JP2009109523A - Voice recognition system and voice recognizer - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a voice recognition system capable of improving availability, and to provide a voice recognizer. <P>SOLUTION: The voice recognition system 1 inputs reproduction signals Sb, Sc output from a television set 22 with a mixer part 12. The mixer part 12 selects and outputs a command signal Sh previously contained in the reproduction signals Sb, Sc for controlling control equipment 20 in a specific mode. In the specific mode, an echo canceling part 13 inputs a voice signal Sa from a voice input part 11a and the command signal Sh from the mixer part 12, and outputs only the command signal Sh of the signals Sa, Sh to a recognition collation processing part 14. The recognition collation processing part 14 collates the command signal Sh and a reference pattern signal and executes control based on the command signal Sh with the control equipment 20 when being collated. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a voice recognition system and a voice recognition device.

  2. Description of the Related Art In recent years, a speech recognition system has been known in which utterance speech from a user is input and a control device is controlled based on the input utterance speech. Such a speech recognition system stores a standard pattern signal in advance. Then, the voice recognition system converts the input utterance voice into a voice signal and collates it with a standard pattern signal. If the degree of coincidence between both is a certain value or more, a vocabulary based on the standard pattern (such as “illumination off”) Judge that there was an utterance. As a result, the speech recognition system controls the control device as instructed by the vocabulary based on the standard pattern.

In addition, some of these voice recognition systems have improved convenience by executing auxiliary processing such as explaining the operation method by playing guidance voice in consideration of user convenience ( For example, see Patent Document 1).
Japanese Patent Laid-Open No. 2001-154689

  However, the conventional speech recognition system still has room for improvement in terms of convenience.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to provide a speech recognition system and a speech recognition apparatus that can further improve convenience.

  A speech recognition system according to the present invention inputs speech speech from a user, controls a control device based on the input speech speech, and plays back content based on a playback signal, and playback device The voice input means for inputting the voice output by the user and the voice by the user's utterance and outputting the voice signal based on the voice, the voice signal from the voice input means, and the collation processing of the standard pattern signal stored in advance Recognition control processing means for performing control, control means for controlling the control device based on the result of verification by the recognition verification processing means, mode selection means capable of selecting one mode from a plurality of control modes for determining the processing method, mode When a specific mode is selected by the selection means, a playback signal is input from the playback device, and, in order to control the control device among the input playback signals, in advance First signal processing means for selecting and outputting a command signal included in the raw signal, and when the specific mode is selected by the mode selection means, the voice signal from the voice input means and the first signal processing means Second signal processing means for inputting a command signal and outputting only the command signal of these signals to the recognition / collation processing means, and the recognition / collation processing means is selected when the specific mode is selected by the mode selection means The command signal from the second signal processing means is compared with the standard pattern.

  According to this speech recognition system, when a specific mode is selected, a speech signal and a command signal are input, and only the command signal among these signals is output to the recognition verification processing means. The recognition verification processing means The signal is compared with the standard pattern signal. For this reason, the speech recognition system controls the control device based on the command signal when the command signal and the standard pattern signal are verified by the recognition verification processing means. That is, if a command signal is incorporated in content data to be played back by a playback device, the control device can be controlled along with playback by the playback device. This makes it possible to change the lighting system to soft lighting similar to a forest bath while playing quiet music that resembles a forest bath environment simply by playing a CD or DVD that records music data for enjoying a forest bath in a simulated manner, for example. Can be made. In addition, it is possible to make the operation explanation by the guidance voice easier to understand by actually controlling the control device while playing the guidance voice by playing a CD or DVD for explaining the operation. Therefore, convenience can be further improved.

  In the speech recognition system according to the present invention, the first signal processing unit selects only the command signal from the playback signal of the content from the playback device when the control mode other than the specific mode is selected by the mode selection unit. And the second signal processing means outputs the reproduction signal based on the reproduction signal from the first signal processing means when a control mode other than the specific mode is selected by the mode selection means. It is preferable to remove the echo component from the audio signal input by the input means.

  According to this voice recognition system, the second signal processing means removes an echo component from the voice signal input by the voice input means when a control mode other than the specific mode is selected by the mode selection means. For this reason, the second signal processing means has an echo cancellation function, and when a control mode other than the specific mode is selected and the user intends to control the control device by the uttered voice, the recognition rate of the uttered voice is improved. Can do.

