JPS63261948A - Voice storage system for conference communication - Google Patents

Abstract

PURPOSE:To retrieve the voice of a specified speaker only, and to enable to reproduce and hear it, and to dispense with a retrieval code by storing the voice of the speaker while coordinating an identification information to it. CONSTITUTION:The voice from TEL 1a-1h is encoded and inputted to a conference communication trunk CFT 4, and the voice of the maximum level is distributed to all other attendants. This voice signal of the maximum level is inputted to the interface part VMI 51 of a voice storage device VM 5 from the CFT 4 through a channel switch TSW 3, and further, stored in a file VMF 54. Besides, the identification information of a speaking TEL is sent from a central control device CC 6 to the VM 5, and a control circuit 52 records it together with an address, to show a storage area in the VMF 54, in an identification information file VIF 53. Then, when a retrieving number and the identification information of the specified speaker are sent from the TEL, the VMC 52 reads the corresponding address from the VIF 53, and reproduces the voice signal in the said area from the VMF 54, and sends it to the TEL through the TSW 3.

Description

[Detailed Description of the Invention] BiJ Field of Sedentary Work The present invention provides up to Ii! This article concerns the voice storage method for conference calls that uses digital exchanges to support universal conference call trunks.

(q Conventional technology) There is a voice mail service that allows you to send a voice message to a voice storage device and listen to the message later.

When using this, if the voice mail is long, the user has to read the whole thing with his right eye and search in order to open the part he needs.

There are similar drawbacks when storing audio from conference calls.
It was not possible to hear the original story of the speaker of the patent.

As a solution to this problem, a method has been proposed in which voice mails are divided into short mails and a search code is attached to each mail (see Japanese Patent Publication No. 58-57850 for an example).

Problems to be Solved by the Invention However, according to the above method, it is inconvenient because the user has to input the search code for the sending terminal each time. Furthermore, search code upper f! for each speaker! :, the same code must be used in expressions uttered by the same speaker.

) The tth method to solve the problem is up to 1 [l! In a digital switching system equipped with a conference call trunk t' of the 111g4 system, when the highest level of audio related to the conference call is interpreted by a comparison means and the selected nth audio is stored in the audio storage device, the audio is emitted. This is an audio storage method for a conference call in which the identification information of the identified speaker is identified by an identification means, and the identification information of the identified speaker is stored in the 11lr period audio storage device in correspondence with the str audio distribution.

(E) Function Since the identification information of the active speaker is recorded in association with the voice, it becomes possible to search for, playback, and listen to only the voice of a specific speaker. Furthermore, there is no need to input a search code every time the transmitting terminal F speaker speaks.

(f) Example Hereinafter, *a example of the present invention will be explained using the drawings.

FIG. 6 shows the overall configuration of an exchange according to an embodiment of the present invention. Telephone 4 (place) (1 &) ~ (lh)
~ (in) audio is from line circuit tLc) (2m)
〜(2h) 〜(zn) encoded as an audio signal, the state of the 4-channel switch (TSW) (3) 2nd transmission ML, 6° position (1a)〜(lh) 〜(in) is LC(21
~(2h) ~C2n) 'tfr' and central control@device (
CC) (6) is monitoring and TEI.

The identification information (for example, each telephone number) of (l&] to (lh) to (1n) is stored in the memory (b) (7).

Mayu, T8W +37 and conference call trunk (CF)
T) (4) and the audio storage device (VM) (53) are connected to the CG (67K ExI!tlj#).

Let's assume that people from (1m3 to (lh)) are participating in a conference call.The audio from (lj) to (1h) is encoded as tL
The specific audio signals (al) to (h') are input manually to 0FT (4).The conference call method is both direct and direct, that is, the voice of the highest level participant is transmitted to all other participants. This is a method of distributing the amount to

FIG. 1 is a block diagram showing the configuration of OFT (4), which performs the main function of conference calls. Incoming channel (CHI) (
Audio signals (&') to (h') input in parallel from 400) to 407 are sent to a selection circuit (SL).

C) Enter 443. Also, autoratio eIglNcscM)
19 audio levels are detected at (410) to (417). Here, CHI (400) to (407) are TSW (
33KmM Sunte25. T E El (1) ~ (l
h) Communication path shape H; 5°SCM (410)
- (417) are circuits that detect the maximum level of the audio signal input within a certain short time (Δt). Audio signal (
The maximum level of the short-time (Δt) circle (a~(h'') is the conference call side - circuit (CFC) - 0JfJ
[S(CMP)14 requires human power. CMPB is further defined as the largest of (a~(h9)).
Let it be #). CFG (c) is the maximum level (d
A signal 2SLC (421) is sent to select the audio signal (this is referred to as td') to wait for.

Figure 42 (aJtI'i, the relationship between the audio signal, its maximum level, and the selected audio signal) 1c Expresses the following case with only audio signals tC & (h) input to the FT (4). 2 and sound P number <'), <h')t1
The horizontal axis is represented by the vertical @t level and the time (13).Here, the symbols 1, 2, 10, and 20 represent the corresponding times (mut).

Comparing the maximum levels ('i), (bl) of the two audio signals (&ri, (h) at time (ΔtlJ173), the maximum level (a1) is larger from the figure. Therefore, this maximum level (al) is The maximum level of
Time l5fi (th t 1) P3O voice mark (”)
(h′l) is the audio signal (d
Selected as

Also, the audio signal (a') which is added to the time (Δtz) circle,
Comparing the maximum levels (a1] and (h;) of 1'!Q (ΔtZ) p'3 audio signal (h'
) (h'2) is the maximum level (dt) Waiting audio signal (
d').

