JP3095484B2 - Audio signal output device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an audio signal output device for outputting an audio signal that can be heard only by a specific user.

[0002]

2. Description of the Related Art In recent years, computers for processing and outputting multimedia such as characters, voices, figures, and images have been advanced, and various application programs have been developed. Language data including characters from conventional numerical data, as well as pattern data such as voice and images, can be handled, high-fidelity sound reproduction and high-quality video output become possible, and 3D graphics and Other than animation,
The reproduction of a spatial sound field has also become possible with the increase in computer power.

[0003] Along with this, the use of visual data on a computer display is increasing, and the problem of increasing the degree of visual fatigue of the user is increasing. Therefore, in recent years, expectations regarding the use of audio data such as voices and acoustic signals capable of distributing the load from visual to auditory are increasing. That is, a technology in which a computer and a user interact with each other using audio data is becoming widespread. In addition, vending machines and cash dispensers at banks, etc.
Services based on voice interaction using audio data are becoming widespread.

However, when such audio data is used, there is an advantage that a message from a computer can be recognized without looking at a display screen.
The disadvantage is that audio messages reach other than the user of the computer. That is, the audio data is necessary information for the user, but merely noise for other people.

In particular, in a small office, office, or home, the current environmental noise alone is a serious problem, and further, it is not preferable to increase unpleasant noise by a synthetic sound from a computer. Therefore, an audio signal output device capable of outputting an audio signal that can be heard only by a specific person is desired.

[0006] As such a technique, a technique has been known in which noise at a specific position is reduced by actively adding a signal having an opposite phase to noise. But,
In this noise prevention technology, since the noise is radiated and propagated to various places, the effective range is limited to measures against compressor noise and measures against duct noise.

Further, there is known a technique of controlling the directivity of a sound field and transmitting sound only in a specific direction. However, in this technique, since only the direction in which the sound is transmitted is controlled, the sound is also transmitted to a person in the same direction as the user.

[0008]

As described above, the conventional noise prevention technology and the directivity control of the sound field cannot transmit an audio signal such as a sound, a buzzer sound, or music only to a specific person. There is a disadvantage that the audio signal is transmitted to a third party who does not need it. Therefore, for example, when a computer transmits a confidential matter to a user by voice, the confidential matter is heard by a third party, which has been a problem in privacy protection and security.

The present invention has been made to solve such a conventional problem, and an object of the present invention is to provide an audio signal output device capable of transmitting an audio signal from a computer only to a specific user. It is in.

[0010]

In order to achieve the above object, the present invention is an audio signal output device for outputting an audio signal to be transmitted to only a specific user in an easily audible manner. Presence position recognition means for recognizing; an input means for capturing an audio signal to be heard by the specific user; and a background sound signal to be heard by another person; and a plurality of output means for outputting the audio signal and the background sound signal. A speaker, a sound field control unit that determines the audio signal output from each speaker and the background sound signal based on a predetermined filter coefficient, and a specific user's presence position information based on the specific user's location information. Filter coefficient optimizing means for determining the filter coefficient so as to maximize the audio signal / background sound signal ratio at the position.

[0011] Further, there is a plurality of the specific users, and the filter coefficient optimizing means determines the filter coefficients so as to maximize the ratio of the audio signal / background sound signal at the position of each user. And

Further, the existence position recognizing means recognizes the existence position of the specific user and the existence position of another person, and the filter coefficient optimizing means comprises an audio signal / background sound signal at the position of the specific user. The filter coefficient is determined so that the ratio of the audio signal / background sound signal is minimized at the position of another person at the maximum.

[0013] Further, the specific user and another person have a wireless card, and the existence position recognizing means recognizes the specific user and another person's existing position by the wireless card. I do.

[0014]

As described above, according to the present invention, sound field control is performed by outputting an audio signal to be transmitted to a user and a background sound signal for masking so as to be inaudible to a person other than the user from a row of speakers. It is. As a result, the audio signal can be clearly transmitted only to a specific user, and the audio signal is masked to other people, so that only the background sound signal (music, natural sound) can be heard.

That is, the noise problem which has conventionally been a problem when using voice or audio media in a small office or room can be replaced with a favorite sound such as comfortable music.

Further, by masking only the confidential matters with the artificially added background sound signal, it is possible to transmit information related to privacy, such as a cash dispenser of a bank or a recitation number of a credit card, by sound. Become.

