JPH03252258A - Directivity reproducing device - Google Patents

Abstract

PURPOSE:To enable a listener to recognize an acoustic image in an accurate position regardless of movement of a sound source by simulating and outputting the input voice whose directivity is reproduced by a directivity reproducing part based on relative position relations between a speaker and the listener which are calculated from position information of both of them. CONSTITUTION:Position information of the speaker and the listener in a conference place or the like are read out from a position storage part 103 via a control part 106 to calculate relative position relations between the speaker and the listener by a phase calculating part 104. A directivity reproducing part 105 processes the voice of the speaker from an input part 101 based on the calculation result to reproduce the directivity, and the artificial voice by which the position of the speaker can be accurately recognized is outputted from a speaker 102, and thus, the conference or the like is smoothly progressed by the directivity reproducing device which enables the listener to recognize the accurate acoustic image regardless of movement of the sound source.

Description

[Detailed Description of the Invention] [Object of the Invention] (Industrial Application Field) The present invention relates to a directional playback device that plays back audio in a specific direction, particularly when there are multiple participants in a conference or the like. However, it concerns a directional playback device that allows the listener to recognize the location of the speaker.

(Prior Art) At international conferences and the like, conference participants listen to the voices of speakers through speakers, headphones, and the like. The audio heard from these speakers and headphones is not only the voice of the speaker. This means that when the speakers and participants speak different languages, the content of the speaker's statement can be interpreted using an interpreter.
If the interpreter has translated it into a language that the participants can understand,
I can hear the voice of the interpreter. In addition, recent video conferencing systems use wired and wireless communication methods to allow conference participants to hold a conference while checking the other party on the TV screen and audio, even if they do not gather in one place. It is appearing.

In this video conference system as well, the voice of the speaker reaches the ears of the participants via an audio output device such as a speaker or headphones. In these cases, especially when the listener listens to the speaker's voice through headphones, the listener can quickly determine the speaker's location because the sound completely ignores the distance and direction between the speaker and the listener. This is because when listening to audio through headphones, the sound image created by the audio generated from the location of the speaker (that is, the sound source) is located at the back of the listener's head.

This sound image reproduces the normal sound source location to the listener.

A circuit according to Japanese Patent Application Laid-Open No. 63-176100 for recognition will be explained with reference to FIG. As shown in the figure, this is a left signal 502 and a right signal 50 output from the amplification output circuit 501.
3 to 600±10 with delay circuits 541 and 542, respectively.
A delay of about 0tis is performed. Furthermore, each of the eight forces is attenuated by about 2±20% in attenuation circuits 551 and 552, and the resulting signal 561.562 is mixed with the left signal 502 and right signal 503 in mixing circuits 571 and 572, respectively. input to headphones 508. Note that the delay time and amount of attenuation can be varied to suit the listener's preference.6 As explained above, if this circuit is used, the left and right signals are mixed with delayed and attenuated inverse signals. , reproduces the sound image in its normal position.

can be recognized.

However, the circuit is a circuit used for one headphone, and cannot be used for multiple sound sources. In addition, the delay time and the amount of attenuation are determined within a certain range, and it is not possible to sufficiently reproduce sound depending on the sound source position. Additionally, the delay time and attenuation amount can be varied outside the above range, but this adjustment must be made by the listener.
There was a problem in that the sound image could not be reproduced completely when the sound source moved.

(Problems to be Solved by the Invention) As described above, the conventional circuit has a problem in that when the sound source moves, the listener cannot reproduce and recognize the sound image in its normal position.

The present invention has been made to solve the above-mentioned problems, and aims to provide a directional reproduction device that allows a listener to always recognize a sound image at an accurate position even if the sound source moves.

[Structure of the invention]

(Means for Solving the Problems) In order to achieve the above-mentioned object, in the directional reproduction device of the present invention.

