JP2008246036A - External noise improved speech voice analysis system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an external noise improved speech voice analysis system capable of improving reliability in the operation of a chaotic speech voice analyzer by simultaneously calculating an index value (reliability information) for evaluating the reliability for a diagnostic value calculated by the chaotic speech voice analyzer. <P>SOLUTION: The external noise improved speech voice analysis system is provided with a user's voice gathering microphone 1 for recording speech voice, an ambient noise gathering microphone 2 for monitoring the noise conditions of a speech environment, a speech voice analyzer (chaotic speech voice analyzer) 5 and an ambient noise analyzer 6. On the basis of a user's voice diagnostic value A from the speech voice analyzer (chaotic speech voice analyzer) 5 and an ambient noise evaluated value B from the ambient noise analyzer 6, the user's voice diagnostic value A from the speech voice analyzer (chaotic speech voice analyzer) 5 is evaluated in a diagnostic value comparing and judging apparatus 7. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an utterance speech analysis system for evaluating a speaker's arousal state (brain active state) from an utterance speech, and more particularly to an external noise-improved utterance speech analysis system for improving external noise of the system. is there.

  The present inventors have already developed and proposed a chaotic speech analysis apparatus and its application in advance (Patent Documents 1 to 4 below).

  The chaos utterance speech analysis device (hereinafter referred to as a chaos theory utterance speech analysis device for the device according to the present invention in order to clarify the distinction from the utterance speech analysis device as a general name) This is a device that evaluates the chaotic nature of speech and calculates a diagnostic value that is considered to be highly correlated with the activity of the cerebral neocortex of the speaker.

  The characteristic of this chaotic utterance speech analysis device is that it returns a single diagnostic value for one utterance speech, and many utterance speech analysis devices and polygraph devices are provided in the general market. Compared to a device that empirically interprets the diagnostic values and graphs and diagnoses the emotion level and the like, it is characterized by easy objective diagnosis that excludes arbitraryness.

In addition, there are some polygraph devices that enable diagnosis that does not require user experience by using knowledge information processing technology, but the knowledge processing system itself has a strong arbitrary structure. The fact that the user's arbitraryness is finally eliminated by such a device, and that the chaotic speech analysis device is not arbitrarily arbitrary in principle are completely different. This is something that should be clearly distinguished in that objective reliability can be achieved.
Japanese Patent Application No. 2006-093267 Japanese Patent Application No. 2006-093268 Japanese Patent Application No. 2006-097389 Japanese Patent Application No. 2006-097391

  However, conversely, the chaos utterance speech analysis device calculates only one diagnosis value for one utterance speech (if there is no misunderstanding, if the signal processing parameter is changed, a different diagnosis value is calculated. Is fixed, only one diagnostic value is calculated.) Since only the diagnostic value is calculated, even if the reliability of the diagnostic value is low due to the poor quality of the input audio signal, etc. In some cases, the same diagnostic value as that when the quality of the speech is sufficiently good is calculated, making it difficult to ensure reliability in the operation of the chaotic speech analysis apparatus.

  Specifically, for example, if the chaotic utterance speech analyzer returns a diagnostic value between 200 and 1000 for each input speech, if the quality of the input speech is sufficiently high, it is sleepy. Situation where the diagnostic value of around 300 is returned in the deflated state, the diagnostic value of 400 to 600 is returned in the office work, etc., and the work must be done despite being tense or fatigued If a diagnostic value of about 700 or more is returned, if the noise having a characteristic frequency spectrum by an air conditioner, a motor, or the like is strong, a low diagnostic value of about 300 or 500 despite the tension. It may return a diagnostic value that is determined to be about a healthy state. In addition, when the level of white noise that does not have a clear frequency spectrum is high, the diagnostic value shifts to a higher level depending on the noise level, and a diagnostic value of 500 or more is obtained despite being ridiculous. Or a diagnostic value of 1000 or more may be calculated if the noise level is extremely high.

