JPH01106098A - Voice recognition system - Google Patents

Abstract

PURPOSE: To generate standard patterns of a microphone speech and a sound- recorded speech by single-time pronunciation and lighten the burden on the registering person by recording a registered speech which is inputted from a microphone in a sound-recording and reproducing device and using this registered speech as a registered speech for generating the standard pattern of the sound-recorded speech. CONSTITUTION: The registered speech is inputted directly to a main body from the microphone 3 and the sound-recording and reproducing device 2 is connected to the main body at this time. Then while this inputted registered speech is recorded in the sound-recording and reproducing device 2, the registered speech from the microphone 3 is analyzed on the main body side to generate the standard pattern, which is stored in a storage device 8. Consequently, the standard pattern generated by the direction input from the microphone 3 and the standard pattern generated from the sound-recorded speech can be both generated by single-time speech registering operation, so the frequency of speech pronunciation by the registering person is reduced to a half to lighten the burden of the speech operation.

Description

[Detailed description of the invention] B) Industrial application fields TECHNICAL FIELD The present invention relates to a speech recognition system.

(b) Conventional technology Voice recognition is performed by recording voice on a recording/playback device such as a tape recorder, playing it back, and inputting the output reproduced voice to a voice recognition device to perform voice recognition and convert it into text. system is being developed (Unexamined Japanese Patent Publication No. 58-15
No. 8736).

The g-voice recorded on a recording/playback device such as a tape recorder is
Because the standard pattern is based on the frequency characteristics of the recording/playback device, a standard pattern created from recorded audio is different from a standard pattern created from audio input directly from a microphone. Therefore,
When making a speech recognition device recognize recorded speech, it is necessary to use a standard pattern created from the recorded speech.On the other hand, when making the speech recognition device recognize the speech input directly from the microphone, it is necessary to use a standard pattern created from the speech input directly from the microphone. Standard patterns must be used. Therefore, in order to be able to recognize both microphone voice and recorded voice, two types of registration operations are required: registration using microphone voice and registration using recorded voice. Conventionally, these two types of registration operations were performed. They had to be done separately.

(c) Problems to be solved by the invention As mentioned above, since the registration operations for microphone voice and recorded voice were performed separately, the same utterance had to be performed twice, which was a great burden for the registrant. there were. The present invention realizes a voice recognition system that can create a standard pattern of microphone voice and bulb voice with one utterance, and can reduce the burden on the registrant.

2) Means to Solve the Problems The voice recognition system of the great invention creates an fM$ pattern while inputting registered voice into the recognition device through the microphone and recording the registered voice into the recording/playback device. Thereafter, the recorded voice is played back and input to the recording/playback device, and a standard pattern of the recorded voice is created.

(e) Function: According to the voice recognition system of the present invention, the registered voice inputted from the microphone is recorded in the recording/playback device, and the recorded voice is used as the registered voice for creating a standard pattern of the recorded voice. , 10 registered voices, it can be seen that a standard pattern created from the microphone input sound and a standard pattern created from the recorded sound are obtained.

DESCRIPTION OF THE PREFERRED EMBODIMENTS In the following, an external view of a desk cheating machine which employs the present invention to create sentences by voice input is shown in the first place, and FIG. 2 shows a functional block diagram of the machine.

In Fig. 2, (1) is a speech recognition section equipped with a circuit in the main body (100) of Fig. 1, and its details are shown in the block diagram of Fig. 3. A pre-processing unit that performs adjustment <11) [Fig. [Fig. 5], a word recognition unit (13) [Fig. 6] that performs word recognition of input speech based on the feature parameters obtained from the extraction unit (12), and a phrase recognition unit (14) [Fig. 7]; It consists of a candidate generation unit (15) that generates candidates for recognized word character strings or recognized syllable characters based on the recognition results from any of these recognition units (13) and (t4).

Furthermore, in Fig. 2, approximately 1, (2) is the main body (1) as shown in Fig. 1.
00) is a mechanically and electrically detachable recording and reproducing device such as a tape recorder, (3) is a headphone type microphone as shown in FIG. 1, and (4) is a recording and reproducing device (2) and a microphone (3). ) and the voice recognition unit (1) (Fig. 8)
It is. (6) is a display device for displaying character strings etc. generated based on the recognition results; (7) is a keyboard for inputting various control signals of the desk cheating machine;
(8) is a storage device such as a magnetic disk device that stores the character strings generated by the digital cheating machine, and (9) is a speaker (10) that uses the character strings in the storage device by rule synthesis.
This is a speech synthesis unit that reads out the text.

Note that (5) is a control section consisting of a microprocessor, which controls the operations of the above-mentioned sections.

There are two ways to create sentences using the desk cheating machine configured as described above, and each will be explained in detail below.

The first method is to input the main voice from the microphone (3) to the voice recognition section 1), perform voice recognition, convert the input voice into a character string, display it on the display device e (6), and store it at the same time. Equipment (8
).

The second method is to record and play the text you want to input in advance! ! (
2), connect this recording and playback device (2) to this device, and input the recorded text to the speech recognition unit (1) to perform speech recognition and convert the input speech into a character string. The results are displayed on the display device (6), and the results are simultaneously stored in the storage device (8).

As mentioned above, there are two ways to input voice i.
Input switching section (4) performs input switching. In addition to switching the input, the input switching unit (4) also switches whether to record the recording signal (a) to the recording/playback device (2) or the audio input from the microphone (3). .

The operations from voice recording to text creation will be explained in detail below.

(i) Voice registration processing Before performing voice recognition, voice registration is performed in order to create a standard pattern of voice necessary for voice recognition.

First, the syllable registration mode will be described.

The standard pattern described here is the speech recognition unit (1)
It serves as a reference pattern during pattern matching in the phrase recognition section (14) of the phrase recognition section (14), and specifically, it is the syllable standard pattern memory (14d) of the phrase recognition section (14) as shown in Fig. 7.
).

