JP7162470B2 - CONVERSATION SOUND LEVEL NOTIFICATION SYSTEM AND CONVERSATION SOUND LEVEL NOTIFICATION METHOD - Google Patents

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a conversation voice level notification system and a conversation voice level notification method for notifying a speaker of a conversation voice level that interferes with the conversation of another person.

In an office space where a large number of workers work, in addition to noise generated by air conditioners and OA (Office Automation) equipment, there are irrelevant conversational voices from other people's telephone calls and meetings. For this reason, the worker cannot concentrate on work due to the above-described conversational voice, and the efficiency of work may decrease. As a result, it becomes a factor that lowers productivity. Also, when the worker himself/herself is having a conversation on the phone or in a meeting, irrelevant voices of other workers from a nearby seat or an adjacent meeting space become a hindrance to the above-mentioned worker's own conversation.

When work is interrupted by irrelevant voices, workers may spontaneously utter loud voices unconsciously in an attempt to continue their own conversations. In addition, due to the interaction of each of a plurality of workers, there are cases where the voices of each other's conversational voices gradually increase unconsciously. As a result, many workers try to talk louder than other workers, which leads to noise in the entire office space.

Therefore, the inventor in Cited Document 1 (Patent No. 6296291) installed a voice sensor in the worker's seat or meeting space in the office space, and the level of the conversation voice measured by the voice sensor We proposed a system and method for notifying the user when it reached a level that inhibited
Japanese Patent No. 6296291

In the conventional technique described in Cited Document 1, the analyzer included in the voice sensor performs octave band filtering with a center frequency of 250 Hz or 500 Hz on the output of the microphone, or filters the characteristics simulating human hearing characteristics. It is audio data. This is a device for extracting only voice from various sounds and measuring its sound pressure level with high accuracy.

However, the above-described simple filtering alone is not necessarily sufficient to extract speech with high accuracy. Sounds other than speech are mixed in with the speech data after the above-described filtering, resulting in erroneous notification by the system. There was a problem that

The present invention solves the above-mentioned problems, and the conversation voice level notification system according to the present invention is provided in each area, which is a worker's seat or a meeting space where workers have a meeting. A-weighted equivalent sound pressure level and an octave band equivalent sound pressure level with a center frequency of 500 Hz are calculated based on the seat detection sensor, the sound sensor arranged for each area, and the sound collected by the sound sensor. If the difference obtained by subtracting the A-weighted equivalent sound pressure level from the octave band equivalent sound pressure level with a center frequency of 500 Hz is equal to or greater than a preset threshold value, the sound collected by the sound sensor does not include voice. a determination unit that determines that the worker is present in the area, and detects whether or not the worker is present in the area based on the output of the presence detection sensor when it is determined that the presence of the worker is included in the determination by the determination unit. a seat detection unit; a voice input unit for inputting voice data of a worker from the voice sensor placed in the region; a speech prediction unit that predicts the speech intensity in another area where another worker is present when the worker is seated based on the distance between the said area and the other area; a voice determination unit that determines whether or not a preset threshold value is exceeded; and a notification device provided for each of the regions, if any of the voice intensities exceeds the preset threshold value, detects the voice intensity. and a voice notification unit for notifying the level of and/or suppression of the voice intensity.

Further, in the conversation voice level notification system according to the present invention, the notification device is a personal computer connected to the conversation voice level notification system, and the voice notification unit displays the voice strength on the display screen of the personal computer. and/or suppression of the voice intensity. Further, in the conversation voice level notification system according to the present invention, the threshold is provided in a plurality of stages, and the voice intensity is set to each stage. It is characterized by performing a different notification each time it exceeds.

Further, in the conversation voice level notification method according to the present invention, the voice input unit is based on the sound collected by the voice sensor arranged in the area that is the worker's seat or the meeting space where the worker has a meeting . Calculate the A-weighted equivalent sound pressure level and the octave band equivalent sound pressure level with a center frequency of 500 Hz, and subtract the A-weighted equivalent sound pressure level from the octave band equivalent sound pressure level with a center frequency of 500 Hz. In the case above, a determination process for determining that the sound collected by the audio sensor includes voice, a presence detection process for detecting presence/absence of the worker in the area; a voice input process for inputting voice data of the worker from the voice sensor; Based on the sound data in each of the areas where the workers are seated, the sound intensity in other areas where other workers are present when the sound propagates is determined based on the distance between the area and the other areas an audio prediction process for predicting; an audio determination process for determining whether any of the audio intensities exceeds a preset threshold; and a voice notification process of notifying the level of the voice intensity and/or suppression of the voice intensity by a notification device provided for each region when a preset threshold value is exceeded.

A conversational sound level notification system and a conversational sound level notification method according to the present invention calculate an A-weighted equivalent sound pressure level and an octave band equivalent sound pressure level with a center frequency of 500 Hz based on sounds collected by a sound sensor. If the difference obtained by subtracting the A-weighted equivalent sound pressure level from the octave band equivalent sound pressure level with a center frequency of 500 Hz is equal to or greater than a preset threshold value, the sound collected by the sound sensor includes voice. I judge. According to the conversation voice level notification system and the conversation voice level notification method according to the present invention, by accurately determining whether or not the sound collected by the voice sensor includes human voice, the system can It is possible to prevent incorrect notifications from being made.

1 is a block diagram showing the configuration of a conversation voice level notification system 1 according to this embodiment; FIG. 3 is a diagram showing a configuration example of a voice/presence data table of a voice/presence data storage unit 116. FIG. FIG. 10 is a diagram showing a configuration example of a distance table showing distances from a seat S_1 to a seat S_2 to a seat S_9 indicated by area identification numbers, stored in a distance storage unit 118; FIG. 10 is a diagram showing a configuration example of a propagation prediction data table showing predicted sound levels of sounds generated by workers at each seat at other seats and evaluation results of the predicted sound levels; 4 is a flowchart showing an operation example of suppressing conversational voice by the conversational voice level notification system 1; FIG. 10 is a diagram showing frequency characteristics of sound pressure levels according to the presence or absence of conversation in a meeting space; FIG. 10 is a diagram showing a frequency distribution of L _diff according to the presence or absence of conversation in a meeting space; FIG. 5 is a diagram showing the relationship between the set value of the threshold _LT and the determination accuracy;

An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a conversational voice level notification system 1 according to this embodiment. In the present embodiment, a configuration for controlling the voice levels of workers from seats S_1 to S_9, which are areas arranged in an office space, will be described as an example. be able to.

