JP2011170113A - Conversation protection degree evaluation system and conversation protection degree evaluation method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To appropriately evaluate a conversation protection degree by a conversation security system without widely varying the evaluation depending on a measuring person. <P>SOLUTION: A conversation protection degree evaluation system is constituted in such a manner that: a listening-prevention sound generation device 1 generates listening-prevention sound, based on a predetermined sound collected by a sound collection microphone 110, and generated listening-prevention sound is outputted from a listening-prevention sound speaker 120, and speech analysis part 22a of a measuring device 2 collects each of mixed sound and original sound; a listening-prevention sound output degree calculation part 22b of the measuring device 2 calculates each of sound pressure levels of the mixed sound and the original sound, and calculates the listening-prevention sound output degree based on difference of the sound pressure levels of the calculated mixed sound and the original sound; and a conversation protection degree calculation part 22c of the measuring device 2 calculates the conversation protection degree from the listening-prevention sound output degree calculated by the listening-prevention sound output degree calculation part 22b by using a correlation expression between the conversation protection degree and the listening-prevention sound output degree. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a conversation protection degree evaluation system and a conversation protection degree evaluation method for evaluating the degree of conversation protection (conversation protection degree) by a conversation security system that protects confidentiality and privacy of conversation leaking to the surroundings. The present invention relates to a conversation protection degree evaluation system and a conversation protection degree evaluation method that can appropriately evaluate the above without causing variation by the senser's sense.

  Conventionally, in an open space such as a hospital, a pharmacy, or a bank, conversations related to privacy are frequently performed. Therefore, for the purpose of protecting the privacy of the conversation parties, a conversation security system has been developed that outputs a masking sound that makes it difficult to hear the conversation voice over the conversation voice.

  In particular, in recent years, a conversation security system that makes it difficult to hear a conversational sound by using a masking sound generated based on the conversational sound, instead of using a voice that is less related to the conversational sound such as white noise or BGM as a masking sound. Has also been developed.

  For example, Patent Document 1 discloses a masking device that generates and outputs a masking sound having a sound pressure peak at a frequency different from a frequency having a sound pressure peak in the conversation sound based on the waveform spectrum of the input conversation sound. Is disclosed.

  That is, the masking device described in Patent Document 1 breaks down the features (phonological characteristics) of conversational speech by perceptually fusing the original conversational speech and masking sound, and a third party recognizes and distinguishes these speeches. Can not be. As a result, the conversation voice becomes unclear for the third party, and the contents of the conversation can be made difficult to hear. In addition, like the masking sound output by the masking device described in Patent Document 1, a masking sound that breaks the phoneme of the original conversational voice may be referred to as “interfering sound” or “hearing sound”.

  By the way, when the masking device described in Patent Document 1 is introduced, after evaluating in advance how much the conversational sound is difficult to hear due to the disturbing sound, the arrangement of the acoustic device such as the microphone and the speaker is determined or disturbed. It is desirable to determine the volume of the sound.

  Specifically, in the past, the effect of the disturbing sound (conversation protection level) was determined by judging how much the person actually heard and understood the sound when the disturbing sound was output in conjunction with the conversation sound. I was evaluating.

Japanese Patent No. 4336552

  However, in the conventional listening experiment, there is a possibility that judgment by the measurer may vary, and a long time is required for the measurement. For this reason, how to realize a conversation protection degree evaluation system or a conversation protection degree evaluation method that can appropriately evaluate the degree of conversation protection by the conversation security system without causing variation by the senser's sense is a major issue. It has become.

  The present invention has been made to solve the above-described problems caused by the prior art, and is capable of appropriately evaluating the degree of conversation protection by the conversation security system without causing variation by the senser's sense. An object is to provide a degree evaluation system and a conversation protection degree evaluation method.

  In the following, a masking sound that breaks the phonological correctness of the original conversational speech is referred to as a “hearing sound”. Further, the conversation protection degree evaluation system and the conversation protection degree evaluation method according to the present invention provide a conversation security system that makes it difficult to hear a conversation by superimposing a hearing loss sound generated by processing the conversation voice on the conversation voice. It is to be provided for.

  In order to solve the above-described problems and achieve the object, the present invention generates a hearing loss sound that makes it difficult to hear the predetermined sound based on the predetermined sound collected by the sound collection unit, and generates the generated hearing loss sound. A hearing-proof sound output means for outputting from a predetermined speaker and a mixed sound that is a sound obtained by superimposing the hearing-proof sound on the predetermined sound or an original sound that is a sound when only the predetermined sound is generated. A sound collecting means for sounding, and a sound pressure level of the mixed sound and the original sound collected by the sound collecting means, respectively, and a difference between the calculated sound pressure level of the mixed sound and the original sound. An output degree calculation means for calculating a hearing loss output degree based on the above, and an expression representing a correlation between the degree of hearing protection and the degree of hearing protection output indicating the proportion of the person who cannot hear the sound Used because was correlation equations, characterized in that a protection degree calculating means for calculating the conversation protection degree from explosion listening output degree calculated by the output degree calculation means.

  Further, in the present invention according to the above invention, the output degree calculation unit specifies a frequency band in which the hearing-proof sound is effectively output based on the waveform spectrum of the mixed sound and the waveform spectrum of the original sound. Calculating the representative value of the sound pressure level of the mixed sound in the specified frequency band as the sound pressure level of the mixed sound, and calculating the representative value of the sound pressure level of the original sound in the specified frequency band. It is calculated as a pressure level.

  In the above invention, the present invention further includes a test speaker that outputs a pre-recorded gender-specific vowel as the predetermined sound, and the output degree calculation means includes a sound pressure level for each vowel in the mix sound. Is calculated as the sound pressure level of the mixed sound, and the representative value of the sound pressure level for each vowel in the original sound is calculated as the sound pressure level of the original sound.

  Further, the present invention is the above invention, wherein the sound collecting means collects ambient sound in a state where the predetermined sound and the hearing-proof sound are not generated as background noise, and the protection degree calculating means includes: Using the correlation formula according to the sound pressure level of the background noise, the conversation protection degree is calculated from the hearing-aid sound output degree calculated by the output degree calculation means and the noise level of the background noise collected by the sound collection means. It is characterized by calculating.

  Further, the present invention provides the hearing-aid sound output level so that the conversation protection level is within a predetermined range when the conversation protection level calculated by the protection level calculation means is outside the predetermined range. It further comprises a setting changing means for changing the set value.

  Further, the present invention is the above invention, wherein the hearing-aid output means collects a predetermined sound uttered by a tester using the sound collection unit and outputs a hearing-aid sound corresponding to the collected sound. Generating and outputting from the predetermined speaker, the sound collecting means collects the predetermined sound uttered by the tester and the hearing-proof sound output by the hearing-proof sound output means, respectively, and the output degree calculating means The predetermined sound collected by the sound collecting means is used as an original sound, and the original sound is synthesized with the hearing-proof sound collected by the sound collecting means, thereby generating a mixed sound.

  Further, the present invention provides a first sound collecting step for collecting a predetermined sound as an original sound, and a hearing-proof sound that makes it difficult to hear the predetermined sound based on the predetermined sound collected by the sound collecting unit. A hearing-proof sound output step of generating and outputting from a predetermined speaker; a second sound-collecting step of collecting a mixed sound that is a sound obtained by superimposing the hearing-aid sound on the predetermined sound; and The sound pressure level of the original sound collected in the sound collecting step and the sound pressure level of the mixed sound collected in the second sound collecting step are calculated, respectively, and the calculated mixed sound and the original sound As an expression representing the correlation between the output degree calculation step for calculating the degree of hearing loss output based on the difference in the sound pressure level, the degree of conversation protection indicating the rate at which a person cannot hear sound, and the degree of hearing loss output. Using the obtained correlation equation, characterized in that the explosion-listening output degree calculated in the output degree calculation step including a protection degree calculation step of calculating the conversation protection degree.