The method for registering voice in this digital cheating machine is to first operate the switch (1491) in Fig. 7 to connect the parameter buffer (14d) and syllable recognition unit (14b).
There are three methods described below.

The first method is to attach a microphone (3) to the main body (100) of the machine.
A method of directly inputting a registered voice, analyzing this registered voice with a voice recognition unit (1), creating a standard pattern, and storing the created standard pattern in a syllable standard pattern memory (14d) and a storage device (8). It is.

The second method is to connect a recording/playback device (2) on which the registered voice has been recorded in advance to the main body (100), input the registered voice by playing back the recorded registered voice, and perform this manual registration. This is a method in which speech is analyzed by a speech recognition unit (1), a standard pattern is created, and the created standard pattern is stored in a syllable standard pattern memory (14d) and a storage device (8>).

The third method is to connect the microphone (3) to the main body (100) of this machine.
At this time, the recording and playback device (2) is connected to the main unit (100), and while recording the input audio to the recording and playback device (2), input the registered audio directly from the main unit (100) side. Then, the registered voice from the microphone (3) is analyzed to create a standard pattern, and the created standard pattern is stored in the storage device (8). Then, when the voice input to this microphone (3) is completed, the recorded voice is subsequently played back by the recording/playback device (2), and this recorded registered voice is reproduced as a voice! ! ! Analyzed by Shikibe (Okibe),
At the same time as creating a standard pattern and storing the created standard pattern in the syllable standard pattern memory (L4d),
This is a method in which the syllable standard pattern of the directly registered voice from the above-mentioned microphone (3) is also stored in the storage device (8).

In this third method, the sound recorded on the recording/playback device (2) is subject to the frequency characteristics of the recording/playback device (2), so the standard pattern created from the recorded sound and the microphone (3) When comparing the standard pattern created from the voice input directly from the standard pattern, differences appear between the two standard patterns. Therefore, when recognizing a recorded voice, it is necessary to use a standard pattern created from the recorded voice, and the microphone (3)
When you want to recognize the voice input directly from the microphone (3)
), it is necessary to use the standard pattern created from the voice input directly from the microphone (3), so by using the method described above, both the standard pattern registered directly from the microphone (3) and the standard pattern created from the recorded voice can be used. can be created and stored with a single voice registration operation. In addition, if the registered voice is recorded in the recording and playback device 2), this recorded voice can be played back and input on the recording machine for which no standard pattern has been created, without requiring the registrant's voice input. Standard patterns can be created simply by Also, if you record the registered voice on the recording and playback device fl (2') and record the text after this registered voice, you can later connect this recording and playback device (2) to the main unit (100) and record it. Just by playing the recorded audio, everything from voice registration to sentence creation can be done automatically.

Note that the registrant's voice input for creating the standard voice pattern is performed by the registrant reading out the characters that the present device displays on the display device (6) in a fixed order.

In addition, a recording and playback device (2
), this recording/playback device (2) can be carried alone, by uttering the voice corresponding to the entry word displayed on the display screen and recording it on the recording/playback device (2), the standard pattern can be reproduced. It becomes possible to create.

As mentioned above, when recording the registered voice for creating a standard pattern into the recording/playback device (2), when creating the standard pattern from the recorded registered voice, the recorded voice may be affected by noise etc. There is a possibility that the corresponding headword may be misaligned, and for the sake of explanation, a case will be described below based on FIG. 9 in which a tape recorder is used as the recording/reproducing device. Figure 9(a) shows the state in which registered voices for standard pattern creation are recorded on a tape recorder.
Registered voice corresponding to the headword “A” ~1ka” ゛゛°〜“
This shows the state of the tape between "E°'" and "O1", where noise is recorded between "E°'" and "O1". As shown in Figure 9 (a), there is a gap between the registered voices [If noiseco registers the voice using the recorded tape, the first recorded sound is "A" and the second recorded sound is recorded. If you decide which syllable the input registered voice corresponds to simply by the order of the sounds recorded on the tape, such as when the sound is "i" Since the headwords are made to correspond to each other, the actual registered voice input and the headwords are misaligned.

Here, Fig. 9(b) shows [Noiseco is mistakenly recognized as voice,
This is a diagram where [noise] has been input in place of the headword ``E,'' and the syllable ``9'' has been input in place of the headword ``O.''

In this way, when creating a standard pattern from registered voices, the recorded voice and headwords may deviate due to the influence of noise, etc., so the types of registered voices are indicated as shown in Figure 9 (C). The character code sounds obtained are recorded in a recording/playback device (2) in correspondence with the registered sounds. With this method, even if [Noiseko] is recorded between "U°' and "E," as described above, a discrepancy between the input sound and the headword can be prevented.

A method of recording a character-code sound of a specific frequency to prevent this deviation will be explained separately for the case where the tape recorder of the recording/reproducing device (2) is a single track and the case where the tape recorder is a multi-track.

First, with reference to FIG. 10, a case will be described in which a recording/playback device with multi-track is used as the recording method.

When a multi-track recording/playback device is used as a recording method, the character code corresponding to the headword is recorded on a track on which no audio is recorded, as shown in FIG. 2(a). In the voice recognition unit (1), this character
From the chord sound, we can know the headword of the input voice, and among the sounds recorded on the audio track, we can identify the sound that is over the sound pressure threshold among the sounds that have been recorded in the section t1 where this character-chord sound has been recorded. Only those that meet the conditions are considered to be audio and analyzed.

Alternatively, as shown in FIG. 2(b), a character code corresponding to the headword is recorded at the beginning and end of the audio, and the character code that indicates the beginning of the audio is selected from among the sounds recorded on the audio track. Among the sounds recorded in the interval t2 between the character and the chord sound indicating the end of the voice,
Only those that satisfy the condition of being equal to or higher than the sound pressure threshold are regarded as voices and analyzed.