  In the speech recognition system according to the present invention, the recognition / collation processing means includes a speech signal from the speech input means and a first standard pattern signal for performing collation processing, and a command signal from the first signal processing means and collation processing. It is preferable to have a second standard pattern signal for performing.

  According to this speech recognition system, the recognition / collation processing means includes the first standard pattern signal for performing collation processing with the speech signal, and the second standard pattern signal for performing collation processing with the command signal. The voice recognition performance can be improved by using each standard pattern signal without using the standard pattern signal in common.

  The speech recognition system according to the present invention further includes volume adjusting means capable of adjusting only the sound volume based on the command signal output from the playback device, and the sound volume adjusting means selects the specific mode by the mode selecting means. In such a case, it is preferable to prohibit the output of sound based on the command signal.

  According to this voice recognition system, when the specific mode is selected by the mode selection unit, the output of voice based on the command signal is prohibited. Here, since the command signal is a part of the reproduction signal, sound is output from the reproduction device. For this reason, the discomfort caused by the user recognizing the sound based on the command signal can be suppressed by prohibiting the output of the sound based on the command signal.

  In the voice recognition system according to the present invention, the playback device plays back and outputs video, the playback signal includes a command signal corresponding to the scene of the output video, and the recognition / collation processing means performs the second signal processing. It is preferable that the control content of the control device is changed by performing a collation process between the command signal from the means and the standard pattern.

  According to this voice recognition system, the reproduction signal includes a command signal corresponding to the scene of the output video, and the recognition / collation processing means performs the collation processing between the command signal from the second signal processing means and the standard pattern. As a result, the control content of the control device is changed. For this reason, when a movie DVD or the like is being reproduced, the brightness of the illumination can be changed for each scene of the movie, and the effect of producing the video output can be enhanced.

  In the voice recognition system according to the present invention, it is preferable that the playback device outputs voice by 5.1 channel, and 0.1 channel is assigned to the playback channel of the command signal.

  According to this voice recognition system, the playback device outputs the voice by 5.1 channel, and 0.1 channel is assigned to the playback channel of the command signal. Therefore, the remaining 5 channels can be used for the output voice, etc. It is possible to suppress a decrease in the effect of the content itself.

  In addition, the speech recognition apparatus according to the present invention inputs speech speech from a user and outputs a control signal for controlling the control device based on the input speech speech, and reproduces content based on the playback signal. Voice input means for inputting the voice output by the playback device and the voice of the user's utterance and outputting a voice signal based on the voice, the voice signal from the voice input means, and the standard pattern signal stored in advance A mode in which one mode can be selected from a plurality of control modes for determining a processing method When a specific mode is selected by the selection means and the mode selection means, a playback signal is input from the playback device, and the control device is controlled from the input playback signal. Therefore, first signal processing means for selecting and outputting a command signal included in the reproduction signal in advance, a voice signal from the voice input means, and a command signal from the first signal processing means are input, A second signal processing unit that outputs only the command signal to the recognition / collation processing unit, and the recognition / collation processing unit receives the command signal from the second signal processing unit when the specific mode is selected by the mode selection unit. And the standard pattern.

  According to this speech recognition apparatus, when the specific mode is selected, the speech signal and the command signal are input, and only the command signal among these signals is output to the recognition verification processing unit. The signal is compared with the standard pattern signal. For this reason, the speech recognition system controls the control device based on the command signal when the command signal and the standard pattern signal are verified by the recognition verification processing means. That is, if a command signal is incorporated in content data to be played back by a playback device, the control device can be controlled along with playback by the playback device. This makes it possible to change the lighting system to soft lighting similar to a forest bath while playing quiet music that resembles a forest bath environment simply by playing a CD or DVD that records music data for enjoying a forest bath in a simulated manner, for example. Can be made. In addition, it is possible to make the operation explanation by the guidance voice easier to understand by actually controlling the control device while playing the guidance voice by playing a CD or DVD for explaining the operation. Therefore, convenience can be further improved.

  ADVANTAGE OF THE INVENTION According to this invention, the speech recognition system and speech recognition apparatus which can improve the convenience more can be provided.

  Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a configuration diagram showing a speech recognition system according to an embodiment of the present invention. The speech recognition system 1 inputs speech speech from a user and controls the control device 20 based on the input speech speech. The speech recognition system 1 includes a speech recognition device 10 and a control device 20.