  In view of the above situation, the present invention simultaneously calculates an index value for evaluating the reliability of the diagnostic value calculated by the chaotic speech analysis apparatus, thereby enabling the operation of the chaotic speech analysis apparatus. An object of the present invention is to provide an external noise-improved utterance speech analysis system capable of improving reliability.

In order to achieve the above object, the present invention provides
[1] In an external noise improvement type speech analysis system, a user speech collection microphone for recording speech speech, an environmental noise collection microphone for monitoring the noise situation of speech environment, a chaotic speech analysis device, An environmental noise analyzer, based on a user voice diagnostic value from the chaotic speech analyzer and an environmental noise evaluation value from the environmental noise analyzer, from the chaotic speech analyzer It is characterized by evaluating a user voice diagnosis value.

  [2] In an external noise improvement type utterance voice analysis system, a user voice collection microphone for recording utterance voice, an environmental noise collection microphone for monitoring the noise situation of the utterance environment, a chaotic utterance voice analysis apparatus, An indicator for evaluating chaotic speech based on a user voice diagnostic value from the chaotic speech analysis device and an environmental noise evaluation value from the environmental noise analysis device And a chaotic utterance voice index value output evaluation device for outputting a value.

  [3] In the external noise-improving speech analysis system according to [1] or [2], the user speech collection microphone for recording the speech is a unidirectional microphone, and the noise situation of the speech environment The microphone for collecting environmental noise for monitoring the above is an omnidirectional microphone.

  [4] In the external noise-improving utterance voice analysis system according to [1] or [2], when the utterance environment is a driver's seat of a car, a user voice collection microphone for recording the utterance voice is a driver's microphone. An environmental noise collecting microphone that is placed at the mouth and monitors the noise situation of the utterance environment is installed at a location away from the driver's seat, and a user voice collection microphone that records the utterance voice and the noise situation of the utterance environment Both of the environmental noise collecting microphones for monitoring the above are microphones having the same directivity characteristics.

  [5] The external noise-improving speech analysis system according to [4], wherein a vibration sensor is provided instead of the environmental noise collection microphone for monitoring the noise state of the speech environment.

  According to the present invention, by simultaneously calculating an index value (reliability information) for evaluating the reliability of the diagnostic value calculated by the chaotic speech analysis apparatus, the operation of the chaotic speech analysis apparatus is performed. Reliability can be improved.

  The external noise-improving utterance speech analysis system of the present invention includes a user speech collection microphone for recording speech speech, an environmental noise collection microphone for monitoring a noise situation in the speech environment, and a speech speech analysis device (chaotic speech). A voice analysis device) and an environmental noise analysis device, and a user voice diagnosis value from the utterance voice analysis device (chaotic utterance voice analysis device) and an environmental noise evaluation value from the environmental noise analysis device. Based on the utterance voice analysis device (chaotic utterance voice analysis device), the user voice diagnosis value is evaluated.

  Hereinafter, embodiments of the present invention will be described in detail.

  FIG. 1 is a block diagram of an external noise improvement type speech analysis system showing a first embodiment of the present invention.

  In this figure, 1 is a user voice collecting microphone for recording speech voice, 2 is an environmental noise collecting microphone for monitoring the noise situation of the speech environment, 3 is an amplifier connected to the user voice collecting microphone 1, 4 is an amplifier connected to the environmental noise collecting microphone 2, 5 is an utterance voice analysis device (chaos chaotic speech analysis device) connected to the amplifier 3, 6 is an environmental noise analysis device connected to the amplifier 4, 7 Is connected to the utterance voice analysis device 5, and the user voice diagnosis value A is input, and also connected to the environmental noise analysis device 6, and the environmental noise evaluation value B from the environmental noise analysis device 6 is input. The determination device 8 is connected to the utterance voice analysis device 5, and the user voice diagnosis value A is input and also connected to the diagnosis value comparison determination device 7, and reliability information of the diagnosis value There diagnostic information display input, and a notification device alerts the like.