In the same way, the maximum level (
d "Ha phonetic name (d ri is selected as shown in Figure 41g2 (bJ).

Here, the explanation will be returned to FIG. 1. Based on the signal from CFC (451), 5LC (transfer) outputs the output circuit (Q
Distribute audio signals (d) with maximum levels (d) to C (430) to (437). QC (430) to (437)
) outputs the audio signal (d') based on the power reduction signal from CFC1451.
467). And the audio signal (d') Ugly name 6
TSWt3)k in the figure, TeLC(2aLJ ~ (2h
) No. 14 in, placed (1 &) ~ (lh) n.

For example, when the maximum level of the voice is set to LD, the TgL that emits the 28th level voice,
(For example, let's say the older brother is le.) The voice is sent. In the OVa,
CMP1441 outputs a large level r* at the 2nd-iIth, and outputs one of 0C (430) to (437) (OC (
Only the 433 horn outputs the loudest level sound No. 1 cc')c- at the second W. Other OCs are fIM signal (d')
Output.

The CFCI45 has the highest level of sound P! 1 signal is detected manually by 9 CH operators. The 6th worst CG (67 is CF)
' After receiving the f# message from T+41, a voice signal of the maximum level is input, and a call R1 is formed at iI.
'rxr,, 1 check, M(71J:ri), read the identification information at that location. That is, CFT(4) and C(4
5) constitutes an identification means t'' for distinguishing speakers who have uttered sounds at an amplified level.

The third villain is V M (
FIG. 5 is a block diagram showing the configuration of 5J. The audio signal waiting for the maximum level (d) mentioned above (d is CFT (41 to TSW
(3Jt mediated L te V M (5) 0i y9 yx-i (V
MI) 611. Further, the cough voice signal passes through the voice storage device control circuit ('/MC) 52 and is accumulated in the voice signal 7 isle (''l/MF) (b). Here, the accumulated voice signal is The emitted TEIC)@betsu news is sent from CG (6) to VMT5J.vmcsz is the identification information?, the address indicating the area where the voice signal is stored in the VMF 541 yen mentioned above, and the job information file (VIP ) 63 @ Figure 84, Figure 5 fl, Son ("2tV IF'531
The internal configuration of VMF (b) is shown in the table below. The aforementioned operation? By doing this, in VIF63+, the speaker ('r
EL) identification information is associated with an address indicating the @ area where the voice signal of the speaker of VMF15JF3 is stored. Also, when storing audio signals in the VMF 54), insert an end code each time the speaker changes. This operation is assumed to be automatically performed by the VMC 53.

I will now explain the procedure for searching for audio. When the terminal transmits the search number and the identification information of the scheduled speaker (for example, the telephone number), the third factor 'I/MC 153 reads the address corresponding to the identification information from the VIP 53 and sends the address. Audio signal IVMF (541
Play in z-order. In the example of the fifth factor in FIG.
(α2), (αS]... are read, and the audio signal (a'l) of the area indicated by the address of 7n etc. of the VMF (b) circle,
(a'to) and (a'zo) are played in order. The reproduced voice M is sent to LCK via TSW (3) in Figure 6.
J and others are decoded. In addition, it is also possible to install a function to cancel the identification of the speaker when searching.For example, by using the method described above, a story with four legs until the middle of the story: Listen only to the ``d'' (hosting a conference call, etc.). Then, listen to the gist of the story.
Speaker identification cancellation No. 1 (for example, send a FB signal) from the place where you want to hear the voice a.Then, only the audio signal of a specific speaker is read out. From that point on, the person who accumulates the signals and reads them out in t'L9 order, that is, searches and collects them, can hear not only a particular speaker but also all speakers.In the above example, You can use two files, VIF (G) and VMF (Found), to associate job information and audio.Next, you can use other methods, such as adding identification information at the beginning of each audio signal. However, the area for recording the audio signal may be divided into several areas for each identification information in advance.

(G) Effects of the Invention As is clear from the above description, the present invention provides a method for selecting and listening to only the voices of the speakers in the yen patent in recording the voices of conference calls.
This is a time-saving way to manually generate speaker identification information (it can be provided on the audio storage system of a conference call).

[Brief explanation of the drawing]

Figure 3 shows the relationship between the audio signal manually input by the trunk, its maximum level, and the audio signal selected. Figure 5 shows P3f of the audio signal file.
1S configuration g' is the front factor, and FIG. 6 shows the entire four configurations of the exchange gate according to the example of the present invention as the front block. CL&) ~(ln)--Telephone (place) S (2a)
~(2n)...Line circuit (LC), (37...
Call path switch f (TSW), (4)... Conference call trunk (CFT) - (400) - (407)... Input channel (CHI), (410) - (417)... Self comparison Circuit (8CM, ■...Selection circuit (5LC), (
430) to (437)...output circuit (QC), 144
1... Comparison part (CMP), (c)... Conference call fee g
1 circuit (CFC), (460) to (467)...ff
l? −? *LE (CHO), t57-f voice recording ffl device (VM). 6υ...Interface part (VMI), ■...Voice storage crack *++yuamu c VMC>, (53
i/R 'r'1lN7 Phil (VIF), C541
...Audio signal file (I/MP), (6)...Central ll1I control device (CC), (7J-・Memory (Incorporated)
.

Claims (1)

[Claims] 1. In a digital exchange equipped with a conference call trunk using a maximum value control system, when storing the highest level of the voice related to the conference call in the voice storage device, the speaker who uttered the voice is identified by an identification means, A voice storage method for a conference call, characterized in that identification information of an identified speaker is stored in the voice storage device in association with the voice.