An input section for inputting the speaker's voice, an output section for storing positional information of the speaker and the listener, and a relative positional relationship between the speaker and the listener calculated based on the positional information of the output section. a phase calculation section; and a directionality reproduction section that reproduces the directionality of the voice in order to make the listener recognize the speaker's position through the voice.

and an output section for outputting audio. The directivity reproduction section causes the output section to reproduce the input audio in a pseudo manner based on the calculation result of the phase calculation section. Note that, in the case of a conference, the input section and the output section are used as windows for conference attendees. Among these, the input section is used to instruct a person who is going to speak among the conference attendees to speak. In addition, instructions such as ending the meeting are also given through the input section. The phase calculation section detects the speaking instruction from the input section and calculates the relative positional relationship between the speaker and each listener. In order to calculate the positional relationship between the speaker and each listener, the output section stores positional information of the speaker and each listener. This position information stores the number and position of each conference attendee, as well as the face orientation of the attendee. The position of each participant is expressed by the distance and phase from the origin of the virtual coordinate axes. Furthermore, a freely settable standard is provided for the direction of each participant, and the direction is stored as a direction relative to this standard. The directivity reproduction section performs processing such as delay, amplification, and attenuation on the audio signals sent to each listener and the speaker's voice input from the input section based on the calculation results of the phase calculation section. obtained by. The control section sends this result to the output section of each listener.

(Operation) According to the directional reproduction device of the present invention configured as described above, the phase calculation section detects the speaker.

Based on this speaker position information and the speaker and listener position information stored in the output unit, the phase calculation unit calculates the relative positional relationship between the speaker and each listener.

The directivity reproduction unit processes the speaker's voice based on this calculation result. The control section sends this to the listener via each output section. As a result, the listener can always accurately recognize the speaker's position (that is, the sound source). For example, in international conferences and video conferences, accurate recognition of speakers allows the conference to proceed smoothly.

(Example) The present invention will be described in detail below with reference to one drawing.

FIG. 1 shows a block diagram of one embodiment of the invention.

Below, multiple people are in different locations due to video conferencing.
The explanation will be given by taking as an example a case where headphones are used as an output section and the voices emitted by each person are transmitted through an input section (for example, a microphone).

In FIG. 1, an input unit 10 is for inputting voices uttered by one speaker and instructions for ending a conference, etc., and is usually a microphone, a keyboard, or the like.

The output unit 102 outputs audio, and may be a wall-embedded speaker or the like, but in this embodiment, a case where headphones are used is considered. Note that since this embodiment considers the case where there are a plurality of listeners, it is assumed that there are a plurality of input sections 101 and output sections 102.

The output unit 103 stores the positions of speakers and listeners. Note that in the case of a conference, if one of the conference participants is a speaker, that is, a speaker, the other conference participants become listeners.

However, this speaker has finished speaking and becomes a listener when other participants speak.

In other words, conference participants can be both speakers and listeners.

For this reason, the output unit 103 stores position information of all conference participants.

Here, consider a virtual conference room as shown in FIG. 2(a). This is used to consider the coordinate axes to be stored in the output unit 103. Note that this coordinate axis is virtual; for example, it may be an actual conference room, or if conference participants are in different locations, such as in a video conference,
You can consider a conference room as a model. therefore,
The origin can also be set to any position. In other words, it is sufficient that the coordinate axes have a range that allows at least the positions of the speaker and the listener to be set. According to FIG. 2(a), in the conference room 201, conference participants 1 to 4 and 5 to 8 are seated at desks 204 with a partition plate 202 as a boundary. The faces of conference participants can be seen by each other through their respective display units 203, such as CRT displays. Note that FIG. 2(b) is an example of this display. This is an example display on the display unit 203-2, and participants 1 to 2
4 is displayed, which is an example of an image that is being viewed by participants 5 to 8. The second display is, for example, a CCD (Charge Coup) attached to the top of the display section 203-1.
These are the results taken by a camera (not shown) or the like and sequentially sent to the display unit 203-2. Then, the voice of the speaker can be heard by the participants through headphones.