The conversation voice level notification system 1 includes a conversation voice level notification device 11, notification devices 21 to 29 arranged at seats S_1 to S_9, and presence detection sensors 31 to 39. , voice sensors 41 to 49 .

Each of audio sensors 41 to 49 includes, for example, a microphone and an analysis device. A microphone has a function of collecting sound in the vicinity where the microphone is installed and converting it into an electric signal. Electrical signals from such microphones are input to an analyzer.

Collected sound data collected by a microphone may include data related to sounds other than human voice. There is a problem that the system gives an erroneous notification when sounds other than voice are mixed in such collected sound data. Therefore, the analysis device performs a process of determining with high accuracy whether or not the sound collected by the microphone includes human speech.

Specifically, the analyzer measures the A-weighted equivalent sound pressure level (L _Aeq ) in a preset time interval width (time slot), that is, the measurement time, and the octave band equivalent sound pressure level with a center frequency of 500 Hz. (also referred to as L _500eq ), it is determined whether or not the measured sound contains speech.

More specifically, the analyzer measures the A-weighted equivalent sound pressure level (L _Aeq ) in a set time interval band and the octave band equivalent sound with a center frequency of 500 Hz based on the sound collected by the microphone. A calculation unit that calculates the pressure level (L _500eq ), and the difference obtained by subtracting the A-weighted equivalent sound pressure level (L _Aeq ) from the octave band equivalent sound pressure level (L _500eq ) with a center frequency of 500 Hz is equal to or greater than a preset threshold , a determination unit that determines that the sound collected by the microphone includes voice. In this specification, the "determination section" and the "audio determination section" have different configurations.

The calculation unit calculates the equivalent sound pressure level of the specified time interval width (time slot) for the A-weighted equivalent sound pressure level (L _Aeq ) and the octave band equivalent sound pressure level (L _500eq ) with a center frequency of 500 Hz. have a function.

The audio sensors 41 to 49 output either the calculated A-weighted equivalent sound pressure level (L _Aeq ) or the octave band equivalent sound pressure level (L _500eq ) with a center frequency of 500 Hz as the sound level. Also, the sound levels output from each of the sound sensors 41 to 49 are transmitted to the conversation sound level notification device 11 . In the present embodiment, either the A-weighted equivalent sound pressure level (L _Aeq ) or the octave band equivalent sound pressure level (L _500eq ) with a center frequency of 500 Hz is defined as the sound level (L), and this sound level (L) is used in formula (2) described later.

The measurement time of the equivalent sound pressure level in the calculator is set according to the required frequency of voice detection. When the measurement time of the equivalent sound pressure level is set long, it may be determined that the sound is not included when the sound is included only in a short time within the measurement time.

The transmission of the sound data collected by the microphone, the calculation of the predicted sound level at other seats using it, and the determination of whether or not the value exceeds the set threshold are performed each time the equivalent sound pressure level is measured. . According to Patent Document 1, it is described that the sound level data is transmitted every one to several seconds. Data transmission and processing loads are small even when the number of points increases.

A sound pressure level meter (noise meter) with an octave analysis function may be used for the microphone and analysis device included in each of the audio sensors (41 to 49), or a MEMS microphone and at least A characteristics and A wireless sensor or the like having an equivalent sound pressure level analysis function of 500 Hz octave band may be used. It is also possible to use a sound pressure level meter application for smartphones and tablet terminals.

A notification device 21, a seat presence detection sensor 31, and a voice sensor 41 are provided at the seat S_1. The notification device 21 notifies the worker sitting at the seat S_1 of the level of the conversational voice supplied from the conversational voice level notification device 11 . The presence detection sensor 31 detects whether or not a worker is present at the seat S_1 and outputs the detection result to the conversation voice level notification device 11 . The voice sensor 41 detects the voice level at the seat S_1, and outputs the detection result of the voice level and the determination result of the determination unit regarding whether the sound collected by the microphone contains voice to the conversation voice level notification device. 11.

In addition, a notification device 22, a seat detection sensor 32, and a voice sensor 42 are provided at the seat S_2. The notification device 22 notifies the worker sitting at the seat S_2 of the level of the conversational voice supplied from the conversational voice level notification device 11 . The presence detection sensor 32 detects whether or not an operator is present at the seat S_2 and outputs the detection result to the conversation voice level notification device 11 . The voice sensor 42 detects the voice level at the seat S_2, and outputs the detection result of the voice level and the determination result of the determination unit regarding whether the sound collected by the microphone contains voice to the conversation voice level notification device. 11.

A notification device 23, a seat detection sensor 33, and a voice sensor 43 are provided at the seat S_3. The notification device 23 notifies the worker sitting at the seat S_3 of the level of the conversational voice supplied from the conversational voice level notification device 11 . The presence detection sensor 33 detects whether or not an operator is present at the seat S_3 and outputs the detection result to the conversation voice level notification device 11 . The voice sensor 43 detects the voice level at the seat S_3, and outputs the detection result of the voice level and the determination result of the determination unit regarding whether the sound collected by the microphone contains voice to the conversation voice level notification device. 11.

A notification device 24, a seat presence detection sensor 34, and a voice sensor 44 are provided at the seat S_4. The notification device 24 notifies the worker sitting at the seat S_4 of the level of the conversational voice supplied from the conversational voice level notification device 11 . The presence detection sensor 34 detects whether or not an operator is present at the seat S_4 and outputs the detection result to the conversation voice level notification device 11 . The voice sensor 44 detects the voice level at the seat S_4, and outputs the detection result of the voice level and the determination result of the determination unit regarding whether the sound collected by the microphone contains voice to the conversation voice level notification device. 11.