  The voice recognition device 10 receives voice and input by a switch operation from a user, and outputs a control signal for controlling the control device 20 according to the received input content. The voice recognition device 10 can select a voice input mode in which the control device 20 can be controlled by voice and a button operation input mode in which the control device 20 can be controlled by switch operation. In the voice input mode, the voice recognition device 10 controls the control device 20 according to the recognized standard pattern when the uttered voice from the user is input and the input uttered voice is recognized as corresponding to the predetermined standard pattern. Output a control signal. In the button operation input mode, the voice recognition device 10 inputs a switch operation from the user, and outputs a control signal for controlling the control device 20 with contents corresponding to the switch operation.

  The control device 20 is an external device that operates according to the content of the control signal from the voice recognition device 10. Specifically, the control device 20 includes five devices, a DVD player 21, a television 22, a bathroom device 23, a ventilation fan 24, and a lighting device 25. The control device 20 operates according to a control signal from the voice recognition device 10 and operates. Stop and so on. For example, the television 22 that is one of the control devices 20 is turned on or changed in channel by a control signal from the voice recognition device 10. Note that the DVD player 21 of the control device 20 reproduces content based on a reproduction signal from the DVD, and outputs audio and video from the television 22. For this reason, the DVD player 21 and the television 22 constitute a playback device. In the present embodiment, the television 22 is configured to perform 2-channel audio output.

  FIG. 2 is an external view showing an installation example of the speech recognition apparatus 10 shown in FIG. As shown in FIG. 2, the speech recognition apparatus 10 is provided in a bathroom, for example. In the bathroom, a DVD player 21 (not shown in FIG. 2), a television 22, a bathroom device 23 (not shown in FIG. 2), a ventilation fan 24, and a lighting device 25 are provided. Further, a controller 11 (described later) that is a component of the speech recognition device 10 is installed in the vicinity of the bathtub 30 in the bathroom.

  1 and 2, the DVD player 21, the television 22, the bathroom device 23, the ventilation fan 24, and the lighting device 25 are given as examples of the control device 20. However, the control device 20 is not limited to this, and the control device 20 is a floor heating device. It may be another device such as a personal computer. The voice recognition device 10 may not be provided in the bathroom, and may be provided in other places such as a bedroom, a living room, a company desk, and a conference room.

  Reference is again made to FIG. As shown in FIG. 1, the speech recognition apparatus 10 includes a controller 11, a mixer unit (first signal processing unit) 12, an echo cancellation unit (second signal processing unit) 13, and a recognition verification processing unit (recognition verification processing unit). ) 14 and a control unit (control means) 15. The controller 11 receives input from the user by voice and switch operation. FIG. 3 is a front view showing details of the controller 11 shown in FIG. As shown in FIG. 3, the controller 11 includes an audio input unit (audio input unit) 11a, an operation button (mode selection unit) 11b, a display unit 11c, and an LED lamp 11d. Note that the LED lamp 11d is not shown in FIG. 1 because the LED lamp 11d is not connected to the other portions 12 to 15 and 20.

  The audio input unit 11a shown in FIG. 3 includes a microphone or the like. The audio input unit 11a receives the audio reproduced by the DVD player 21 and output from the television 22 and the audio from the user's utterance, and the audio signal Sa based on these inputs. Is output. The operation button 11b receives a switch operation by the user. The display unit 11c is configured by an LCD or the like, and displays operation statuses and the like (for example, bath temperature and current time) of the various control devices 20. The LED lamp 11d indicates to the user whether the current voice input mode or the button operation input mode is selected. This LED lamp 11d is composed of three LEDs. For example, when one is lit, it is in the voice input mode, when the other is lit, it is in the button operation input mode, and the remaining one is lit. In some cases, the configuration indicates that both modes are used together.

  The various operation buttons 11b will be specifically described. The various operation buttons 11b include a priority button 11b1, a follow-up button 11b2, a bath automatic button 11b3, a call button 11b4, a controller on / off button 11b5, a menu button 11b6, a confirm button 11b7, a return button 11b8, and a cross key 11b9.

  The priority button 11b1 is a button used when it is desired to set a hot water supply temperature or a shower temperature in the bathroom. In general, water and hot water are used not only in the bathroom but also in the kitchen. For this reason, even if the hot water supply temperature and the shower temperature of the bathroom device 23 are set, if water or hot water is used in other places, the actual hot water supply temperature or the shower temperature may be shifted. Therefore, by depressing the priority button 11b1, it is possible to give priority to the bathroom over other places, and to prevent the actual hot water supply temperature and shower temperature from being displaced. When the priority button 11b1 is pressed, a priority mark (not shown) is displayed on the display unit 11c.