Next, participant position information will be explained using participant 1 as an example. Participant 1's position information is expressed by distance γ1 from the origin of the coordinate axis to participant 1 and phase θ. In addition to this, the location information also includes the direction in which the participant 1 is facing when seated. FIG. 2(c) is a diagram showing a part of the correspondence between FIG. 2(b) and FIG. 2(a). The direction indicated by 251 in FIG. 2(c) indicates the direction of the participant 1, and 2 in FIG. 2(c)
Reference numeral 52 indicates the headphones mentioned above. Note that information about which direction the vertical direction of a straight line connecting both ears of the participant points is included and stored in the position information based on the participant position (γ
1. θ, ) represents only the location information of where each conference participant is in the conference room, and this is to avoid the possibility that the direction of each participant's ear from which the sound is emitted cannot be determined. This is also necessary to make it possible to freely set the origin of the coordinate axes mentioned above. In other words, as shown in Figure 2(a), the coordinate axis (
For example, instead of setting a conference room in an area of +, +), a circular conference room is set and the origin of the coordinate axis is set at the center of the circle. At this time, if we do not have information about not only the position of each participant but also the direction of the participant, it is necessary to have information about the direction of each participant's audio input (i.e., ears) and output (i.e., mouth). This is because we do not know whether the For this reason, it is necessary to know which direction the vertical direction of the straight line connecting the participants' ears faces. This results in
The position of the listener's ears and mouth can also be approximately determined. The above explanation has been given using participant 1 as an example, but this will be done for all conference participants.

FIG. 3 shows an example of a storage format in which the position information is stored in the output unit 103. The participant number 301 identifies a conference participant, and is a number that each conference participant has. Therefore, if the total number of conference participants is n, there are also n pieces of location information. Furthermore, location information is
It consists of a participant number 301, each participant [302, which will be described later], and a participant direction 303. And output section 10
3 is.

This location information is stored. Among these, the participant position 302 consists of the distance γ from the origin of the coordinate axes described above and the phase θ. In other words, the participant position 302 of participant 1 is (γ1.θ□).

In other words, to all participants of participant i [302 is (γ1.θ engineering)
becomes. Next, the participant direction 303 indicates which direction the vertical direction of the straight line connecting both ears of the participant points to, as indicated by the direction 251 in FIG. 2(C). Note that this direction is displayed in the same format as the phase θ of the participant position 302. This is indicated by a "+" (plus) for counterclockwise rotation around the participant, and a "-" (minus) for clockwise directions. However, "r+" (plus) can be omitted and is not described in this embodiment.

In other words, θiJ of the participant direction 303 indicates which direction the vertical direction of the straight line connecting both ears of participant i faces. Further, θij and θnn are also based on the above description format. Here, in this embodiment, the position information is stored in the output unit 103 before the start of the conference, but the participant direction 303 can be changed sequentially. This corresponds to a change in speaker, but this is allowed only if
Processing stored in the output unit 103 and phase calculation unit 1 described later
04 and the directivity reproduction unit 105 must be performed at high speed and must be within a range that does not affect the processing.

Here, the range that does not have any influence is the range in which the processing has caught up with the speed of the speaker's voice, and this includes the time used for extraction or input processing of the participant's direction, and the range in which this information is transferred to the output unit 103. It is determined by the time used to memorize the data. It is assumed that the position information of conference participants to be stored in the output section 103 described above is manually entered by each participant from the input section 101. Note that the position of the participant can also be extracted by image recognition and stored in the output unit 103. Furthermore, in the case of assuming a conference room, it is also possible to arrange participants at arbitrary positions in the conference room and use the positions as position information.

The rank calculation unit 104 detects speakers from among conference participants. This can be done by having the person who wants to speak input an instruction to speak from the input unit 101 and detecting this input.

In addition to this, VOX (Voice eXchan)
In the case of a microphone equipped with a voice detection system, detection can also be performed based on its operating state. and,
When a speaker is detected, the position information from the output unit 103 is used to calculate the relative distance and phase to the speaker for each listener.

