JPH0287861A - Echo erasing device for multi-spot conference - Google Patents

Abstract

(57) [Abstract] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] "Industrial Application Field" This invention relates to an echo canceling device for multi-point conferences that eliminates indoor echo signals that cause howling and cause hearing impairment in multi-point conferences. It is.

``Prior Art'' With the spread of voice conferencing, it is desired to provide a public address communication device that has excellent simultaneous call performance and reduces echo interference.

There is an echo cancellation device that satisfies this requirement.

Figure 3 shows an example of the configuration of a multipoint communication conference system.
It shows the ground communication conference system. Ground A microphone 1
The audio signal input from the ground B is amplified by the speaker 6 at the ground B and the speaker 10 at the ground C, and similarly, the audio signal input from the microphone 5 at the ground B is amplified from the speaker 11 at the ground C and the speaker 2 at the ground A. The audio signal input from the microphone 9 at the ground C is amplified by the speaker 3 at the ground A and the speaker 7 at the ground B, thereby conducting a multi-point communication conference. Here, 4
, 8.12 are echo cancellers for multi-point conferences, which connect speakers 2 and 3 of ground A to microphone 1, speakers 6.7 of ground B to microphone 5, and speaker 1 of ground C.
0.11 to the microphone 9 is eliminated, making it possible to realize a natural loudspeaker call without howling or echo.

Figure 4 is a block diagram showing an example of a conventional echo cancellation device for multipoint conferences. This shows a case where there is one transmitting channel. An example of this type of structure is the IEICE Transactions '86/10 VOI.

J69-A No. 1O' multi-channel adaptive digital filter. In FIG. 4, reference numeral 13 represents an echo canceling device for multi-point conferences, which corresponds to 4, 8° 12 in FIG. 14. ．． 14□2..., 14x is a receiving channel, 151.15t,... 15° is a speaker, 16 is a microphone, 1717□,...
・,17. There are two pseudo-echo paths, 1 is an input 1 output adder, and 19 is a subtracter. In this construction method [(
The successive correction of the 12 echo paths is performed as follows.

The input signal at one point in time to channel 14 is
m (t), and the number of taps of the first channel is NK, then the input signal vector Xx(t) of the first channel is X x (t) = (x m(t) + x h(t-
1) + ”・+ x t+(t-NK+1))”k=1
．． 2. ..., K T: can be expressed as vector transposition. If we define a vector obtained by further combining this vector Xg(t) as the input signal vector X(t) of the echo cancellation device for multipoint conferences, then
It is expressed as T(LL X z"(t), -, X K"(L))". The i-th coefficient of the pseudo echo path 178 of the channel corresponding to the input signal is hk.

Then, the pseudo echo path of the channel h * (L)
is 1”l K(t)-(hh, +(t), hh,z(t
), ...+11, is expressed as □(L))''. If we define the vector that combines these as the pseudo-reflection path h (L) of the echo cancellation device for multipoint conferences, we get h (t) =(h +”(L)+h z”(t
), .
)・X (t)/X (t)TX (t)y D) =
h T(t) ・X(t)e (t) = d (t
) y (t) d (t): Can be determined as the true room echo signal.

With the above configuration, it is possible to simultaneously cancel the room echo signals generated by the K received signals.

``Problem to be Solved by the Invention J'' However, since the same number of pseudo echo paths 17. are required as the number of channels, the hardware scale is twice that of monaural, and the convergence speed is also significantly reduced. was there.

The present invention has been made in view of the above problems, and an object of the present invention is to provide an echo canceling device for multipoint conferences that has a small hardware scale and a high convergence speed.

"Means for Solving the Problem" This invention has m (1≦m<K, the number of channels) pseudo-reverse channels, and selects m reception channels from the one with the highest reception signal level. and a memory circuit for erasing the room echo signal and storing the transfer characteristic of the echo path of the K channel, and in conjunction with the selection of the receiving channel, the transfer characteristic stored in the memory circuit is erased. It is characterized by replacing it with a pseudo echo path.

Since the present invention is configured as described above, it is possible to reduce the hardware scale and at the same time improve the convergence speed.

``Example'' In multi-point communication conferences, it is rare for all destinations to transmit at the same time, and usually one or several destinations transmit,
The remaining destinations will receive the call.

This means that it is rare for reception signals to arrive from all of the reception channels, and it is usually sufficient to cancel the room echo signals of one or several channels. This fact is used to reduce the hardware scale and improve convergence speed.

FIG. 1 shows one embodiment of the present invention in the case of canceling the room echo signal of one receiving channel.
Parts corresponding to those in the figure are given the same reference numerals. In FIG. 1, 20 is a switch for selecting one of the receiving channels and connecting it to the pseudo echo path 21, 21 is the pseudo echo path, and 22 is a memory circuit 23. a switch for selecting one from and connecting it to the pseudo echo path 21; 23.
.

23□,...23. 24. A storage circuit for storing the transfer characteristics of the pseudo echo path; 24, a receiving channel selection circuit connected to the receiving channels, detecting the maximum level and controlling the switches 20 and 22; 25. .