[0017]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a configuration of a first embodiment of an audio signal output device to which the present invention is applied. As shown, the audio signal output device includes an acoustic data input unit 1, a sound field control unit 2, a position recognition unit 3, an optimization processing unit 4, a plurality of output signal generation units 5 to 5, It is also roughly divided into a plurality of amplifiers 6 to 6 and speakers 7 to 7.

The audio data input section 1 is for receiving an audio signal such as an audio signal or music, and is for receiving a message to be transmitted to a user.
And a background sound input unit 1b for taking in background sounds such as music, birdsong, river babbling and the like.

The sound field control unit 2 performs a filtering process so that an audio signal is maximized and a background sound signal is minimized at a specific user position.
And a minimizing unit 2b.

The position recognizing unit 3 recognizes the position of a user to whom a voice message is to be transmitted and the position of another person, and supplies them to the optimization processing unit 4.

The optimization processing unit 4 obtains a filter coefficient when the sound field control unit 2 performs the maximization and minimization processing based on the position data of the user and other people.

The output signal generation unit 5 synthesizes the audio signal and the background sound signal filtered by the sound field control unit 2 to generate an output signal. These are amplified by the amplifier 6 and then output from the speaker 7. Is output.

Next, the operation of the first embodiment will be described. Figure 2 is an explanatory diagram conceptually showing the operation principle of the present invention, the audio signal transfer characteristic h _a between a speaker and the user 11's and is set to be the maximum. On the other hand, the background sound signal is set as the transmission characteristic h _b between the speaker and the user 11 is minimized.

The control of such directivity is shown in FIG.
As shown in, it can be appropriately adjusted transfer characteristics h ₁ to h _N between the speakers 7-7 and the user 11, to transfer characteristics h _a in FIG. 2, h _b a maximum respectively minimum It is.

Now, the signal (audio signal and background sound signal) x ^P (n) at the position of the user 11 uses the audio signal a (n) and the background sound signal b (n) as sound sources, and a plurality of speakers 7 to 7. 7 is expressed by the following equation (1).

[0026]

(Equation 1)

Here, i is the speaker number (i = 1 to 1)
N) and a '(n) in equation (1) is maximized, b'
transfer characteristics h _a (n) is in order to minimize, may be obtained a h _b.

In FIG. 1, first, the position recognition unit 3 recognizes the position of the user to whom the audio signal is to be transmitted. As a recognition method at this time, for example, a floor mat is arranged on the floor surface, and the position of the user can be known by the load applied to the floor mat. In addition, the user can be photographed by using a camera, thereby detecting the position of the user, and recognizing the position of the user by using an infrared sensor, a wireless card, a laser card, or the like. Furthermore, it is also possible to arrange a plurality of microphones in a room and to know the position of the user based on the direction of the voice uttered by the user.

[0029] Then, the optimization processing unit 4, based on the location data of the user, the transfer characteristics h _a shown in equation (1),
Find the optimal value of h _b .

At this time, there are various methods for finding the optimum value as well-known techniques. However, due to recent advances in simulation technology and computer power, design problems are replaced with search problems while changing parameters sequentially. , The optimum value can be obtained. Thus, by simulation techniques, transfer characteristics h _a, the optimum value of h _b can be easily obtained. Also, have been made many studied analytical sound field control, it is possible to determine transfer characteristics h _a, a h _b in such a deterministic manner.

On the other hand, the audio data input unit 1 is supplied with an audio signal and a background sound signal, and is input to the audio input unit 1a and the background sound input unit 1b, respectively. Of these, the audio signal is a message that is actively transmitted to the user, and may include real-time voice, synthetic voice, and recorded voice. As described above, the background sound signal is desirably music, birdsong, babbling of a river, and the like. Both the audio signal and the background sound signal have a sampling frequency of 20 [kHz] and a quantization bit number of 16 bits. Digital signals can be used. Then, the audio signal and the background sound signal are output to the sound field control unit 2.

[0032] In the sound field control unit 2, transmission characteristics obtained by the optimizing section 4 h _a, based on h _b, the input audio signal, the background sound signals, performs a filtering process. That is, based on the maximum processing unit 2a in transfer characteristics h _a, we obtain an output in order to maximize the audio signal at a location of the user for each speaker 7, and the minimum processing unit 2
In b, an output is obtained for each speaker 7 based on the transfer characteristic h _b in order to minimize the background sound signal at the position where the user exists.