Alternatively, as shown in FIG. 5C, a character code corresponding to the headword is recorded at the beginning of the audio. Among the sounds recorded on the audio track, among the sounds recorded in the section t3 from the character-code sound indicating the type of sound to the character-code sound corresponding to the next headword,
Only those that satisfy the condition of being equal to or higher than the sound pressure threshold are regarded as voices and analyzed.

The second method is a single-track recording/playback device (in the case of 2), the character code corresponding to the headword is expressed as a sound outside the voice analysis frequency band, and is recorded along with the voice on the track where the voice is recorded. The method for recording character chord sounds in this case is the same as in the multi-track case described above.
2. Among the sounds completely recorded in the interval t3, only those that satisfy the same conditions as described above are regarded as voices and analyzed.

However, except in the case of the example shown in FIG.
It is not necessary to use a sound outside the voice analysis frequency band as a chord sound.

Next, alphabets, numbers, parentheses, punctuation marks, etc. are preliminarily stored in the word dictionary (13) of the word recognition unit (13) as shown in Figure 6.
The word standard pattern corresponding to the word registered as a character in d) is stored in the word standard pattern memory (13) in the figure.
c) Register.

First, by a predetermined operation, the parameter buffer (13g) shown in FIG. 6 and the word standard pattern memory (13c) are connected by the switch <13sl), and the word registration mode is set.

Next, alphabets, numbers, parentheses, punctuation marks, etc. are displayed on the display device 6) of the main body of the apparatus (100), and the operator inputs the corresponding pronunciation by voice.

The speech recognition section (1) analyzes this speech and registers the word standard pattern in the word standard pattern memory (13c).

Voice recognition becomes possible through the operations described above.

However, independent word/attached word dictionary (14e) and word dictionary 1
If you want to recognize a word that is not in (L3d), use the independent word
Although the words to be recognized are registered in the adjunct word dictionary (14e), it is necessary to register the words to be recognized in the word dictionary (13d) and the word standard pattern in the word standard pattern memory (L3c). However, independent word/attached word dictionary (14
Whether to register words in e) or register words and word standard patterns in the word dictionary (13d) and word standard pattern memory (13c) depends on whether the user wants the word to be recognized as a phrase utterance, or whether the word Determine whether you want it to be recognized as a vocalization.

In addition, if the word is in the independent word/adjunct word dictionary (14e) but not in the word dictionary <13d) and you still want to recognize it with word recognition, you can add the word to the word dictionary (13d) and word standard pattern memory (13c). , it is necessary to register words and word standard patterns.

The method for registering arbitrary words will be described below.

To register words, use the independent word/attached word dictionary (14a)
, register the character string in the word standard pattern memory (13c), and register the word standard pattern in the word standard pattern memory (13c).
There are two methods for registering character strings in the word dictionary (13d).

A dictionary of independent words and adjunct words! When registering in (14e), speak the word you want to register and input it into the device.

At this time, this device recognizes this voice using the voice recognition section (1) and displays the recognition result on the display device (6). If the result is correct, the user presses the specified key on the keyboard (7>) and registers the uttered voice as the character string displayed on the display device (6) in the independent word/adjunct word dictionary (14Q).
If the recognition result displayed on the display device (6) is not correct, the recognition result displayed on the display device (6) is corrected using the syllable correction function of the present device, or the word to be registered is reuttered. If the re-uttered result is incorrect, it is corrected again using the syllable correction function of this device. The above operations are repeated until the character string displayed on the display device (6) matches the word to be registered.

When registering a word in the word standard pattern memory (13c) and word dictionary (13d), first the display device (
6) Display the character string you want to register correctly. Next, the correctly recognized character strings and word standard patterns are stored in a word dictionary (
13d) and Ia words! M quasi-pattern memory (13c)
Register each.

In addition, recognizing natural voice input is
Considering the current level of speech recognition technology, this is unreasonable. Since the current level of speech recognition technology is limited to continuous syllable utterance input, an example of continuous syllable utterance input will be described below.

In the case of continuous syllable speech input, the above procedure is the same, but in the case of continuous syllable speech input manually, the R-word standard pattern is also a continuous syllable speech pattern, so the word you want to register can be recited naturally. A word standard pattern is created by natural utterance, and the word standard pattern and character string are registered in a word standard pattern memory (13C) and a word dictionary (13d), respectively.

Through the above operations, the data necessary for creating sentences using voice recognition has been registered.

(i) Text creation An example of text creation will be described below.

Also, when performing a recognition operation, the switch (L3sl) of the word recognition unit (13) is set to the parameter buffer (13a).
) and the word determination section (13b), and the switch (14sl) of the phrase recognition section (14) is set so as to connect the parameter buffer (14a) and the syllable recognition section (14b).

There are three ways to write sentences.

The first method is an online recognition method in which the text to be created is directly input into the main body of the device by voice from the microphone (3).

The second method is to use a recording/playback device that records the text (2).
This is offline recognition in which the device is connected to the device, the recorded text is played back, and the text is recognized.

First, an example of online recognition will be described.

In the case of online recognition, sentences uttered in units of phrases or words are input directly from the microphone 3) to this device, so by performing the specified operation, the microphone (3) and the voice recognition unit Connect (1).

Also, if you want to record the audio input from the microphone (3) on the recording/playback device (2),
) to the main unit, and connect the input switching section (4) to the microphone 3.
) and the recording terminal of the recording/playback device (2).

At the same time, the input switching section (4) functions to record a beep sound indicating a phrase or word break when a silence detection signal is input from the feature extraction section (12) as described later.

During speech recognition, the word recognition unit (13) and the phrase recognition unit (1
4) is running.

The audio input from the microphone (3) is processed by the preprocessing unit (11).
The input voice is processed to have characteristics suitable for voice analysis (for example, when the sound pressure of the input voice is low, the sound pressure is amplified using an amplifier), and the feature extraction unit (12) sent to.

In the feature extraction section (12), as shown in FIG.
).