A notification device 25, a seat presence detection sensor 35, and a voice sensor 45 are provided at the seat S_5. The notification device 25 notifies the worker sitting at the seat S_5 of the level of the conversational voice supplied from the conversational voice level notification device 11 . The presence detection sensor 35 detects whether or not an operator is present at the seat S_5 and outputs the detection result to the conversation voice level notification device 11 . The voice sensor 45 detects the voice level at the seat S_5, and outputs the detection result of the voice level and the determination result of the determination unit regarding whether the sound collected by the microphone contains voice to the conversation voice level notification device. 11.

A notification device 26, a seat presence detection sensor 36, and a voice sensor 46 are provided at the seat S_6. The notifying device 26 notifies the worker sitting at the seat S_6 of the level of his/her own conversational voice supplied from the conversational voice level notifying device 11 . The presence detection sensor 36 detects whether or not an operator is present at the seat S_6 and outputs the detection result to the conversation voice level notification device 11 . The voice sensor 46 detects the voice level at the seat S_6, and outputs the voice level detection result and the determination result of the determination unit regarding whether the sound collected by the microphone contains voice to the conversation voice level notification device. 11.

A notification device 27, a seat detection sensor 37, and a voice sensor 47 are provided at the seat S_7. The notification device 27 notifies the worker sitting at the seat S_7 of the level of the conversational voice supplied from the conversational voice level notification device 11 . The presence detection sensor 37 detects whether or not an operator is present at the seat S_7 and outputs the detection result to the conversation voice level notification device 11 . The voice sensor 47 detects the voice level at the seat S_7, and outputs the voice level detection result and the determination result of the determination unit regarding whether the sound collected by the microphone contains voice to the conversation voice level notification device. 11.

A notification device 28, a seat presence detection sensor 38, and a voice sensor 48 are provided at the seat S_8. The notification device 28 notifies the worker sitting at the seat S_8 of the level of the conversational voice supplied from the conversational voice level notification device 11 . The presence detection sensor 38 detects whether or not an operator is present at the seat S_8 and outputs the detection result to the conversation voice level notification device 11 . The voice sensor 48 detects the voice level at the seat S_8, and outputs the detection result of the voice level and the determination result of the determination unit regarding whether the sound collected by the microphone contains voice to the conversation voice level notification device. 11.

A notification device 29, a seat detection sensor 39, and a voice sensor 49 are provided at the seat S_9. The notification device 29 notifies the worker sitting at the seat S_9 of the level of the conversational voice supplied from the conversational voice level notification device 11 . The presence detection sensor 39 detects whether or not an operator is present at the seat S_9 and outputs the detection result to the conversation voice level notification device 11 . The voice sensor 49 detects the voice level at the seat S_9, and outputs the detection result of the voice level and the determination result of the determination unit regarding whether the sound collected by the microphone contains voice to the conversation voice level notification device. 11.

The conversation voice level notification device 11 includes a voice input unit 111, a presence detection unit 112, a voice prediction unit 113, a voice determination unit 114, a voice notification unit 115, a voice/seat data storage unit 116, a propagation prediction data storage unit 117, and Each distance storage unit 118 is provided.

The voice input unit 111 receives the voice level information supplied from each of the voice sensors 41 to 49 and the determination information regarding whether voice is included, and inputs the voice/seat data in the voice/seat data storage unit 116 . Write it to the table and store it.

The presence detection unit 112 stores the information on whether or not the seat is present, which is supplied from each of the presence detection sensor 31 and the presence detection sensor 39, in the voice/presence data table of the voice/presence data storage unit 116. be written to and memorized.

FIG. 2 is a diagram showing a configuration example of the voice/presence data table of the voice/presence data storage unit 116. As shown in FIG. In FIG. 2, area identification information for identifying each seat, voice level data (dB) of the conversational voice at the seat indicated by this area identification information, and determination as to whether the sound collected by the microphone contains voice Information is stored in association with presence/absence information indicating whether or not there is a worker at the seat indicated by the area identification information. The audio level data is the level of conversational audio detected by the audio sensors (audio sensors 41 to 49) provided in the corresponding area. The determination information is information relating to whether or not a human voice is included, and is indicated as "included" when included, and indicated as "not included" when not included. The presence/absence information is information (presence/absence information) of a worker detected by the presence detection sensors (presence detection sensors 31 to 39) provided in the corresponding area. "Present" is indicated when there is an operator, and "Absent" is indicated when there is no operator.

Returning to FIG. 1, the voice input unit 111 sequentially inputs voice level data indicating the strength level (speech intensity) of the conversational voice supplied from each of the voice sensors 41 to 49 at a predetermined cycle, and It is written and stored in the voice level data column of the voice/seat data table in the voice/seat data storage unit 116 corresponding to the area identification number of the seat where the sensor is arranged.

The occupancy detection unit 112 stores the occupancy information, which is supplied from each of the occupancy detection sensors 31 and 39 and indicates whether or not the occupancy is present, in the voice/occupancy data storage unit 116 . It is written and stored in the presence/absence information column of the presence data table.

The speech prediction unit 113 obtains a prediction of how strong the conversational speech at each seat has speech level data at other seats using the following equations (1) and (2).

In equation (1), L _p is the sound pressure level (dB) at the prediction point, L _w is the power level (dB) of the conversational voice, and r is the location (region) of the speaker (worker). is the distance (m) to the predicted point (area where other workers exist), R is the room constant (R = Sα/1-α)), and S is the surface area of the room (office space) (m2) and α is the average sound absorption coefficient of the room (office space). Here, α is denoted by α bar (− above α) in the formula. In the equation (2), L is the measured speech level (sound pressure level, dB), and d is the distance (m) from the speaker (worker) to the speech sensor (speech sensor). be.

