JP2015191219A - Voice processing system, voice processing method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a voice processing system, a voice processing method, and a program which can accurately detect a voice during output of a response signal even when the response signal and the input voice are the voice.SOLUTION: The voice processing system includes: response selection means for selecting a response voice; band selection means for selecting at least a part of a band of the band excluding the frequency band of the selected response voice; and voice detection means for detecting the voice of the input signal at least in the part of the selected band.

Description

  The present invention relates to a voice processing system, a voice processing method, and a program.

  In the voice response / voice dialogue, there is a barge-in function for interrupting and inputting a voice while outputting a response signal or a response voice. The barge-in function is used, for example, when rephrasing due to misrepresentation or misrecognition of voice recognition, when performing the next input immediately without waiting for a response, or when re-executing because of a change in mind. However, there is a problem that the performance of voice detection and voice recognition deteriorates when the response signal goes around the microphone for voice input.

  On the other hand, there is a method of paying more attention to the frequency characteristics of the response signal and performing greater weighting on the frequency band with less response signal (Patent Document 1). The method described in Patent Document 1 weights each frequency band based on a system sound of multiple channels, that is, a response signal (for example, beep sound or music), and applies a smaller weight to a band that is likely to be included in the response signal. / Non-voice discrimination. Therefore, the method described in Patent Document 1 can suppress degradation in performance such as voice detection when the overlap of frequency bands between the response signal and the input voice is small.

JP2012-189907

When the response signal and the input voice are voices such as a human voice as in the voice dialogue system, the frequency bands of the response voice and the input voice overlap greatly. Therefore, the technique described in Patent Document 1 has a problem that the performance of voice detection and the like is lowered.

[Object of invention]
An object of the present invention is to provide a voice processing system, a voice processing method, and a program capable of accurately detecting a voice while outputting a response signal even when the response signal and the input voice are voices.

  The present invention provides response selection means for selecting response voice, band selection means for selecting at least a part of a band excluding the frequency band of the selected response voice, and at least part of the selected band. Is a voice processing system including voice detection means for detecting voice of the signal input in.

  The present invention provides a response selection step of selecting a response voice, a band selection step of selecting at least a part of a band excluding a frequency band of the selected response voice, and at least a part of the selected band And a voice detection step of performing voice detection of the signal input in.

The present invention provides a response selection step of selecting a response voice, a band selection step of selecting at least a part of a band excluding a frequency band of the selected response voice, and at least a part of the selected band Is a program that causes a computer to execute a voice detection step of performing voice detection of a signal input in.

In the speech processing system according to the present invention, even when the response signal and the input speech are speech as in the speech dialogue system, speech detection can be performed with high accuracy during the output of the response signal.

It is a hardware block diagram concerning the 1st embodiment of the present invention. It is a block diagram concerning the 1st embodiment of the present invention. 3 is a flowchart according to the first embodiment of the present invention. It is a block diagram concerning the 2nd embodiment of the present invention. It is a conceptual diagram concerning the 2nd Embodiment of this invention. It is a flowchart which concerns on the 2nd Embodiment of this invention. It is a block diagram concerning the 3rd embodiment of the present invention. It is a flowchart which concerns on the 3rd Embodiment of this invention.

Embodiment 1. FIG.
FIG. 1 is a hardware configuration diagram of a speech processing system 1 according to the first embodiment of the present invention. As shown in FIG. 1, the audio processing system 1 includes a CPU 10, a memory 12, a hard disk drive (HDD) 14, a communication interface (IF) 16 that performs data communication via a network (not shown), a display device 18 such as a display, and the like. The input device 20 includes a pointing device such as a keyboard and a mouse. These components are connected to each other through the bus 22 and input / output data to / from each other. Note that the hardware configuration of the voice processing system 1 is not limited to this configuration, and can be changed as appropriate. Further, the voice processing system 1 does not need to be configured by one computer system, and may be configured by a plurality of computer systems.

  FIG. 2 is a block diagram showing the configuration of the speech processing system according to the first embodiment of the present invention.

