JPS61224647A - Conference talking system - Google Patents

Abstract

PURPOSE:To attain less background noise and to prevent changeover of voice sent to each subscriber from being caused frequently by selecting a subscriber having the maximum voice level not being an effective talker among subscribers sending a voice in the voice presence region and using it as the effective talker when the number of effective talkers does not reach the designated value. CONSTITUTION:A sound volume detecting circuit 7 supervises always the sound volume of each subscriber and informs the result of comparison with a threshold value of presence/absence voice designated from a microprocessor 8 in advance to the microprocessor (or the result is read by the microprocessor 8). Further, the round volume detection circuit 7 itself stores the presence/absence of voice at each subscriber and starts a timer corresponding to the subscriber provided to a timing circuit 9 as to the subscriber changing from the voice presence region into the voice absence region. The microprocessor 8 uses the information from the sound volume detection circuit 7, the information from the timing circuit 9 and the information stored in a memory 10 so as to discriminate an effective talker, a selection circuit 4 is drived, the voice signal of the effective talker is sent as it is to an adder circuit 5 and the voices of the other subscribers are not sent.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Application of the Invention] The present invention relates to a conference call/system in which calls are made by multiple parties at the same time.

[Background of the invention]

In a conference call, multiple speakers may speak at the same time.
It is necessary to determine and control which of the outputs of the microphones that collect the voices of each speaker should be transmitted to others, and this method is called the voice addition method. The conventional addition method is "Conference call method in digital exchanges" written by Konyo et al.
．． 1652), there are the following methods.

(a) N-+ addition method (C) Maximum value + 1 addition method (α) When there are 8 participants in a conference, all the voices are added, and the voice of a certain person is subtracted from it. f
Vj stop) This is a system in which the voice is sent to the person, and there are problems such as ψ bias due to background noise such as room noise. Method (j) is a method in which the volumes of conference participants are compared at regular intervals, the subscriber with the highest volume is regarded as the effective speaker, and the voice is sent to all the subscribers, and (C ) is a method in which the person with the highest sound and the next highest sound and sound are considered to be effective speakers, and the sum of the four results is sent to the Zen Ministry. (Although both (C) and (C) improve the drawbacks of (α), the subscriber whose volume reaches the maximum value changes frequently when two people or more than 5 Å speak at the same time, or when no one speaks. There is a risk that the noise caused by speaker switching may be unpleasant to the ears.

[Purpose of the invention]

SUMMARY OF THE INVENTION An object of the present invention is to provide a conference call system in which background noise is low and audio signals sent to each subscriber are not frequently switched.

[Summary of the invention]

The present invention monitors the voices of subscribers participating in a conference at regular intervals and compares them with a preset threshold for determining voice/silence, and for subscribers within one silent area, It is not recognized as a new valid speaker, and when the voice of an effective speaker changes from a voice area to a silent area, the voice is recognized as valid speech as long as the voice remains in the silent area within a predetermined time. The system is characterized in that the subscriber is regarded as a subscriber and the voice is transmitted to each subscriber.

[Embodiments of the invention]

Hereinafter, the present invention will be explained in detail using examples. 1 and 2 show an embodiment of the present invention, with FIG. 1 being a system configuration diagram and FIG. 2 being a flowchart of a program for valid speaker determination. In FIG. 1, voices from a subscriber attending a conference are converted into digital signals through a transmitter of a telephone 1 and an intabace circuit 2, and are multiplexed by a multiplexer 5.

This output is input to the volume detection circuit 7 and the selection circuit 4, and the volume 1 output circuit 7 constantly monitors the volume of each subscriber and compares the volume with the sound/silence threshold specified by the microprocessor 8 in advance. Notify microprocessor 8 (or read this result)
. In addition, the volume detection circuit 7 itself also stores the presence/absence of sound for each subscriber, and starts a timer corresponding to the subscriber provided in the timing circuit 9 for a subscriber who changes from a sound area to a silent area (no sound). For a subscriber who changes from the area to the active area, the timer is stopped and the false timer value is also reset). The microprocessor 8 receives information from the volume detection circuit 7, information from the timing circuit 9, and memory 10.
The effective speaker is determined by a method to be described later based on information stored in the system, and the selection circuit 4 is driven to send the voice signal of the valid speaker as it is to the adder circuit 5, while the voice signals of other subscribers are not transmitted. Instead, send the silent pattern "0". The adder circuit 5 adds the voice signals of these valid speakers and distributes them to each subscriber, and subtracts the voice signal of the valid subscriber himself for those to be sent to the valid subscriber (the signal for the subtraction is The component designation information is sent by the microprocessor 8 to the adder circuit 5.The output is level-controlled by the level control circuit 6, sent to each subscriber by the demultiplexer 5, and transmitted to each telephone set 7 via the interface circuit 2. is sent to the handset, and a conference call is realized.