In this filtering process, the amount of calculation increases as the number of speakers 7 increases, and if the length of the impulse response is increased, the convolution operation takes a long time. However, since the calculation in this case is a large amount but a simple formula operation, it is desirable to use a processor capable of executing the product-sum calculation in parallel, a DSP board having a parallel operation unit, or the like.

The output signal generator 5 combines the output signals from the sound field controller 2, and the combined signal is amplified by the amplifier 6 and then output from the speaker 7. Thus the output signal outputted from the speakers 7-7, the audio signals S ₁ at the position of the user 11 and the background sound signal S ₂
Since the ratio S ₁ / S ₂ to the maximum is maximized, the user can hear the message by the audio signal.

As described above, in the first embodiment, each speaker 7 is set so that the audio signal is maximized and the background sound signal is minimized at the position of the user to whom the message is to be transmitted.
7 are adjusted so that the user can recognize the message from the computer.

When the above-mentioned music or natural sound is used as the background sound, a comfortable background sound is mainly heard at a position where another person is present, so that the audio signal does not cause discomfort.

FIG. 4 is a schematic diagram for explaining a second embodiment of the present invention. In this example, the location of the user P _1, and to detect the locations of other people O _1, O _2, on the basis of the present position data, to determine a transfer characteristic h ₁ ~h _N.

Now, the user P ₁ and the others O ₁ and O ₂
When it is desired to transmit a message to the three persons, the transfer function h is set so that the audio signal is maximum and the background sound signal is minimum at all of the existing positions of P ₁ , O ₁ , and O _2.
Seeking ₁ to h _N, it may be adjusted to output of the speakers 7-7. However, in this case, control becomes more difficult than in the case of a single user, so that the S / N ratio of a message received by the user generally decreases.

On the other hand, if a message based on an audio signal is transmitted only to the user P ₁ and it is not desired to hear the message to other persons O ₁ and O ₂ , the audio is transmitted at the point P ₁ . maximum signal, the minimum background sound signal, O _1, minimizing the audio signal at point O _2, may be obtained transfer characteristics h ₁ to h _N in order to maximize the background sound signal.

In this way, when a confidential matter is to be conversed between the user and the computer, the confidentiality is improved without being heard by another person. Further, as shown in FIG. 5, in the audio signal output, when outputting the confidential signal S _3, if caused to superimpose the dummy information S ₄ to the background sound signals, only the third portion of the confidential S ₃ Can be hard to hear.

This is because, for example, when a name, an account number, and a recitation number are output by voice at a bank or the like, it is possible to make this part (confidential matter) inaudible to others. In particular, in various speech rule synthesizing systems such as sentence / speech conversion and concept / speech conversion, such processing can be performed if the importance of security is added to the data contents.

FIG. 6 is a flowchart showing an operation for making only the confidential matter inaudible to other people. First, when the position of a user to whom a message is to be transmitted is recognized (step ST1), a message to be transmitted is given (step ST2), and it is determined whether or not this message is a confidential matter (step ST3). .

If it is a confidential matter (YES in step ST3), the ratio of the audio signal / background sound signal at the position of the user is maximized to make it difficult for others to hear (step ST4). If it is a confidential matter (NO in step ST3), a normal message is output (step ST3).
5). This operation is repeated until the message ends (step ST6).

In the above example, the plurality of speakers 7 to
Although the impulse response of the sound field control between the user 7 and the user P ₁ and the other persons O ₁ and O ₂ is fixed, the introduction of an adaptive control method that converts every moment will introduce more precision. High sound field control becomes possible.

FIG. 7 is an explanatory view showing a third embodiment of the present invention. A plurality of speakers 7 are provided around a room. Each speaker 7 can be controlled by a computer, and can output an audio signal and a background sound signal by the above-described digital signal processing.

Each of the users P _{1 to} P in the room
_4, ID information, i.e., possession of a radio card information and microphone indicating whether the user is who is shark, for example, if you want to talk and the user P ₁ and P ₂ are wiped in this wireless card Conversation is possible with the microphone. That is, the location of each user P _1, P ₂ is recognized by the radio card each user P _1, P ₂ possessed, for example, when the user P ₁ utters into the microphone, the voice is taken into the computer as the audio signal at the position of the user P ₂ is maximum, the audio signal from the speaker 7, and a background sound signal is output.

