Speech systems and alarms

Abstract

To provide an utterance system and an alarm device which can, with a simple and inexpensive configuration, achieve an utterance suppressing function, such as suppressing the output of sound information for example, when the sound information outputted from a speaker regardless of intention and situation of a user is offensive to the user in the utterance system equipped with a sound output processing part which sequentially outputs the sound information from a speaker.SOLUTION: A system 1 provided in an alarm device 50 includes: an audio command output part 20 which outputs an utterance suppression command in accordance with the input of the audio satisfying a specific determination condition for the utterance suppression; and an utterance suppression control part 13 for executing the utterance suppression control for suppressing the output of sound information by a sound output processing part 23 when the utterance suppression demand is outputted by the audio command output part 20.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a speech system including an audio output processing unit that sequentially outputs audio information from a speaker, and to an alarm device including the speech system. [Background technology]

[0002] There is known a speech system equipped with a voice output processing unit that sequentially outputs voice information from a speaker (in this application, outputting voice information from a speaker may be referred to as "speech") (see Patent Documents 1 and 2). The speech system described in Patent Document 1 is a so-called smart speaker system that has a voice recognition function that identifies the user's voice in addition to an audio output processing unit, and is configured to recognize voice from the user, collect desired content information from external information sources such as the Web in response to instructions obtained by identifying the voice, convert the collected content information into voice information and output it sequentially from the speaker. The speech system described in Patent Document 2 is a language education device that outputs by voice an output sentence that is generated in response to an input sentence entered by a learner from a keyboard, and has a speech repetition function that outputs by voice the most recent output sentence again when the learner enters a repeat request code by operating the keyboard. The speech system described in Patent Document 3 is a system that has a voice output processing unit as well as a voice recognition function that identifies the user's voice, and is capable of engaging in dialogue with the user.The system is configured to recognize voice from the user and suppress the output of voice information in response to a command obtained by identifying the voice. There is also known an alarm device that monitors abnormal conditions at a user's location, which is equipped with a speech system that has an audio output processing unit that sequentially outputs notification information regarding the abnormal condition as audio information from a speaker (see Patent Document 4). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2020-021040 A [Patent Document 2] Japanese Patent Application Publication No. 03-226783 [Patent Document 3] JP 2017-161637 A [Patent Document 4] JP 2020-155956 A Summary of the Invention [Problem to be solved by the invention]

[0004] In particular, in configurations such as alarm devices equipped with speech systems that output audio information from a speaker regardless of user requests, it is desirable to adopt a speech suppression function that suppresses the output of audio information at that time when the audio information output from the speaker regardless of the user's intentions or situation is unpleasant to the user. In a speech system that accepts input operations from a keyboard, such as the speech system described in Patent Document 2 above, if a user inputs a request code for speech suppression by key input operation, the output of voice information at that time can be suppressed. However, this requires a key input device such as a keyboard, and there is a problem that if the device is installed in a place that is out of reach or if the user's hands are full, the repeat request code cannot be input.

[0005] In addition, speech systems with a voice recognition function, such as the speech systems in Patent Documents 1 and 3 above, can recognize the user's voice and suppress the output of voice information at that time in accordance with the command that recognizes the voice, but there is a problem in that the system becomes expensive and large-scale because it is necessary to be equipped with a relatively large-scale voice recognition function. In view of this situation, the main object of the present invention is to provide a technology that can realize a speech suppression function with a simple and inexpensive configuration in a speech system equipped with an audio output processing unit that sequentially outputs audio information from a speaker, such as when audio information output from the speaker regardless of the user's intentions or situation is unpleasant to the user. [Means for solving the problem]

[0006] A first characteristic configuration of the present invention is a speech system including a voice output processing unit that sequentially outputs voice information from a speaker, an audio command output unit that outputs a speech suppression command in response to input of audio that satisfies a specific speech suppression determination condition; and a speech suppression control unit that executes speech suppression control to suppress output of audio information by the audio output processing unit when the speech suppression command is output by the acoustic command output unit.