図３は、距離記憶部１１８に記憶されている、領域識別番号の示す席Ｓ＿１から席Ｓ＿２から席Ｓ＿９までの距離を示す距離テーブルの構成例を示す図である。図３は、席Ｓ＿１を基準とした距離テーブルであるが、同様に、席Ｓ＿２から席Ｓ＿９の各々を基準とした距離テーブルがある。したがって、距離記憶部１１８には、８個の距離テーブルが予め書き込まれて記憶されている。図３において予測点までの距離が、席Ｓ＿１から音声レベルデータを予測する他の席までの距離（ｒ）を示している。また、距離記憶部１１８には、室（オフィス空間）の表面積Ｓ（ｍ²）及び室の平均吸音率α（αの上部にバー）とが予め書き込まれて記憶されている。また、距離記憶部１１８には、作業者（話者）から音声センサまでの距離ｄ（ｍ）が記憶されている。

FIG. 3 is a diagram showing a configuration example of a distance table showing the distances from the seat S_1 to the seat S_2 to the seat S_9 indicated by the area identification numbers, stored in the distance storage unit 118. As shown in FIG. Although FIG. 3 is a distance table based on seat S_1, there are similar distance tables based on each of seats S_2 to S_9. Therefore, eight distance tables are written in advance and stored in the distance storage unit 118 . In FIG. 3, the distance to the prediction point indicates the distance (r) from the seat S_1 to another seat whose voice level data is to be predicted. Further, the distance storage unit 118 pre-writes and stores the surface area S (m ² ) of the room (office space) and the average sound absorption coefficient α of the room (a bar above α). Further, the distance storage unit 118 stores the distance d (m) from the operator (speaker) to the voice sensor.

Returning to FIG. 1, the speech prediction unit 113 obtains the strength level L _W of the conversational speech using equation (2). At this time, the voice prediction unit 113 substitutes the voice level data L read from the voice/seat data storage unit 116 and the distance d read from the distance table into the equation (2), and calculates the strength level L of the conversation voice. ask for _W.

Further, the speech prediction unit 113 reads the surface area S and the average sound absorption coefficient α from the distance storage unit 118, and substitutes the read surface area S and the average sound absorption coefficient α into the formula “R=Sα/(1−α)”. Calculate the room constant R by

Then, the speech prediction unit 113 substitutes the calculated intensity level L _W and the room constant R, and the distance r (m) read from the distance table of the distance storage unit 118 into the equation (1) to obtain the result of propagation prediction. , the predicted voice level _Lp of the conversation voice at the other seat is calculated.

FIG. 4 is a diagram showing a configuration example of a propagation prediction data table showing the predicted sound level of the sound generated by the worker at each seat at another seat and the evaluation result of this predicted sound level. This propagation prediction data table is written and stored in the propagation prediction data storage unit 117 .

In the propagation prediction data table of FIG. 4, the region identification number, the prediction voice level, and the voice level comparison result are associated and stored. In the predicted voice level column, the highest predicted voice level among the predicted voice levels obtained by predicting the voice level when the conversation voice at the seat corresponding to the area identification number propagates to other seats is extracted and written. there is For example, in the predicted voice level column for seat S_1 in the propagation prediction data table of FIG. An expected audio level is written. Similarly, the predicted voice level column of each of the seats S_2 to S_9 in the propagation prediction data table in FIG. 4 is written with the highest predicted voice level among the predicted voice levels of the other seats. In addition, in the voice level comparison result column, when the predicted voice level of another seat described in the predicted voice level column is equal to or higher than a preset threshold value, a "warning level" is stored. If the predicted voice level of the other seat described in the column of predicted voice level is less than the threshold, "non-warning level" is stored.

Therefore, when "warning level" is stored, it indicates that the voice level of the conversation voice at the seat indicated by the written area identification number has an adverse effect on the other seats. On the other hand, if the "non-warning level" is stored, it indicates that the voice level of the conversation voice at the seat indicated by the written area identification number does not adversely affect the other seats. there is The threshold value for evaluating the predicted voice level is set to a voice level extracted by the majority of workers as having no influence through experiments using a plurality of workers. In other words, if this threshold is exceeded, there will be workers who feel that it will have an adverse effect.

Returning to FIG. 1, the voice prediction unit 113 calculates the predicted voice level Lp of the conversation voice at another seat, and outputs the calculated predicted voice levels to the voice determination unit 114 in the order of the area identification numbers.

Here, when calculating the predicted voice level, the voice prediction unit 113 refers to the voice/seat data table of the voice/seat data storage unit 116 to determine whether the worker is present in two or more areas. Only the predicted voice level of the conversation voice for the seats in the areas where the two people are seated may be calculated.

The voice determination unit 114 extracts the highest predicted voice level among the predicted voice levels at other seats obtained from the respective conversation voices of each seat supplied from the voice prediction unit 113 for each seat of the region identification number. . Then, for each area identification number seat, the voice determination unit 114 stores the highest predicted voice level at the other seat obtained from the conversation voice of each seat in the predicted voice level column of the propagation prediction table of the propagation prediction data storage unit 117 . be written to and memorized.

Further, the voice determination unit 114 determines whether or not the predicted voice level of each of the other seats written in the predicted voice level is equal to or higher than a preset threshold value. Then, when the measured voice level of each of the other seats is equal to or higher than a preset threshold value, the voice determination unit 114 stores "warning level ” to be memorized. On the other hand, when the predicted sound level is less than the preset threshold value, the sound determination unit 114 writes “non-warning level” to the sound level comparison result column of the propagation prediction table of the propagation prediction data storage unit 117 . Memorize.

Further, when extracting the highest predicted voice level among the predicted voice levels at other seats obtained from the conversation voices at each seat, the voice determination unit 114 extracts the voice/presence data stored in the voice/seat data storage unit 116 . The seat data table may be referred to, and the highest predicted voice level among the predicted voice levels of the seat where the user is present may be extracted.

That is, if there is no worker sitting at another seat (if absent), there is no need to consider the sound level for that seat.

The voice notification unit 115 refers to the propagation prediction table of the propagation prediction data storage unit 117 and extracts the area identification number in which the "warning level" is written. Then, the voice notification unit 115 notifies the notification device corresponding to the seat of the extracted area identification number of information indicating that the voice level of the conversation voice should be suppressed.