Therefore, a conversation between the users P ₁ and P ₂ is possible, and the contents of this conversation are different from those of the other users P ₃ and P ₂ .
It will not be asked to P _4. That is, a hands-free telephone system can be configured.

Further, a plurality of microphones can be arranged in the room, and the voices of the users P _{1 to} P ₄ can be detected by each microphone to detect the user who has spoken. That is, if the uttered voice is detected by a plurality of microphones based on the directivity of the microphone, the position of the uttered user can be easily recognized. Such an acoustic signal pickup method is described in, for example, "Measurement of Sound Source Position and Waveform Using Many Sensors: Journal of the Acoustical Society of Japan, Vol.
pp. 268-273, Hitoshi Nagata, Masato Abe, Kenichi Kido ".

Therefore, for example, if the user P ₁ is the user P ₂
When the user P ₂ answers “Yes” to the user, the locations of the users P ₁ and P ₂ are recognized, and the user P 2 is operated by the speakers 7 and the microphone. conversation between ₁ and P ₂ becomes possible. In addition, since the other users P ₃ and P ₄ only hear the background sound, secrecy is ensured.

According to this method, without a wireless card,
Conversations between users can be held hands-free. If comfortable music or the like is selected as the background sound, the efficiency of work and thinking can be improved even in an office factory or the like.

Further, if the above-mentioned example is further extended, even if the user is not in the same room, if the sound field is controlled in each room, a handless voice conversation can be realized via the communication network.

As a method of detecting the position of the user, images captured by a plurality of cameras can be used.
In this case, since the position of the user's ear can be recognized, the sound field control can be performed with higher accuracy.

FIG. 8 is an explanatory view showing a fourth embodiment of the present invention. In this embodiment, sound field control is performed in consideration of the positions of both ears of the user. In the above-described example, the position of the user is indicated by dots for simplicity, but the position of the user is actually represented as a fixed area. Further, it is more difficult to control sound field control when the control position is a region than when it is a point.
Therefore, if the position of the user's ear can be recognized, the control position can be treated as a point, so that a more accurate sound field control can be performed.

[0054] In FIG. 8, the transfer function h _ARR as the sound pressure of the audio signal for the right ear P _R and the left ear P _L of the right speaker 7 users than _R 11 is maximized, determining the h _ARL. Also, the background sound signal, P _R, h _BRR as the sound pressure at P _L is the minimum, determining the h _BRL.

[0055] Then, similarly for the left speaker 7 _L, h _ALR as sound pressure of the audio signal for the right ear P _R and the left ear P _L of the user 11 is maximum, seek h _ALL, background sound signals P _R, as the sound pressure at the minimum P _L h _BLR, seek h _BLL.

As described above, considering the position of the ear,
More accurate sound field control becomes possible, and it is more effective to examine the position of the ear with high accuracy by image processing when the position of the user 11 is roughly determined.

It is also possible to approximately estimate the position of the ear from the direction in which the user 11 is facing. Furthermore, when the user 11 is sitting in a chair or knowing his / her height in advance, it is possible to actively use information on the direction and height of the face to guide the user to easily determine the position of the ear. It is also effective to perform calibration and measure the ear position as accurately as possible so that the S / N ratio is as large as possible.

FIG. 9 is a block diagram showing a fifth embodiment of the present invention. In this example, a speaker for outputting a sound signal is configured separately in a speaker column 7 _B for outputting speaker column 7 _A and the background sound signal to output the audio signal. With this configuration, it is easy to design filter coefficients for maximizing filtering and minimizing filtering.

That is, in each of the above-described embodiments, when designing the transfer characteristic from each spatially fixed speaker 7 to the user position, both the audio signal and the background sound signal are output from all speakers 7. However, they are separate in this example. Thus, the speakers 7 _A to 7-
_A, 7 design of the transfer characteristic h _i between the _B to _7-B and the user 11 is facilitated.

[0060] Also, in the figure, the transfer characteristic h _a for maximizing filtering all speaker column 7 (e.g. 64
Outputs sound speaker column 7 _A of the pieces) (for example, 32) as the target, the transfer characteristic h _b of minimization filtering is performed using the speaker column 7 _B (e.g. 32).

