JPS6149680B2 - - Google Patents

Abstract

PURPOSE:To secure a smooth editing of the aural information by combining the aural input signal and the control signal to obtain the composite signal at the mixer and then the output being recorded to the same medium in the form of the double- system information of the original signal and the editing information. CONSTITUTION:Voice input part 1 converts the aural signal into the electric signal, and the control signal is generated from keyboard 2. These two signals are combined through mixer 4 to obtain the composite signal. This composite signal is recorded 6 and 7 to the same medium in the form of the information of the double-system of the original and editing information to carry out reproduction through the medium. Here, the control signal forms part of the composite signal in the form of the signal showing the unit section of the aural information, the signal showing the insertion start, the signal showing the correction start and the signal showing the end of editing respectively. And the aural information to be edited is held between the signal showing the start of insertion or correction and the signal showning the end of editing to be recorded and reproduced as the editing information. In this way, a smooth editing is ensured for the aural information.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a voice information editing method used for voice recognition.

Traditionally, speech recognition devices such as voice typewriters were intended to convert human speech directly into text, but Japanese information input devices that combine a single-phonetic recognition device and a kana-kanji conversion device have been designed to convert cassette tapes, etc. A system is being considered in which a voice is injected into a computer, a recognition process is performed at a processing center, and the results are returned as printed matter. In this case, an important issue is how to correct and edit audio input errors.

An object of the present invention is to provide an audio information editing device suitable for the above-mentioned audio information input system.

In the present invention, for example, delimiter signals are recorded in advance for each sentence, and editing information is recorded along with control signals such as insertion correction in an independent channel for editing.
These control signals are used to control rewinding, fast forwarding, channel switching, etc., and to perform editing such as insertion and correction in units of sections.

An embodiment of the present invention will be described below using a tape recorder as an example.

FIG. 1 is a block diagram showing an embodiment of the recording unit of the audio information editing device according to the present invention. Input audio is input from an audio input unit 1 consisting of a microphone and an audio amplifier, and control signals for editing are input. is inputted from the keyboard 2, converted into an audio signal such as a tone burst by the encoder 3, and combined with the audio signal by the synthesizer 4 to form a composite signal. This signal is recorded on different tracks on the same tape by a recording system 6 for original information or a recording system 7 for edited information by a changeover switch 5.

FIG. 2a shows an example of original information recorded on a tape, and FIG. 2b shows an example of edited information recorded on different tracks of the same tape. As the original information shown in FIG. 2a, for example, voice information B1, B2 is recorded separated into one sentence unit, and delimiter signals A1, A2, . . . are recorded as the delimiters. Below, regarding this information, audio information B1
An example of editing in which D1 is inserted before and replaced with B2 and D2 will be described.

As editing information, as shown in FIG. 2b, audio information D1 to be inserted is recorded on an editing track between a control signal C1 indicating the start of insertion and a control signal E1 indicating the end of editing. The audio information D2 to be replaced is similarly recorded between the control signal F1 indicating the start of correction and the control signal E2 indicating the end of editing. At this time, the control signal Ai indicating the break
includes information indicating the number of the corresponding audio information, and the editing control signals Ci, Ei, and Fi also include information indicating the number of the audio information to be edited. This editing information is recorded as follows. In other words, the original information is played back in order, and the insertion/
When you are at the position where you want to make a correction, turn the changeover switch 5 in Figure 1 to the editing information side, press the insert or correction key on the keyboard 2, and then record the voice to be inserted/corrected from the audio input section 1. 2
Press the end editing key. At this time, the encoder 3 encodes the signal from the keyboard 2 together with the number of the audio information in the immediately preceding delimiter signal Ai.

FIG. 3 is a block diagram showing an embodiment of a reproducing section for decoding the editing information recorded in this manner.
The original information is applied to the decoder 11 via the reproduction system 10, and after being separated into audio and control information, the control information is led to the control section 14. Similarly, the editing information passes through the reproduction system 12 and is separated into audio and control signals by the decoder 13, and then the control signals are applied to the control section 14. The control section 14 includes, for example, a tape start control section 15, a fast forward control section 17, and a rewind control section 1.
It generates signals to be sent to each of the control sections 8 and performs the following control according to the control signals in the editing information.