  The chasing button 11b2 is a button used when raising the temperature of hot water that has become cold in the bathtub 30. When the tracking button 11b2 is pressed, a tracking mark (not shown) is displayed on the display unit 11c.

  The bath automatic button 11b3 is a button used when hot water is poured into the bathtub 30 with the set amount and temperature of hot water. When the automatic button 11b3 is pressed, an automatic mark (not shown) is displayed on the display unit 11c.

  The call button 11b4 is a button used when talking with a kitchen remote controller installed outside the bathroom, for example, in a kitchen. When the call button 11b4 is pressed, a call mark (not shown) is displayed on the display unit 11c.

  The controller on / off button 11b5 is a button for turning on / off the power of the controller 11 itself. When the power is turned off by the controller on / off button 11b5, the display on the display unit 11c is erased.

  The menu button 11b6 is a button for setting the operation of the control device 20 by manual input. When the button 11b6 is pressed, a plurality of operation items (for example, lighting off, ventilation fan off, television power on, television channel +1, etc.) of each control device 20 are displayed on the display unit 11c. The user selects one action item by operating the cross key 11b9 from the plurality of action items.

  The confirmation button 11b7 is a button that is pressed when the control device 20 executes the operation of the operation item selected by operating the cross key 11b9. The return button 11b8 is a button used to return the screen displayed on the display unit 11c to the previous state. For example, when only about three action items can be displayed on the display unit 11c, the display screen can be shifted to the next screen and a new action item can be displayed by operating the cross key 11b9. In this state, if the return button 11b8 is pressed, the transitioned screen can be restored and the operation items of the previous screen can be displayed on the display unit 11c.

  The cross key 11b9 is a button used for setting the temperature of the hot water supply temperature or shower temperature, setting the amount of hot water, and the like. The cross key 11b9 is also used to select an operation item displayed on the display unit 11c.

  Furthermore, in this embodiment, by operating the operation button 11b of the controller 11, the voice input mode and the button operation input mode can be selected. Specifically, the user can switch between the voice input mode and the button operation input mode by operating the menu button 11b6 and selecting an input mode displayed on the display unit 11c.

  Reference is again made to FIG. The mixer unit 12 inputs the reproduction signals Sb and Sc from the television 22 and outputs them to the echo canceling unit 13 and outputs a mixing signal Sd obtained by mixing the reproduction signals Sb and Sc from the television 22. .

  The echo cancel unit 13 receives the reproduction signals Sb and Sc from the television set 22, that is, the mixing signal Sd from the mixer unit 12, and removes echoes based on the input mixing signal Sd. That is, the voice input to the voice input unit 11a includes both the user's uttered voice and the echo from the device such as the television 22. The echo canceling unit 13 is configured to remove a signal component corresponding to an echo from the audio signal Sa composed of the uttered voice and the echo based on the mixing signal Sd. Note that the echo canceling unit 13 has a limit in performance, and the echo canceling unit 13 cannot remove the echo completely, and generally removes the echo to some extent.

  The recognition / collation processing unit 14 stores in advance a voice signal Sa from the voice input unit 11a, more specifically, a voice signal Se from which echoes have been removed to some extent by the echo canceling unit 13 (hereinafter referred to as echo cancellation signal Se). This is used for collation with the standard pattern signal. In addition, when the echo removal signal Se matches the standard pattern signal, the recognition verification processing unit 14 outputs a recognition result signal Sf to that effect to the control unit 15. More specifically, for example, when the utterance content from the user is “TV power on”, the recognition / collation processing unit 14 uses the “TV power on” standard in which the echo removal signal Se based on the utterance content is registered in advance. It is determined whether or not the pattern signal is matched. Then, when the recognition / collation processing unit 14 determines to match the standard pattern signal of “TV power on”, it outputs a recognition result signal Se for turning on the power of the television 22 to the control unit 15.

  The control unit 15 controls the operation of the control device 20. For example, when receiving a recognition result signal Se for turning on the power of the television 22 from the recognition / collation processing unit 14, the control unit 15 outputs a control signal Sg for turning on the power to the television 22. Thereby, the power supply of the television 22 is turned on.