25□,...,25. are loss insertion circuits inserted in each receiving channel.

Next, its operation will be explained. The reception channel selection circuit 24 compares the signal levels of the K reception channels and determines the channel with the maximum signal level, and then switches the switch 2.
0 to connect the channel with the maximum signal level to the pseudo echo path 21, and at the same time, operate the switch 22 to connect the pseudo echo path transfer characteristic storage circuit 23 corresponding to the channel with the maximum signal level. 21 is connected, and the transfer characteristic stored in the storage circuit 23° is replaced with the pseudo echo path 21. By inserting losses in the loss insertion circuit 25 to channels other than the channel with the maximum signal level, it is possible to reduce the indoor noise coming from each channel.

When the channel with the maximum signal level changes due to a change in the transmitting destination, the transfer characteristics of the pseudo echo path 21 are first stored in the pseudo echo path transfer characteristic storage circuit 23. , the switch 20 is operated to connect the channel with the new maximum signal level to the pseudo echo path 21, and at the same time, the switch 22 is operated to connect the pseudo echo path transmission corresponding to the channel with the maximum signal level. The pseudo echo path 21 is connected to the characteristic storage circuit 23, and the transfer characteristic stored in the storage circuit 23A is replaced with the pseudo echo path 21.

When receiving signals arrive at two or more receiving channels, the loss insertion circuit 25g inserts loss into channels other than the channel with the maximum signal level, thereby eliminating howling and echoes caused by channel signals other than the channel with the maximum signal level. prevent.

FIG. 2 shows an embodiment of the present invention in the case of canceling room echo signals of m receiving channels (2≦m<K), and the parts corresponding to FIGS. 1 and 4 are are given the same symbol. In FIG. 2, 261. ..., 26
．． 27. a switch for selecting m out of the number of receiving channels; 27. ,...271 are m pseudo echo paths, 2
8 is an m-input l-output adder, 291. . . , 29m is a switch for selecting m memory circuits 231 from the memory circuits 231.

Next, its operation will be explained. The reception channel selection circuit 24 compares the signal levels of the K reception channels and determines the m channel from the one with the highest signal level, and then
The switch 261 is operated to select the m channel and the pseudo echo path 27.m channel from the one with the highest signal level. At the same time as connecting switch 29. is operated to connect the pseudo echo path transfer characteristic storage circuit 23g and the pseudo echo path 27 y corresponding to the m channels in descending order of the signal level. 271. By inserting a loss in the channels other than the m channels having the highest signal level using the loss insertion circuit 25, it is possible to reduce the indoor noise coming from each channel.

When the m channel changes from the one with the highest signal level due to a change in the destination, the pseudo echo path 27. After storing the transmission characteristics in the pseudo echo path transfer characteristic storage circuit 238 corresponding to the m channel from the one with the highest signal level immediately before, the switch 26□ is operated to store the m channel and the pseudo echo corresponding to the new maximum signal level. Road 2
7. At the same time as connecting switch 29. are operated to select the pseudo echo path transfer characteristic storage circuit 23 corresponding to the m channels from the one with the highest signal level. and pseudo anti-G path 27
The transfer characteristics stored in the storage circuit 23 are transferred to the pseudo echo path 27. Replace with

When a received signal arrives at m+1 or more receiving channels, a loss insertion circuit of 25 degrees inserts a loss into channels other than m from the one with the highest signal level, so that the signal from the highest signal level is inserted into the channels other than m. Prevent howling and echo caused by channel signals.

"Effects of the Invention" As explained above, there are m (1≦m<K, the number of channels) pseudo echo paths, and the m reception channels with the highest reception signal level are selected and the indoor K The hardware scale can be reduced compared to the case of using multiple pseudo-echo paths, and at the same time, the convergence speed can be improved.

[Brief explanation of drawings]

Fig. 1 is a block diagram showing an embodiment of the present invention in the case of canceling the room echo signal of one reception channel, and Fig. 2 shows the cancellation of the room echo signal of m reception channels (2≦m<K). FIG. 3 is a block diagram showing an example of the configuration of a multipoint communication conference system (3 sites), and FIG. 4 is an example of a conventional echo cancellation device for multipoint conferences. FIG. Patent applicant Nippon Telegraph and Telephone Corporation

Claims (1)

[Claims] (1) A pseudo echo path is generated by sequentially estimating the impulse response of the echo path from the multi-channel received signal and the echo signal after passing through the echo path of the received signal, and the multi-channel received signal is In an echo cancellation device for multipoint conferences that cancels the echo signal by subtracting from the echo signal a pseudo echo signal obtained by inputting the pseudo echo path, m (1≦m<K, K is the number of channels ), select m reception channels from the one with the highest reception signal level and cancel the room echo signal,
K that stores the transfer characteristics of the echo paths of K channels, respectively.
1. An echo canceling device for a multipoint conference, comprising a memory circuit, and replacing the transfer characteristics stored in the memory circuit with a pseudo echo path in conjunction with the selection of the reception channel.