a reduction display (114), the decided voiceless sound signal is saved in a buffer (118) and a return to processing for reading an input signal is made. When it is decided that the signal is not the voiceless sound signal in the 110 and a reduction display of the signal saved in the buffer so far is made, the reduction display is made (122). Then the signal which is decided not as the voiceless sound signal in the 110 is displayed regardless whether or not the reduction display is made (126). Consequently, the speech signal waveform for a long period of time can be displayed and edited without remarkably losing characteristic information on the speech signal waveform.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for displaying and editing a digital voice waveform, and more particularly to a voice editor for editing voice for multimedia processing and the like.

[0002]

2. Description of the Related Art When displaying a digital voice waveform, the length of the waveform to be displayed is determined by the sampling frequency. Also, the length of a waveform that can be displayed at one time is limited by the resolution of the display that displays it. For example,
If the digital voice to be displayed is for a long time like narration, the entire waveform cannot be displayed at one time, making it difficult to perform editing operations. There is "thinning-out display" as a method for solving such a problem. This is an operation of reducing the time axis and lengthening the displayable waveform length by extracting and displaying signals at regular intervals instead of displaying each digital signal for each sampling frequency. As a publicly known example of effectively performing this thinning-out display, there is a matter described in JP-A-1-120598. This enables thinning display by extracting and displaying feature points for each constant section of the digital signal.

[0003]

The conventional method has the following problems. That is, in the thinned-out display in which signals are extracted at regular intervals, the signal waveform obtained by the display may be significantly different from the actual signal waveform. This is because, for example, when a waveform that fluctuates drastically in the above-mentioned certain section is input, one of the signal groups becomes a signal representative of that section, and the fluctuation of the signal can no longer be observed at the display stage. .

On the other hand, in the above-mentioned known example, by extracting and displaying the characteristic points for each fixed section, it is possible to perform the thinning display without damaging the characteristics of the waveform. In this method, it is possible to prevent the displayed waveform from being significantly different from the input, but even if there is a very important signal to the observer in a certain section, it is represented by the feature points, so that it is accurate. It does not display a waveform.

An object of the present invention is to provide an audio editor capable of displaying an audio signal in a longer section. Another object of the present invention is to provide an audio editor that does not drop important audio signals. Still another object of the present invention is to provide a voice editor capable of grasping a voice and a voice following it, that is, a silent state between sentences.

[0006]

SUMMARY OF THE INVENTION The present invention is to reduce and display a portion where a signal regarded as a silent state continues for a certain period. Also, for the section to be reduced and displayed,
This is a display that can be distinguished from other sections. Furthermore, for the section to be reduced and displayed, its length, maximum value,
It provides information such as the minimum value.

[0007]

By reducing and displaying the section of the audio signal which is considered to be in the silent state, it is possible to display the audio signal waveform for a long period of time. Further, by reducing the section of the signal which is considered to be in the silent state and performing thinning display, it is possible to prevent the loss of other important signals by the thinning operation. Further, by displaying the section to be thinned and displayed so as to be distinguishable from other sections, it is possible to grasp the section in which the thinning operation is performed. In addition, the section to be thinned out is displayed so that it can be distinguished from other sections, and information such as its length, maximum value, and minimum value is given, so that the voice and the following voice, that is, the space between sentences, It is possible to grasp the silent state of the in detail.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described in detail below.
FIG. 1 is a flow chart showing an embodiment of a processing procedure of a non-voice section thinning-out display in the present invention. FIG. 2 is a block diagram showing the configurations of a voice signal input device, a voice editor, and a voice signal output device according to the present invention.

In FIG. 2, an audio signal input device is an audio signal file device 1 for recording and extracting an audio signal.
0, a microphone / CD / cassette tape 11 for inputting an audio signal, and an audio signal input control device 12 for controlling the input audio. The voice editor 13 includes an editing unit 14 for editing a voice signal, a display unit 15 for displaying a waveform of the edited voice signal, a reproducing unit 16 for reproducing, a mouse 17, and a keyboard 18. Here, the display unit 15
Includes a silent interval thinning-out display device 19. Furthermore, the audio signal output device includes a display device 20 that outputs a signal from the display unit 15 to a screen, an audio signal output control device 21 that reproduces a signal from the reproduction unit, and a speaker 22.
Consists of.