  According to the present invention, the hearing-proof sound output means generates the hearing-aid sound that makes it difficult to hear the predetermined sound based on the predetermined sound collected by the sound collecting unit, and outputs the generated hearing-aid sound from the predetermined speaker. The sound collecting means collects a mixed sound that is a sound in which a hearing-proof sound is superimposed on a predetermined sound or an original sound that is a sound when only the predetermined sound is generated, and an output degree calculating means In addition to calculating the sound pressure level of the mixed sound and the original sound collected by the sound collecting means, the degree of hearing loss output is calculated based on the difference between the calculated sound pressure level of the mixed sound and the original sound, thereby protecting the sound. The degree calculation means uses a correlation formula obtained in advance as an expression representing the correlation between the degree of conversation protection and the degree of hearing loss output, which indicates the rate at which a person cannot hear speech, and the output degree calculation means Because from the calculated anti-listening output degree it was decided to calculate the conversation protection degree, an effect that can be properly evaluated without variations conversation protection degree according to the conversation security system by the sense of the measurer is caused.

  Further, according to the present invention, the output degree calculation means specifies the frequency band in which the hearing-proof sound is effectively output based on the waveform spectrum of the mixed sound and the waveform spectrum of the original sound, and mixes in the specified frequency band The representative value of the sound pressure level of the sound is calculated as the sound pressure level of the mixed sound, and the representative value of the sound pressure level of the original sound in the specified frequency band is calculated as the sound pressure level of the original sound. There is an effect that the output degree can be calculated with higher accuracy.

  Further, according to the present invention, the test speaker outputs a pre-recorded gender-specific vowel as a predetermined sound, and the output degree calculation means calculates the representative value of the sound pressure level for each vowel in the mix sound. As the sound pressure level is calculated, the representative value of the sound pressure level for each vowel in the original sound is calculated as the sound pressure level of the original sound. By doing so, there is an effect that it is possible to obtain the degree of conversation protection in accordance with the actual environment.

  Further, according to the present invention, the sound collecting means collects ambient sounds in a state where the predetermined sound and the hearing-proof sound are not generated as background noise, and the protection degree calculating means adjusts the sound pressure level of the background noise. By using the corresponding correlation equation, the conversation protection level is calculated from the hearing-aid sound output level calculated by the output level calculation means and the background noise level collected by the sound collection means. There is an effect that it is possible to appropriately obtain the degree of conversation protection that varies accordingly.

  Further, according to the present invention, when the setting change means has the conversation protection degree calculated by the protection degree calculation means outside the predetermined range, the output level of the hearing loss sound is adjusted so that the conversation protection degree is within the predetermined range. Since the setting value is changed, it is possible to reduce the labor required for adjusting the output level of the hearing loss sound.

  Further, according to the present invention, the hearing-aid sound output means collects a predetermined sound uttered by the tester using the sound collecting unit, and generates a hearing-aid sound corresponding to the collected sound to generate a predetermined sound. The sound collecting means outputs the predetermined sound uttered by the tester and the hearing-proof sound output by the hearing-proof sound output means, respectively, and the output degree calculating means outputs the predetermined sound collected by the sound collecting means. The original sound is used as the original sound and the mixed sound is generated by synthesizing the original sound with the hearing-proof sound collected by the sound collecting means. Therefore, the test sound is prepared without preparing the pre-recorded sound. An effect is obtained in that the degree of conversation protection can be obtained using the voice uttered by the person on the spot.

FIG. 1 is a diagram showing an outline of a conversation protection degree evaluation method according to the present invention. FIG. 2 is a diagram illustrating a preferred arrangement example of the conversation protection degree evaluation system according to the present embodiment. FIG. 3 is a block diagram illustrating the configuration of the hearing-aid generator and the measuring device according to the present embodiment. FIG. 4 is a diagram for explaining a calculation procedure of the hearing-aid sound output degree. FIG. 5 is a diagram showing the measurement result of the hearing-proof sound output degree at the listening position 1 and the experimental result of the listening experiment. FIG. 6 is a diagram showing a result of averaging the correlation between the hearing-proof sound output level and the conversation protection level at each of the listening positions 1 to 3. FIG. 7 is a sequence diagram showing a processing procedure executed between the hearing-proof sound generating device and the measuring device. FIG. 8 is a flowchart showing the processing procedure of the hearing-aid sound output degree calculation processing. FIG. 9 is a flowchart showing a processing procedure of sound pressure level calculation processing. FIG. 10 is an explanatory diagram in the case of obtaining the hearing-aid output level using the tester's raw voice. FIG. 11 is an explanatory diagram in the case of adjusting the volume of the hearing loss sound using the conversation protection degree obtained in the actual environment.

  Exemplary embodiments of a conversation protection degree evaluation system and a conversation protection degree evaluation method according to the present invention will be described below in detail with reference to the accompanying drawings.

  First, prior to detailed description of the embodiment, an outline of the conversation protection degree evaluation method according to the present invention will be described with reference to FIG. FIG. 1 is a diagram showing an outline of a conversation protection degree evaluation method according to the present invention.

  The conversation protection degree evaluation method according to the present invention is generally a method for mechanically evaluating how difficult the conversation content is to be heard by the conversation security system. Here, the conversation security system is a system that makes it difficult for a third party to hear the contents of the conversational sound by superimposing and outputting a predetermined hearing-proof sound generated based on the conversational sound with the conversational sound.

  Specifically, as shown in the figure, in the conversation security system, a sound collecting microphone installed in the vicinity of the speaker collects the conversation voice of the speaker (in this case, the test voice) to prevent hearing loss. Output to the generator. In the conversation security system, the hearing-proof sound generating device generates a hearing-proof sound based on the conversation sound collected by the sound collecting microphone, and the generated hearing-proof sound is transmitted from the hearing-proof sound speaker installed in the vicinity of the listener. Output.

  Here, the hearing-proof sound generating device generates, for example, a sound having a sound pressure peak at a frequency different from a frequency having a sound pressure peak in the conversational sound as a hearing-proof sound. The hearing-aid generator can also generate a sound having a modified spectrum envelope obtained by inverting the spectrum envelope of a conversational sound as a hearing-aid. In this way, in the conversation security system, the hearing aid sound based on the conversation sound is generated and output, so that the third person cannot distinguish and recognize these sounds. As a result, the conversation voice becomes unclear for the third party, and the contents of the conversation can be made difficult to hear.

  In the case of evaluating such a conversation security system, conventionally, evaluation is performed by judging the extent to which a person has actually heard and understood the sound when hearing-proof sound is output in conjunction with the conversation sound. However, in such a listening experiment, there is a possibility that judgment by the measurer may vary, and a long time is required for the measurement.

  Here, the inventors of the present application have found that the hearing-aid sound output level and the conversation protection level are in correlation with each other through measurement experiments and listening experiments. Therefore, in the conversation protection level evaluation method according to the present invention, the conversation protection level is calculated from the hearing loss output level by using a correlation formula between the hearing loss output level and the conversation protection level.

  That is, the speech protection degree evaluation method according to the present invention has a main feature in that the hearing protection sound output degree is calculated by measurement, and the conversation protection degree is calculated using the hearing protection sound output degree and the correlation equation. Here, the degree of conversation protection indicates the rate at which a person cannot hear sound.

  Specifically, as shown in the figure, in the conversation protection degree evaluation method according to the present invention, a predetermined test voice is output from a test speaker in a state where the conversation security system is not operated, and a third party Sound is collected using a measurement microphone installed at a position (measurement position) that is assumed to be present. In the conversation protection degree evaluation method according to the present invention, the measurement device stores the test sound collected by the measurement microphone as the original sound.

  Subsequently, in the conversation protection degree evaluation method according to the present invention, the same test sound as described above is output from the test speaker while the conversation security system is operated. As a result, the hearing loss sound based on the test sound is output from the hearing loss speaker. Then, in the conversation protection degree evaluation method according to the present invention, the voice including the test voice and the hearing loss is collected using the measurement microphone, and the measurement device stores the collected voice as the mixed sound.