[0007] According to this configuration, in cases where the voice information output from the speaker regardless of the user's intention or situation is unpleasant to the user, the user can simply emit a sound that satisfies a specific speech suppression judgment condition without using a key input device, and the audio command output unit can output a speech suppression command in association with the input of the sound. Therefore, when the speech suppression command is output by the audio command output unit, the speech suppression control unit executes the speech suppression control, so that the output of the voice information at that time can be suppressed (in this application, suppressing the output of voice information in this manner may be referred to as "speech suppression"). Furthermore, a speech suppression function for suppressing speech as described above can be realized by simply adopting a simple and rational configuration such as the audio command output unit that identifies the input of a sound that satisfies a specific condition, without requiring an expensive and large-scale voice recognition function. Therefore, the present invention makes it possible to provide a technology that can realize a speech suppression function in a simple and inexpensive configuration in a speech system equipped with an audio output processing unit that sequentially outputs audio information from a speaker, such as when audio information output from the speaker regardless of the user's intentions or situation is unpleasant to the user.

[0008] A second characteristic configuration of the present invention is that the acoustic command output unit identifies an acoustic input having a sound pressure level equal to or higher than a predetermined speech suppression set sound pressure level as an acoustic input that satisfies the speech suppression determination condition.

[0009] According to this configuration, by simply adopting a simple and rational configuration in which only the sound pressure level of the input sound is recognized in the sound command output unit, it is possible to output a speech suppression command when the sound pressure level of the input sound reaches or exceeds a predetermined speech suppression set sound pressure level. This allows the user to suppress the output of voice information from the speaker by simply taking a simple action such as emitting a sound that is equal to or exceeds the speech suppression set sound pressure level, such as speaking loudly or clapping their hands.

[0010] A third characteristic configuration of the present invention is that the acoustic command output unit identifies consecutive inputs of acoustic sound having a sound pressure level equal to or higher than a predetermined speech suppression setting sound pressure level a predetermined number of times for re-speech as acoustic input that satisfies the speech suppression judgment condition.

[0011] According to this configuration, the sound command output unit recognizes the sound pressure level of the input sound and counts the number of consecutive inputs of sound having a sound pressure level equal to or higher than a predetermined speech suppression set sound pressure level, and when the sound pressure level of the input sound reaches or exceeds the predetermined speech suppression set sound pressure level for a predetermined number of consecutive times, a speech suppression command can be output. This allows the user to suppress the output of audio information from the speaker by taking a simple action such as continuously emitting a sound equal to or higher than the speech suppression set sound pressure level, such as continuously uttering a loud voice or continuously clapping hands, and when a loud sound is input one-off without being continuous against the user's intention, it is possible to prevent an erroneous speech suppression command from being output, thereby preventing unnecessary execution of speech suppression control.

[0012] A fourth characteristic configuration of the present invention is that the acoustic command output unit corrects the speech suppression set sound pressure level based on an output sound pressure level from the speaker or an input sound pressure level from the surrounding environment.

[0013] According to this configuration, the speech suppression set sound pressure level, which is the criterion for determining whether the input sound satisfies the speech suppression judgment condition, is not fixed to a constant value, but is corrected based on the output sound pressure level indicating the volume of the sound output from the speaker or the input sound pressure level indicating the volume of the sound input from the surrounding environment, for example, by changing the speech suppression set sound pressure level proportionally to the output sound pressure level or the input sound pressure level. This makes it possible to reliably avoid erroneous judgment that the input sound is erroneously judged as the input sound that satisfies the speech suppression judgment condition when the sound output from the speaker or the sound of the surrounding environment is input regardless of the user's intention.

[0014] A fifth characteristic configuration of the present invention is a method for identifying a speech priority of the voice information, The speech suppression control unit prohibits execution of the speech suppression control when the voice output processing unit outputs the voice information having a high speech priority.

[0015] According to this configuration, for audio information that needs to be reliably recognized by the user, such as alarm information transmitted from an alarm device, and for which speech suppression is not desirable, by setting the speech priority high, even if the speech suppression command is output by the acoustic command output unit when the audio information is output, the execution of speech suppression control is prohibited, and the audio information can be output so that it can be clearly heard by the user without being suppressed.