For example, when the notification device is a display device, the voice notification unit 115 displays a message such as "Please lower the conversation volume a little more, as it may disturb other people", and compares the predicted voice level and the threshold. It is configured to be displayed side by side and to notify the suppression of speech in conversation. In the case of a display device, a configuration using a display device connected to a personal computer placed at a seat may be used.

Also, in the present embodiment, one threshold has been described, but a plurality of thresholds to be compared with the predicted speech level, for example, caution thresholds and warning thresholds, may be set. The caution threshold is set below the warning threshold. The speech determination unit 114 extracts a predicted speech level equal to or higher than the caution threshold and less than the warning threshold from the predicted speech level supplied from the speech prediction unit 113, and extracts the lowest predicted speech level among the extracted predicted speech levels. do.

In this case, a caution level column is newly set in the propagation prediction table of the propagation prediction data storage unit 117 . Then, in the propagation prediction table of the propagation prediction data storage unit 117, the voice determination unit 114 writes “caution level” to the voice level comparison result (caution level) of the area identification number corresponding to the seat having the caution level. memorize with

Further, the voice determination unit 114 extracts a predicted voice level equal to or higher than the alarm threshold from the predicted voice levels supplied from the voice prediction unit 113, and extracts the lowest predicted voice level among the extracted predicted voice levels.

In the propagation prediction table of the propagation prediction data storage unit 117, the sound determination unit 114 writes and stores the "warning level" for the sound level comparison result (warning level) of the area identification number corresponding to the seat with the warning level. Let In addition, when the predicted sound level supplied from the sound prediction unit 113 is less than the caution threshold, the sound determination unit 114 determines " "Non-Caution Level" and "Non-Warning Level" are written and stored. Further, the voice determination unit 114 writes “non-caution level” to the voice level comparison result (caution level) in the area identification number indicated by the propagation prediction table “warning level” of the propagation prediction data storage unit 117 .

The voice notification unit 115 refers to the propagation prediction table of the propagation prediction data storage unit 117 and extracts the area identification number in which the "caution level" is written. Then, the voice notification unit 115 notifies the notification device corresponding to the seat of the extracted area identification number of information indicating that the voice level of the conversation voice should be suppressed. That is, the voice notification unit 115 displays the notification content such as "The conversation is getting louder. Please do not raise your voice any more as it will disturb other people" on the display screen provided in the notification device. Perform display processing. In addition to the notification contents, the voice notification unit 115 displays the predicted voice level and the threshold side by side on the display screen of the notification device, and notifies the worker to suppress the voice level increase in conversation. .

Also, the voice notification unit 115 refers to the propagation prediction table of the propagation prediction data storage unit 117 and extracts the region identification number in which the "warning level" is written. When the predicted voice level is "warning level", the voice notification unit 115 displays "Please lower the voice of the conversation a little more", etc. It may be configured to notify suppression.

Further, each of the notification device 21 to the notification device 29 may be composed of a set of LED (Light Emitting Diode) lamps of three colors of blue, yellow and red as a display device.

In this case, when the predicted audio level is neither "caution level" nor "warning level," the audio notification unit 115 turns on the blue LED while turning off the yellow and red LEDs. In addition, when the predicted audio level is the “caution level”, the audio notification unit 115 turns on the yellow LED while turning off the blue and red LEDs. When the predicted sound level is "warning level", the voice notification unit 115 turns on the red LED while turning off the blue and yellow LEDs.

In addition, the conversation voice level notification device 11 in the conversation voice level notification system 1 of the present embodiment is a position detection type facility control system that detects the presence or absence of a worker by a sensor or the like and controls air conditioning, lighting, etc. For example, it may be configured by being combined with a smart workplace system. In this case, each of the occupancy detection sensor 31 to the occupancy detection sensor 39 in this embodiment is configured to use sensors in the position detection type equipment control system. At this time, the functions of the conversational voice level notification device 11 of the present embodiment may be incorporated as software into the core control device (for example, a server) of the position detection type equipment control system. In addition, the conversation voice level notification device 11 of the present embodiment uses the network of the position detection type equipment control system, the notification device 21 to the notification device 29, the presence detection sensor 31 to the presence detection sensor 39, and the voice sensor 41 to the voice sensor 49 may be configured to transmit and receive information respectively.

Further, in the position detection type equipment control system, if the lighting device of each seat can be controlled, the conversation voice level notification device 11 may be configured to use the control function of this lighting device. That is, when the "warning level" is reached, the voice notification unit 115 repeats turning on/off the lighting device (or changing the illuminance of the lighting) at a predetermined cycle, and the "warning level" is reached. It may be configured to notify that there is something.

In addition, if the notification level has a plurality of stages of "caution level" and "warning level", the voice notification unit 115 changes the frequency of the cycle for turning on/off (or changing the illuminance of the lighting) for each level. and may be configured to notify the worker.

Also, each of the notification device 21 to the notification device 29 in the present embodiment may be replaced with a smart phone or the like carried by the worker. In this case, it is necessary to install the following applications on the smartphone. That is, when the conversation voice level notification system 1 reaches the "caution level" or "warning level", the smartphone carried by the worker in the corresponding seat is sent an e-mail or via a LAN to drive the vibration function of the smartphone. Alternatively, the display screen may be configured to display the already-described "warning level" or the like.

In addition, the presence detection sensor 31 to the presence detection sensor 39 that detect whether the worker is present or absent may be configured to use the position information notification function by BYOD (Bring Your Own Device) such as a smartphone. . That is, the presence detection unit 111 obtains the position information from the smart phone carried by each worker, and the position of the position information associated with the worker identification information carried by the smart phone is detected by the worker. It is determined whether or not the position is the seat indicated by the identification information. If the position information of the smartphone and the position of the seat are the same, the presence detection unit 111 determines that the seat is present. It is written and stored in the voice/seat data table of the seat data storage unit 116 .