Next, the operation of displaying the audio signal waveform in the silent interval thinning-out display device 19 of FIG. 2 will be described based on the flowchart of FIG. Here, it is assumed that the input audio signal is obtained from the audio signal file device 10 and the output waveform is output to the display device 20.

First, as an initial condition, a threshold value for determining whether or not the input signal is silent, and a period during which the silent signal continues to be reduced are set. In addition, the counter that measures the continuous silence signal is initialized (zero, etc.), and the flag that records whether or not to reduce the silence interval is set to "Do not reduce" to save the continuous silence signal. A buffer to be placed is prepared (102).

Next, the input signal is converted into the audio signal file device 1.
Read from 0 (104). Therefore, it is determined whether or not the input signal is the end (106), and if the input signal is the end, a series of processing is ended (108). If the input signal is not the end, it is compared with the threshold of the silence signal to determine if the input signal is silence (110). When it is determined that the signal is a silent signal, the counter for measuring the continuing silent signal is incremented by 1 (112). When the number of continuous silent signals reaches the reduced display (114), the flag for recording whether or not the reduced display is performed is set to "reduce" (116). Next, the input signal determined to be silent is saved in the buffer (118), and the input signal is read (10
Return to 4).

When it is judged in 110 that the signal is not a silent signal, it is judged whether or not the signal saved in the buffer up to the present time is reduced and displayed (120). The reduced display is performed only when the reduced display is necessary (122). Thereafter, regardless of whether or not the reduced display is performed, the signal determined as not being a silent signal in 110 is displayed (12
4), the flag is set to "not be reduced" (12)
6). In addition, the counter for measuring the continuous silent signal is cleared (128) and the input signal is read (104)
Return to.

FIG. 3B shows the signal waveform shown in FIG. 3A in a reduced size by the method shown in the flowchart of FIG. 2 according to the present invention. Here, it is assumed that the threshold value for determining the silent signal is 30, and the silent duration time for reducing is 31.

Below, the input signal shown in FIG.
The process of displaying the reduced image will be described. In the section of the input signal 32, at 110, the signal value is larger than the threshold value 30, and therefore it is determined that the signal is not a silent signal. Further, since the silent signal to be reduced and displayed is not saved in the buffer during that time, the same waveform as the input signal waveform is output as in 35.

On the other hand, in the section of the input signal 33, at 110, the signal value is smaller than the threshold value 30, so it is judged as a silent signal. In addition, the silent continuation section is equal to the section length of 33, which is the length up to the reduced display.
It is displayed in reduced size.

Further, the sections of the input signal 34 are 32, 35.
Similar to the case of, the same waveform 37 as the input waveform is output.

In the example shown in the flow chart of FIG. 1, it is necessary to predict how long the silent signal will continue and prepare a buffer to save the silent signal. FIG. 4 shows a flow chart in the case where the same number of buffers as the continuous number of is executed.

First, as in the case of FIG. 1, as an initial condition, a threshold value for determining whether or not the input signal is silent, and a period during which the silent signal for reducing display is set are set. In addition, the counter that measures the continuous silence signal is initialized (such as zero), and the flag that records whether or not to display the silence section in a reduced size is set to "Do not display in a reduced size" to save the continuous silence signal. Prepare a buffer to store. In addition, a buffer counter is prepared and initialized (zero or the like) here (202).

Next, the input signal is converted into the audio signal file device 1
Read from 0 (204). Therefore, it is determined whether or not the input signal is the end (206), and if it is the end, a series of processing is ended (208). If the input signal is not the end, it is compared with the threshold of the silence signal to determine if the input signal is silence (210). When it is determined that the signal is a silent signal, the counter for measuring the continuous silent signal and the buffer counter are incremented by 1 (21
2). When the number of continuous silent signals reaches the reduced display (214), the flag for recording whether or not the reduced display is performed is set to "reduce display" (216). The signal saved in the buffer up to the present time is reduced and displayed (21
7). Next, the counter value for the buffer is cleared (21
8) The input signal determined to be silent is saved in the buffer (220). If it is not necessary to display the reduced image at 214, only an operation of saving the silent signal in the buffer is performed (220). After that, the process of reading the input signal (20
Return to 4).