[0016] A sixth characteristic configuration of the present invention lies in that the audio command output unit receives audio input from the speaker by using the speaker as a microphone.

[0017] According to this configuration, since the acoustic command output unit determines whether or not to accept a command based solely on the acoustic characteristics, it is possible to reduce the cost and simplify the system by using a speaker capable of inputting recognizable sounds for simple acoustic characteristics as a microphone, without using a high-performance microphone that enables voice recognition.

[0018] A seventh characteristic configuration of the present invention is a method for producing a speech re-utterance command by the audio command output unit when audio satisfying a specific re-utterance determination condition is input, The voice output processing unit executes a re-speech target setting process for sequentially setting voice information being output or having been output as a re-speech target, and a re-speech control unit executes a re-speech control for causing the voice output processing unit to output the voice information of the re-speech target again from the beginning when the re-speech command is output by the acoustic command output unit.

[0019] According to this configuration, when the user cannot hear the voice information output from the speaker, the user can simply emit a sound that satisfies a specific re-speaking judgment condition without using a key input device, and the audio command output unit can output a re-speaking command in association with the input of the sound. Meanwhile, the re-speaking control unit executes the re-speaking target setting process, so that the voice information being output or output by the voice output processing unit is sequentially set as a re-speaking target. Therefore, when the audio command output unit outputs the re-speaking command, the re-speaking control unit executes the re-speaking control, so that the voice information that has already been output from the speaker and set as the re-speaking target at that time can be output again from the beginning to the speaker (in the present application, the re-output of the already output voice information in this manner may be referred to as "re-speaking"). Furthermore, the re-speaking function for causing re-speaking as described above can be realized by simply adopting a simple and rational configuration such as the audio command output unit that identifies the voice information in association with the input of the voice that satisfies a specific condition.

[0020] An eighth characteristic configuration of the present invention is that the acoustic command output unit outputs the speech resumption command only during a speech resumption command reception time period from the time point when the speech suppression command is output until a predetermined time has elapsed.

[0021] According to this configuration, the re-speech command is output only during the re-speech command reception time period from the time the speech suppression command is output until a specified time has elapsed. Therefore, if the user is unable to hear the audio information due to speech suppression being performed, the audio information that has already been output from the speaker at that time and is set as the re-speech target can be re-spoken.

[0022] A ninth characteristic configuration of the present invention is a method for producing a speech information output by the speech output processing unit, the speech information including an introductory portion and a main text following the introductory portion, The acoustic command output unit outputs the speech suppression command while the introductory portion is being output, and outputs the re-speech command while the main text is being output.

[0023] According to this configuration, if sound satisfying specific speech suppression judgment conditions is input during the speech of the introduction, a speech suppression command is output in response to the input of that specific sound, and on the other hand, if sound satisfying specific speech suppression judgment conditions is input during the speech of the main text, which has a higher speech priority than the introduction, a re-speech command is output in response to the input of that specific sound. Therefore, if the user is unable to hear the audio information during the introductory part of the speech due to speech suppression, the audio information that has already been output from the speaker at that time and is set as the target for re-speech can be re-spoken.

[0024] A tenth characteristic configuration of the present invention is an alarm device for monitoring an abnormal state of a user's location, A speech system having any one of the first to ninth characteristic configurations, The audio output processing unit sequentially outputs notification information relating to the abnormal state as the audio information from the speaker.

[0025] According to this configuration, in an alarm device in which an audio output processing unit sequentially outputs notification information regarding abnormal conditions at the user's location as audio information from a speaker, the alarm device is equipped with the speech system of the present invention described above, and therefore can achieve the same effects as the speech system. [Brief description of the drawings]