Next, an operation example of the conversation voice level notification system 1 will be described with reference to the drawings. FIG. 5 is a flow chart showing an operation example of suppressing conversational voice by the conversational voice level notification system 1 . A case will be described below in which workers are present at each of the seats S_1 and S_6, and workers are absent from each of the seats S_2 to S_5 and S_7 to S_9. Further, in the following description, the operation will be described with a configuration that utilizes the worker presence/absence information in the voice/presence data table of the voice/presence data storage unit 116 .

Step F1:
After a predetermined period, the voice input unit 112 sequentially inputs the voice level data of the conversation voices output from the voice sensors 41 to 49 from the seats S_1 to S_9.

Then, the voice input unit 112 associates the input voice level data from the voice sensors 41 to 49 with the area identification information of the seat, and sequentially records the voice/seat data table of the voice/seat data storage unit 116 . is written and stored in the voice level data column.

Step F2:
The analysis device included in each of the audio sensors 41 to 49 measures the A-weighted equivalent sound pressure level (L Aeq ) in a preset time interval width (time slot) and the octave band equivalent sound pressure level (L _Aeq ) with a center frequency of 500 Hz ( L _500eq ), it is determined whether or not the measured sound contains speech.

Based on the sound collected by the microphone, each analysis device measures the A-weighted equivalent sound pressure level (L _Aeq ) in the set time interval band and the octave band equivalent sound pressure level (L _500eq ) with a center frequency of 500 Hz. and

Then, based on the calculated A-weighted equivalent sound pressure level (L _Aeq ) and the octave band equivalent sound pressure level (L _500eq ) with a center frequency of 500 Hz, the determination unit determines that the difference L _diff (equation (3)) is , When the set threshold L _T or more is reached, voice is included in the time range in which the A _-weighted equivalent sound pressure level (L Aeq ) and the octave band equivalent sound pressure level (L _500eq ) with a center frequency of 500 Hz are measured. I judge. That is, when the inequality of formula (4) holds, it is determined that the voice (by a human voice) is included.
L _diff =L _500eq -L _Aeq (3)
L _diff ≧L _T (4)
If the result of the determination is NO, the process returns to step F1 and loops. On the other hand, if the determination result is YES, the process proceeds to step F3. As described above, in this embodiment, it is determined with high accuracy whether or not a human voice is included, and then the steps after F2 are executed. can be prevented. Moreover, since the second and subsequent steps are executed only when the result of the determination becomes YES, unnecessary processing load is not imposed on the system.

Step F3:
Next, the presence detection unit 111 detects presence/absence information (presence information) output from each of the presence detection sensors 31 to 39 at the same period as the audio input unit 112 inputs the audio level data. (Information indicating either seat or absence) is entered sequentially.

Then, the occupancy detection unit 111 associates the inputted occupancy information from each of the occupancy detection sensors 31 to 39 with the area identification information of the seat, and sequentially stores the information in the voice/occupancy data storage unit 116 . It is written and stored in the presence/absence information column of the voice/occupation data table.

Next, the presence detection unit 111 refers to the voice/seat data table of the voice/seat data storage unit 116 and extracts the seat of the region identification information indicating the presence of the seat.

At this time, if there is no seat occupied by the worker in the voice/seat data table, or if there is only one seat occupied by the worker, the presence detection unit 111 detects other seats. Since the conversation voice does not affect the worker, the process proceeds to step F7.

On the other hand, if there are two or more seats where the worker is present in the voice/seat data table, the presence detection unit 111 determines that the conversation voice affects other workers and performs processing. Proceed to step F4.

Step F4:
Next, the speech prediction unit 113 refers to the speech/seat presence data table of the speech/seat presence data storage unit 116, and extracts the seat of the area identification information indicating the seat presence. Here, the speech prediction unit 113 extracts that the seats S_1 and S_6 are occupied.

Then, the voice prediction unit 113 performs voice prediction based on each of the seat S_1 and the seat S_6 where the seat is present.

That is, the speech prediction unit 113 reads the data of the distance r from the seat S_1 to the seat S_6 from the distance table corresponding to the seat S_1 in the distance storage unit 118 .

The voice prediction unit 113 also reads the voice level data of the seat S_1 from the voice/seat data table of the voice/seat data storage unit 116 .

Then, the voice prediction unit 113 predicts the propagation of the conversation voice of the worker at the seat S_1 from the read data of the distance r and the voice level data of the seat S_1 using each of the formulas (1) and (2). Then, the predicted voice data for the seat S_6 is calculated. The voice prediction unit 113 writes the predicted voice data into the column of the predicted voice level of the seat S_1 in the propagation prediction data table of the propagation prediction data storage unit 117 for storage.

Similarly, the speech prediction unit 113 reads data of the distance r from the seat S_6 to the seat S_1 from the distance table corresponding to the seat S_6 in the distance storage unit 118 .

The voice prediction unit 113 reads the voice level data of the seat S_6 from the voice/seat data table of the voice/seat data storage unit 116 .

Then, the voice prediction unit 113 predicts the propagation of the conversation voice of the worker at the seat S_1 from the read data of the distance r and the voice level data of the seat S_6, using each of the equations (1) and (2). Then, the predicted voice data for the seat S_1 is calculated. The voice prediction unit 113 writes the predicted voice data into the column of the predicted voice level of the seat S_6 of the propagation prediction data table of the propagation prediction data storage unit 117 to store it.

Step F5:
The voice determination unit 114 refers to the propagation prediction data table of the propagation prediction data storage unit 117 and determines whether or not the predicted voice level exceeds the threshold.

Then, when the predicted speech level at seat S_6 of the conversational speech at seat S_1 exceeds the threshold, the speech determination unit 114 adds " "Warning level" is written and memorized. On the other hand, if the predicted speech level at seat S_6 of the conversational speech at seat S_1 does not exceed the threshold, the speech determination unit 114 adds " "Non-warning level" is written and stored.

In addition, when the predicted speech level at seat S_1 of the conversational speech at seat S_6 exceeds the threshold, the speech determination unit 114 adds " "Warning level" is written and memorized. On the other hand, if the predicted speech level at seat S_1 of the conversational speech at seat S_6 does not exceed the threshold, the speech determination unit 114 adds " "Non-warning level" is written and stored.