Next, a method of determining an effective speaker mainly performed by the microprocessor 8 will be explained. FIG. 2 is a 70-chart showing a 6-finger program for this purpose, and this program is executed every fixed period of time, for example, every 32 milliseconds. In the same figure, steps 100 to 105 are the parts for determining whether or not the subscriber who was determined to be an effective speaker in the previous cycle is also an effective speaker in this cycle.
(determined as 0) The subscriber who is determined to have a voice in this cycle as well (determined in step 101) is designated as a valid speaker in this cycle in step 105. The subscriber who is determined to be silent in this cycle in step 101 moves to step 102, and if the time after the change from sound to silence is less than a certain tKK (for example, 0.5 seconds), the subscriber also moves to step 105 and moves to step 105 in this cycle. is also an effective speaker,
If the above time is greater than or equal to the above time, the subscriber is treated as an invalid speaker and the number of valid speakers is decreased by 1 (step 104). When the above operations are performed for all subscribers determined to be effective speakers in the previous cycle, the process moves from step 105 to step 106. Note that in the above, step 101 is performed.

The sound/silence determination and the elapsed time from the time of silence at step 102 are determined based on the information from the volume detection circuit 7 and timing circuit 9 shown in FIG. 1, respectively.

In the process described above, if there is a silent area for a certain period of time among valid speakers, that person is treated as an invalid speaker, so the number of effective speakers may decrease but will not increase. do not have. On the other hand, steps 106 to 110 are a part in which a subscriber to be newly designated as a valid speaker is selected from invalid speakers in this cycle, and the number of valid speakers at that time is the specified value No (for example, N
0 = 3), that is, when the number of valid speakers is -1 at system startup or in step 104, the process proceeds from step 106 to step 107 and selects the subscriber with the highest audio level from among the invalid speakers. do. Next, it is determined in step 108 whether or not the subscriber is active, and if the subscriber is active, the subscriber is designated as a valid speaker in step 109, the number of valid speakers is increased by 1 in step 0, and the process returns to step 106. Return.

When the number of effective speakers reaches No or there is no new active subscriber, the process moves from step 106 or 108 to step 111, and the subscriber determined to be an effective speaker in this cycle is processed by the first microprocessor 8. The selection circuit 7 and the addition circuit 5 are notified, and the above-mentioned addition and subtraction of the audio is performed, and the memory 10
The subscriber determined to be a valid speaker is stored.

As shown in the above example, first of all, until now, most of the invalid speakers are silent even if they have the maximum fi, in other words, they are not considered effective speakers unless they exceed a certain threshold. Unnecessary speaker switching due to ground noise etc. does not occur, and since the number of effective speakers is limited as in the conventional system, no complications in conversation occur. Furthermore, background noise itself is not mixed in. In addition, even if a subscriber who has been recognized as an effective speaker becomes silent, the subscriber remains an effective speaker unless this state continues for a certain period of time (so, if two or more than 3 people speak at the same time, Priority is given to the subscriber who was currently speaking, and this also prevents frequent speaker switching.

It should be noted that although there is a method of realizing the sound/silence determination function of the volume detection circuit 7 and the timing circuit 90 function of the sound volume detection circuit 7 shown in FIG. [Effects of the Invention] As explained above, according to the present invention, it is possible to eliminate background noise and realize an easy-to-use conference call in which unnecessary speaker switching does not occur frequently.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a system configuration showing an embodiment of the present invention, and FIG. 2 is a flowchart showing an example of an effective speaker determination program executed by the microprocessor in FIG. 1... Telephone 2... Interface circuit 5... Multi, demultiplexer 4... Selection circuit 5... Addition; Itching circuit 6... Level controller 7... Volume detection circuit 8... Microprocessor 9...timing circuit 10...memory-1

Claims (1)

[Claims] In a conference call system in which calls are made between one, three or more subscribers at the same time, the volume of each subscriber is detected at regular intervals, and the sound volume is detected in a sound area where the volume is higher than a set threshold. Is there?
Providing a volume determining means for determining whether the sound is in a silent region smaller than the threshold value, and a limiting means for restricting the number of effective speakers whose sound is transmitted to all subscribers to a predetermined value or less, If the number of effective speakers is less than the specified value, the subscriber with the highest audio level is selected from among the subscribers who are not effective speakers and who send audio in the active area, and the subscriber is selected as an effective speaker. A conference call method characterized by: 2. A timer is added to measure the elapsed time after the voice level of each subscriber changes from a sound area to a silent area, and the elapsed time is predetermined even if the voice of the effective speaker is in the silent area. If you speak within the set time, you will be considered a valid speaker.
In addition, if the voice of the effective speaker becomes a sound region again while measuring the elapsed time, the elapsed time measured up to that point is reset and the speaker is considered to be the effective speaker. Conference call method described.