[0026] [Figure 1] Basic configuration diagram of a speech generation system in the first embodiment [Diagram 2] FIG. 1 shows a process flow of the first embodiment. [Diagram 3] Basic configuration diagram of a speech generation system in the second embodiment [Figure 4] FIG. 11 is a diagram showing a processing flow of the second embodiment. [Diagram 5] FIG. 13 is a diagram showing a processing flow of the third embodiment. [Figure 6] FIG. 13 is a diagram showing a processing flow of the fourth embodiment. [Figure 7] Basic configuration diagram of another type of speech system DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0027] [First embodiment] A first embodiment of the present invention will be described with reference to FIG. The speech system 1 of this embodiment (hereinafter referred to as "this speech system") is a system provided in an alarm device 50 that is installed in a user's residence and monitors abnormal conditions in the residence, such as a gas alarm, residential fire alarm, or fire alarm, and is configured to convert content information such as a weather forecast collected by the audio information distribution system 30 from an external information source, the web 32, into audio data in addition to notification information regarding abnormal conditions detected by the alarm device 50, and to sequentially output the converted audio information from the speaker 24. Note that, in this embodiment, an example in which the speech system 1 is provided in the alarm device 50 will be described, but the speech system 1 may be provided independently or in a separate device.

[0028] The speech system 1 includes a voice output processing unit 23 that sequentially outputs (utters) voice information from a speaker 24. In detail, the speech system 1 includes a speaker 24 that converts an input voice signal into voice and outputs it to the outside (for example, the user side), a voice output processing unit 23 that converts the input voice information into a voice signal and outputs it to the speaker 24, a control unit 10 that controls the operation of these components, and a voice information storage unit 26 that can appropriately store the voice information output from the speaker 24. Furthermore, the control unit 10 functions as a speech control unit 11, and the speech control unit 11 executes speech control for sequentially outputting voice information collected by the voice information distribution system 30 in response to a user request or other conditions to the voice output processing unit 23, and for sequentially outputting the voice information to the speaker 24 and uttering it. The voice information distribution system 30 can be configured as a server system installed on the administrator's side separate from the user's residence, and the speech system 1 and the voice information distribution system 30 are configured to be capable of communicating with each other through a communication line such as the Internet network.

[0029] Furthermore, the speech system 1 is provided with an acoustic command output unit 20 that outputs a specific command in response to an input of acoustic that satisfies a specific judgment condition (in the present application, such an input of acoustic may be referred to as an "input of specific acoustic.") The acoustic that is the subject of judgment by the acoustic command output unit 20 is acoustic excluding that output from the speaker 24 as voice information during speech. The audio command output unit 20 is configured to use the speaker 24 as the microphone 21 and to receive an input of an audio signal from the speaker 24. When the speaker 24 is used as a microphone, a known microphone / speaker element can be used as the speaker 24. Also, the speaker 24 can be configured to switch between a state in which it functions as an original speaker and a state in which it functions as the microphone 21 using a switching element. As shown in FIG. 7, instead of using the speaker 24 as the microphone 21, a separate microphone 21 may be provided and an input of an acoustic signal may be received from the microphone 21.

[0030] 1, the audio command output unit 20 can identify an input of audio having a sound pressure level equal to or higher than a predetermined set sound pressure level, or an input of audio having a sound pressure level equal to or higher than a predetermined set number of times in succession as the input of the specific audio. For example, when an audio having a sound pressure level equal to or higher than the set sound pressure level is input to the microphone 21 as a result of the user speaking loudly or clapping hands, or when an audio having a sound pressure level equal to or higher than the set sound pressure level is input to the microphone 21 in succession as a result of the user speaking loudly or clapping hands continuously, the audio command output unit 20 identifies the input of the audio as the input of the specific audio, and outputs a specific command.

[0031] In order to accurately determine whether or not to output the above-mentioned specific command, the maximum sound pressure level of the user's voice or the sound of the surrounding environment input under normal circumstances can be learned, or the sound pressure level of the sound actually input in a specified setting mode can be recorded, and the above-mentioned set sound pressure level can be appropriately set to a sound pressure level that exceeds the learned or recorded sound pressure levels. Furthermore, the acoustic command output unit 20 can automatically correct the set sound pressure level based on the output sound pressure level, which is the sound pressure level of the voice information output from the speaker 24, or the input sound pressure level, which is the sound pressure level of the sound input from the surrounding environment to the microphone 21. That is, the set sound pressure level, which is the criterion for determining whether or not to issue the specific command, is not fixed to a constant value but is changed proportionally to the output sound pressure level or the input sound pressure level. This makes it possible to reliably avoid erroneous determination that, when sound output from a speaker or sound from the surrounding environment is input regardless of the user's intention, the input sound is erroneously determined to be an input sound that satisfies the re-speaking determination condition. In this embodiment, whether or not to output the specific command is determined based on the sound pressure level of the input sound, but the determination may be based on other sound characteristics, such as the frequency of the sound. Furthermore, when a set number of consecutive inputs of sound that satisfy a specific determination condition are identified as the input of the specific sound, the determination conditions set for each input of sound may be different. For example, consecutive inputs of loud or high-pitched sound and quiet or low-pitched sound may be recognized as the input of the specific sound.