At the same time, the analysis east position determination unit (12b) of the feature extraction unit (12)
), based on the analysis results of the analysis unit (Ha), it was found that the silent interval after each syllable or phrase was uttered, and the beep sound recorded after each phrase or word was uttered (details will be explained later in offline recognition). ) shown in the embodiment, and when a silent section is detected, a silent section detection signal (b) is generated.

The parameter buffer (12c) which receives the silent section detection signal (b) sends the stored characteristic parameters to the word recognition section (13) and phrase recognition section (14), and erases the stored contents.

The feature parameters that cannot be input to the word recognition unit (13) are stored in the parameter buffer (13a) shown in FIG. The word determination unit (13b) compares the feature parameters stored in the parameter buffer (13a) with the word standard pattern memory (13c), and selects the feature parameters stored in the parameter buffer (13g>) and words with a large likelihood. Words with standard patterns are added to the word dictionary (13d
), and sends the character strings of the selected words and their likelihood values to the candidate creation section (15).

On the other hand, the feature parameters that have been input to the clause recognition unit (14) are stored in the parameter buffer (14a). In the syllable recognition unit (14b), the parameter buffer (14a)
The characteristic parameters stored in the syllable standard pattern memory (14d) are compared, and the characteristic parameters stored in the parameter buffer (14a) are compared.
A 3-phrase judgment unit (14c) converts the feature parameters stored in
In 4c), the input syllable string and the independent word/adjunct word dictionary (1
Compare the words registered in 4e), create multiple sets of phrases with a high likelihood by combining independent words and attached words, and send the character strings of the created phrases and their likelihood values to the candidate creation section (15). send.

The candidate creation section (15) selects a plurality of strings with a large likelihood from the input strings, and uses the likelihood value and the data sent from the word recognition section (13) or the phrase recognition section (14). A code indicating the data is added and stored. at the same time,
A signal is sent to the control unit (5) to display the character string with the greatest likelihood on the display device. When the control unit (5) receives this signal, it adds a delimiter mark "r'7" after the character string with the highest likelihood, and for example, for the input sentence in FIG. 14(a), ) is displayed on the display device in the format shown below. At the same time, the candidate creation section (15) controls the control section (
5), a signal is sent to cause the content stored in the candidate selection unit (15) to be stored in the storage device t(8). The control unit (5) receives this signal and stores the character string stored in the candidate creation unit (15) in the storage device (8) in the form of a code representing a delimiter added after the character string. The character strings stored in this external storage device are assumed to be the primary W and draft of the word processor. - A floppy disk is used inside the shell, and at this time the storage device (8)
The format of the file must match the file format of the word processor.

Further, the signal generating section (42) of the input switching section (4) shown in FIG. 8, which receives this silent section detection signal, generates a beep sound representing the break between clauses or words of the sentence, and switches the beep sound. Enter in <41). The switch (41) connects a circuit so that the audio input from the microphone (3) and the beep sound input from the signal generator (42) are recorded on the recording/playback device (2). (2) A beep sound is recorded in the silent period that is considered to be a break between the phrases or i-words of the sentence recorded in step (2).

Next, an example of offline recognition will be described.

For offline recognition, this device is equipped with a recording/playback device (2).
Since the text is created by inputting and reproducing the recorded voice, the text is first recorded on the recording/playback device (2).

In addition, in order to input audio from the recording and playback device (2),
The input switching section (4) connects the recording/reproducing device (2) and the voice recognition section 1).

When recording sentences, the system vocalizes phrases or words, creating silent intervals between phrases and words. In addition, when using a dedicated recording/playback device (2) for this device as shown in Figure 1,
In order to clearly mark the divisions between phrases and words, beep sounds indicating divisions can be played on the recording/playback device (2) or this daylight.
Separator key (71) that has been set on the machine
) to record.

In addition, the registered words are uttered in units of S words, but if the recording/playback device (2) has a character sound generation function and has a character corresponding to the word you want to input, it will be uttered instead of voice. You may also record the character sounds.

In addition, signals indicating the beginning and end of a sentence are recorded together with the voice in order to prevent noise recorded between sentences at the beginning of each sentence from being mistakenly recognized as voice.

However, the method of recording this signal changes depending on whether the recording/reproducing device (2) is a multi-track system or not, as described in the audio registration section. FIG. 11 shows a multi-track system and FIG. 12 shows a single-track system. FIGS. 11(a) and 12(a) show a method of detecting a section in which a sound such as a DTMF signal is recorded as an audio region.

Figures 11(b) and 12(a) show that a sound such as a DTMF signal is recorded before a sentence begins, and then recorded again when the sentence ends, and the section between these two signals is recorded. This is a method of detecting it as a voice area.

Furthermore, in the case of the single track system shown in FIG. 12, a DTMF signal or the like outside the audio band is used, considering that the audio section and the sound such as the DTMF signal overlap.

Furthermore, when recognizing a sentence, the sections t4 and t5 before and after the recorded signal are sampled, and in order to determine whether or not it is voice, it is not always necessary to identify the beginning of the sentence and the beginning of the signal, and the end of the sentence and the end of the signal. They do not need to match, so recognition is possible even if there is a slight lag between the timing at which the sentence is uttered and the timing at which the key is pressed.

Next, the recording and playback device (2) is connected to the main body of the device and the recorded voice is played back to perform recognition processing, but before the recorded voice is recognized, the recognition speed mode and the playback speed of the recorded voice are set. There is a fast recognition mode that speeds up the recognition time to shorten the recognition time, a mode that uses normal playback speed for recognition, or a double playback recognition mode when you have extra time and need a high recognition rate. Set it to the FL mode.

First, an example of the fast listening recognition mode will be described.

In fast listening recognition mode, the playback speed of the recorded audio is increased, so the characteristics of the input audio are different from the standard pattern created from the registered audio played back at the normal playback speed. Even if you input speeded up audio,
Speech recognition cannot be performed accurately.