  Here, in the present embodiment, the operation button 11b6 can select one control mode from a plurality of control modes for determining the processing method of the speech recognition apparatus 10. Specifically, the control mode includes a demonstration mode (specific mode), a healing mode (specific mode), and a normal mode. The user operates the operation button 11b to select one of these three control modes. The control mode can be selected. Specifically, the user can select one control mode by pressing the menu button 11b6, selecting one of the modes displayed on the display unit 11c, and pressing the confirm button 11b7. it can. The demo mode is a mode for playing a guidance voice or a guidance video for explaining the operation of the voice recognition device 10, and the healing mode is setting the environment in the bathroom to obtain a healing effect. Mode.

  Further, in the present embodiment, when the demo mode or the healing mode is selected and a predetermined DVD is reproduced, the following operation is performed. FIG. 4 is a diagram showing the operation of the voice recognition system 1 when the demo mode or the healing mode is selected and a predetermined DVD is reproduced. In this case, one of the reproduction signals Sb and Sc from the television 22 (for example, the R-side audio signal of the two-channel audio signals) is generated by a command signal Sh for controlling the control device 20. It shall be configured.

  First, as shown in FIG. 4, when the demo mode or the healing mode is selected, the control unit 15 outputs the switching signal Si to the mixer unit 12 and the switching signal Sj to the echo cancellation unit 13. Is output. Thereby, the mixer unit 12 and the echo cancel unit 13 perform the following operations.

  That is, the mixer unit 12 cuts the audio signal Sc without mixing the audio signal Sc and the command signal Sh, and selects and outputs only the command signal Sh. As a result, only the command signal Sh is output to the echo cancellation unit 13.

  Next, the echo cancel unit 13 receives the audio signal Sa input from the audio input unit 11 a and the command signal Sh from the mixer unit 12, and outputs only these signals to the recognition / collation processing unit 14.

  Thereby, the recognition / collation processing unit 14 performs collation processing between the command signal Sh and the standard pattern signal. The command signal Sh is configured to cause the control device 20 to perform predetermined control in advance. That is, the command signal Sh corresponding to the speech voice such as “Dark the light” is recorded on the healing mode DVD. For this reason, the recognition / collation processing unit 14 performs a collation process between the command signal Sh and the standard pattern signal, thereby determining that the command signal Sh matches the standard pattern signal of “darkening the light”. A recognition result signal Sf to that effect is output. As a result, the lighting device 25 is lit darkly, and the surrounding environment can be adjusted and the healing direction can be further improved as compared with the case where a video or sound having a healing effect from the television 22 flows.

  The demonstration mode DVD records a command signal Sh corresponding to an utterance voice such as “ventilator fan off”. For this reason, the recognition / collation processing unit 14 performs a collation process between the command signal Sh and the standard pattern signal to determine that the command signal Sh matches the standard pattern signal of “ventilator fan off”, and recognizes that fact. The result signal Sf is output. As a result, the operation of the ventilation fan 24 is stopped, and the ventilation fan 24 can be actually turned off while a guidance voice or guidance video for turning off the ventilation fan 24 is played, and an impressive operation explanation can be given to the user.

  Furthermore, in the present embodiment, the television 22 includes volume adjusting means that can adjust only the volume of the sound based on the command signal Sh among the sounds output to the user. That is, the DVD includes the command signal Sh, and if the sound is output as it is on the television 22, a sound such as “darken the light” flows from the television 22 and the healing effect is diminished. Become. However, by providing a volume adjusting means that can cut or extremely reduce the volume of the voice based on the command signal Sh, the voice based on the command signal Sh is recognized by the user, giving unpleasant feeling. Can be suppressed.

  Specifically, the volume adjusting means is configured as follows in the case of 2-channel audio output. For example, when music is output from the L side and a command signal Sh is output from the R side, the sound volume adjusting unit is configured to output the sound only from the L side and cut the audio output from the R side. Similarly, in the case of 5.1 channel audio output or 6.1 channel audio output, a command signal Sh is assigned to a certain channel. Then, the sound volume adjusting means cuts the sound output from the channel to which the command signal Sh is assigned.

  In the case where the above operation is performed, the recognition / collation processing unit 14 includes a first standard pattern signal for performing collation processing with the voice signal Sa (specifically, the echo removal signal Sd) from the voice input unit 11a, and a mixer. It is desirable to have a command signal Sh from the unit 12 and a second standard pattern signal for performing a collation process. This is because the voice recognition performance can be improved by using the dedicated standard pattern signals without using both standard pattern signals in common.