  As shown in FIG. 2, the voice processing system according to the first embodiment includes response selection means 111, band selection means 121, and voice detection means 131.

  The response selection unit 111 selects a response voice whose frequency band to be used in advance is known, and notifies the band selection unit 121 of the response voice. The response voice is a voice output from the voice processing system. The response voice is, for example, a voice having a content that responds to a certain input voice.

  The band selection unit 121 selects a band excluding the frequency of the response voice selected by the response selection unit 111, and notifies the voice detection unit 131 of band information that is information on the selected band. The band selection unit 121 may select a band excluding at least a part of the response voice frequency. For example, the band selection unit 121 may select a band excluding frequencies that contain more response voices from the response voice frequencies.

  The voice detection means 131 performs voice detection on the input signal using the band information. The voice detection unit 131 may perform voice detection using at least a part of the selected band.

  Next, the operation of the first embodiment for carrying out the present invention will be described in detail. FIG. 3 is a flowchart illustrating an example of the operation of the first embodiment.

  The response selection unit 111 selects a response voice whose frequency band to be used in advance is known, and notifies the voice detection unit 131 of the selected response voice (step 101). The band selection unit 121 selects a band excluding at least a part of the frequency band of the response voice selected by the response voice selection unit 111 and notifies the voice detection unit 131 of band information that is information on the selected band. (Step 102). The voice detecting means 131 performs voice detection in at least a part of the selected band (step 103).

  If the voice detecting means 131 detects voice, the process returns to step 101 (step 104).

  According to the present embodiment, since the voice of the input signal is detected using a band other than the frequency band of the response voice, the voice of the input signal can be detected even while the response signal is being output.

Embodiment 2. FIG.
FIG. 4 is a block diagram showing a configuration of a voice processing system according to the second embodiment of the present invention. The speech processing system according to the second embodiment of the present invention corresponds to the first example. The voice processing system according to the second embodiment of the present invention may be a voice interaction device.

  The voice processing system according to this embodiment includes a voice response device 2, a response voice storage unit 212, an output unit 162, and an input unit 172. Is provided. The voice response device 2 includes a response selection unit 112, a band selection unit 122, a voice detection unit 132, a voice recognition unit 142, and a voice reproduction unit 152.

  The response selection unit 112 selects a response voice from one or more response voices whose frequency bands to be used in advance stored in the response voice storage unit 212 are known. Further, the response selection unit 112 notifies the selected response voice to the band selection unit 122 and the voice reproduction unit 152.

  The band selection unit 122 selects a band excluding the frequency band of the response voice selected by the response selection unit 112 and notifies the voice detection unit 132 of the band information.

  The band selection unit 122 may notify the band excluding the entire band including the response voice as band information, or the band excluding the band including the response voice in a certain time unit as the band information together with the time information. You may be notified. For example, as shown in FIG. 5, the band selection unit 122 may extract a band including the response voice per unit time (for example, for each processing frame) and set the band on the opposite side as band information. At that time, the response selection unit 112 may select the response voice so as to use the frequency band used immediately before as long as possible. Thereby, when the same user performs voice input continuously, the frequency bands of the response voice and the input voice overlap less. The technique of using the band information excluding the entire band including the response voice as the band information can perform voice detection at a low cost. On the other hand, the method of changing the band information in a fixed time unit can perform voice detection with high accuracy.

  Further, the band selection unit 122 may divide a frequency band that is a target of voice detection into a plurality of subbands in advance, and discretely select the corresponding subbands. Furthermore, the band selection unit 122 performs weighting according to the amount of response sound included in each subband. The band selection unit 122 may reduce the weighting of subbands with a larger amount. Since the technique for performing voice detection for each subband is a known technique, a description thereof is omitted here.

  The voice detection unit 132 receives input voice from an input unit 172 (to be described later) and band information from the band selection unit 122, and performs voice detection of the input voice.

  When the sound detection unit 132 detects a sound, the sound detection unit 132 may notify the sound reproduction unit 152 described later. The audio reproduction means 152 may stop the audio reproduction upon receiving the notification. As a result, the voice processing system according to the present embodiment stops the playback of the response voice as soon as the voice detection is correctly performed, so that subsequent processing such as voice detection and voice recognition can be performed with higher accuracy. It becomes.