Therefore, in order to accurately recognize audio that has been played back at a faster speed, the sampling frequency is changed. Examples of such methods are described below.

The sampling frequency control unit (12d) of the feature extraction unit 12) in FIG. speed) to sample and analyze the audio. Feature extraction part (12
) The subsequent processing is the same as in the example for online recognition, except that a beep sound is added to the recorded sentence of the recording/playback device #(2) to clearly indicate the break between phrases and words. When the feature extractor (12) detects the beep sound, the feature extractor (12) extracts the beep sound detection signal (mouth) instead of the silent section detection signal (b).
occurs. When the received signal is not a silent section detection signal (b) but a beep sound detection signal (mouth'), the one-word generator (42) of the input switching unit (4) detects the segment or word break of the sentence. No beep sound will be generated to indicate this.

Further, when the speech recognition unit (1) recognizes a character-sound indicating a word, it outputs a word corresponding to the character-sound as a recognition result.

Next, an example of the twice playback recognition mode will be described.

In this mode, the recorded audio is first played back and input to the device.

At this time, the preprocessing unit (11) of the speech recognition unit (1) reads all the sound pressure fluctuations of the recorded voice, and stores this data in the sound pressure fluctuation memory (llb) shown in Fig. 4. Play the recorded audio and input it to this device. At this time, the preprocessing unit (11) uses the h data stored in the sound pressure fluctuation storage memory (llb>) to determine the input sound pressure to the feature extraction unit (12) as shown in FIG. The amplification factor of the AGC circuit (lla) is adjusted to match the appropriate level. That is, the gain G is the fixed gain A multiplied by the control voltage V (variably adjusted).

Further, as another example of the twice playback recognition mode, a multiple playback recognition mode can also be considered. This involves playing and inputting a recorded sentence many times, changing the recognition method in the speech recognition unit (1) each time, comparing the recognized results, and selecting the one with the greatest likelihood of certainty. This is the way to do it.

In addition, the audio for registration is not recorded on recording/playback device #(2), and depending on the recording/playback device (2), the frequency characteristics when the playback speed is increased are different from those when the playback speed is normal. or when recognizing sentences recorded on a recording/playback device (2) that has different frequency characteristics from the recording/playback device (2) used to create the standard voice pattern, or when creating a standard voice pattern. Although it has the same frequency characteristics according to the standard as the recording and playback device (2) that was used, the actual frequency characteristics are different from the recording and playback device (2) that was used to create the standard audio pattern due to the influence of errors in the parts used etc. When the recording and reproducing device (2) recognizes recorded sentences, a function for correcting the influence of frequency characteristics described below is used.

First, a reference sine wave signal, which is a reference when measuring the frequency characteristics of the recording/playback device (2), is generated by the reference signal generator (42).
and record it on a recording/playback device (2). After that, the recorded reference sine wave signal is reproduced and input to the present device. The speech recognition unit (1) analyzes the input reference sine wave signal and determines the difference in frequency characteristics between the recorded reference sine wave signal and the reference sine wave signal generated by the reference signal generation unit (42). , a correction is made to reduce the difference in frequency characteristics between the recorded reference sine wave signal and the reference sine wave signal generated by the reference signal generator (42). A large number of means for applying the correction can be considered depending on the feature extraction method of the feature extraction section (12) of the speech recognition section (1). For example, as shown in FIG.
If an analog filter bank system is used, which consists of a parallel connection of an amplifier (AMP) and an amplifier (AMP),
By adjusting the amplification factor of AMP), the output from the filter is adjusted so as to reduce the difference in frequency characteristics from the reference sine wave signal generated by the reference signal generator (42). In addition, if a digital filter is used as the feature extraction method of the feature extraction unit (12), the parameters that determine the characteristics of the digital filter may be changed. Various methods can be considered corresponding to the feature extraction method 12).

By the above operations, the text input by voice has been converted into a kana string. A correction method when the kana string-converted text differs from the input text will be described below, using FIG. 14 and classifying each case into error. Correct it using the following steps.

Figure 14 (a) is the input text, Figure 14 (b) is the input voice, Figure 14 (c) is the recognition result, Figure 14 (d) to (h) are the correction process,
Figure <i) shows the correction results.

First, we will discuss a correction method when a word uttered is incorrectly recognized as a phrase. As shown in the same figure (C), what is uttered as the single aid ('' is the phrase ``shi-''
If the word is recognized as incorrect, first move the cursor (X) to the part of the incorrect word [(d) I in the same figure. Next, press the word next candidate key (72) to display the next candidate for Ugo [FIG. 8(d) 8F. If this result is correct, proceed to the next part to be corrected; if this result is incorrect, press the word next candidate key (72) again to display the next word candidate. Repeat this operation until the correct answer is displayed.

Next, we will discuss how to correct a phrase uttered as a phrase that is misrecognized as a word.If the phrase uttered as the phrase ``i'' is recognized as the word ``''E'', first move the cursor (X) Move to the incorrect clause. Next, press the phrase next candidate key (73) to display the next phrase candidate. If this result is correct, proceed to the next modification section.

If this result is incorrect, the phrase next candidate key (73)
Press to display the next phrase option. Repeat this operation until the correct answer is displayed.

It is also possible to display the previous candidate for a word by pressing the pre-word candidate key (74), and the previous candidate for each phrase by pressing the pre-phrase candidate key (75).

If the correct answer cannot be obtained using the above two correction methods, correction may be performed in units of syllables, or words, phrases, or syllables may be re-inputted.

In addition, in order to avoid recognizing phrases as words or recognizing words as phrases again during re-voice input, the candidate generation section (15) is configured to only recognize the recognition results sent from the word recognition section (13). Regarded as recognition result, phrase recognition unit (14)
The recognition results sent from the computer can be controlled from the outside to be ignored.

In addition, the candidate generation unit (15) is configured to consider only the recognition results sent from the phrase recognition unit (14) as recognition results, and to ignore the recognition results sent from the word recognition unit (13). More control.