Then, the voice determination unit 114 advances the process to step F6.

Step F6:
The voice notification unit 115 refers to the propagation prediction data table of the propagation prediction data storage unit 117, and selects the region identification number corresponding to the region identification number of either or both of the seats S_1 and S_6 for which the predicted voice level is equal to or higher than the threshold. to extract

Then, the voice notification unit 115 sends information indicating that the voice level of the conversation voice should be suppressed to the notification device (either or both of the notification device 21 and the notification device 26) corresponding to the seat of the extracted area identification number. is notified, and the process proceeds to step F1.

Step F7:
The voice notification unit 115 does not notify any area of the information indicating that the voice level of the conversational voice is to be suppressed, and advances the process to step F1.

According to the present embodiment described above, since a warning is given to a worker whose conversational voice is loud that it is interfering with other workers' conversations, there is an effect of encouraging the worker to suppress his or her own voice.

Further, according to the present embodiment, when the increase in the voice level of each worker who is on the phone at different locations in the same office space or between groups who are talking with each other exceeds the respective thresholds, a warning is issued. Therefore, there is an effect that it is possible to prevent a phenomenon in which each other's voice gradually becomes louder due to interaction.

In addition, according to the present embodiment, as described above, since the voice of each worker is suppressed below a certain level (threshold value), it is possible to reduce the noise in the office space and reduce the stress of the workers. This has the effect of making it possible to improve the productivity of each worker.

In addition, this embodiment can perform propagation prediction and evaluation of conversation voice only for areas where other workers are present by linking with a position detection type facility control system. Therefore, according to the present embodiment, by avoiding unnecessary notifications without predicting and evaluating the conversation voice in an area where the worker is absent, the effectiveness of notifications such as warnings can be ensured, and the conversation can be stopped. There is an effect of preventing unnecessary discomfort from being given to the worker who is doing this.

In addition, since hearing deteriorates with age, it becomes difficult for elderly people to hear conversations, and this tendency becomes more pronounced as the surrounding noise becomes louder.

Therefore, according to the present invention, since the office space can be made quieter as described above, it is possible to reduce the difficulty of hearing conversations of elderly people working in this office, and reduce the stress of elderly workers who are difficult to hear. and improve the productivity of older workers.

By recording a program for realizing the functions of the conversational voice level notification system 1 of the present invention in a computer-readable recording medium and causing a computer system to read and execute the program recorded in the recording medium, Control may be performed to suppress the conversation voice of the operator. It should be noted that the "computer system" referred to here includes hardware such as an OS and peripheral devices.

Also, the "computer system" includes a WWW system provided with a home page providing environment (or display environment). The term "computer-readable recording medium" refers to portable media such as flexible discs, magneto-optical discs, ROMs and CD-ROMs, and storage devices such as hard discs incorporated in computer systems. In addition, "computer-readable recording medium" means a volatile memory (RAM) inside a computer system that acts as a server or client when a program is transmitted via a network such as the Internet or a communication line such as a telephone line. , includes those that hold the program for a certain period of time.

Further, the above program may be transmitted from a computer system storing this program in a storage device or the like to another computer system via a transmission medium or by a transmission wave in a transmission medium. Here, the "transmission medium" for transmitting the program refers to a medium having a function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line. Further, the program may be for realizing part of the functions described above. Further, it may be a so-called difference file (difference program) that can realize the above-described functions in combination with a program already recorded in the computer system.

As described above, the conversational voice level notification system 1 and the conversational voice level notification method according to the present invention are based on the sound collected by the voice sensor, the A-weighted equivalent sound pressure level and the octave band equivalent sound pressure level with a center frequency of 500 Hz. is calculated, and if the difference obtained by subtracting the A-weighted equivalent sound pressure level from the octave band equivalent sound pressure level with a center frequency of 500 Hz is equal to or greater than a preset threshold value, the sound collected by the sound sensor includes voice. It is determined that As a result, according to the conversation voice level notification system 1 and the conversation voice level notification method according to the present invention, by accurately determining whether or not the sound collected by the voice sensor includes human voice, the system can It is possible to prevent erroneous notification.

Further, according to the conversational voice level notification system 1 and the conversational voice level notification method according to the present invention, it is possible to detect voices with high accuracy, and the system may give erroneous notifications due to various sounds other than voices in the office space. can prevent

Further, according to the conversational voice level notification system 1 and the conversational voice level notification method according to the present invention, the voice sensor only needs to have the function of analyzing the equivalent sound pressure level of the A characteristic and the 500 Hz octave band at the minimum. can be manufactured cheaply.

Further, according to the conversational voice level notification system 1 and the conversational voice level notification method according to the present invention, the method for determining whether or not the measured sound contains voice is simple subtraction and comparison calculations. system is unnecessary.

Further, according to the conversation voice level notification system 1 and the conversation voice level notification method according to the present invention, the necessary data is only the A characteristic and the equivalent sound pressure level of the 500 Hz octave band, and the data when using a wireless sensor etc. The amount of transfer is small, and the load on the communication and processing of the system is small.

Further, according to the conversation voice level notification system 1 and the conversation voice level notification method according to the present invention, the frequency of voice detection can be arbitrarily set by setting the equivalent sound pressure level measurement time. By appropriately setting the measurement time, it is possible to prevent the system from being activated by a short conversation and sending excessive notifications. In addition, when the measurement time is set long, it is possible to reduce the load of processing such as communication and prediction calculation of the system.

Next, the technical grounds for determining whether voice is included or not included in the conversation voice level notification system 1/conversation voice level notification method according to the present invention as described above will be described.

FIG. 6 shows the frequency characteristics of the sound pressure level when a conversation is taking place in a meeting space in an office and when no conversation is taking place. For the measurement of the frequency characteristics in FIG. 6, the octave band equivalent sound pressure level was continuously measured for one minute. A total of 5,622 samples were measured, of which 1,406 samples were conversational, and 4,216 samples were not. FIG. 6 shows the average value and the average value±standard deviation of the measurement results in the cases where conversation was taking place in the meeting space and where no conversation was taking place. Note that the sound pressure level at each frequency is a value weighted by the A characteristic.