  Subsequently, in the conversation protection level evaluation method according to the present invention, the hearing loss output level is calculated based on the difference between the sound pressure levels of the original sound and the mixed sound (see (1) in the figure). Here, the degree of hearing-proof sound output is represented by the ratio of the difference between the sound pressure level of the mixed sound and the sound pressure level of the original sound to the sound pressure level of the original sound. A specific procedure for calculating the hearing-aid sound output level will be described later in the embodiment.

  In the conversation protection degree evaluation method according to the present invention, the measurement device calculates the conversation protection degree from the hearing-aid sound output degree using a correlation equation obtained in advance by a listening experiment (see (2) in the figure).

  As described above, in the conversation protection level evaluation method according to the present invention, the original sound and the mixed sound are collected using the measurement microphone, and the measurement device outputs the hearing-aid sound output level based on the difference between these sound pressure levels. And the conversation protection degree is calculated from the hearing-aid sound output degree using the correlation equation obtained in advance. Therefore, it is possible to appropriately evaluate the effect of the conversation security system, that is, the degree of conversation protection without variation depending on the feeling of the measurer.

  Moreover, in the conversation protection degree evaluation method according to the present invention, the degree of conversation protection, that is, how difficult it is to hear the contents of the conversation voice is mechanically determined numerically. Therefore, for example, when the introduction of a conversation security system is recommended in sales or the like, the effect of the conversation security system can be presented numerically to the customer, so a sales promotion effect can also be expected.

  Incidentally, the degree of conversation protection has a correlation not only with the degree of hearing-proof sound output but also with the sound pressure level of background noise (hereinafter referred to as “background noise level”). Here, background noise refers to ambient sounds in a state where no test sound or hearing-proof sound is generated.

  For this reason, in the conversation protection level evaluation method according to the present invention, the degree of conversation protection is calculated from the hearing loss output level and the background noise level using a correlation formula indicating the correlation between the hearing loss output level, the background noise level, and the level of conversation protection. To do. Thus, by calculating the conversation protection level in consideration of not only the hearing-proof sound output level but also the background noise level, it is possible to appropriately obtain the conversation protection level that varies according to the noise level of the background noise.

  The voice used as the test voice may be any voice, but it is particularly preferable to use a vowel for each gender. Specifically, in the conversation protection degree evaluation method according to the present invention, a male and a female each have five vowels “A”, “E”, “I”, “O”, and “U” separated by a predetermined interval. A total of 10 pronunciationed syllables are used as test sounds. In this way, by using the vowels of men and women with the highest appearance frequency in the conversational voice, it is possible to obtain the degree of conversation protection in accordance with the actual environment. In addition, discrete vowels have a clear feature (formant) by spectral analysis, and have an advantage that they are suitable for measuring the degree of hearing loss output.

  In this case, the measurement is performed using the voice recorded in advance as the test voice. However, the measurement is not limited to this, and a live voice of a person (tester) can be used as the test voice.

  Here, when a raw voice is used, it is difficult for the tester to utter the same voice twice at the same volume. Therefore, in the conversation protection degree evaluation method according to the present invention, the tester can generate a mixed sound by mechanically synthesizing the test sound uttered by the tester and the hearing aid sound generated by the hearing aid sound generation device. The degree of conversation protection can be calculated by uttering the test voice once. Details of this point will be described later in the embodiments.

  Further, in the method for evaluating the degree of conversation protection according to the present invention, not only simply calculating the degree of conversation protection, but also determining the optimum output level (volume) of the hearing loss sound from the calculated degree of conversation protection and automatically adjusting it. You can also. As a result, the labor required for adjusting the volume of the hearing loss can be reduced.

  Below, the Example about the conversation protection degree evaluation system and the conversation protection degree evaluation method which applied the conversation protection degree evaluation method demonstrated using FIG. 1 is described in detail.

  FIG. 2 is a diagram illustrating a preferred arrangement example of the conversation protection degree evaluation system according to the present embodiment. As shown in the figure, the conversation security system is installed at, for example, consultation counters of hospitals and financial institutions so that third parties waiting in the waiting space cannot hear the contents of the conversation in the booth. To do.

  Specifically, the conversation security system includes a hearing-proof sound generating device 1, a sound collecting microphone 110, and hearing-proof sound speakers 120a and 120b. Here, the sound collecting microphone 110 is a sound collecting unit for collecting conversation voice in a booth surrounded by partitions, and is installed in the vicinity of a position (speaker position) where a speaker will sit. . In the case shown in the figure, the sound collecting microphone 110 is installed on a counter table in the booth. The sound collecting microphone 110 converts the collected sound into a digital signal by A / D (Analog / Digital) conversion, and then outputs the digital signal to the hearing aid sound generator 1.

  The hearing-proof sound generating device 1 is a device that generates a hearing-proof sound based on the conversational sound input from the sound collecting microphone 110, and is generated at a timing at which the generated hearing-aid sound and the conversational sound are perceived as a synthesized sound. The hearing sound is output from the hearing-proof sound speakers 120a and 120b.

  The hearing-proof sound speakers 120a and 120b are sound output units that output the hearing-proof sound generated by the hearing-proof sound generating device 1, and are installed in a place (here, a waiting space) where a third person will be. Specifically, the hearing-proof sound speakers 120a and 120b are installed in a partition facing the waiting space of each booth, and output the hearing-proof sound toward the waiting room.

  On the other hand, the measurement system for evaluating the degree of conversation protection by the conversation security system includes the measurement device 2, measurement microphones 210 a to 210 c, and a test speaker 220. The measurement microphones 210a to 210c are sound collection units that collect test sound, synthesized sound of test sound and deaf sound, background noise, and the like instead of the ear of the measurer, and a place where a third party will be (here, , Waiting space).

  Specifically, the measurement microphone 210a is installed at the listening position 1 in front of the hearing-proof sound speaker 120a, and the measurement microphone 210c is installed at the listening position 3 in front of the hearing-proof sound speaker 120b. The measurement microphone 210b is located at the listening position 2 that is an intermediate position between the measurement microphone 210a and the measurement microphone 210b and that is also in front of the booth entrance. The measurement microphones 210 a to 210 c convert the collected sound into a digital signal by A / D (Analog / Digital) conversion, and then output the digital signal to the measurement apparatus 2.

  The measuring device 2 is a device that calculates the degree of conversation protection based on the sound input from the measuring microphones 210a to 210c. Specifically, the measuring device 2 calculates the hearing-aid output level based on the difference between the sound pressure levels of the original sound and the mixed sound input from the measurement microphones 210a to 210c. And the measuring apparatus 2 calculates a conversation protection degree from the calculated hearing-aid sound output degree using the correlation formula calculated | required beforehand by listening experiment.

  The measuring device 2 calculates the degree of conversation protection at the listening position 1 based on the sound input from the measurement microphone 210a, and calculates the degree of conversation protection at the listening position 2 based on the sound input from the measurement microphone 210b. Assume that the degree of conversation protection at the listening position 3 is calculated based on the sound input from the measurement microphone 210c.

  The test speaker 220 is an audio output unit that outputs a test audio in accordance with an instruction from the measurement device 2. Specifically, the test speaker 220 is installed at a speaker position in the booth, and outputs a test sound toward a sound collecting microphone on the counter table and a waiting space located at an extension thereof.

  Here, the test speaker 220 outputs vowels for each gender recorded in advance as test voices at predetermined intervals. Specifically, a total of 10 syllables, each of which is pronounced by a man and a woman, with five vowels “a”, “e”, “i”, “o”, and “c” at predetermined intervals, are output as test sounds. . In this way, by using the vowels of men and women with the highest appearance frequency in the conversational voice, it is possible to obtain the degree of conversation protection in accordance with the actual environment. In addition, since discrete vowels have clear characteristics (formants) based on spectrum analysis, they are suitable for measuring the degree of hearing loss output.

  Next, the configuration of the hearing-aid generator 1 and the measuring device 2 according to the present embodiment will be described with reference to FIG. FIG. 3 is a block diagram showing the configuration of the hearing-aid generator 1 and the measuring device 2 according to the present embodiment. In the drawing, only components necessary for explaining the characteristics of the hearing-aid generator 1 and the measuring device 2 are shown, and descriptions of general components are omitted.