The above-mentioned next candidate key is a key having the function described below, and will be explained using FIG. 15.

In the speech recognition unit (1) of this device, word recognition and phrase recognition run in parallel, and as mentioned above, both word and phrase recognition results are obtained. The phrase next candidate key (73) is used to display the recognition results on the display device 6) in descending order of the word recognition processing results, in descending order of likelihood. The key to display the recognition result on the display device is the word next candidate key (72), which displays the recognition result with one more likelihood than the recognition result currently displayed on the display device on the display device (6). The keys to be displayed are the pre-phrase candidate key and the pre-bunset candidate key.

FIG. 15 shows the candidate buffer (15a) of the candidate creation section (15).
). In this figure, the recognition result in the - position is 'tango', and the (# word) indicates that this is the recognition result sent from the word recognition unit (13>). Similarly, the recognition result of the third place is ``Tango is J, and this is the recognition result sent from the phrase recognition unit (14), which is expressed as (phrase), and the recognition result of the third place is `` Tangoni” and
This is a recognition result sent from the phrase recognition section (14), which is expressed as ku bunsetsu), and the fourth recognition result is 'tankou'', which is sent from the word recognition section (13). It is expressed by (word) that it is the result of recognition that has come.

It is now assumed that 'Tango' is being displayed on the display device (6). In this state, when the phrase next candidate key (73) is pressed, "Tango wo" is displayed on the display device (6). Further, when the next word candidate key (72) is pressed, ゝtanko'' is displayed on the display device (6).

In addition, when the display device (6) is displaying ``tankou'', if you press the pre-word candidate key (74), the display device (6) will display ``matango,'' and the pre-word candidate key (74) will be displayed. key (7
3) When you press ``Matagoni'' is displayed on the display device (6).

Next, we will discuss how to modify an entire passage at once.

In the example in Figure 14(e), is the word "T" incorrectly written as r A? 2
This is a well-known example. First, move the cursor to the word you want to correct [see (e) i in the same figure.

Next, press the word next candidate key (72) to display the next word candidate [FIG. 4(e) il]. If this result is correct, proceed to the next modification section. If this result is incorrect, press the Komoro next candidate key (72) to display the next candidate for the word. Repeat this operation until the correct answer is displayed. If the correct answer is not displayed, re-speak and re-enter. By pressing the previous word candidate key (74), candidates for the word displayed immediately before can be displayed.

Next, we will discuss how to correct an entire word at once.

The example in FIG. 14(f') is an example in which the phrase ``gamen no'' is incorrectly recognized as ``1 gainen no''. First, move the cursor to the phrase you want to modify [Figure (r) il.

Next, press Phrase Next Candidate -<73) to display the next bunsetsu candidate [Figure (r) il, if this result is correct, proceed to the next correction part, if the result of this 4 is incorrect, Press the phrase next candidate key (73) to display the next phrase candidate.

Repeat this operation until the correct answer is displayed. If the correct answer is not displayed, re-speak and re-enter. By pressing the previous clause candidate key (75), display the candidate clause for the previous clause displayed. You can also do it.

Next, a method for correcting syllables will be described.

The example in FIG. 14(h) is an example in which the phrase ``Onseite'' is incorrectly recognized as ``Onkeide''. In this example, the syllable ``ke'' is to be corrected to ``ke''. By pressing the syllable next candidate key (76), the syllable closest to the syllable of the part you want to correct will be displayed [Figure (h) i]'. If the correct answer is displayed, move on to the next part to be corrected. Moving. If this result is incorrect, press the syllable next candidate key again to display the next syllable candidate.

Repeat B's operation until the correct answer is displayed. If the correct answer cannot be displayed, re-enter the answer by speaking again. If the re-input results are incorrect, correct them again using the steps above. Repeat the operation ′- until the correct answer is displayed.

Furthermore, by pressing the previous syllable candidate key (77), the previous syllable candidate can be displayed.

If you want to delete a syllable, move the cursor to the syllable you want to modify and press the delete key (78>) to delete it.

When you want to insert a syllable, move the cursor to the syllable you want to modify and press the insert key (79) to insert it.

Next, using FIG. 16, the several syllable correction method will be described.

In this example, the input sentence "Kaijiyo" in Figure (a) is incorrectly recognized as Figure (b > 1 Ganjo).In this case, first move the cursor (X) to the syllable you want to correct. [
Figure (c)], the word “kai” is used again and again. Such re-voiced input speech can be recognized by the speech recognition unit (1), and the recognition result is displayed on the display device (6>).If the recognition result is correct,
Proceed to the next correction part, if as shown in the same figure (d),
If you misrecognize "kai" as "kae", if it is an RL word, press the word secondary candidate key (72). If it is a 0 phrase, press the phrase next candidate key (73). Since this is an example of the case of
If you press the next word candidate key (72) in the state of d),
First, the control unit (5) selects from the word dictionary (13d) the recognition result "Kansho" in the same figure (b) before correction and the recognition result "1 Kaejiyo" in the same figure (d) after re-voicing. Compare and find the same parts. Next, the control unit (5) selects the same 81 minutes from the word dictionary <13d).
The same figure (f) is a word dictionary (13d
), and (g) in the same figure shows words with "C-yo" selected from the memory contents. Next, the control unit (5) calculates the likelihood between the word written in FIG. Figure (e)].

Next, we will discuss the case of correcting recognition boundary errors of phrases or words.

The example in Figure 14 (g) incorrectly recognizes that there is a silent section marked Co3 between the phrase ``Bunsyowo'' and ``1'', and! This is an example of a sentence being misrecognized as being divided into two parts, such as "AM'bun" and the phrase "shouwo."

In this case, the recognition boundary error must be corrected, but if you want to delete the recognition boundary delimiter, move the cursor (X) to the recognition boundary delimiter you want to delete [see figure (g)
il, press the delete key (78) [(g) il in the same figure. If you want to insert a recognition boundary delimiter, move the cursor (X) to the syllable at the position you want to insert and press the insert key (79).