From FIG. 6, it can be seen that the 500 Hz octave band sound pressure level is a prominent frequency characteristic when speech is included in one minute when the equivalent sound pressure level is measured. On the other hand, if there is no speech, ie no speech, no such feature is shown. The above measurement results show that speech has a predominant frequency component in the 500 Hz octave band.

Conversely, when the 500 Hz octave band sound pressure level is dominant in the frequency characteristics of the measured sound, it suggests that the measured sound is highly likely to contain speech.

Now consider the relationship between the A-weighted sound pressure level and the 500 Hz octave band sound pressure level. Assuming that the measured sound has frequency components only in the 500 Hz octave band, the difference L _diff is 3.2 dB from the definition of the A-weighted sound pressure level.

In reality, sound has frequency components in bands other than the 500 Hz octave band, so the A-weighted sound pressure level and L _diff are not exactly 3.2 dB, but the closer L _diff is to 3.2 dB, the more measured. It means that the 500 Hz octave band sound pressure level of the sound produced is outstanding.

L _diff was calculated from the measurement results shown in FIG. 6, and the frequency distribution of L _diff when conversation was taking place in the meeting space and when no conversation was taking place are shown in FIG.

From FIG. 7, it is shown that L _diff is larger, ie, closer to _3.2 dB, when conversation is taking place compared to when no conversation is taking place.

The determination accuracy changes depending on the set value of the threshold L _T for determining that the sound measured from L _diff contains speech. FIG. 8 shows the relationship between the set value of the threshold L _T and the determination accuracy.

In the measurement example shown here, when the threshold L _T is set to −1 dB, the determination accuracy of both conversation and non-conversation is about 90%.

If the threshold value L _T is set to a large value, the probability of erroneously determining that the measured sound does not contain speech increases even though the measured sound contains speech. On the other hand, if the threshold L _T is set to a small value, the probability of erroneously determining that the measured sound does not contain speech but contains speech increases. The threshold L _T must be set taking into account the influence of ambient noise other than voice.

In the measurement example shown here, the equivalent sound pressure level is measured for one minute, but the measurement time of the equivalent sound pressure level is not limited to one minute. The equivalent sound pressure level measurement time is set according to the required frequency of sound detection, but generally it is appropriate to set it to about 10 seconds to 10 minutes. When the measurement time of the equivalent sound pressure level is set long, it may be determined that the sound is not included when the sound is included only in a short time within the measurement time.

Based on the above technical grounds, the conversational voice level notification system/conversational voice level notification method according to the present invention performs voice detection, and in principle, complicated data processing is unnecessary. is clear.

1... Conversation voice level notification system 11... Conversation voice level notification devices 21, 22, 23, 24, 25, 26, 27, 28, 29... Notification devices 31, 32, 33, 34, 35, 36, 37, 38, 39... presence detection sensors 41, 42, 43, 44, 45, 46, 47, 48, 49... voice sensor 111... presence detection unit 112... voice input Unit 113: Voice prediction unit 114: Voice determination unit 115: Voice notification unit 116: Voice/seat data storage unit 117: Propagation prediction data storage unit 118: Distance storage unit

Claims (4)

A presence detection sensor arranged for each area that is a worker's seat or a meeting space where the worker has a meeting;
an audio sensor arranged for each region;
Based on the sound collected by the audio sensor, an A-weighted equivalent sound pressure level and an octave band equivalent sound pressure level with a center frequency of 500 Hz are calculated. a determination unit that determines that the sound collected by the sound sensor includes voice when the difference obtained by subtracting the sound pressure level is equal to or greater than a preset threshold;
a presence detection unit that detects whether or not the worker is present in the region based on the output of the presence detection sensor when the determination unit determines that the presence of a voice is included;
a voice input unit for inputting voice data of a worker from the voice sensor arranged in the area;
From the sound data in each of the areas where the workers are present, the sound intensity in other areas where other workers are present when sound is propagated is determined from the area and the other areas. a speech prediction unit that predicts based on the distance at which
a voice determination unit that determines whether any of the voice intensities exceeds a preset threshold;
and a voice notification unit that, when any of the voice intensities exceeds a preset threshold, notifies the level of the voice strength and/or suppression of the voice strength by a notification device provided for each of the regions. A conversation voice level notification system characterized by: The notification device is a personal computer connected to the conversation voice level notification system, and the voice notification unit causes the display screen of the personal computer to notify the level of the voice intensity and/or the suppression of the voice intensity. 2. The conversational voice level notification system according to claim 1, wherein: 3. The conversation voice level notification system according to claim 1, wherein said threshold is provided in a plurality of stages, and a different notification is made each time said voice intensity exceeds each stage. Based on the sound collected by the sound sensor arranged in the area where the voice input unit is the worker's seat or the meeting space where the worker has a meeting , the A-weighted equivalent sound pressure level and the octave with a center frequency of 500 Hz If the difference obtained by subtracting the A-weighted equivalent sound pressure level from the octave band equivalent sound pressure level with a center frequency of 500 Hz is equal to or greater than a preset threshold, the sound collected by the sound sensor is calculated. a determination process for determining that the sound contains speech;
a presence detection process in which the presence detection unit detects whether or not the worker is present in the area based on the output of a presence detection sensor arranged in each area;
a voice input process of inputting voice data of a worker from the voice sensor;
The voice prediction unit calculates the voice strength in other areas where other workers are present when the voice propagates from the voice data in each of the areas where the workers are present. a speech prediction process for predicting based on distances between regions and other regions;
an audio determination process in which an audio determination unit determines whether any of the audio intensities exceeds a preset threshold;
When any of the sound intensities exceeds a preset threshold, the sound notification unit notifies the level of the sound intensity and/or the suppression of the sound intensity by the notification device provided for each of the regions. A conversation voice level notification method, comprising: a notification process;

Images