  That is, when there is no dedicated standard pattern signal and collation processing is performed using a common standard pattern signal, a standard pattern signal that matches both the voice signal Sa and the command signal Sh based on the uttered voice by the user is prepared. Is required. Here, since the audio signal Sa and the command signal Sh are not exactly the same, the matched standard pattern signal has an intermediate property between them. As a result, the matching rate between the audio signal Sa and the standard pattern signal and the matching rate between the command signal Sh and the standard pattern signal are reduced. However, in the case of having a dedicated standard pattern signal, it is possible to prevent the matching rate from being lowered and improve speech recognition performance.

  Next, detailed operation of the speech recognition system 1 will be described with reference to a flowchart. FIG. 5 is a flowchart showing details of the operation of the speech recognition system 1 shown in FIG. The process shown in FIG. 5 is repeated until the power of the speech recognition apparatus 10 is turned off.

  As shown in FIG. 5, first, the control unit 15 determines whether or not the demonstration mode is selected and started (S1). If it is determined that the demo mode has been started (S1: YES), the process proceeds to step S3.

  On the other hand, when determining that the demo mode has not been started (S1: NO), the control unit 15 determines whether the healing mode has been selected and started (S2). If it is determined that the healing mode has been started (S2: YES), the process proceeds to step S3. On the other hand, when it is determined that the healing mode has not been started (S2: NO), the process proceeds to step S1.

  In step S3, the control unit 15 displays the current mode (that is, demo mode or healing mode) of the display unit 11c (S3). Then, the control unit 15 transmits a switching signal Si to the mixer unit 12 (S4). Thus, the mixer unit 12 does not mix and output the audio signals Sb and Sc, but selects and outputs only the command signal Sh.

  Next, the control unit 15 transmits a switching signal Sj to the echo cancellation unit 13 (S5). As a result, the echo cancellation unit 13 outputs the input command signal Sh to the recognition / collation processing unit 14 as it is without executing the echo cancellation processing.

  Further, the control unit 15 transmits a control signal Sg to the television 22 and switches so that the sound of the command signal Sh is not output (S6). This prevents voice output based on the command signal Sh and prevents a healing effect and the like from being reduced.

  Next, the control unit 15 determines whether or not there is a cancel operation in the demo mode or the healing mode (S7). That is, the control unit 15 determines whether or not there has been an operation for shifting to the normal mode. If it is determined that there has been an operation for shifting to the normal mode (S7: YES), the control mode shifts to the normal mode, and the process proceeds to step S9. On the other hand, when it is determined that there has been no operation for shifting to the normal mode (S7: NO), the control unit 15 determines whether or not the DVD playback has ended (S8). If it is determined that the DVD playback has not ended (S8: NO), the process proceeds to step S7.

  On the other hand, if it is determined that the DVD playback has ended (S8: YES), the process proceeds to step S9. In step S9, the control unit 15 displays the current mode (that is, the normal mode) of the display unit 11c (S9). Then, the control unit 15 transmits a switching signal Si to the mixer unit 12 (S4). Thus, the mixer unit 12 mixes and outputs the audio signals Sb and Sc without selecting and outputting only the command signal Sh.

  Next, the control unit 15 transmits a switching signal Sj to the echo cancellation unit 13 (S5). As a result, the echo cancellation unit 13 does not output only the command signal Sh, but executes echo cancellation processing.

  Furthermore, the control part 15 transmits control signal Sg with respect to the television 22, and returns an audio | voice output to a normal state (S6). Thereafter, the process shown in FIG. 5 ends.

  Thus, according to the speech recognition system 1 and the speech recognition apparatus 10 according to the present embodiment, when the specific mode is selected, the speech signal Sc and the command signal Sh are input, and the signals Sc and Sh Only the command signal Sh is output to the recognition / collation processing unit 14, and the recognition / collation processing unit 14 performs collation processing between the command signal Sh and the standard pattern signal. For this reason, when the command signal Sh and the standard pattern signal Sc are collated by the recognition collation processing unit 14, the voice recognition system 1 controls the control device 20 based on the command signal Sh. That is, if the command signal Sh is incorporated in the data of the content played back by the playback device, the control device 20 can be controlled along with playback by the playback device. This makes it possible to change the lighting system to soft lighting similar to a forest bath while playing quiet music that resembles a forest bath environment simply by playing a CD or DVD that records music data for enjoying a forest bath in a simulated manner, for example. Can be made. Further, it is possible to make the operation explanation by the guidance voice easier to understand by actually controlling the control device 20 while playing the guidance voice by playing a CD or DVD for explaining the operation. Therefore, convenience can be further improved.