[0032] The acoustic command output unit 20 can output, as the specific command, a speech suppression command that serves as a trigger for the speech suppression control unit 13, which will be described later, to execute speech suppression control. In other words, the acoustic command output unit 20 recognizes an input of acoustic sound having a sound pressure level equal to or higher than a predetermined speech suppression set sound pressure level, or an input of acoustic sound having a sound pressure level equal to or higher than a predetermined speech suppression set number of times as the input of the above-mentioned specific acoustic sound, and can output a speech suppression command in conjunction with the input of that acoustic sound.

[0033] The control unit 10 functions as a speech suppression control unit 13 that executes a predetermined speech suppression control. In the speech suppression control, when the acoustic command output unit 20 outputs a speech suppression command, the output of audio information (speech) by the audio output processing unit 23 is suppressed. That is, in a case where the voice information output from the speaker 24 regardless of the user's intention or situation is unpleasant to the user, for example, if the user inputs the specific sound while the voice information is being output, the sound command output unit 20 outputs a speech suppression command in response to the input of the sound. Then, the speech suppression control unit 13 executes speech suppression control, and the output of the voice information at that time is suppressed. Furthermore, as a specific method for suppressing the output of audio information, a method for temporarily stopping the output of audio information is adopted. Specifically, the output of audio information from the speaker 24 is stopped, and when a predetermined speech stop time has elapsed from the time of the stop, the audio information is output from the beginning. In this embodiment, the audio information is output again from the beginning after a pause, but when outputting again, it is possible to adopt a method of skipping the already output portion and outputting the audio information from the continuation, or of lowering the volume of the speaker 24 for the audio information to be output again and for audio information to be output thereafter. Also, as another method of suppressing the output of audio information, it is possible to adopt a method of lowering the volume of the speaker 24 while continuing the output of the audio information without stopping it. Furthermore, when speech suppression is performed by stopping the output of specific audio information or reducing the volume by speech suppression control, the same speech suppression may be performed on all audio information output thereafter, but it may also be configured to perform speech suppression only on the output of audio information of the same genre or in the same output time period as the audio information for which speech suppression is performed. When speech suppression is performed on the output of audio information of the same genre, audio guidance may be performed to inform the user of this. The output of audio information of the same genre may be stopped on the speech system 1 side, or on the audio information distribution system 30 side.

[0034] The voice information output by the voice output processing unit 23 is configured as information that can identify the speech priority. That is, the speech priority is set high for voice information that needs to be reliably recognized by the user, such as the notification information sent from the alarm device 50, while the speech priority is set low for voice information related to general content information. The speech suppression control unit 13 prohibits the execution of the above-mentioned speech suppression control when audio information with high speech priority is output by the audio output processing unit 23. This allows the audio information with high speech priority to be output without being suppressed so that the user can clearly hear it. Specific embodiments will now be described with reference to the drawings.

[0035] (First embodiment) The first embodiment is an example in which the specific sound is input during the utterance of the voice information A, as shown in FIG. In this case, a speech suppression command is outputted along with the input of the specific sound during the speech of the speech information A, and speech suppression control is executed. Therefore, the speech of the speech information A is temporarily stopped at the time of input of the specific sound, and when a predetermined speech stop time Tx has elapsed from the time of the stop, the speech information A whose speech has been temporarily stopped is respoken from the beginning. Furthermore, if the above-mentioned specific sound is input during the repetition of the next voice information A, a speech suppression command is output, the speech of the voice information A is temporarily stopped, and the voice information A is repetitively spoken after a predetermined speech stop time Tx has elapsed from the point of the stop.