  In the following, any hearing-aid speaker among the hearing-aid speakers 120a and 120b is simply referred to as “a hearing-aid speaker 120”, and any measurement microphone among the measurement microphones 210a to 210c is simply referred to as a “measurement microphone 210”.

  As shown in FIG. 3, the hearing-aid generator 1 according to the present embodiment includes an output audio amplifying unit 11, a control unit 12, and a storage unit 13. In addition, the control unit 12 includes an audio analysis unit 12a, a hearing-proof sound generation unit 12b, an output control unit 12c, and a setting change unit 12d, and the storage unit 13 stores a volume setting value 13a. .

  On the other hand, the measuring apparatus 2 according to the present embodiment includes a display unit 21, a control unit 22, and a storage unit 23. The control unit 22 includes a voice analysis unit 22a, a hearing-aid sound output level calculation unit 22b, a conversation protection level calculation unit 22c, and a setting change instruction unit 22d. The storage unit 23 stores spectrum data 23a.

  Below, each component of the hearing-aid generator 1 is demonstrated first. The output sound amplifying unit 11 is a processing unit that amplifies a hearing-proof sound signal, which is an output sound, in accordance with an instruction from the output control unit 12c and adjusts the volume. The output sound amplifying unit 11 also performs a process of outputting the amplified hearing loss sound to the hearing loss sound speaker. Note that the output sound amplifying unit 11 outputs the hearing loss sound from the hearing aid sound speaker 120 at a timing instructed by the output control unit 12c (for example, timing within “30 msec” from the speaker's utterance).

  The control unit 12 is a processing unit that performs processing such as spectrum analysis of sound collected by the sound collecting microphone 110, generation of hearing loss, output control of hearing loss, and setting change of the hearing loss volume. The sound analysis unit 12 a is a processing unit that performs spectrum analysis of the sound collected by the sound collection microphone 110.

  Specifically, the voice analysis unit 12a performs analysis such as FFT (Fast Fourier Transform) / DFT (Discrete Fourier Transform) on the voice signal extracted in fragments from the input voice (conversation voice) every predetermined time. Processing is performed to analyze the characteristics of the waveform spectrum such as sound pressure level and frequency distribution. In this embodiment, as a feature of the waveform spectrum of the conversational speech, a formant (a frequency component in which energy is particularly concentrated in the spectrum of the speech waveform, the first formant and the second formant in order from the lowest). ) ...) and formant spacing characteristics, energy concentration characteristics on formants, or frequency band characteristics.

  The hearing-proof sound generating unit 12b is a processing unit that generates a hearing-proof sound to be output based on the conversational sound analyzed by the sound analyzing unit 12a. Specifically, the hearing-proof sound generating unit 12b is based on the characteristics of the waveform spectrum of conversational speech (characteristics of formant-to-formant spacing, characteristics of how energy is concentrated on formants, or characteristics of frequency bands). Determine the frequency band and sound pressure level of the output hearing loss. More specifically, the hearing-proof sound generation unit 12b generates a spectrum having a sound pressure peak at a frequency different from a frequency having a sound pressure peak in the conversational voice.

  The output control unit 12c is a processing unit that controls to output a hearing-aid sound corresponding to the spectrum generated by the hearing-aid sound generation unit 12b. That is, the output control unit 12c processes the conversational sound collected by the sound collection microphone 110 based on the spectrum generated by the hearing-proof sound generation unit 12b to generate “hearing sound”.

  Then, the output control unit 12c instructs the output sound amplifying unit 11 to output the hearing aid sound at a timing when the generated hearing aid sound and the conversational sound are perceived as a synthesized sound. Further, the output control unit 12c determines the volume of the hearing loss sound with reference to the volume setting value 13a stored in the storage unit 13, and instructs the output audio amplification unit 11 to output the hearing loss sound at the determined volume. This process is also performed.

  As described above, since the hearing aid sound determined based on the characteristics of the waveform spectrum of the conversation voice is dynamically output following the conversation voice in real time, the optimum hearing aid sound is activated for the conversation voice. Can be output automatically. In addition, since the hearing person's understanding of the meaning is prevented by outputting the hearing-aid sound determined based on the characteristics of the waveform spectrum of the conversation voice, the hearing-aid effect can be increased with a volume lower than that of the conversation voice. The details of this point are described in, for example, Japanese Patent Application Laid-Open No. 2006-243178.

  Note that the hearing sound generator 1 may generate a sound having a modified spectral envelope obtained by inverting the spectral envelope of the conversational sound as a hearing-proof sound. In such a case, the same effect as described above can be obtained. .

  The setting change unit 12d is a processing unit that changes the volume setting value 13a stored in the storage unit 13 in accordance with an instruction from the setting change instruction unit 22d of the measuring device 2. The setting changing unit 12d can change the volume setting value 13a in accordance with not only an instruction from the setting change instructing unit 22d but also an operation on an operation unit (not shown).

  Next, the configuration of the measuring apparatus 2 according to the present embodiment will be described. The display unit 21 is a display device such as a monitor that displays various types of information. Specifically, a selection screen or the like for allowing the user to select whether or not to automatically set the degree of conversation protection and the hearing-aid volume measured by the measurement device 2 is displayed.

  The control unit 22 is a processing unit that executes processing such as analysis of the sound collected by the measurement microphone 210, calculation processing of the hearing-aid sound output level, calculation processing of the conversation protection level, and setting change instruction of the hearing-aid volume. The voice analysis unit 22 a is a processing unit that performs spectrum analysis of the voice collected by the measurement microphone 210. The voice analysis unit 22a is also a processing unit that stores an analysis result by spectrum analysis in the storage unit 23 as spectrum data 23a.

  Here, the measurement microphone 210 generates a mixed sound that is a sound obtained by superimposing a hearing loss sound on the test sound output from the test speaker 220 and an original sound that is a sound when only the test sound is generated. We are going to collect sound.

  Specifically, the control unit 22 of the measuring apparatus 2 causes the test speaker 220 to output the test sound twice. At this time, the control unit 22 of the measuring device 2 instructs the hearing-aid generating device 1 not to generate the hearing-aid sound when the first test sound is output, and generates the hearing-aid sound when the second test sound is output. Instruct the device 1 to generate a hearing-proof sound.

  As a result, the measurement microphone 210 collects and stores the original sound that is the sound when only the test sound is generated and the mixed sound that is the sound obtained by superimposing the hearing loss sound on the test sound. The sound analysis unit 22a performs spectrum analysis on the original sound and the mixed sound collected by the measurement microphone 210, respectively.

  The voice analysis unit 22a performs analysis processing such as FFT / DFT on the voice signal obtained by extracting each syllable of the test voice every unit time to analyze the characteristics of the waveform spectrum such as the sound pressure level and the frequency distribution. The analysis result is stored in the storage unit 23 as spectrum data 23a. That is, the waveform spectrum for each syllable unit time in the mixed sound and the waveform spectrum for each syllable unit time in the original sound are stored in the storage unit 23 as spectrum data 23a.

  The hearing-aid sound output degree calculation unit 22b is a processing unit that calculates the hearing-aid sound output degree based on the difference between the sound pressure levels of the original sound and the mixed sound. Here, the calculation procedure of the hearing-aid sound output degree by the hearing-aid sound output degree calculation part 22b is demonstrated using FIG.

  FIG. 4 is a diagram for explaining a calculation procedure of the hearing-aid sound output degree. In addition, (A) in the figure shows a waveform spectrum at a certain moment in a “U” syllable emitted by a woman, and (B) in the figure shows an output time zone and an effective frequency band of each syllable of the test voice. An example is shown.

  As shown in FIG. 4A, the hearing-aid sound output degree calculation unit 22b first performs a process of determining an effective frequency band for each syllable included in the test sound. Specifically, the hearing-aid sound output degree calculation unit 22b determines a frequency band of 1000 Hz centering on a frequency at which the difference in sound pressure level between the mixed sound and the original sound is maximum as an effective frequency band. For example, in the case shown in FIG. 5A, the frequency band of 1750 to 2750 Hz is determined as the effective frequency band.