  In addition, when the normal mode other than the demo mode and the healing mode is selected by the operation button 11b, the echo cancel unit 13 removes the echo component from the audio signal Sa input by the audio input unit 11a. Thus, the echo cancellation unit 13 has a co-cancelling function, and when the normal mode other than the demo mode and the healing mode is selected and the user intends to control the control device 20 by the uttered speech, the recognition rate of the uttered speech Can be improved.

  Further, since the recognition / collation processing unit 14 includes the voice signal Sc and the first standard pattern signal for performing the collation processing, and the command signal Sh and the second standard pattern signal for performing the collation processing, both standard patterns The voice recognition performance can be improved by using each dedicated standard pattern signal without sharing the signal.

  Further, when the demo mode or the healing mode is selected by the operation button 11b, the output of sound based on the command signal Sh is prohibited. Here, since the command signal Sh is a part of the reproduction signal Sb, audio is output from the television 22. For this reason, the discomfort caused by the user recognizing the sound based on the command signal Sh can be suppressed by prohibiting the output of the sound based on the command signal Sh.

  As mentioned above, although the speech recognition system and the speech recognition apparatus according to the present invention have been described based on the embodiments, the present invention is not limited thereto, and modifications may be made without departing from the spirit of the present invention. Good.

  For example, in the present embodiment, the DVD player 21 and the television 22 are given as an example of a playback device. However, the present invention is not limited to this, and the playback device may play only music and audio, such as a CD player. Other devices such as a personal computer may be used.

  In the present embodiment, the reproduction signal Sb includes the command signal Sh corresponding to the scene of the output video, and the recognition / collation processing unit 14 collates the command signal Sh from the echo cancel unit 13 with the standard pattern. By performing the process, the control content of the control device 20 is changed. For this reason, when a movie DVD or the like is being reproduced, the brightness of the illumination can be changed for each scene of the movie, and the effect of producing the video output can be enhanced.

  In the present embodiment, the speech recognition system 1 in the case of 2-channel audio output has been described. However, the present invention is not limited to this, and a playback device outputs audio through 5.1 channels, and command signal Sh is output to 0.1 of them. Is preferably assigned. This is because the remaining five channels can be spent on output audio and the like, and the reduction of the effect of the content itself can be suppressed.

  Further, in the present embodiment, the echo canceling unit 13 is provided, and in the demo mode and the healing mode, the echo canceling unit 13 inputs the audio signal Sa from the audio input unit 11a and the command signal Sh from the mixer unit 12, and the command Although only the signal Sh is output, it is not particularly limited to the echo cancel unit 13 as long as it can input the audio signal Sa and the command signal Sh and output only the command signal Sh. If the echo cancel unit 13 is not provided, the echo cannot be removed in the normal mode, and the voice recognition rate may be reduced. Therefore, an utterance button is provided in the controller 11 and the voice from the television 22 is pressed when the utterance button is pressed. It is preferable that the output is cut (that is, the mute state is set). This is because it is possible to prevent a decrease in the speech recognition rate due to the sound from the television 22.

  Further, in the present embodiment, in the demo mode and the healing mode, as described with reference to FIG. 4, since the signal of the user's speech is not transmitted to the recognition verification processing unit 14, the control by speech is performed. The device 20 cannot be operated. For this reason, when the controller 11 is provided with an utterance button, it is preferable that the demo mode and the healing mode are temporarily stopped or stopped by returning to the normal mode when the utterance button is pressed in the demonstration mode and the healing mode. This is because the control device 20 can be operated by speech even in the demonstration mode and the healing mode.

  Furthermore, in the present embodiment, one control mode can be selected from a plurality of control modes by operating the operation button 11b6. However, the present invention is not limited to this, and the control mode can be selected by voice. Also good. Furthermore, when a command signal Sh for selecting a control mode is recorded on a DVD or CD, and the DVD or CD is played on the DVD player 21 or music player, one control mode is automatically selected. It may come to be.

  Further, in the present embodiment, the voice input mode and the button operation input mode can be switched by operating the operation button 11b. However, the present invention is not limited to this, and during the voice input mode, the button operation input mode is set by utterance. You may make it shift to.