[0036] Second Embodiment A second embodiment of the present invention will be described with reference to FIG. The present embodiment differs from the above-described first embodiment only in that the acoustic command output unit 20 outputs a speech resumption command as the specific command separately from the speech suppression command, and in that the control unit 10 functions as the speech resumption control unit 12 separately from the speech suppression control unit 13. Therefore, in the following description, the same components as those in the above-described first embodiment are denoted by the same reference numerals in the drawings, and detailed description thereof will be omitted.

[0037] The acoustic command output unit 20 can output, as the specific command, a talk resumption command that serves as a trigger for the talk resumption control unit 12, which will be described later, to execute talk resumption control. In other words, the acoustic command output unit 20 recognizes an input of acoustic having a sound pressure level equal to or higher than a predetermined set sound pressure level for re-speech, or an input of acoustic having a sound pressure level equal to or higher than a predetermined set sound pressure level for re-speech a predetermined number of times consecutively for re-speech, as an input of the above-mentioned specific acoustic, and can output a re-speech command in conjunction with the input of that acoustic.

[0038] The control unit 10 functions as a re-talk control unit 12 that executes a predetermined re-talk control. In the re-talk control, a re-talk target setting process is executed in which the audio output processing unit 23 sequentially sets audio information being output or having been output as an object for re-talk, and when a re-talk command is output by the acoustic command output unit 20, the audio information to be re-talked is output (re-talked) from the beginning by the audio output processing unit 23. That is, by executing the re-speaking target setting process, the voice information A being outputted by the voice output processing unit 23 or having been outputted is sequentially set as a re-speaking target and stored in the voice information storage unit 26. For example, if the user cannot hear the voice information output from the speaker 24 regardless of the user's request, and the user inputs the specific sound while the voice information A is being output, a re-speak command is output by the sound command output unit 20 in association with the input of the sound. Then, the re-speak control unit 12 executes the re-speak control, and the voice information that has already been output from the speaker 24 and set as the re-speaking target at that time is extracted from the voice information storage unit 26, and the extracted voice information A is re-spoken from the beginning again from the speaker 24.

[0039] In this re-speech control, when re-speaking the voice information to be re-spoken, the voice information can be output in a high recognition state where the user can easily recognize the voice information. Specifically, in this embodiment, as the high recognition state, the voice information is re-spoken in a state where a re-speech message is added to the beginning to make the user aware that the voice information is being re-spoken. Note that as the high recognition state, the voice information can be re-spoken in a state where the user can easily hear the voice information, such as by increasing the volume of the speaker 24 or slowing down the playback (speech) speed.

[0040] In the above-mentioned re-speech target setting process, the voice information being output by the voice output processing unit 23 is set as the re-speech target, and updating of the re-speech target is prohibited for a predetermined time from the end of output of the voice information. That is, in a case where a plurality of pieces of voice information are output continuously by the voice output processing unit 23, when the voice information being output is sequentially set as the re-speaking target, the set voice information as the re-speaking target is maintained as the re-speaking target for a predetermined time from the end of the output of the voice information. As a result, even if the output of a certain voice information is finished, if a re-speaking command is output immediately after the end of the output, the next voice information being output at that time is not re-speaked, but the voice information before the output is finished is re-speaked. Therefore, the user can emit a sound for outputting a re-speaking command in good time to re-speak the voice information that has already been output. It is also possible to configure the system so that the update of the re-speaking target is prohibited from the time when the output of the voice information ends until the output of the next voice information starts. In this case, even if the output of some voice information ends, if a re-speaking command is output before the next voice information is output, the previous voice information whose output has ended will be re-speaked. It is also possible to configure the system so that the voice information being output is set as the re-speaking target, and the re-speaking target is reset when the output of that voice information ends. In this case, the voice information will be re-speaked only if a re-speaking command is output while some voice information is being output.