  In this way, the hearing-aid output level calculation unit 22b determines the effective frequency band for each syllable (see (B) in the figure). Here, as shown in (B) of the figure, the time width of the output time zone of each syllable and the bandwidth of the effective frequency band are constant for each syllable. For example, the time width of the output time zone is constant at 80 ms, and the bandwidth of the effective frequency band is constant at 1000 Hz. As described above, by making the output time band and the effective frequency band constant, it is possible to simplify the processing for calculating the hearing-aid sound output degree described later. The output time zone is an output time zone of each syllable when the start time of test voice output is set to 0 ms.

式（１）によって表される計算式を用いて防聴音出力度合（Ｂｒ）［％］を算出する。ここで、式（１）中の「Ｐｄ（ｍｉｘ）［ｄＢ］」は、ミックス音の音圧レベルの代表値（ここでは、合計値）をあらわしており、「Ｐｄ（ｏｒｇ）［ｄＢ］」は、オリジナル音の音圧レベルの代表値（ここでは、合計値）をあらわしている。 Subsequently, the hearing-aid output level calculation unit 22b

A hearing loss output degree (Br) [%] is calculated using a calculation formula represented by Expression (1). Here, “Pd (mix) [dB]” in the expression (1) represents a representative value (here, a total value) of the sound pressure level of the mixed sound, and “Pd (org) [dB]”. Represents the representative value (here, the total value) of the sound pressure level of the original sound.

式（２）によってあらわされる計算式を用いて算出される。ここで、式（２）中の「ｆｐｔ（ｉ）」は、音節（ｉ）における音圧レベルの代表値（ここでは、合計値）をあらわしている。また、式（２）中の「Ｎ」は、音節の個数をあらわしている。ここで、本実施例では、音節の数は１０個であるため、「Ｎ」は「１０」となる。また、本実施例において音節（０）〜音節（９）は、それぞれ図４の（Ｂ）に示した音節（女性発話の「ア」〜「ウ」および男性発話の「ア」〜「ウ」）にそれぞれ対応している。 The representative value of the sound pressure level of the mixed sound (Pd (mix) [dB]) is

It is calculated using a calculation formula represented by the formula (2). Here, “fpt (i)” in the equation (2) represents a representative value (here, a total value) of the sound pressure level in the syllable (i). Further, “N” in equation (2) represents the number of syllables. Here, in this embodiment, since the number of syllables is 10, “N” is “10”. In this embodiment, the syllable (0) to syllable (9) are the syllables shown in FIG. 4B (female utterances “A” to “U” and male utterances “A” to “U”). ) Respectively.

式（３）によってあらわされる計算式を用いて算出される。ここで、式（３）中の「ｆｐ（ｉ）」は、出力時間（ｊ）における有効周波数帯域の音圧レベルの代表値（ここでは、合計値）をあらわし、「ｔ」は、出力時間帯の開始時間をあらわし、「Ｔ」は時間をあらわしている。 Further, the representative value of sound pressure level (fpt (i) [dB]) in syllable (i) is

It is calculated using a calculation formula represented by the formula (3). Here, “fp (i)” in Expression (3) represents a representative value (here, a total value) of the sound pressure level in the effective frequency band at the output time (j), and “t” is the output time. The start time of the belt is represented, and “T” represents the time.

式（４）によってあらわされる計算式を用いて算出される。ここで、式（４）中の「ｗｐ（ｋ）」は、周波数（ｋ）におけるパワーをあらわし、「ｆ」は、有効周波数帯域の始点となる周波数をあらわし、「Ｆ」は、周波数をあらわしている。 In addition, the representative value (fp (i) [dB]) of the sound pressure level in the effective frequency band at the output time (j) is

It is calculated using a calculation formula represented by the formula (4). Here, “wp (k)” in Expression (4) represents power at the frequency (k), “f” represents the frequency that is the starting point of the effective frequency band, and “F” represents the frequency. ing.

  On the other hand, the hearing-aid sound output degree calculation unit 22b calculates the representative value (Pd (org)) of the sound pressure level of the original sound using the same calculation formulas as the formulas (2) to (4). Then, by substituting the calculated Pd (mix) [dB] and Pd (org) [dB] into Equation (1), the hearing-aid sound output degree (Br) [%] is calculated.

  That is, the deafness sound output degree calculation unit 22b obtains the area of the region surrounded by the waveform spectrum of the mixed sound, the waveform spectrum of the original sound, and the effective frequency band for each syllable (see FIG. 4A). Is added for a certain time (80 ms). Then, the hearing-aid sound output degree calculation unit 22b performs this process for all syllables, and calculates the ratio of the total value of each syllable (that is, Pd (mix) −Pd (org)) to Pd (org). Calculate as the output level.

  As described above, the hearing-aid sound output degree calculation unit 22b identifies the effective frequency band in which the hearing-aid sound is effectively output based on the waveform spectrum of the mixed sound and the waveform spectrum of the original sound, and in the identified effective frequency band The representative value of the sound pressure level of the mixed sound is calculated as the sound pressure level of the mixed sound, and the representative value of the sound pressure level of the original sound in the effective frequency band is calculated as the sound pressure level of the original sound. The output degree can be calculated with higher accuracy.

  In this embodiment, since the time width of the output time zone and the bandwidth of the effective frequency band are constant for each syllable, Pd (mix), Pd (org), fpt (i) or fp (i) is calculated. In this case, it is possible to simply represent the sum of the sound pressure levels without performing a process of calculating an average value from the sum of the sound pressure levels. Therefore, it is possible to simplify the calculation process of the hearing loss output degree.

  Here, as shown in the equation (1), the ratio of Pd (mix) −Pd (org) to Pd (org) is defined as the hearing-aid sound output degree, but is not limited to this. The ratio of (mix) -Pd (org) to Pd (mix) may be used as the hearing-aid output level.

  In addition, here, the case where the hearing-aid sound output degree calculation unit 22b automatically determines the effective frequency band has been described. However, the present invention is not limited to this, and the measurer actually confirms the waveform spectrum to check the effective frequency band. May be determined manually.

  Returning to FIG. 3, the conversation protection degree calculation unit 22c will be described. The conversation protection level calculation unit 22c is a processing unit that calculates the conversation protection level from the hearing protection sound output level calculated by the hearing protection sound output level calculation unit 22b. Specifically, the conversation protection degree calculation unit 22c calculates the conversation protection degree by using a correlation expression obtained in advance as an expression representing the correlation between the conversation protection degree, the hearing-aid sound output degree, and the background noise level.

  Here, a correlation between the conversation protection level, the hearing-aid sound output level, and the background noise level will be described with reference to FIGS. 5 and 6. FIG. 5 is a diagram showing the measurement result of the hearing-proof sound output degree at the listening position 1 and the experimental result of the listening experiment. FIG. 6 is a diagram showing a result of averaging the correlation between the hearing-proof sound output level and the conversation protection level at each of the listening positions 1 to 3.

  5A shows the measurement result of the hearing-aid sound output degree at the listening position 1, and FIG. 5B shows the experimental result of the listening experiment at the listening position 1. FIG. Here, the vertical axis of the graph shown in FIG. 6A represents the hearing loss output degree (%), and the horizontal axis represents a value obtained by standardizing the volume of the hearing loss. In addition, the vertical axis of the graph shown in FIG. 5B represents the degree of conversation protection (%), and the horizontal axis represents a value obtained by standardizing the volume of the hearing protection sound.

  First, the outline of the listening experiment will be described. In the listening experiment, the speech data for the word intelligibility test is used, the words are randomly output from the test speaker 220 installed at the speaker position, and the measurer positioned at the listening position outputs the word output from the test speaker 220. Listening to Then, such a listening is performed at each noise level (here, 36 dB, 39 dB, 42 dB, 45 dB) and each listening position (here, listening positions 1 to 3), and the rate at which the measurer cannot hear the word correctly Are respectively calculated as conversation protection degrees.

  For example, the conversation protection degree “60%” means that when 10 words are output from the test speaker 220, the measurer can hear 4 words and cannot hear 6 words. It shows that there is. That is, the degree of conversation protection indicates the rate at which a person cannot hear a voice.