However, as will be described later, the delimiter beep of the recording/playback device (2) and the delimiter added to the recognition result stored in the storage device (e) 8), so you must take precautions. Therefore, at this time, the storage device (8) indicates that the delimiter has been inserted and deleted.
8).

For example, the sentence shown in Figure 14(g)i is
g) As shown in i, it is assumed that the storage device (8) is fully stored. If the sentence in (g)i is modified as shown in (g)i, the delimiter ``ma'' stored in the storage device (8) will be changed to the symbol rv.The symbol ``vJ'' will be changed to the delimiter ``vJ''. This symbol indicates that "Symbol 17" has been deleted, and is not used as a symbol to indicate a recognition unit.
2) This symbol is used only for control.

With such a configuration, even after delimiting the delimiter r, the beep sound recorded in the recording/playback device (2) and the symbol rv stored in the storage device (8) will be retained.

rv, it is possible to control both devices while maintaining synchronization.

The above is an example where the delimiter "7" is deleted, but the same idea can be applied when it is inserted. In other words, a specific symbol that is not used as a control signal and represents only a delimiter may be inserted in place of the delimiter "1ma".

By the above correction procedure, as shown in FIG. 14(i),
Correct the text.

After the recognition boundary error has been corrected, the recognition unit for which the recognition boundary error has been corrected is corrected according to the correction procedure. In the case of correction by re-voicing, the voice of anyone who has registered the standard pattern can be recognized, so corrections can be made even if one is not the person who recorded the text.

The method for correcting kana string sentences has been described above, and the following functions are provided to assist in correction.

By moving the cursor on the character string displayed on the display device (6) and scrolling the display screen, it is possible to sequentially display memorized sentences from the storage device (8) on the display screen, but at this time, the memorized sentences can be displayed on the display screen in sequence. The audio corresponding to the part can be played back from the recording/playback device (2).

It also has the opposite function to the above-mentioned function, and has a recording/playback device (
2), a character string corresponding to the portion that has been completely reproduced is displayed on the display device (6).

In addition, in both of the above methods, the time before the delimiter recorded in the recorded text and the delimiter recorded on the display side are used as timing signals for synchronization, and the playback of the recording and playback device (2) is performed. and the display are controlled so that they operate in synchronization with each other. Also, the keyboard (7
), or when No. 6 to stop the playback has been inputted from the recording/playback device (2), the playback is stopped and the scrolling of the display or the movement of the cursor is also stopped.

The above synchronization function between playback and display of the recording/playback device (2) allows character strings to be checked while listening to the playback sound, making it easier to find corrections.

As for the synchronization method described here, the character string in the storage device (8) corresponding to the part where the playback is completed is transferred to the display device (6).
), and how to display the kana column of the part that is delayed by one delimiter from the part corresponding to the part that has been played completely on the display device (
6) is a display method.

In this case, when the display is stopped for correction, the corrected part of the recorded voice has already been played back, so in order to play the corrected part again, it is necessary to locate the beginning of the part to be corrected from the reproduced text. Therefore, when this method is adopted, the recording/reproducing device (2) is provided with a function to automatically pack and rack up to the previous delimiter when the display is stopped.

In addition, when a tape recorder is used as the recording/playback device (2), the playback section is controlled by the rotation of a motor, and the beginning of the section corresponding to the correction section may not be accurately located due to slack in the tape, etc. There is.

In such a case, record the input audio for a long time using PCM recording or ADPCM recording, and if you want to listen back to the input audio, use PCM recording or ADPCM recording.
Adds a function to listen back to P'CM recorded audio.

FIG. 17 shows an embodiment of the above-mentioned functions, and numbers 01 to 05 in FIG. 0, which is a diagram of a PCM data memory for storing PCM recording data, indicate addresses. The input voice is ``1 circle, which is 1 unit, which is 1 unit, 1 unit, and 1 unit, which is 1 unit, 1 unit, and 1 unit.'' It is a text.

When the above audio is input, the PCM data memory <
In the DM), the voice “Washiwa °°” up to the first silent section is stored at address 01. The voice “ten” up to the second silent section is stored at address 02. The fifth silent section is stored at address 05. The voice "te-"' up to the section is completely memorized. At this time, the PCM address pointer (AP) stores the address of the data that is stored first among the data stored in the PCM data memory. In this example, 01 is stored.

At this stage, the PCM data memory is full.

Next, when audio is input, the PCM data memory (
9. Store the input voice at the address of the first data stored in the DM).
In this example, "Kamen no" is stored at address 01 where "Washiwa" was stored. At this time, the PCM address pointer (AP) stores the address of the first data stored in the PCM data memory (DM). In this example, 02 is stored. Ru.

In this state, when reproducing the contents of the PCM data memory (DM), the contents are reproduced starting from the address pointed to by the PCM address pointer (AP). In this example, 02, 03,
It will be played back in the order of 04, 05, and 01.

With this method, it is possible to listen back to the audio accurately and quickly as many times as desired.

In addition, the recognition unit delimiter on the screen (X
) and press the recording audio cue key (70), the recording/playback value r! corresponding to the recognition unit indicated by the cursor is displayed. It has a function of searching the t(2) side delimiter sound part from the recorded sentence and reproducing the sentence that follows. An example of such functionality is shown below.

To confirm the recognized sentences, the recognition results are stored in a storage device (8)
The data is read out from the beginning and displayed on the display device (6) from the beginning.

At this time, in FIG. 19, the delimiter counter (5a) of the control unit (5>) counts the number of delimiters read from the storage device (8).If the read recognition result is incorrect, Place the cursor on the incorrect part and press the cue key.The control unit (5) causes the recording and playback device (2) to record and play back the text whose sound is cut out in fast forward playback mode.Feature extraction unit (12) Beep sound counter <12a)
counts beeps without marking the breaks in a sentence input from the recording/playback device (2).