[0041] The sound command output unit 20 can distinguish between a re-speech command and a speech suppression command in response to the input of the specific sound. For example, a re-speech judgment condition including a set sound pressure level for re-speech and a set number of times for re-speech and a speech suppression judgment condition including a set sound pressure level for speech suppression and a set number of times for speech suppression can be set as separate conditions, but in this embodiment, the re-speech judgment condition and the speech suppression judgment condition are set as the same condition, and the sound output unit 20 is configured to distinguish between a re-speech command and a speech suppression command and output them at the timing when sound satisfying the condition is input. Specifically, if the input time of the specific sound is within a re-speech command reception time period from the output time of the speech suppression command until a predetermined time has elapsed, the acoustic command output unit 20 outputs a re-speech command in response to the input of the specific sound, and if the input time is within a time period other than the re-speech command reception time period, the acoustic command output unit 20 outputs a speech suppression command in response to the input of the specific sound. With this configuration, if a user wishes to check again the voice information that has been suppressed from being uttered, a re-speaking command can be output in response to the input of a specific sound during the re-speaking command reception time period, so that the voice information that has been set as the target for re-speaking can be re-spoken. Specific embodiments will now be described with reference to the drawings.

[0042] Example 2 The second embodiment is an example in which, as shown in FIG. 4, the specific sound is input during the utterance of the voice information A, and further the specific sound is input during the re-speak command reception time period Tz starting from that time point. In this case, a speech suppression command is outputted along with the input of the specific sound during the speech of the speech information A, and speech suppression control is executed. Therefore, the speech of the speech information A is temporarily stopped at the time of input of the specific sound, and if there is no second input of the specific sound, for example, the speech of the speech information A that has been temporarily stopped is respoken from the beginning at the time when a predetermined speech stop time Tx has elapsed from the time of the stop. However, as shown in the figure, since the specific sound is input during the re-speech command reception time period Tz from the output of the speech suppression command until a predetermined time has elapsed, the re-speech suppression command is output with the input of the specific sound, and re-speech suppression control is executed. Also, during the re-speech command reception time period Tz, the speech information A during which speech has stopped is set as the re-speech target. Therefore, at the time of input of the specific sound, the speech information A during which speech has stopped and is set as the re-speech target is re-spoken from the beginning with an appropriate re-speech message attached. In this embodiment, the re-speech command receiving time period Tz is set to the same time period as the speech stop time period Tx, but the re-speech command receiving time period Tz may be set to a time period shorter than the speech stop time period Tx.

[0043] Example 3 As shown in Fig. 5, the third embodiment is an example of a case where, in a case where speech information A consisting of an introduction part followed by a main text is spoken, the specific sound is input during the speech of the introduction part of the speech information A. In the speech information A, the introduction part is information that is not particularly important to the user and has a low speech priority, whereas the main text is information that is more important to the user than the introduction part and has a high speech priority. The audio command output unit 20 is configured to output a speech suppression command in response to the input of the specific sound when the specific sound is input during the speech of the introduction part of the speech information A, and to output a respeak command in response to the input of the specific sound when the specific sound is input during the speech of the main text of the speech information A. For example, as shown in the figure, when the specific sound is input during the speech of the introductory portion of the speech information A, a speech suppression command is output in response to the input of the specific sound, and speech suppression control is executed. Thus, at the time of input of the specific sound, the speech of the speech information A is temporarily stopped, and at the time when a predetermined speech stop time Tx has elapsed from the time of the stop, the speech information A whose speech has been temporarily stopped is respoken from the beginning of the introductory portion. Furthermore, if the above-mentioned specific sound is input again during the re-speech command reception time period Tz starting from the output of the speech suppression command, then, as in the above-mentioned fifth embodiment, a re-speech suppression command is output in response to the input of the specific sound during the re-speech command reception time period Tz (not shown), and re-speech suppression control is executed, so that the speech information A that has been set as the re-speech target and is currently stopped is re-spoken from the beginning, with an appropriate re-speech message attached. In this embodiment, the re-speech command receiving time period Tz is set to the same time period as the speech stop time period Tx, but the re-speech command receiving time period Tz may be set to a time period shorter than the speech stop time period Tx.