  Next, the measurement result of the hearing-aid sound output degree and the experimental result of the listening experiment will be described with reference to FIG. As shown in FIG. 5A, at the listening position 1, the hearing loss output and the hearing loss output level are in a proportional relationship, and it can be seen that the higher the hearing loss output, the higher the hearing loss output level. Similarly, as shown in FIG. 5B, it can be seen that the degree of conversation protection is proportional to the hearing-aid sound output, and the degree of conversation protection increases as the hearing-aid sound output increases.

  Thus, it can be seen that, in an environment where the background noise level is equal, the hearing protection sound output level and the conversation protection level are correlated, and the conversation protection level can be calculated using the background noise level and the hearing protection sound output level.

  In addition, the measurement of the hearing-aid sound output level and the listening experiment were also performed at the listening position 2 and the listening position 3. As a result, not only the typical speech protection area where the test sound does not reach directly at the front position of the hearing-proof sound speaker 120 as in the listening position 1 and the listening position 3, but also in the front as in the listening position 2. There was a correlation between the degree of hearing loss output and the degree of conversation protection even in the area where the test sound was received directly and the hearing loss was received from the side rather than from the front.

  Therefore, a general-purpose correlation expression independent of the listening position is obtained by obtaining these correlation expressions using the correlation between the conversation protection level, the hearing sound output level, and the background noise level at each of the listening positions 1 to 3. It becomes possible.

  Therefore, when the correlation between the hearing protection sound output level and the conversation protection level at each of the listening positions 1 to 3 was averaged, a result as shown in FIG. 6 was obtained. Specifically, as shown in the figure, as a result of linear approximation of the data for each background noise level, “y = 1.964x−38.613” (when the background noise level is 36 dB), “y = 1. 9917x-22.694 ”(when the background noise level is 39 dB),“ y = 2.209x−11.764 ”(when the background noise level is 42 dB) and“ y = 2.552x + 6.99084 ”(the background noise level is 45 dB). Two approximations were obtained. Note that “y” is the degree of conversation protection [%], and “x” is the degree of hearing loss output [%].

  Here, if it is considered that a slight difference between the inclinations of the four approximate expressions is an error, the average value “2” of the inclinations of the approximate expressions is set as the inclination of the correlation expression between the hearing-aid sound output degree and the conversation protection degree. be able to.

  Further, using the result shown in FIG. 6, the intercept “5n-216” of the correlation equation is obtained from the relationship between the background noise level and the conversation protection level at a predetermined hearing loss output level. “N” is the background noise level.

  As described above, the correlation equation “y = 2x + 5n−216” of the conversation protection level, the hearing-aid sound output level, and the background noise level is obtained. As described above, this correlation equation is a general-purpose correlation equation that does not depend on the listening position and is obtained by using the correlation between the degree of conversation protection, the degree of hearing protection sound output, and the background noise level at each of the listening positions 1 to 3. It is.

  The conversation protection degree calculation unit 22c calculates the conversation protection degree using the correlation equation. Specifically, when the conversation protection level calculation unit 22c receives the hearing protection sound output level from the hearing protection sound output level calculation unit 22b, the conversation protection level calculation unit 22c displays the received hearing protection sound output level and the background noise level measured by the measurement microphone 210. The degree of conversation protection is calculated by substituting into the correlation equation.

  In this way, the conversational protection level calculation unit 22c uses the correlation formula obtained in advance as a formula representing the correlation between the conversational protection level, the hearing-aid sound output level, and the sound pressure level of the background noise, and the hearing-aid output level calculation unit 22b. Since the conversation protection level is calculated from the hearing-aid sound output level and the background noise level calculated by the above, it is possible to appropriately obtain the conversation protection level that varies according to the background noise level.

  When the conversation protection degree calculation unit 22c calculates the conversation protection degree, the conversation protection degree calculation unit 22c displays the calculated conversation protection degree on the display unit 21. When the conversation protection level calculation unit 22c calculates the conversation protection level, the conversation protection level calculation unit 22c also performs a process of passing the calculated conversation protection level to the setting change instruction unit 22d.

  The setting change instruction unit 22d is a processing unit that issues a change instruction for adjusting the output level (volume) of the hearing loss sound to the optimum value to the setting change unit 12d of the hearing loss sound generator 1. For example, the setting change instruction unit 22d changes the setting so as to increase the setting value of the hearing loss volume when the conversation protection level received from the conversation protection level calculation unit 22c is lower than a predetermined range (for example, 40 to 60%). Instruct the unit 12d. Similarly, the setting change instruction unit 22d instructs the setting change unit 12d to decrease the set value of the hearing loss volume when the conversation protection level received from the conversation protection level calculation unit 22c is higher than a predetermined range.

  Note that the setting change instruction unit 22d specifically determines the hearing loss volume such that the conversation protection level is within a predetermined range from the hearing loss output level and the background noise level, and indicates that it should be changed to the determined hearing loss volume. You may make it instruct | indicate with respect to the setting change part 12d.

  As described above, when the conversation protection level received from the conversation protection level calculation unit 22c is outside the predetermined range, the setting change instruction unit 22d sets the output level of the hearing loss so that the conversation protection level is within the predetermined range. Since it is instructed to change the value, the labor required for adjusting the output level of the hearing-proof sound can be reduced.

  Next, specific operations of the hearing-aid generator 1 and the measuring device 2 according to the present embodiment will be described with reference to FIG. FIG. 7 is a sequence diagram showing a processing procedure executed between the hearing-proof sound generating device and the measuring device. In the figure, the processing procedure until the conversation protection level calculated by the conversation protection level calculation unit 22c is displayed on the display unit 21 is shown. However, after the conversation protection level is displayed, the setting change instruction unit 22d is displayed. However, you may instruct | indicate the change of a hearing loss volume with respect to the setting change part 12d.

  As shown in FIG. 7, when the measurement is started, the measurement apparatus 2 measures the background noise level (step S101). Specifically, in the measurement apparatus 2, background noise is collected by the measurement microphone 210, and the voice analysis unit 22a specifies a background noise level that is a noise level of the collected background noise.

  Subsequently, in the measuring apparatus 2, the control unit 22 causes a test sound to be output from the test speaker 220 (step S102). Thereby, the test speaker 220 outputs vowels for each gender at predetermined intervals as test voices.

  Subsequently, in the measurement apparatus 2, the test sound is collected by the measurement microphone 210, and the sound analysis unit 22a performs spectrum analysis on the collected test sound, and stores the analysis result as spectrum data of the original sound. 23 (step S103).

  Subsequently, in the measuring device 2, the control unit 22 instructs the hearing aid sound generator 1 to start outputting the hearing aid sound (step S104). Then, the hearing-aid generator 1 receives the hearing-aid output start instruction from the measuring device 2 and starts the hearing-aid output process (step S105).

  Subsequently, in the measuring apparatus 2, the control unit 22 causes a test sound to be output from the test speaker 220 (step S106). Thereby, the test sound similar to that in step S102 is output from the test speaker 220.

  Subsequently, in the hearing-proof sound generating device 1, the test sound is collected by the sound collecting microphone 110, and the sound analyzing unit 12a performs spectrum analysis of the collected test sound (step S107), and the hearing-proof sound generating unit 12b and The output control unit 12c generates a hearing loss sound based on the analysis result by the voice analysis unit 12a (step S108).

  In the hearing-aid generator 1, the output sound amplifying unit 11 prevents the hearing-aid sound at the hearing-aid volume instructed from the output control unit 12c and at the timing when the hearing-aid sound and the test sound are perceived as a synthesized sound. The sound is output from the listening speaker 120 (step S109).

  Subsequently, in the measurement apparatus 2, a synthesized sound of the test sound and the hearing aid sound is collected by the measurement microphone 210, and the sound analysis unit 22a performs spectrum analysis on the collected synthesized sound and mixes the analysis results. It memorize | stores in the memory | storage part 23 as spectrum data of a sound (step S110). And in the measuring apparatus 2, the hearing-aid sound output degree calculation part 22b performs a hearing-aid sound output degree calculation process (step S111). Such hearing-aid sound output degree calculation processing will be described with reference to FIG.