When it is judged in 210 that the signal is not a silent signal, it is judged whether or not the signal saved in the buffer up to the present time is reduced and displayed (222). Only when the reduced display is necessary, the reduced display is performed (224). Then, regardless of whether or not the reduced display is performed, the signal determined to be not the silent signal in 210 is displayed (22).
6). In addition, the flag is set to "not display reduced" (22
8). The counter for measuring the continuous silent signal is cleared (230), and the process returns to the process of reading the input signal (204).

In the above-mentioned example, the number of buffers required is smaller than that of the method shown in the flowchart of FIG. 1, and a fixed number of buffers may be prepared. However, when the silent signal is for a long time, the method shown in the flowchart of FIG. 1 saves the silent signal in which the silent section continues in the buffer and finally displays them in a reduced size, so that the silent section reaches a reduced display. The reduction ratio of the silent section is increased as compared with the method shown in the flowchart of FIG. 4 in which the reduced display is repeated every time the length becomes longer.

102, 110, 2 in the above embodiment
As an example of the method of changing / setting the threshold values relating to 02 and 210, a signal waveform currently displayed for setting the threshold value or a waveform displayable from the waveform is used. As a procedure, first enter the mode to change / set the threshold value from the menu. Next, the waveform is used as a guide on the screen on which the signal waveform is displayed, and is set / changed with at least one of the mouse and the keyboard.
In this embodiment, since the displayed waveform is used as a guide, the user can easily set the threshold visually.

As another embodiment, a threshold value is designated by a numerical value and set / changed. In this case, it is possible to make detailed settings and changes as compared with the above case. Further, the method of setting / changing the duration of the silent signal for reducing the display of 102, 114, 202, and 214 in the above-described embodiment is also displayed as in the embodiment of the threshold setting / changing method. The same effect can be obtained by using a signal waveform or by setting a numerical value.

Further, as one embodiment of the reduced display method related to 122 and 224 in the above-mentioned embodiment, the signal waveform is displayed separately from other signal waveforms which are not reduced. This includes the reduced waveform, the scale of the reduced section, the area of the reduced section, and at least one color, linetype, and background of the scale on the scale of the reduced display. It can be implemented by distinguishing at least one of them from the others. In addition to this, at least one of the reduced waveform, the scale of the reduced display section, the area of the reduced display section, and the scale on the reduced display section scale has a start point and an end point of the section. It is possible to put it on by making a mark that can be recognized. In addition to this, it is possible to implement by distinguishing the character font of the scale written on the scale of the reduced display section from other fonts. In addition to this, it can be performed by displaying a waveform figure processed from the displayed waveform in the reduced display section, such as thinning display, symbol display or zero display (display the signal to be displayed as zero) Is. In FIG. 5, the background of the section to be reduced and displayed in the above embodiment is shaded (51), a boundary line is provided at the start point 52 and the end point 53 of the reduced and displayed section, and symbols are further added to the waveform display. The characteristic points of the waveform are extracted and displayed (54)
Here is an example.

Further, an embodiment for compensating for the lack of recognition items in the waveform observation in the reduced display section 122 and 224 in the above-mentioned embodiment and other sections will be described. In this embodiment, it is assumed that at least one of the recognition items that can be recognized in normal waveform observation such as the length, the value of the maximum and the minimum of the section and the position of the displayed waveform can be recognized.

This is a display of a matter to be recognized by a quantitative numerical value, and a matter to be recognized is recognized on at least one of the screens displayed in reduced size or the screens displayable in association therewith. It can be recognized by performing at least one method of display of possible symbols, enlarged display of a section to be recognized, and waveform display without reduced display. In addition to the embodiment of FIG. 5, FIG. 6 quantitatively displays the position and the amplitude value at the point where the amplitude of the waveform is maximum in the above-mentioned embodiment (6
1), the position and amplitude value are quantitatively displayed at the minimum point (62), the real time of the reduced display section is displayed (63), and the true waveform of the reduced display section is additionally displayed. This is an example of (64).