[0044] Example 4 As shown in Fig. 6, the third embodiment is an example of a case where, in the case of uttering voice information A consisting of an introduction part followed by a main text, the specific sound is input during utterance of the main text of the voice information A. In the voice information A, the introduction part is information that is not particularly important to the user and has a low utterance priority, while the main text is information that is more important to the user than the introduction part and has a high utterance priority. The audio command output unit 20 is configured to output a speech suppression command in response to the input of the specific sound when the specific sound is input during utterance of the introduction part of the voice information A, and to output a respeak command in response to the input of the specific sound when the specific sound is input during utterance of the main text of the voice information A. The acoustic command output unit 20 is configured to output a re-speak command only during the utterance of the main text, which has a higher speech priority than the introduction, when uttering the audio information A consisting of an introduction portion followed by a main text. For example, as shown in the figure, when the specific sound is input during the utterance of the main text of the voice information A, a re-speaking command is output in response to the input of the specific sound, and re-speaking control is executed. Also, during the utterance of the voice information A, the voice information A being uttered is set as a re-speaking target. Therefore, at the time of input of the specific sound, the voice information A being uttered that is set as a re-speaking target is re-speaked from the beginning with an appropriate re-speaking message attached.

[0045] Also, as shown in FIG. 6, a human presence sensor 25 is provided to detect a nearby user, and when the specific sound is input while voice information is being spoken, the audio command output unit 20 is configured to determine that the user wishes to resume speaking when the human presence sensor 25 detects the user and output a speech resume command, and to determine that the user wishes to suppress speech when the human presence sensor 25 does not detect the user and output a speech suppression command. [Explanation of symbols]

[0046] 1. Speech system 10 Control section 11 Speech control unit 12 Re-talk control unit 13 Speech suppression control unit 20. Acoustic command output section 21. Mike 23 Audio output processing section 24 Speakers A. Audio information Tx speech pause time Tz Re-talk command reception hours

Claims

1. A speech system including a voice output processing unit that sequentially outputs voice information from a speaker, an audio command output unit that outputs a speech suppression command in response to input of audio that satisfies a specific speech suppression determination condition; a speech suppression control unit that executes speech suppression control to suppress output of audio information by the audio output processing unit when the speech suppression command is output by the acoustic command output unit, A speech system in which the acoustic command output unit identifies consecutive inputs of acoustic sound having a sound pressure level equal to or higher than a predetermined speech suppression setting sound pressure level a predetermined number of times as an input of acoustic sound that satisfies the speech suppression judgment condition.

2. 2. The speech system according to claim 1, wherein the acoustic command output unit corrects the speech suppression set sound pressure level based on an output sound pressure level from the speaker or an input sound pressure level from a surrounding environment.

3. The voice information is information capable of identifying speech priority, The speech system according to claim 1 , wherein the speech suppression control unit prohibits execution of the speech suppression control when the audio output processing unit outputs the audio information having a high speech priority.

4. 4. The speech system according to claim 1, wherein the audio command output unit receives audio input from the speaker by using the speaker as a microphone.

5. the acoustic command output unit outputs a re-speaking command in response to an input of acoustic data satisfying a specific re-speaking determination condition; The speech system according to any one of claims 1 to 4, further comprising a re-speech control unit that executes a re-speech target setting process in which the audio output processing unit executes a re-speech target setting process for sequentially setting audio information being output or having been output as a re-speech target, and executes a re-speech control process in which the audio output processing unit outputs the audio information of the re-speech target again from the beginning when the audio command output unit outputs the re-speech command.

6. 6. The speech system according to claim 5, wherein the acoustic command output unit outputs the speech resumption command only during a speech resumption command reception time period from the time point when the speech suppression command is output until a predetermined time has elapsed.

7. the audio information output by the audio output processing unit comprises an introduction followed by a main text, 7. The speech system according to claim 5, wherein the acoustic command output unit outputs the speech suppression command while the introductory portion is being output, and outputs the re-speech command while the main text is being output.

8. An alarm device for monitoring an abnormal condition at a user's location, comprising: A speech generating system according to any one of claims 1 to 7, The audio output processing unit sequentially outputs notification information relating to the abnormal state as the audio information from the speaker.

Images