  FIG. 2 is an operational flowchart of the external noise-improved utterance speech analysis system according to the first embodiment of the present invention.

  As is clear from this figure, in this system, the comparison determination calculation unit 11 generates the reliability information C in consideration of the user voice diagnosis value A and the environmental noise evaluation value B. If the reliability information C is valid by the operation reference unit 12, the notification information D such as a warning is output, and if it is invalid, the re-diagnosis E is requested.

  An utterance voice analysis device (chaotic utterance speech analysis device) is provided along with a user voice collection microphone 1 that records utterance speech and an environmental noise collection microphone 2 that monitors the noise situation of the utterance environment when recording the utterance speech. ) The user voice diagnosis value A by 5 and the environmental noise evaluation value B from the environmental noise analyzer 6 are simultaneously evaluated. Then, a reliability index value (reliability) for evaluating the reliability of the user speech diagnosis value A calculated by the speech speech analysis device (chaotic speech speech analysis device) 5 in consideration of the environmental noise evaluation value B. Sex information) C was calculated at the same time. Thereby, even when the reliability of the user voice diagnosis value A of the utterance voice analysis device (chaotic utterance voice analysis device) 5 is low, the user voice diagnosis value A is evaluated and compensated. .

  The diagnostic value from the utterance voice analysis device (chaos chaotic utterance voice analysis device) 5 is output as it is. When the environmental noise evaluation value B of the environmental noise analysis device 6 is large, the supervisor manually The determination based on the diagnostic value may be stopped (warning or other notification).

  When combining the user sound collection microphone 1 for recording speech speech as described above with the environmental noise collection microphone 2 for monitoring the noise situation of the speech environment, when both must be installed close to each other If the microphone 1 for collecting user's voice for recording speech is a unidirectional microphone and the microphone 2 for collecting environmental noise for monitoring the noise situation in the speech environment is a non-directional microphone, the microphone can be used sufficiently. .

  FIG. 3 shows an experimental example (part 1) of the external noise-improved utterance voice analysis system of the present invention, FIG. 4 shows an experimental example (part 2) of the external noise-improved utterance voice analysis system of the present invention, and FIG. It is a figure which shows the experimental example (the 3) of an external noise improvement type | formula speech analysis system.

  In FIG. 3, the solid line A indicates the speech waveform of the user's mouth, and the thin line B indicates a slightly strong noise. If it is this level, the speech waveform of the mouth can be adopted and the speech waveform can be evaluated.

  In FIG. 4, the solid line C indicates the speech waveform of the user's mouth, and the thin line D indicates weak noise. If it is such a mild weak noise, there is no problem in evaluating the speech waveform of the mouth.

  In FIG. 5, the solid line E indicates the speech waveform of the user's mouth, and the thin line F indicates strong noise. As described above, when there is strong noise, it is necessary to improve the external condition and to request rediagnosis in order to employ the speech waveform.

  FIG. 6 is a block diagram of an external noise improvement type speech analysis system showing a second embodiment of the present invention.

  In this figure, 21 is a vehicle, 22 is a driver's seat, 23 is a user voice collecting microphone for recording speech, 24 is a driver, 25 is a speech analysis device (chaos chaotic speech analysis device), and 26 is The environmental noise collecting microphone 27 monitors the noise situation of the utterance environment, 27 is an environmental noise analyzing device connected to the environmental noise collecting microphone 26, and 28 is outputted from the utterance voice analyzing device (chaotic utterance speech analyzing device) 25. The chaos utterance speech index value output device outputs the index value by evaluating the chaos theory utterance speech analysis in consideration of the environmental noise evaluation value output from the environment noise analysis device 27 in the diagnosis value.

  In this way, in a uniform noise environment such as the driver's seat 22 of the vehicle 21, an environment in which the noise situation of the utterance environment is monitored with the user voice collecting microphone 23 that records the utterance voice as the mouth of the driver 24. If the noise collecting microphone 26 is installed at a location away from the driver's seat 22, both may be microphones having the same directivity characteristics.