  In addition, when the band information received from the band selection unit 122 is weighted for each subband, the voice detection unit 132 may change the threshold for detection according to the weight. For example, the voice detection unit 132 may use a detection result in a subband having a large weight as a result with higher reliability. Thereby, the voice detection means 132 can detect the voice with higher accuracy. The voice recognition means 142 recognizes the voice input by the input means 172 described later. Further, the response selection unit 112 selects a response voice based on the voice recognition result of the voice recognition unit 142.

  The sound reproducing means 152 causes the output means 162 to reproduce the response sound selected by the response selecting means 112.

  The output unit 162 is a unit that outputs a signal such as voice. The output unit 162 may be a speaker, for example.

  The input means 172 is a means for inputting a signal such as voice. The input unit 172 may be a microphone, for example.

  The response voice storage unit 212 is a unit that stores the response voice.

  Next, the operation of the second embodiment for carrying out the present invention will be described in detail. FIG. 6 is a flowchart illustrating an example of the operation of the second embodiment.

  The response selection unit 112 selects a response voice whose frequency band to be used in advance is known from the response voice storage unit 212, and notifies the voice playback unit 152 and the band selection unit 122 of each of them (step 201). Answer selection means 112, for example, may be selected response voice suitable for the start of the conversation, such as "Hello" at system startup. The band selection unit 122 selects a band excluding the frequency band of the response voice notified from the response selection unit 112, and notifies the voice detection unit 132 of the band information (step 202). The sound reproducing means 152 reproduces the response sound notified from the response selecting means 112 on the output means 162 (step 203).

  The voice detection unit 132 receives the input voice from the input unit 172 and the band information from the band selection unit 122, and performs voice detection of the input voice (step 204). When the voice detection unit 132 detects voice, the voice recognition unit 142 performs voice recognition using the voice detection result, and returns to step 201 (steps 205 and 206).

  According to the present embodiment, since the voice of the input signal is detected in a band other than the frequency band of the response voice, the voice of the input signal can be detected even while the response signal is being output. In particular, when the frequency band of the response voice and the input voice is likely to overlap, the voice can be detected with higher accuracy by changing the voice detection band according to the time change of the response voice band. Further, by selecting the response voice so as to use the frequency band used immediately before as much as possible, it is possible to avoid frequency band overlap more accurately when the same user performs voice input continuously.

Embodiment 3. FIG.
FIG. 7 is a block diagram showing the configuration of a speech processing system according to the third embodiment of the present invention. The speech processing system according to the present embodiment corresponds to the second example.

  The voice processing system of this embodiment includes a voice response device 3, a response voice storage unit 213, an output unit 163, and an input unit 173. The voice response device 3 includes a response selection unit 113, a band selection unit 123, a voice detection unit 133, a voice recognition unit 143, a voice reproduction unit 153, and a scenario reference unit 183.

  The response selection unit 113, the band selection unit 123, the voice detection unit 133, the voice reproduction unit 153, the output unit 163, the input unit 173, and the response voice storage unit 213 are respectively the response selection unit 112, the band selection unit 122, and the voice detection unit. 132, the sound reproduction means 152, the output means 162, the input means 172, and the response sound storage means 212 have the same functions, and thus description thereof is omitted.

  The voice recognition unit 143 recognizes the voice input from the input unit 173 using the voice detection result notified from the voice detection unit 133. The voice recognition unit 143 notifies the scenario reference unit 183 of the recognition result.

  The scenario reference unit 183 refers to the scenario storage unit 223 and notifies the response selection unit 113 of a scenario corresponding to the recognition result notified from the voice recognition unit 143.

  The response selection unit 113 selects a response voice corresponding to the notified scenario.

  The scenario storage unit 223 stores a scenario indicating the content of a response corresponding to the voice recognition result.

  Next, the operation of the third embodiment for carrying out the present invention will be described in detail. FIG. 8 is a flowchart illustrating an example of the operation of the third embodiment.