When the value of the beep sound counter (12e) becomes one smaller than the value of the delimiter counter (5a) mentioned above, the comparison circuit (5b) outputs the icon number (c) to the recording and reproducing device (2). and stop playback.

In addition, in order to confirm the recognition result and the corrected text, the data stored in the storage device (8) can be displayed on the display device (6).
It is very tiring for the eyes because the user has to visually follow and read the string displayed on the display screen.

In view of this, this device has a function to read out the character strings stored in the storage device (8) that can store the recognition results using a voice synthesis function. You can do it by listening to the sound.

In this case as well, the delimiter is used as a timing signal for synchronizing the speech synthesis section (9), the storage device (8), the recording and reproducing device (2), and the display device (6).

In other words, the character string corresponding to the part that the speech synthesis unit (9) reads out from the storage device (8) is displayed on the display device (6>), and at the same time, the recording part is cued up from the recording/playback device (2>). With this method, when an error is discovered by the voice synthesis function and the voice synthesis function is stopped to correct it, neither the display on the display device mF (6) nor the recorded part on the recording/playback device (2) will be displayed. Errors are shown and corrections can be made immediately.

The method of synchronization described here is to display the kana column of the storage device corresponding to the part being read aloud by the speech synthesis function on the display device (6), and at the same time display the text recorded on the recording/playback device (2). There are two methods: one is to play back the syllable part that corresponds to the part that is being read aloud by the speech synthesis function, and the other is to play back the part of the sentence recorded in the recording and playback device (2) that is one delimiter later than the part that corresponds to the part that is being read aloud by the speech synthesis function. be. In the latter case, when voice synthesis is stopped for correction, the recording and reproducing device (2) has stopped before the part to be corrected, so if it is played back in this state, the corrected part of the audio can be immediately played back. In the former case, when voice synthesis is stopped for correction, the corrected part of the recorded voice has already been played back, so it is necessary to backtrack in order to play the corrected part again. Therefore, when adopting the former method, when the display is stopped, the recording and playback device (2) is automatically activated.
It is preferable to have the function of backtracking to the previous delimiter.

Above, an example has been described in which the recognition result is stored in the storage device (8), but below, as another example, a method of storing the recognition result in the recording/reproducing device (2) will be described.

The recognition result stored in the storage device (8) is recorded in the recording/playback device (2) where the original text was recorded. In this way, the original text and recognition results can be recorded on the same recording medium, so
It becomes easier to manage the original text and recognition results.

Also, by recording the recognition results in the recording/reproducing device (2) while reproducing and inputting the recording 1-sentence, an external storage device becomes unnecessary.

In either case, a multi-track recording/playback device (
By using 2), it is possible to store the recognition result in a track on which no audio is recorded while reproducing the recorded audio.

Below is a margin (G) Effects of the Invention Q A standard pattern created from audio recorded on a sound reproducing device and a standard pattern created from audio input directly from a microphone differ by the frequency characteristics of the recording and reproducing device.

Therefore, when recognizing a recorded voice, it is necessary to use a standard pattern created from the recorded voice, and when recognizing a voice input directly from a microphone, it is necessary to use a standard pattern created from the recorded voice.
It is necessary to use a standard pattern created from the input voice. Therefore, the speech recognition device must have an internal standard pattern, but according to the system of the present invention, both the standard pattern created by direct input from the microphone and the standard pattern created from the recorded voice can be registered in a single voice registration process. Since it can be created manually, the number of vocalizations required by the registrant can be cut in half, reducing the burden of voice work.

[Brief explanation of the drawing]

Figure 1 is an external view of a daytime cheating machine that employs the voice recognition system of the present invention, Figure 2 is a configuration diagram of the daytime cheating machine, Figure 3 is a configuration diagram of the voice recognition section (1), and Figure 4 is a block diagram of the preprocessing section (11), FIG. 5 is a block diagram of the feature extraction section (t2), FIG. 6 is a block diagram of the word recognition section (13), and FIG. 7 is a block diagram of the phrase recognition section (14). Fig. 8 is a block diagram of the input switching section (4), Fig. 9 is a diagram showing the relationship between headwords, recording methods, and characters and sounds, and Fig. 10 is a relation between character and sound recording methods and voice sections. Figure 11 is a diagram showing the recording method when the recording/playback device is a multi-track system, Figure 12 is a diagram showing the recording method when the recording/playback system is a single-track system, and Figure 13 is an example of a frequency correction circuit. Figure 14 is a diagram showing corrections made in the case of misrecognition, Figure 15 is a diagram showing the candidate buffer (15a) in the candidate creation section (15), and Figure 16 is an example of correcting several syllables in the case of misrecognition. Figure shown, 1st
Section 7 is an explanatory diagram of the PCM recording method, FIG. 18 is an explanatory diagram of AGC operation, and FIG. 19 is an explanatory diagram of a delimiter counter. (1)...Speech recognition unit, (2)...Recording and playback device,
(3)...Microphone, (6)...Display device, (7)...
...Keyboard, (8)...Storage device, (11)...
・Pre-processing unit, (12)...Feature extraction unit, (13)...
・Word recognition unit, (14>... phrase recognition unit, (lla)
...Variable gain amplifier, (llb)...Sound pressure variation memory.

Claims (1)

[Claims] (1) Create standard patterns from a recording and playback device that can record and play back the voice input from the microphone, the voice input directly from the microphone, and the playback voice obtained from the recording and playback device, and use these as standard patterns. It is equipped with a voice recognition device that stores patterns, and when the registered voice is input into the voice recognition device from the microphone to create a standard pattern, this registered voice is also recorded on the recording and playback device, and the voice recognition device can record the registered voice from the microphone input. After the registration is completed, the recorded voice is played back to the voice recognition device and registered, so that in addition to the standard pattern of the voice input from the microphone, the standard pattern of the voice input from the recording and playback device is reproduced. A speech recognition system characterized by the ability to obtain