It is a block diagram which shows the speech recognition system which concerns on embodiment of this invention. It is an external view which shows the example of installation of the speech recognition apparatus shown in FIG. It is a front view which shows the detail of the controller shown in FIG. It is a figure which shows operation | movement of the speech recognition system when demo mode or healing mode is selected and predetermined | prescribed DVD is reproduced | regenerated. It is a flowchart which shows the detail of operation | movement of the speech recognition system shown in FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Voice recognition system 10 Voice recognition apparatus 11 Controller 11a Voice input part 11b Operation button (mode selection means)
11c Display unit 12 Mixer unit (first signal processing means)
13 Echo cancellation unit (second signal processing means)
14 recognition / collation processing unit (recognition / collation processing means)
15 Control unit (control means)
20 Control equipment 21 DVD player (playback equipment)
22 Television (playback equipment)
23 Bathroom equipment 24 Ventilation fan 25 Lighting equipment

Claims (7)

A speech recognition system that inputs speech speech from a user and controls a control device based on the speech speech input,
A playback device for playing back content based on the playback signal;
Voice input means for inputting the voice output by the playback device and the voice of the user's utterance and outputting a voice signal based on the voice;
Recognition collation processing means for performing collation processing between a voice signal from the voice input means and a standard pattern signal stored in advance;
Control means for controlling the control device based on the collation result by the recognition collation processing means;
Mode selection means capable of selecting one mode from a plurality of control modes for determining a processing method;
When a specific mode is selected by the mode selection means, a command signal included in the playback signal is selected and output in advance to control the control device from the playback signal of the content from the playback device. First signal processing means;
When a specific mode is selected by the mode selection means, a voice signal from the voice input means and a command signal from the first signal processing means are input, and only the command signal among the signals is recognized and verified. Second signal processing means for outputting to the processing means,
The speech recognition system, wherein the recognition / collation processing means performs collation processing between the command signal from the second signal processing means and the standard pattern. When the control mode other than the specific mode is selected by the mode selection unit, the first signal processing unit does not select and output only the command signal from the playback signal of the content from the playback device, but the playback Output signal,
When the control mode other than the specific mode is selected by the mode selection unit, the second signal processing unit is an audio signal input by the audio input unit based on a reproduction signal from the first signal processing unit. The speech recognition system according to claim 1, wherein an echo component is removed from the speech recognition system. The recognition collation processing means includes a first standard pattern signal for performing collation processing with the voice signal from the voice input means, and a second standard pattern for performing collation processing with the command signal from the first signal processing means. The speech recognition system according to claim 1, further comprising: a signal. A volume adjusting means capable of adjusting only the volume of the sound based on the command signal output from the playback device;
4. The sound volume adjusting unit according to claim 1, wherein, when a specific mode is selected by the mode selecting unit, the sound output based on the command signal is prohibited. 5. Speech recognition system. The playback device plays back and outputs video,
The playback signal includes the command signal corresponding to an output video scene,
The control content of the control device is changed when the recognition / collation processing unit performs a collation process between the command signal from the second signal processing unit and the standard pattern. 5. The speech recognition system according to any one of 4 above. The playback device outputs audio through 5.1 channel,
The voice recognition system according to any one of claims 1 to 5, wherein 0.1 channel is assigned to a reproduction channel of the command signal. A speech recognition device for inputting a speech voice from a user and outputting a control signal for controlling a control device based on the inputted speech voice,
A voice input means for inputting a voice output by a playback device that plays back content based on a playback signal and a voice of a user's utterance and outputting a voice signal based on the voice;
Recognition collation processing means for performing collation processing between a voice signal from the voice input means and a standard pattern signal stored in advance;
Control means for controlling the control device based on the collation result by the recognition collation processing means;
Mode selection means capable of selecting one mode from a plurality of control modes for determining a processing method;
When a specific mode is selected by the mode selection unit, a command signal included in the reproduction signal is selected and output in advance to control the control device from the content reproduction signal from the reproduction device. First signal processing means;
A second signal processing means for inputting a voice signal from the voice input means and a command signal from the first signal processing means, and outputting only the command signal of these signals to the recognition verification processing means,
The recognition / collation processing means performs a collation process between a command signal from a second signal processing means and the standard pattern when the specific mode is selected by the mode selection means. Voice recognition device.

Images