The directivity reproduction unit 105 reproduces the directivity of the voice in order to make the listener recognize the position of the speaker through the voice. For this purpose, the relative distance and phase between the speaker and each listener calculated by the phase calculation unit 104 are used. Then, based on this calculation result, the directionality of the audio is reproduced.

The control unit 106 controls the entire device.

This is a conventional control section with an added function of sending the sounds for each listener reproduced by the directivity reproduction section 105 to the output section 102 of each listener.

FIG. 4 is a flowchart for explaining the operation of the control unit 106. The control unit 106 stores position information of conference participants in the 1 output unit 103, and starts operating when an instruction to start the conference is input from the input unit 1.01. First, the output unit 103 sends the position information of all conference participants to the phase calculation unit 104 (step S401). Next, the phase calculation unit 104 detects the speaker (step S4).
02), the person who wants to speak among the participants uses the input section 10
When an instruction to speak from 1 is input, a voice input device such as a microphone as one of the input units 101 starts operating, and the 0 phase calculation unit 104 stops accepting instructions to start operation of other microphones. By checking from which input unit 101 this operation start instruction was input and checking the corresponding position information, it is possible to detect who is the speaker. Further, the phase j peak portion 104 extracts the position information of the listener i from the speaker position information and the listener position information identified in step S402 (step S403), and extracts the relative position information between the speaker and the listener i. The distance and phase are calculated (step S404), and this calculation process will be described later. This calculation result is sent to the directivity reproduction unit 105 (step S405). The directivity reproduction unit 105 reproduces the directivity of the speaker's voice for the listener i using this calculation result (step 8406). Note that this reproduction process will be described later. and.

This reproduction result is sent to the listener's power unit 102 (step 4
07).

The above processing (steps 8404 to S407) is performed for all listeners (step 8408).
409), once the reproduced audio has been sent to all listeners, it is determined whether the conference is over (step S410).
If there is an instruction to end the conference via 1, the process ends. If there is no instruction to end the meeting, step 5402
Return to , and detect the speaker again.

Next, step 5404 in FIG. 4, that is, the process of calculating the distance and phase difference between the speaker and listener i by the phase calculation unit 104 will be described. For example, assume that the position information of the speaker, listener A, and listener B are shown below.

Participant number O Participant position (γ., θ− Participant direction θ. .

Participant number A Participant position (γA, θA) Participant direction θA^ Participant number B Participant position I! (γB, θB) Participant direction
θBB It is assumed that the above position information is stored in the output unit 103. The control unit 106 sends this position information to the phase calculation unit 104. Then, the phase calculation unit 104 calculates the distance and phase difference between the speaker and listeners A and B, respectively, as follows.

■ In the case of speaker and listener A, the distance between the speaker and listener A is γ, and A is the phase difference θ. ^ is θ. A=θ. . -θ^^ ■ In the case of speaker and listener B The distance between the speaker and listener B is γ. B is γ. ^=JcγBginθa−7. ．． sinθo)”+
(γBcoaθB−γ, cosθo) 2 phase difference θ. B is θ. B=θ. . -θBB Next, step 5406 in FIG. 4, that is, the directivity reproduction section 1
The reproduction process of 05 will be explained in detail. This is step 5
In step 405, when the calculation result from the phase calculation unit 104 is sent, the operation starts. Note that the directivity reproduction unit 105 expresses the distance and direction between the speaker and each listener by applying processing such as attenuation, amplification, and delay to the speaker's voice. We will discuss the case where distance and direction are expressed by changing. First, to express distance, the volume is reduced in inverse proportion to the square of the distance, for example. This can be expressed by the following formula.

Here, γij in equation (2) represents the distance between speaker i and listener j, and α represents a proportionality constant, which can be determined depending on the size of the conference room, etc., but can be freely set. At the same time, if the volume is too low or too high, listeners may not be able to hear the sound or feel uncomfortable, so the range of sound pressure can be adjusted.