When detecting the control signal C1 indicating the start of insertion, when detecting the signal C1 in Figure 2b, switch to audio information and take D1, and when detecting E1 indicating the end of editing, detect A1 corresponding to it. The tape is rewound until the tape is rewound, the tape is switched to the original audio information, and after the tape is stopped, the tape is started and B1 is selected.

When the control signal F1 indicating the start of correction is detected When the signal F1 in Figure 2b is detected, the audio information is switched to the editing information side, D2 is substituted for B2 for that part, and the end signal E2 is detected. Then, fast forward until the next corresponding A3 is detected,
After stopping, switch to the original information side, start, and proceed to the next B
Play 3. Note that if no audio is recorded in D2, that portion of B2 will be deleted.

Note that there may be cases where the audio signal in the edited information is longer than the audio signal in the original information to be edited, but by performing control while comparing the numbers in the control signals that indicate breaks, it is possible to play back in the correct order. is being used.

In other words, the control unit 14 constantly compares the number indicating a break in the original information with the number in the editing information, and determines that when a number in the editing information arrives, the same number in the original information has already arrived before that. If so, a rewind control signal is issued, and vice versa, a fast forward control signal is issued.

Further, the numbers of the audio information to be edited may be recorded together at the beginning of the tape.

As described above, the present invention facilitates the editing of audio information by providing editing tracks and editing control signals.

Therefore, according to the present invention, when correcting recording errors in a system that records audio onto a cassette tape and obtains printed output as Japanese text, it is possible to correct recording errors by starting over from the beginning or by inputting correction information all at the end. It is possible to obtain a correctly corrected final result at the audio level without the inconvenience of having to make a mistake. Even when simply obtaining audio output, it is possible to obtain complete audio output.

The present invention is not limited to the embodiments described above, and various modifications are possible. For example, the original information and the edited information may include a signal indicating the type of character. After inputting audio information into individual syllables and performing monosyllable recognition to obtain kana sentences,
In a system that converts sentences containing kanji through kana-kanji conversion, the original information, as shown in Figure 4 a and b,
Each of the editing information tracks may be divided into a track for signals G1 to G5 indicating the type of character and a track for audio information. in this case,
The signal indicating the division of audio information may be placed on the same track as the signal indicating the type of character, as shown in FIGS. 5a and 5b.

As a modification of this, the audio information and the signal indicating the type of character may be multiplexed, for example, by frequency division, and the original information and edited information may be recorded on each track as shown in FIGS. 6a and 6b.

In other words, in Figure 6 a and b, the upper half is, for example, 4KHz.
All you have to do is record at the above speed and record the lower half at 4KHz or lower.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing an embodiment of the recording section according to the present invention, FIG. 2 a and b are diagrams showing an example of the recording state on the tape according to the present invention, and FIG. The figures showing the embodiment, FIGS. 4a and 4b, 5a and 5b, and 6a and 6b are views showing other examples of the recording state on the tape according to the present invention, respectively. 1...Voice input section, 2...Keyboard, 3...
Encoder, 4...Synthesizer, 5...Switch,
6, 7...recording system, 10,12...playback system, 1
1, 12...decoder, 14...control unit, 15...
... tape start control section, 16 ... tape stop control section, 17 ... fast forward control section, 18 ... rewind control section.

Claims (1)

[Claims] 1. Recording on a medium as two types of information, original information and editing information each including audio information and a control signal, and performing editing processing while reproducing the original information and editing information from this medium. In the method, the control signal is added to the audio information as a signal indicating a division of one unit of audio information, a signal indicating the start of insertion, a signal indicating the start of correction, and a signal indicating the end of editing, An audio information editing method, characterized in that the audio information is recorded and reproduced as the editing information while being sandwiched between a signal indicating the start of the insertion or correction and a signal indicating the end of the editing. 2. Insert the audio information to be edited between the signal indicating the start of insertion and the signal indicating the end of editing before the corresponding audio information in the original information, and insert the signal indicating the correction and the end of editing. 2. The audio information editing method according to claim 1, wherein the audio information to be edited between the signal shown and the audio information to be edited is replaced with the audio information in the corresponding original information and reproduced. 3. The audio information editing according to claim 1, wherein the signal indicating the break, the signal indicating the start of insertion, and the signal indicating the start of correction specify a break number of the audio information. Method. 4. After editing the audio information specified by the control signal in the original information corresponding to the control signal in the editing information, automatically return to the editing starting point specified by the control signal in the original information. An audio information editing method according to claim 1, characterized in that: