JPH07123157A - Voice input device for multi-spot conference - Google Patents

Abstract

PURPOSE:To provide the voice input device for multi-spot conference where the noise component due to the environmental noise in a conference room included in a synthesized voice is suppressed. CONSTITUTION:An output control part 301 controls output of the voice signal of a speaker and the environmental noise in the conference room. An environmental noise extracting part 302 extracts the environmental noise in each conference room from the output of the output control part 301 before the start of conference. An inverted signal generating part 303 inverts the polarity of the noise signal extracted by the environmental noise extracting part 302 to generate the inverted noise signal. An inverted signal storage part 304 stores the inverted noise signal after sampling it with a prescribed period. A voice synthesizing part 305 electrically synthesizes the inverted noise signal read out from the inverted signal storage part 304 and the conference voice including the environmental noise signal from the output control part 301. A noise control part 306 changes and holds the phase of the sampled inverted noise signal from the inverted signal. storage part 304 so that the output level of the voice synthesizing part 305 is lowest.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice input device for multipoint conferences, and more particularly to a multipoint voice input device for inputting voices in each conference room through a communication network between conference rooms at three locations. The present invention relates to a point conference audio input device.

[0002]

2. Description of the Related Art A conventional speech synthesis method in a multipoint conference inputs a microphone for voice input and a speaker for voice output to each conference room in which a conference attendee is present, and a voice from the microphone is input to a switching network side. It was done by installing a voice input device. Further, in an image conference system that requires video, a television camera and a monitor television are added to each conference room.

Then, the voice from the voice input device in each conference room is input to the multipoint conference voice synthesizer via the exchange network and synthesized.

Regarding the image system, although the image signal from each conference room is sent to and distributed to other conference rooms, the monitor TV screen is divided into a plurality of screens, or an arbitrary selection is instructed. Images are also sent and distributed to each conference room.

On the other hand, regarding the voice system, a voice synthesizing method for preventing the wraparound due to the return of the voice to the speaker itself is adopted, and in this voice synthesizing, the sum component of the voices of the attendees of the conference is the speaker. It was sent and distributed to all attendees except myself.

FIG. 3 is a block diagram showing an example of the configuration of a voice synthesis section in a general multipoint conference system.

Referring to FIG. 3, the multipoint conference system includes a plurality (N rooms) of conference rooms 31, 32, ..., 3 with attendees.
N is connected to the multipoint conference speech synthesizer 1 through the switching network 2.

The voices v1, v2, ..., VN from the voice input devices of the respective conference rooms 31, 32, ..., 3N are input to the multipoint conference voice synthesizer 1 through the switching network 2, and the multipoint conference voice synthesizer 1 For one conference room, only the voice of the speaker himself in that room is subtracted, and the sum component of the voices from all other conference rooms is transmitted and distributed.

For example, although the speaker's upstream voice v1 is output from the conference room 31 to the switching network 2 side, the downstream voice from the multipoint conference voice synthesizer 1 to the conference room 31 is not included in the voice v1. The sum components of the sounds v2, ..., VN from the chambers 32, ..., 3N are output. The transmission and distribution of the upstream and downstream voices are the same for the other conference rooms 32, ..., 3N.

The voice synthesis method for preventing the wraparound due to the return of voice to the speaker itself is applied to either analog voice or digital voice.

[0011]

In this conventional multipoint conference speech synthesizer, all the conference rooms except the conference room in which the sum of the voices of the attendees of the conference is the speaker themselves in the speech synthesis. Sent to and distributed to.

Considering the environment of the conference, the voice from one attendee (speaker) is distributed to all the other attendees, but the voice from the speaker is the conference with the speaker. Environmental noise in the room (noise generated by an air conditioner, etc.) is included and transmitted.

Now, the number of conference rooms with attendees N = 8
In the case of the image / speech conference in the case of, the image signal from one conference room and the voice signal v of the speaker are transmitted including the environmental noise n of the image / speech in the room, and the multipoint conference speech synthesizer speaks. For a person who has a voice signal from another attendee excluding the voice signal v from this speaker, and the presence of another attendee 7
It provides the environment noise of the two conference rooms with voice synthesis.

Therefore, in order to simplify the explanation, assuming that the level of the environmental noise in each conference room is uniform, the environmental noise level is about 7 times the level of the environmental noise n in the conference room in which the speaker is present. The synthesized voice including "" will be distributed to the speakers.

As a matter of course, when the number of conference rooms N in which attendees are present is small, this synthesized noise is so small that the conference is not hindered. There was a problem that the voice became difficult to hear.

On the other hand, with respect to the image signal, since the image noise corresponding to the environmental noise in the audio signal is the noise mixed or induced in the image equipment and the transmission path system of the image signal, the noise level is very low, It is not noticeable due to the characteristics of the system.

Therefore, an object of the present invention is to provide a voice input device for a multipoint conference in which a noise component due to environmental noise in a conference room included in a synthesized voice is suppressed.

[0018]

According to the present invention, in a multipoint conference voice input device for conducting a video or audio conference between conference rooms at least at three points through a communication network, environmental noise components in each of the conference rooms. There is provided a voice input device for multipoint conference, which is provided with an environment noise canceling means for canceling.

The environmental noise canceling means extracts an environmental noise in each of the conference rooms before the start of the conference, and an inverted noise signal obtained by inverting the polarity of the environmental noise signal extracted by the environmental noise extracting unit. An inverted signal generation unit for generating
An inverted signal storage unit that stores the inverted noise signal after sampling it at a predetermined cycle, and a voice synthesis unit that synthesizes the sampled inverted noise signal from the inverted signal storage unit and the conference voice including the environmental noise. And the phase of the sampled inverted noise signal from the inverted signal accumulating unit is changed so as to minimize the output level of the voice synthesizing unit, the phase amount of the minimum point is held, and the conference including the environmental noise is held. A voice input device for multipoint conference, comprising: a noise control unit that controls which of the noise extraction unit and the voice synthesis unit the voice is output to.

Further, the voice synthesizer receives the conference voice including the environmental noise as a first input, and outputs the sampled inverted noise signal accumulated in the inverted signal accumulator as a second input.
Input, and these first and second inputs may be combined electrically or acoustically.

Furthermore, the voice synthesizer may be constituted by a differential amplifier for extracting and amplifying only the difference signal between the first and second inputs.

[0022]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of a voice input device for multipoint conference according to the present invention, and FIG. 2 is a waveform diagram showing input and output waveforms of the voice synthesizing section in FIG.

Referring to FIG. 1, the audio input device 3 for multipoint conference of the present embodiment has the conference rooms 31, 32, 32 shown in FIG.
.., 3N, and an output control unit 301 that controls output of a voice signal (conference voice) from a speaker input from a microphone (not shown) and environmental noise in the conference room.
An environmental noise extraction unit 302 that extracts the environmental noise in each conference room at multiple points from the output of the output control unit 301 before the conference starts;
Inversion signal generation unit 3 that generates an inversion noise signal in which the polarity of the environment noise signal extracted by this environment noise extraction unit 302 is inverted.
03, an inversion signal accumulating section 304 that accumulates the inversion noise signal from the inversion signal generating section 303 after sampling at a predetermined cycle, and inversion after sampling read from the inversion signal accumulating section 304 through the inversion signal generating section 303. A voice synthesis unit 305 for electrically synthesizing a noise signal and a conference voice including the environmental noise signal from the output control unit 301;
A noise control unit 306 that performs control to select either the sampled inverted noise signal read from the inverted signal storage unit 304 or the inverted noise signal extracted by the environmental noise extraction unit 302.

Further, the noise control unit 306 is the voice synthesis unit 30.
Inversion signal accumulator 30 so that the output level of 5 is minimized.
4 changes the phase of the sampled inverted noise signal to hold the phase amount of the minimum point, and sends a control signal to the output control unit 301 to input the input from the microphone to the noise extraction unit 302 or the voice synthesis unit 305. Control which is output.

Based on the control signal from the noise control unit 306, the output control unit 301 supplies the conference voice containing the original environmental noise to the + input of the voice synthesis unit 305, and the inverted signal generation unit 303 the inverted signal accumulation unit 303. The sampled inverted noise signal stored in 304 is read out, and the speech synthesis unit 305
To the negative input, and the speech synthesizer 305 takes out only the difference signal of both inputs and cancels the noise component.

Next, the operation of this embodiment will be described.

First, before the conference is started, the noise control unit 306 is provided so that the environmental noise extraction unit 302 can extract the noise component input from the microphone (not shown) in each conference room.
Operate the voice control switch (not shown) inside.

The noise control unit 306 is set to a scrambler mode for reducing the influence of noise extracted by the environmental noise extraction unit 302 by operating the voice control switch, and sends a control signal to the output control unit 301. .

With this control signal, the environmental noise extraction unit 30
Reference numeral 2 extracts environmental noise from the output of the output control unit 301 and inputs it to the inverted signal generation unit 303.

The inverted signal generator 303 inverts the polarity of the input environmental noise signal, samples the inverted noise signal at a predetermined cycle, and then stores it in the inverted signal storage 304.

The inverted signal storage unit 303 is the inverted signal storage unit 3
The sampled inverted noise signal stored in 04 is read out at a predetermined cycle and relayed to the negative input of the voice synthesis unit 305 as a noise signal source for comparison.

The noise control unit 306 is an inverted signal generation unit 303.
Through the voice synthesizing unit 30 by changing the read phase of the inverted noise signal after sampling from the inverted signal accumulating unit 304 through
The phase at which the combined output level of 5 is minimized is detected, and this phase amount is held.

Once the noise control unit 306 holds this phase amount, it sends a control signal to the output control unit 301 to prohibit the environmental noise from being input to the environmental noise extraction unit 302 again, and completes the scramble operation.

After the completion of this scramble operation, when the voice control switch in the noise control section 306 is returned to the call mode, the noise control section 306 sends a control signal to the output control section 301 to send the conference voice (including the environmental noise). 2) is supplied to the + input of the voice synthesis unit 305.

On the other hand, the inverted noise signal (shown in FIG. 2) stored in the inverted signal storage unit 304 is input to the negative input of the voice synthesis unit 305 according to the phase previously held in the noise control unit 306. It continues to be supplied via the generation unit 303.

The voice synthesizer 305 is a differential amplifier which amplifies only the difference between two inputs (+ input and −input) with a predetermined amplification degree.

As shown in FIG. 2, since the environmental noise in the + input and the inverted noise signal in the − input have the same waveform amplitude and phase and opposite polarities, the speech synthesis unit 305 cancels the environmental noise component. And outputs only the conference audio.

The output of the voice synthesizing unit 305 is 1 for multipoint conference.
The voice of the conference room is input to the multipoint conference voice synthesizer 1 (shown in FIG. 3) via the exchange network 2 (shown in FIG. 3).

FIG. 1 shows each conference room with attendees of a multipoint conference.
If the multipoint conference voice input device 3 shown in FIG. 2 is provided, only the conference voices from the conference rooms 31, 32, ..., 3N are input to the N input terminals of the multipoint conference voice synthesizer 1. Therefore, the level of the environmental noise EMN = (N-1) n transmitted and distributed to any one conference room is very small because the level of the environmental noise n is almost zero.

The environmental noise EMN is dominated by the level of the environmental noise n in the conference voice output from each conference room in which attendees are present. According to the present embodiment, the environmental noise EMN should be extremely small. You can

In the present embodiment, it is assumed that the voice synthesis is performed electrically. However, the sound signal is acoustically scrambled by using a speaker or the like by a similar method, and the acoustic signal is reached when the acoustic level becomes the minimum. Can be converted into an electric signal, stored in the inverted signal storage section, read out, and speech-synthesized. The same effect can be obtained.

[0043]

As described above, the present invention cancels out environmental noise components in each conference room in a multipoint conference voice input device for performing a video or audio conference between conference rooms at N points through a communication network. As a result, the level of the environmental noise EMN transmitted and distributed from the multipoint conference speech synthesizer to each conference room can be reduced to almost zero, which has the effect of significantly improving the intelligibility of the conference voice. .

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of a voice input device for multipoint conference according to the present invention.

FIG. 2 is a waveform diagram showing input and output waveforms of a voice synthesis unit in FIG.

FIG. 3 is a block diagram showing an example of a configuration of a voice synthesis part in a general multipoint conference system.

[Explanation of symbols]

 1 Multipoint conference voice synthesizer 2 Switching network 3 Multipoint conference voice input device 31, 32, ..., 3N Conference room 301 Output controller 302 Environmental noise extractor 303 Inverted signal generator 304 Inverted signal accumulator 305 Speech synthesizer 306 Noise control unit

Claims (4)

[Claims] 1. A multi-point audio input device for performing a video or audio conference between conference rooms at least at three points through a communication network, comprising environmental noise canceling means for canceling environmental noise components in each of the conference rooms. A voice input device for multipoint conference, characterized by: 2. The environmental noise canceling means extracts an environmental noise in each of the conference rooms before the conference starts, and
An inversion signal generation unit that generates an inversion noise signal in which the polarity of the environment noise signal extracted by the environment noise extraction unit is inverted, an inversion signal accumulation unit that accumulates the inverted noise signal after sampling it at a predetermined cycle, and A voice synthesizing unit for synthesizing the sampled inversion noise signal from the inversion signal storage unit and the conference voice including the environmental noise, and an output from the inversion signal storage unit so that the output level of the voice synthesis unit is minimized. Noise control that changes the phase of the inverted noise signal after sampling to hold the phase amount of the minimum point and controls which of the noise extracting unit and the voice synthesizing unit outputs the conference voice including the environmental noise. The multi-point conference audio input device according to claim 1, further comprising: 3. The speech synthesizer receives the conference speech including the environmental noise as a first input and the sampled inverted noise signal stored in the inverted signal storage as a second input. The voice input device for multipoint conference according to claim 2, wherein the first and second inputs are combined electrically or acoustically. 4. The voice input for multipoint conference according to claim 2, wherein the voice synthesizer is a differential amplifier that extracts and amplifies only the difference signal between the first and second inputs. apparatus.