  Subsequently, in the measuring apparatus 2, the conversation protection degree calculation unit 22c uses the correlation equation “y = 2x + 5n−216” obtained in advance, and the hearing loss output degree calculated in step S111 and the darkness measured in step S101. The degree of conversation protection is calculated from the noise level (step S112). Then, in the measuring device 2, the conversation protection level calculation unit 22c displays the calculated conversation protection level on the display unit 21 (step S113), and the process ends.

  Next, the hearing-aid output level calculation process in step S111 will be described with reference to FIG. FIG. 8 is a flowchart showing the processing procedure of the hearing-aid sound output degree calculation processing.

  As shown in FIG. 8, when the hearing-aid sound output degree calculation process is started, the hearing-aid sound output degree calculation unit 22b first extracts each spectrum data of the mixed sound from the storage unit 23 (step S201), and the sound pressure level calculation process. Is performed (step S202). Such sound pressure level calculation processing will be described with reference to FIG.

  Subsequently, the hearing-aid sound output degree calculation unit 22b determines the representative value of each sound pressure level calculated by the sound pressure level calculation process as the sound pressure level Pd (mix) of the mixed sound (step S203). Here, the total value of each sound pressure level is used as a representative value.

  Subsequently, the hearing-aid sound output degree calculation unit 22b extracts each spectrum data of the original sound from the storage unit 23 (step S204), and performs a sound pressure level calculation process (step S205). The sound pressure level calculation process is the same as the sound pressure level calculation process in step S202.

  Subsequently, the hearing-aid sound output degree calculation unit 22b determines the representative value (here, the total value) of each sound pressure level calculated by the sound pressure level calculation process as the sound pressure level Pd (org) of the original sound (step). S206).

  Then, the hearing-aid output level calculation unit 22b substitutes the sound pressure level Pd (mix) of the mixed sound determined in step S203 and the sound pressure level Pd (org) of the original sound determined in step S206 into the equation (1). Thus, the hearing-aid sound output level is calculated (step S207), and the hearing-aid output level calculation process is terminated.

  Next, the sound pressure level calculation process shown in step S202 will be described with reference to FIG. FIG. 9 is a flowchart showing a processing procedure of sound pressure level calculation processing.

  As shown in FIG. 9, when the sound pressure level calculation process is started, the hearing-aid output level calculation unit 22b extracts spectrum data for one syllable from a total of ten syllables, and an effective frequency for the syllable. A band is determined (step S301). Specifically, the hearing-aid sound output degree calculation unit 22b determines a frequency band of 1000 Hz centering on a frequency at which the difference in sound pressure level between the mixed sound and the original sound is maximum as an effective frequency band.

  Subsequently, the hearing-aid sound output degree calculation unit 22b adds the sound pressure level at each frequency within the effective frequency band (step S302). Specifically, this processing is performed using Equation (4).

  Subsequently, the hearing-aid sound output degree calculation unit 22b determines whether or not the process of step S302 has been performed for all the times in the output time zone (step S303), and if not (step S303, No), The process of step S302 is repeated for the remaining time. Specifically, this process is performed using Expression (3).

  On the other hand, when it is determined that the process of step S302 has been performed for all the times in the output time zone (step S303, Yes), the hearing-aid output level calculator 22b performs the processes of steps S301 to S303 for all syllables. (Step S304, No), the process of Steps S301 to S303 is repeated for the remaining syllables. Specifically, this processing is performed using Expression (2).

  On the other hand, if it is determined that the processing of steps S301 to S303 has been performed for all syllables (step S304, Yes), the hearing-aid output level calculation unit 22b ends the sound pressure level calculation processing.

  As described above, in this embodiment, the hearing-aid generator generates the hearing-aid based on the predetermined sound collected by the sound-collecting microphone, and outputs the generated hearing-aid from the hearing-aid speaker. The sound analysis unit of the measuring device collects the mixed sound and the original sound. In this embodiment, the hearing-aid sound output degree calculation unit calculates the sound pressure levels of the mixed sound and the original sound, and outputs the hearing-aid sound based on the difference between the calculated sound pressure level of the mixed sound and the original sound. Calculate the degree.

  In this embodiment, the conversation protection degree calculation unit calculates the conversation protection degree from the hearing protection sound output degree calculated by the hearing protection sound output degree calculation unit using a correlation formula between the conversation protection degree and the hearing protection sound output degree. It was decided. Therefore, it is possible to appropriately evaluate the degree of conversation protection by the conversation security system without variation due to the senser's sense.

  By the way, in the above-described embodiment, the case where the sound recorded in advance as the test sound is output from the test speaker 220 has been described. However, the present invention is not limited to this, and the voice of a person (tester) is tested. It can also be used as voice. In such a case, it is difficult for the tester to utter the same voice twice at the same volume, so it is desirable to be able to calculate the degree of conversation protection only by the tester uttering the test voice once.

  Therefore, this point will be described below with reference to FIG. FIG. 10 is an explanatory diagram in the case of obtaining the hearing-aid output level using the tester's raw voice. In addition, (A) in the figure shows a state in which the hearing-aid generating device 1 and the measuring device 2 each collect the test sound uttered by the tester, and (B) in the same figure shows the original sound and the prevention sound. It shows how the mix sound is mechanically synthesized from the listening sound.

  As shown in FIG. 10A, the tester utters a test voice (for example, “a, e, i, o, u”) at the speaker position shown in FIG. (See A-1)). Subsequently, the hearing-aid generator 1 generates a hearing-aid based on the test sound as in the above-described embodiment (see (A-2) in the figure). However, here, the hearing-aid generator 1 temporarily stores the generated hearing-aid without outputting it. Further, in the measuring apparatus 2, the test sound collected by the measurement microphone 210 is stored as the original sound, as in the above-described embodiment (see (A-3) in the figure).

  Subsequently, as shown in FIG. 5B, the hearing-aid generator 1 outputs the temporarily stored hearing-aid from the hearing-aid speaker 120 (see (B-1) in the figure). ), The measuring device 2 collects the hearing-aid sound using the measuring microphone 210 (see (B-2) in the figure).

  Then, the measuring device 2 generates a mixed sound by synthesizing the original sound stored in (A-3) of FIG. 5 and the hearing-proof sound collected by the measurement microphone 210 ((B-3) of FIG. 5). reference). The subsequent steps are the same as in the above-described embodiment, and the hearing protection sound output level can be calculated from the mixed sound and the original sound, and the conversation protection level can be calculated from the calculated hearing loss sound output level.

  In this way, the hearing-proof sound generating device collects the predetermined sound uttered by the tester using the sound collecting microphone, generates the hearing sound corresponding to the collected sound, and outputs it from the predetermined speaker. The sound analysis unit 22a of the measuring device 2 collects the predetermined sound uttered by the tester and the hearing-aid sound output by the hearing-aid generation device, and the hearing-aid sound output degree calculation unit calculates the predetermined sound collected. The sound is an original sound, and the original sound is synthesized with the hearing-proof sound collected by the measurement microphone 210 to generate a mixed sound.

  Therefore, it is possible to obtain the degree of conversation protection using the voice uttered by the tester on the spot without preparing the voice recorded in advance as the test voice.

  So far, the case where the degree of conversation protection is evaluated in the measurement environment has been described. However, the present invention is not limited to this, and the degree of conversation protection can be increased even in the actual environment after the introduction of the conversation security system. Further, it is possible to dynamically change the hearing-aid volume according to the evaluation result.

  Hereinafter, this point will be described with reference to FIG. FIG. 11 is an explanatory diagram in the case of adjusting the volume of the hearing loss sound using the conversation protection degree obtained in the actual environment. Note that (A) in the figure shows a state in which the transfer function H of the conversation protection area by the conversation security system is calculated by prior measurement. Further, (B) in the figure shows a state in which the degree of hearing protection sound is adjusted by calculating the degree of conversation protection from the conversation voice in an actual environment.