Further, in order to express the direction, the sound pressure on the left and right sides of the output section 102 is changed in correspondence with the phase difference, for example, as follows.

This formula ■ equalizes the sound pressure on the left and right ears when the speaker is in front of the listener, equalizes the sound pressure on the right ear when the speaker is on the right as viewed from the listener, and equalizes the sound pressure on the left ear when the speaker is on the left. Make it expensive. Note that the above-described processing is performed based on location information of the speaker and the listener.

As described above, according to this embodiment, participants listen to voices that change depending on the distance and direction to the speaker, so they can listen without using headphones.

It is the same, and a sense of realism can be maintained. At international conferences,
A sense of realism can be maintained by changing the sound pressure of the simultaneous interpreter's voice depending on the distance and direction from the speaker. Also, IC
If individual participants can be identified using a personal authentication device such as a card, audio can be sent only to specific participants.

Therefore, it is possible to transmit private messages or simultaneous interpretation without having other participants hear them.

Further, the output unit 102 of this embodiment is not limited to headphones, but may be a micro-sized speaker that can be inserted into the ear, a speaker embedded in a wall, or the like.

In short, various changes and modifications can be made without departing from the spirit of the invention.

〔Effect of the invention〕

As described above, according to the present invention, a listener can accurately recognize the speaker's location by listening to audio based on location information of the speaker and the listener.

This will help the meeting proceed smoothly.

Furthermore, it is possible to create a sense of realism in video conferences and the like.

[Brief explanation of drawings]

Fig. 1 is a block diagram of an embodiment of the present invention, Fig. 2(a)
is a conceptual diagram showing an example of a video conference in an embodiment of the present invention, FIG. 2(b) is an example of the screen display in FIG. 2(a), FIG. A diagram showing a part of the correspondence with FIG. 2(b), FIG. 3 is an example of the storage format of position information stored in the output section in FIG. 1, and FIG. 4 is an example of an embodiment of the present invention. FIG. 5 is a conceptual diagram of a conventional circuit for reproducing and recognizing a sound source using a sound image when the sound source does not move. 101... Input section 102... Output section 1
03...Part of position 5ba 104...Phase calculation unit 105...Directivity reproduction unit

Claims (1)

[Claims] (1) An input section for inputting the speaker's voice; a position storage section for storing position information of the speaker and the listener; A phase calculation unit that calculates the relative positional relationship with the listener, a directional reproduction unit that reproduces the directionality of the voice so that the listener can recognize the speaker's position by voice, and an output that outputs the voice. A directional reproduction device, comprising: a directional reproduction unit, wherein the directional reproduction unit causes the output unit to reproduce input audio in a pseudo manner based on a calculation result of the phase calculation unit. (2) The directional playback device according to claim 1, wherein the input section and the output section are provided for the number of speakers and listeners present. (3) The directional reproduction device according to claim 2, wherein the directional reproduction section reproduces the directionality of sound for the output section of each listener. (4) The directional reproduction device according to claim 3, wherein the control section sequentially sends the results of the directional reproduction section to the output section of each listener. (5) The directional reproduction device according to claim 1, wherein the position storage unit stores the positions of the speaker and the listener in terms of distance and phase from the origin of the coordinate axes. (5) The directional reproduction device according to claim 1, wherein the position storage section further stores a vertical direction of a straight line connecting both ears of the listener. (7) The directional playback device according to claim 2, wherein the output section uses headphones that output audio from left and right sides. (8) The directional playback device according to claim 2, wherein the position storage section stores only position information of the speaker. (9) The directional playback device according to claim 2, wherein the position storage section stores position information of all attendees at a meeting or the like. (10) The directional playback device according to claim 2, wherein the input unit includes a keyboard or the like for giving instructions such as making a statement and ending the meeting. (11) The directional reproduction device according to claim 10, wherein the phase calculation section detects a speaker based on a speaking instruction from the input section.