FIG. 10 is a block diagram showing a sixth embodiment of the present invention. In this embodiment, an adaptive algorithm is introduced into the optimization filtering process. That is, in the first embodiment shown in FIG. 1, the control condition is determined from the position where the user is present and the like, and the filter coefficient of the sound field control unit 2 is optimized. However, such a deterministic algorithm cannot cope with changes in the surrounding acoustic transfer function due to the presence of reflections and obstacles, sound absorbing conditions, movement of people or objects, and the like. In this case, an adaptive algorithm is effective.

As shown in FIG. 10, microphones are installed at the target position 21 of the user and the target position of another person, and the sound from each microphone is switched by the switch SW1, and each of the adaptation processing units 13 and 14 is performed. To supply. Then, the adaptation processing unit 13 maximizes the audio signal, and the adaptation processing unit 14 minimizes the background sound signal.

Although schematically shown in the example of FIG. 10, in an actual application, the initial optimization processing is first performed by simulating the acoustic environment including the room, the computer, the speaker, and the user in the computer. After that, it is possible to cope with a complex acoustic environment using an adaptation algorithm. However, in this case, since the position of the microphone and the position of the person do not always match, sound field control that takes this difference into full consideration is possible.

Various adaptation algorithms have been proposed. For example, see "Bernard Widrow, Samuel D. Stearn
s “Adaptive Signal Processing”, Prentice Hall (198
5) ". In particular, chap.13 and ch
ap.14, `` Adaptive Arraysand Adaptive Beam forming
The technique described in "" is used.

[0065]

As described above, according to the present invention, the position where the user to whom the audio signal is to be transmitted is recognized, and at this position the audio signal is maximized and the background sound signal is minimized. Controls signal output. Therefore, the audio signal from the computer can be transmitted only to a specific user, and the background sound can be heard by other people, so that voice conversation can be performed without disturbing other people such as noise.

Further, by recognizing not only the user's location to be transmitted but also the location of another person, and controlling to minimize the audio signal and maximize the background sound signal at the location of the other person, Higher precision control is possible, and the confidentiality between the computer and the user is improved.

As a result, since the media for dialogue with the computer, which has conventionally tended to be visually predominant, can be distributed to the auditory sense, the effect of significantly improving the operability of the computer can be obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram showing the configuration of a first embodiment of an audio signal output device to which the present invention has been applied.

FIG. 2 is an explanatory view conceptually showing the operation principle of the first embodiment.

FIG. 3 is an explanatory diagram illustrating transfer characteristics between each speaker and a user.

FIG. 4 is a schematic view showing a second embodiment of the present invention.

FIG. 5 is a time chart when dummy information is superimposed on a confidential matter portion.

FIG. 6 is a flowchart showing the operation of the second embodiment.

FIG. 7 is an explanatory view showing a third embodiment of the present invention.

FIG. 8 is an explanatory diagram showing a fourth embodiment of the present invention.

FIG. 9 is an explanatory view showing a fifth embodiment of the present invention.

FIG. 10 is an explanatory view showing a sixth embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF REFERENCE NUMERALS 1 acoustic data input unit 2 sound field control unit 3 position recognition unit 4 optimization processing unit 5 output signal generation unit 6 amplifier 7 speaker 11 user

Claims (4)

(57) [Claims] 1. An audio signal output device for outputting an audio signal to be transmitted to only a specific user in an easily audible manner, comprising: an existence position recognizing means for recognizing the existence position of the specific user; Audio signal,
And input means for capturing a background sound signal to be heard by another person, the audio signal, a plurality of speakers that output the background sound signal, and the audio signal and the background sound signal output from each speaker, Sound field control means for determining based on a predetermined filter coefficient, and the filter coefficient for maximizing an audio signal / background sound signal ratio at a specific user position based on the specific user's location information. And a filter coefficient optimizing means for determining the audio signal. 2. The method according to claim 1, wherein the specific user is a plurality of users, and the filter coefficient optimizing means determines the filter coefficient so as to maximize an audio signal / background sound signal ratio at each user position. The audio signal output device according to the above. 3. The existence position recognizing means recognizes the existence position of the specific user and the existence position of another person, and the filter coefficient optimizing means comprises an audio signal / background sound signal at the position of the specific user. 3. The audio signal output device according to claim 1, wherein the filter coefficient is determined so as to maximize a ratio of the audio signal / background sound signal at a position of another person. 4. The specific user and another person have a wireless card, and the existence position recognizing means recognizes the specific user and another person's existence position by the wireless card. 3. The audio signal output device according to 3.