  As shown in FIG. 11A, the measurement apparatus 2 first outputs a test sound for measuring transfer characteristics from the test speaker 220 (see (A-1) in FIG. 11). Subsequently, in the hearing-proof sound generating device 1, the sound analysis unit 12a performs spectrum analysis of the test sound collected by the sound collecting microphone 110, and transmits the spectrum data a, which is the analysis result, to the measuring device 2 (the same figure). (See (A-2)). Further, in the measurement apparatus 2, the voice analysis unit 22a performs spectrum analysis of the test voice collected by the measurement microphone 210 and stores the spectrum data b that is an analysis result.

  Then, in the measuring apparatus 2, the voice analysis unit 22a calculates the transfer function H between the speaker position and the listening position using the spectrum data a and the spectrum data b (see (A-3) in the figure).

  Subsequently, as shown in FIG. 5B, when a speaker utters in an actual environment, first, the hearing-aid generator 1 collects conversational speech by the sound collecting microphone 110 (FIG. 5). (See (B-1)). In the hearing-proof sound generating apparatus 1, the voice analysis unit 12 a performs spectrum analysis of the collected conversational voice and transmits spectrum data as an analysis result to the measurement apparatus 2. Then, the measuring device 2 generates the original sound at the listening position by multiplying the spectrum data received from the hearing-proof sound generating device 1 by the transfer function H (see (B-2) in the figure).

  In addition, the hearing-aid generator 1 generates a hearing-aid based on the collected conversational sound as a provisional hearing-aid (see (B-3) in the figure), and the spectrum data of the generated provisional hearing-aid is measured by the measuring device 2. Send to. And in the measuring apparatus 2, the audio | voice analysis part 22a produces | generates a mix sound by synthesize | combining the original sound produced | generated using the transfer function H, and provisional hearing loss (refer (B-4) of the figure).

  Subsequently, the measuring apparatus 2 uses the original sound shown in (B-2) of FIG. 5 and the mixed sound shown in (B-4) of FIG. The listening output level is calculated and the conversation protection level is calculated (see (B-5) in the figure). In the measuring device 2, the setting change instruction unit 22d instructs the hearing-aid sound generator 1 to change the hearing-aid volume according to the calculated degree of conversation protection (see (B-6) in the figure).

  In the hearing-proof sound generating device 1, the setting changing unit 12d adjusts the hearing-aid volume in accordance with an instruction from the setting change instruction unit 22d, and the output sound amplifying unit 11 confirms the hearing-proof sound whose volume has been adjusted. It outputs from the hearing-proof sound speaker 120 as an audible sound (refer (B-7) of the figure).

  Thus, by obtaining the transfer function H in advance, the original sound and the mixed sound at the listening position can be predicted. Thus, the conversation protection degree can be calculated before the timing for outputting the hearing protection sound, and the hearing protection sound can be output after adjusting the volume according to the conversation protection degree. That is, the degree of conversation protection can be evaluated even in an actual environment after the conversation security system is actually introduced, and the hearing loss volume can be dynamically changed according to the evaluation result.

  As described above, the conversation protection degree evaluation system and the conversation protection degree evaluation method according to the present invention are useful when it is desired to appropriately evaluate the conversation protection degree by the conversation security system without causing variation due to the sense of the measurer, In particular, it is suitable for evaluating the degree of conversation protection by a conversation security system that generates and outputs hearing-proof sound based on conversational sound.

DESCRIPTION OF SYMBOLS 1 Deaf sound generator 11 Output sound amplification part 12 Control part 12a Sound analysis part 12b Deaf sound generation part 12c Output control part 12d Setting change part 13 Memory | storage part 13a Volume setting value 110 Sound collection microphone 120a, 120b Hearing sound speaker 2 Measuring apparatus 21 Display unit 22 Control unit 22a Audio analysis unit 22b Deafness sound output level calculation unit 22c Conversation protection level calculation unit 22d Setting change instruction unit 23 Storage unit 23a Spectrum data 210a to 210c Measurement microphone 220 Test speaker

Claims (7)

Hearing-aid sound output means for generating a hearing-aid sound that makes it difficult to hear the predetermined sound based on the predetermined sound collected by the sound collecting unit, and outputting the generated hearing-aid sound from a predetermined speaker;
Sound collecting means for collecting a mixed sound that is a sound in which the hearing-proof sound is superimposed on the predetermined sound or an original sound that is a sound when only the predetermined sound is generated;
While calculating the sound pressure level of the mixed sound and the original sound collected by the sound collecting means, respectively, the degree of deafness sound output is calculated based on the difference between the calculated sound pressure level of the mixed sound and the original sound. An output degree calculating means for calculating;
Using a correlation formula obtained in advance as a formula representing the correlation between the degree of conversation protection indicating the rate at which a person cannot hear speech and the level of hearing-aid sound output, the hearing-aid sound output level calculated by the output level calculation means A conversation protection degree evaluation system comprising: a protection degree calculation means for calculating a conversation protection degree. The output degree calculation means includes
Based on the waveform spectrum of the mixed sound and the waveform spectrum of the original sound, the frequency band in which the hearing protection sound is effectively output is specified, and the representative value of the sound pressure level of the mixed sound in the specified frequency band is The conversation protection according to claim 1, wherein the sound pressure level of the mixed sound is calculated and a representative value of the sound pressure level of the original sound in the specified frequency band is calculated as the sound pressure level of the original sound. Degree evaluation system. A test speaker that outputs pre-recorded gender-specific vowels as the predetermined voice;
The output degree calculation means includes
The representative value of the sound pressure level for each vowel in the mixed sound is calculated as the sound pressure level of the mixed sound, and the representative value of the sound pressure level for each vowel in the original sound is used as the sound pressure level of the original sound. The conversation protection degree evaluation system according to claim 1, wherein the conversation protection degree evaluation system is calculated. The sound collecting means collects the surrounding sound as background noise in a state where the predetermined sound and the hearing-proof sound are not generated,
The degree of protection calculating means
Using the correlation formula according to the sound pressure level of the background noise, the conversation protection degree is calculated from the hearing-aid sound output degree calculated by the output degree calculation means and the noise level of the background noise collected by the sound collection means. The conversation protection degree evaluation system according to claim 1, wherein the conversation protection degree evaluation system is calculated. A setting changing means for changing a setting value of the output level of the hearing-proof sound so that the conversation protection level is within a predetermined range when the conversation protection level calculated by the protection level calculation unit is out of a predetermined range; The conversation protection degree evaluation system according to claim 1, wherein the conversation protection degree evaluation system is provided. The hearing-proof sound output means includes
Collecting a predetermined voice uttered by the tester using the sound collection unit, generating a hearing-proof sound corresponding to the collected voice and outputting it from the predetermined speaker,
The sound collecting means includes
Collecting the predetermined sound spoken by the tester and the hearing-aid sound output by the hearing-aid sound output means,
The output degree calculation means includes
The mixed sound is generated by combining the predetermined sound collected by the sound collecting means as an original sound and synthesizing the original sound with the hearing-proof sound collected by the sound collecting means. The conversation protection degree evaluation system according to any one of 1 to 5. A first sound collecting step for collecting a predetermined sound as an original sound;
A hearing-proof sound output step of generating a hearing-aid sound that makes it difficult to hear the predetermined sound based on the predetermined sound collected by the sound collecting unit and outputting it from a predetermined speaker;
A second sound collecting step of collecting a mixed sound that is a sound obtained by superimposing the hearing-proof sound on the predetermined sound;
The sound pressure level of the original sound collected in the first sound collecting step and the sound pressure level of the mixed sound collected in the second sound collecting step are calculated, respectively, and the calculated mixed sound and An output level calculation step of calculating a hearing loss output level based on a difference in sound pressure level of the original sound;
Using a correlation formula obtained in advance as a formula representing the correlation between the degree of conversation protection indicating the rate at which a person cannot hear speech and the degree of hearing protection sound output, the conversation is determined from the degree of hearing loss output calculated in the output level calculation step. A method for evaluating a degree of protection of conversation, comprising a step of calculating a degree of protection for calculating a degree of protection.

Images