  First, a response sound is selected from one or more response sounds whose frequency bands to be used in advance stored in the response sound storage unit 213 are known. Further, the response selection unit 113 notifies the selected response sound to the band selection unit 123 and the audio reproduction unit 153. (Step 301).

  The band selection unit 123 selects a band excluding the frequency band of the response voice notified from the response selection unit 113, and notifies the voice detection unit 133 of the selected band (step 302).

  The voice reproduction means 153 reproduces the response voice notified from the response selection means 113 by the output means 163 (step 303).

  The voice detection unit 133 receives the input voice from the input unit 173 and the band information from the band selection unit 123, performs voice detection, and when the voice is detected, notifies the voice recognition unit 143 of the detection result (step). 304).

  When the voice detection unit 133 detects a voice, the voice recognition unit 143 performs voice recognition using the voice detection result. Further, the voice recognition unit 143 notifies the scenario reference unit 183 of the voice recognition result (steps 305 and 306). Since the technology for performing speech recognition using the speech detection result is a well-known technology, description thereof is omitted here.

  The scenario reference unit 183 refers to the scenario storage unit 223, and if there is a response content corresponding to the voice recognition result, notifies the response selection unit 113, and returns to step 301 (steps 307 and 308).

  According to the present embodiment, even in a voice processing system that responds based on a scenario, voice detection of an input signal can be performed even while a response signal is being output.

In the plurality of flowcharts used in the above description, a plurality of processes are described in order, but the execution order of the processes executed in each embodiment is not limited to the description order. In each embodiment, the order of the illustrated steps can be changed within a range that does not hinder the contents. Moreover, each above-mentioned embodiment and modification can be combined in the range with which the content does not conflict.

DESCRIPTION OF SYMBOLS 1 Voice processing system 2 Voice response apparatus 3 Voice response apparatus 10 CPU
12 Memory 14 HDD
16 Communication IF
18 Display device 20 Input device 22 Bus 111, 112, 113 Response selection means 121, 122, 123 Band selection means 131, 132, 133 Voice detection means 142, 143 Voice recognition means 152, 153 Voice playback means 162, 163 Output means 172 , 173 Input means 183 Scenario reference means 212, 213 Response voice storage means 223 Scenario storage means

Claims (10)

A response selection means for selecting a response voice;
Band selection means for selecting at least a part of the band excluding the frequency band of the selected response voice;
A speech processing system comprising speech detection means for performing speech detection of a signal input in at least a part of the selected band.
2. The voice processing system according to claim 1, wherein said band selecting means selects a band excluding a frequency band of said selected response voice on a time unit basis.
The band selection unit weights the subband so that the smaller the amount including the selected response voice in the subband is, the smaller the value is.
The voice processing system according to claim 1, wherein the voice detection unit performs voice detection of the input voice based on a value weighted for each subband.
The voice processing system according to any one of claims 1 to 3, wherein the response selection unit selects the response voice so that a large number of frequency bands including the immediately previous response voice are included. A response selection step of selecting a response voice;
A band selection step of selecting at least a part of the band excluding the frequency band of the selected response voice;
And a voice detection step of performing voice detection of a signal input in at least a part of the selected band.
6. The voice processing method according to claim 5, wherein the band selection step selects a band excluding a frequency band of the selected response voice in units of time.
In the band selection step, the subband is weighted so as to be a smaller value as the amount including the selected response voice in the subband is larger.
The speech processing method according to claim 5 or 6, wherein the speech detection step performs speech detection of the input speech based on a value weighted for each subband.
A response selection step of selecting a response voice;
A band selection step of selecting at least a part of the band excluding the frequency band of the selected response voice;
A program for causing a computer to execute a sound detection step of performing sound detection of a signal input in at least a part of the selected band.
9. The program according to claim 8, wherein said band selecting step selects a band excluding a frequency band of said selected response voice in units of time.
In the band selection step, the subband is weighted so as to be a smaller value as the amount including the selected response voice in the subband is larger.
The program according to claim 8 or 9, wherein the voice detection step performs voice detection of the input voice based on a weighted value for each subband.

Images