Further, as a method of recognizing the length of the reduced display section, it is possible to recognize it by providing a step in at least one of the above-mentioned reduced display portion and the other portion. In FIG. 7, the length is determined according to the real time of the section in which the boundary line between the start point 51 and the end point 52 is reduced and displayed in the embodiment of FIG. 5, and the scale is set to enable estimation of the real time. This is an example of (71, 72) additionally displayed.

As an application example of the above-described embodiment, the component of the waveform signal reduced and displayed by the audio editor is recorded in the audio signal file device in place of the input signal waveform, thereby saving the storage capacity of the file device. By pre-storing in the form of a signal corresponding to the reduced display, it is possible to shorten the time until the voice signal is input and the reduced display is performed. It also improves the efficiency of signal transmission in communications such as networks. Furthermore, as another method of recording the reduced waveform, the section to be reduced and the reduced display mode are added to the input signal waveform and recorded, or separately recorded, so that the true input signal is held and the reduced display is performed. The time can be reduced.

Further, another application example is shown in FIG. This is a true waveform display (81, 82, 8) of the desired observation section.
3 () is an enlarged display (84, 85, 86) of a section (sound section) which is not a silent section. However, the start point of the voiced section is 87, and the end point is 88. Further, in the enlarged voiced section, the waveform display as shown in FIG. 3b, FIG. 5, FIG. 6 and FIG. 7 can be performed similarly to the silent section described above, and the waveform information can be obtained.
As a result, not only detailed information on the voiced section can be obtained, but also the appearance of the entire desired section can be observed.

[0031]

As described above, according to the present invention,
In the display of the audio signal including the silent section, the section determined to be the silent section is reduced and displayed, whereby the signal waveform over a long period can be displayed. Further, since the waveform characteristic of the display waveform is obtained, the audio signal editing becomes easy.

[Brief description of drawings]

FIG. 1 is a flowchart showing an example of a processing procedure of the present invention.

FIG. 2 is a block diagram showing configurations of a voice signal input device, a voice editor, and a voice signal output device according to the present invention.

FIG. 3 is an example of an input voice signal waveform for explaining the embodiment shown in FIG. 1 and an example of an output voice signal waveform when the processing procedure shown in FIG. 1 is performed.

FIG. 4 is a flowchart showing another embodiment of the processing procedure of the present invention.

5A and 5B show an example in which a reduced display section is distinguished from other sections and a reduced display example is applied to the output audio signal waveform shown in FIG. 3B.

FIG. 6 is an example to which an embodiment for compensating for the lack and lack of waveform information of the output audio signal waveform shown in FIG. 5 is applied.

FIG. 7 is a diagram in which an embodiment is applied to the output audio signal waveform shown in FIG. 5 so that the real time of the reduced display section can be recognized.

FIG. 8 shows, as an application example of the embodiment of the present patent, a true audio signal output waveform and a voiced signal which are enlarged and displayed.

[Explanation of symbols]

10 ... Voice signal file device, 11 ... Microphone, C
D, tape recorder, 12 ... Voice signal input control device,
13 ... Voice editor, 14 ... Editorial department,
Display unit 16 Playback unit 17
..Mouse 18..Keyboard, 19..Thinning display unit, 20..Display device, 21 ..
Audio signal output controller 22 ... Speaker

 ─────────────────────────────────────────────────── ─── Continuation of front page (72) Inventor Toshio Moriya 1099, Ozenji, Aso-ku, Kawasaki-shi, Kanagawa Ltd. Hitachi, Ltd. System Development Laboratory

Claims (4)

[Claims] 1. A sound signal is input, a function of displaying and editing a waveform of the sound signal is provided, and the sound signal is decimated and displayed, and the decimated display is performed by reducing a section in which a silent signal continues. And a voice editor. 2. A user is provided with at least one of a threshold value for determining whether or not there is a silence signal and a duration of the silence signal for reducing display for the purpose of prompting the determination of the silence section. The audio editor according to claim 1, wherein the audio editor can be set by. 3. The voice editor according to claim 1, wherein the portion where the thinning display is performed is displayed separately from other portions. 4. The voice editor according to claim 1, wherein a numerical value indicating the characteristic of the displayed waveform is displayed.