  In particular, when engine noise or road surface noise is a problem in the driver's seat of an automobile, the same diagnostic value can be evaluated even if a vibration sensor or the like is used as an environmental noise collecting microphone for monitoring the noise situation of the speech environment. It is possible to do.

  In addition, this invention is not limited to the said Example, Based on the meaning of this invention, a various deformation | transformation is possible and these are not excluded from the scope of the present invention.

  The utterance speech analysis system with improved external noise according to the present invention can be used as an utterance speech analysis system that can improve the reliability of the index value even in an utterance environment in which external noise is a problem.

1 is a block diagram of an external noise improvement type speech analysis system showing a first embodiment of the present invention. FIG. It is an operation | movement flowchart of the external noise improvement type | mold speech analysis system which shows 1st Example of this invention. It is a figure which shows the experiment example (the 1) of the external noise improvement type | mold speech analysis system of this invention. It is a figure which shows the experiment example (the 2) of the external noise improvement type | mold speech analysis system of this invention. It is a figure which shows the experiment example (the 3) of the external noise improvement type | mold speech analysis system of this invention. It is a block diagram of an external noise improvement type speech analysis system showing a second embodiment of the present invention.

Explanation of symbols

1,23 Microphone for collecting user's voice to record uttered voice 2,26 Microphone for collecting environmental noise to monitor the noise situation of utterance environment 3 Amplifier connected to microphone for collecting user's voice 4 Connect to microphone for collecting environmental noise 5,25 utterance speech analyzer (chaotic utterance speech analyzer)
6, 27 Environmental noise analysis device 7 Diagnostic value comparison / determination device 8 Diagnostic information display and warning and other notification device 11 Comparison determination operation unit A User voice diagnostic value B Environmental noise evaluation value C Reliability information 12 Operation reference unit D Notification Information E Re-diagnosis 21 Vehicle 22 Driver's seat 24 Driver 28 Chaotic utterance speech index value output device

Claims (5)

(A) a microphone for collecting user voice for recording uttered voice;
(B) an environmental noise collecting microphone for monitoring the noise situation of the speech environment;
(C) a chaotic speech analysis device;
(D) an environmental noise analyzer,
(E) a user voice diagnostic value from the chaotic utterance speech analyzer based on a user voice diagnostic value from the chaotic utterance speech analyzer and an environmental noise evaluation value from the environmental noise analyzer Speech noise analysis system with improved external noise, characterized by (A) a microphone for collecting user voice for recording uttered voice;
(B) an environmental noise collecting microphone for monitoring the noise situation of the speech environment;
(C) a chaotic speech analysis device;
(D) an environmental noise analyzer,
(E) Chaos theory that outputs an index value that evaluates chaotic speech based on a user speech diagnostic value from the chaotic speech analysis device and an environmental noise evaluation value from the environmental noise analysis device An external noise-improved utterance speech analysis system, characterized by comprising a speech utterance index value output evaluation device.   3. The external noise improving utterance voice analysis system according to claim 1 or 2, wherein the user voice collection microphone for recording the utterance voice is a unidirectional microphone, and environmental noise collection for monitoring a noise situation of the utterance environment. A speech analysis system with improved external noise, characterized in that the microphone is an omnidirectional microphone.   In the external noise improvement type utterance voice analysis system according to claim 1 or 2, when the utterance environment is a driver's seat of an automobile, a user voice collection microphone for recording the utterance voice is arranged in a driver's mouth, An environmental noise collection microphone for monitoring the noise situation of the speech environment is installed at a location away from the driver's seat, and a user voice collection microphone for recording the speech voice and an environmental noise collection for monitoring the noise situation of the speech environment An external noise improvement type speech analysis system characterized in that both microphones are microphones having the same directivity characteristics.   5. The external noise-improved utterance voice analysis system according to claim 4, further comprising a vibration sensor instead of the environmental noise collection microphone for monitoring the noise situation of the utterance environment.