JPS6046625A - Echo canceller - Google Patents

Abstract

PURPOSE:To increase an echo suppressing effect and to reduce power consumption by deriving an approximate form to the trasmission function of an echo cource and constituting a circuit generating said transmission fuction by an RC primary circuit, a multiplier and an adder. CONSTITUTION:An input signal X(t) is inputted to a course passing through the multiplier 12-1, the RC primary circuit network consisting of a resistor 10 and a capacitor 11 and a course passing through a multiplier 12-1 and outpus of respective courses are added by an adder 13 to generate a false echo Y'(t). After being inverted at the phase, the false echo Y'(t) is added to an echo Y(t) by an adder 5 and erased. The zero point of the transmission function of an echo canceller 18 can be made to coincide with the zero point of the transmission function of the echo course by controlling the application of tap factors d0, D1.

Description

DETAILED DESCRIPTION OF THE INVENTION (Technical Field) The present invention relates to an echo canceller that cancels signals leaking from a 4-wire side transmitting section to a receiving section generated in a 2-wire to 4-wire conversion section in a dinotal subscriber line bidirectional transmission system. .

(Prior Art) Recently, various devices have been digitized due to dramatic advances in integrated circuit technology. For telephone subscriber lines, various fields are considering a transition from the conventional analog transmission system, which mainly uses the voice band (4 kHz band), to a high-speed digital transmission system that aims to provide a wider range of services. FIG. 1 shows a block configuration of a conventional digital subscriber line bidirectional transmission system.

In the figure, 1 is the 4-wire side input end, 2 is the 4-wire side output end, and 3 is the 4-wire side input end.
4 is a hybrid circuit, 4 is a pseudo echo generating circuit of an echo canceller, and 5 is an adder. The echo leaking from the input terminal 1 on the 4-wire side to the output terminal 2 and the output of the pseudo echo generating circuit 4 are inverted in phase and added by an adder 5, thereby forcibly canceling the echo. FIG. 2 shows a pseudo echo generating circuit of a conventional echo canceller.

In the same figure, 7-1+ 7-2 y...l 7-(N-
1) is a delay element, 8-1°8-2. ..., 8-N is a coefficient. , hl, . . h(N-1), and 9 is an adder. The 4th line side transmission signal x(t) is transmitted through the delay element 7-1
and are sequentially input to delay elements 7-2. ....

7-(N-1), and at the same time passes through the taps coming out between each delay element and the corresponding multiplier 8.
-1. . . , 8-N, each coefficient hk (k=0 , ., N-
1) is multiplied and the result is added using an adder to create a pseudo echo? (t) is generated. When the conventional pseudo-echo generation circuit 4 is configured with an analog circuit, it is not suitable for pLSI because it requires an analog delay line, and when it is configured with a dinotal circuit, it is necessary to use a high-speed subscriber line as described above. Considering digitalization, there were problems in increasing the processing speed and achieving low power consumption.

The object of the present invention is to provide an echo canceller that has high processing speed, low power consumption, and is suitable for LSI implementation due to the primary circuit network, multipliers, and adders. The present invention will be explained in detail below using the drawings.

(Structure of the Invention) In an echo canceller that cancels signals on an echo path and a pseudo-echo path, a human input signal is applied to one of two multipliers having coefficients that minimize the echo cancellation error, and an RC primary circuit is applied to the other multiplier. This echo canceller has an adder that applies an input signal via a network and adds the outputs of both multipliers, and means that cancels the echo to the output of the 4-wire side and the output of the adder.

(Explanation of Examples) FIG. 3 is a circuit diagram showing one embodiment of the present invention.

In the same figure, 10 is a resistance R (Ω), and 11 is C (F).
capacitor, 12, t12-2 is the coefficient d. l
d, a multiplier, 13 an adder, 14 a multiplier, 15
is an accumulation adder, 16 is a multiplier with coefficient α, 17-, .

17-2 is a terminating resistor of the subscriber line R (Ω), and 18 is an echo canceller. The transfer function H(S) of the echo path shown in FIG. 3 can be expressed by equation (1). Here S is the complex frequency, ZIN(
s is the impedance seen from the 2-wire/4-wire converter to the 2-wire side (subscriber line side). Also, if the time function h(t) corresponding to the transfer function I((S) is h(t) -L"[l'H(S) :) ・-...
...... It is expressed as (2). Here t −
] indicates an inverse lath transformation. The input signal on the 4-wire side is x(t)
Then, the echo y'(t) is y(t) -x(t) * h(t) ・・・・・・・・・
......(3). * indicates a convolution operation. At the same time, the input signal X(t) is echo canceller 1
Also enter 8.

This input signal x(t) is processed in the echo canceller 18 as follows, and is processed as a pseudo echo? (t) is generated. The input signal x(t) is input to a path passing through a multiplier 12-1, an RCI circuit network consisting of a resistor 10 and a capacitor 11, and a path passing through a multiplier 12-2, and the output of each path is inputted by an adder 13. Added together, pseudo echo? (t) occurs. After the pseudo echo 9(t)U phase inversion, the adder 5
is added to echo 3/(t), and echo y(t) is deleted. The signal e(t) shown at the output terminal 2 on the 4-wire side is e(t) -y(t) -9(t) ...
...(4), which indicates the echo cancellation error. The principle of the present invention will be explained below.

Impedance function ZIN(S)l/1 viewed from the 2-wire/4-wire converter to the 2-wire side (subscriber line side) is expressed by the following equation.

・・・・・・(5) Here z　−(7’　l　)2 f mouth ct-]-(R+G-.

LC Q-qp R32VeC γ-(R+5L)(G+SC): Propagation constant R of subscriber line
, L, C, G: Primary constant of subscriber line l: Crossed by subscriber line length. Therefore, the transfer function H(s) of the echo path is (1)
.

From equation (5)... (6). In the actual line = (-'-)2S (S + 2α) -...
...... (7) πU It can be regarded as formula (7). Using equation (7), equation (6) is shown as follows.

00) is the minimum value of k for which the expression is an imaginary root. Furthermore (8)
In the formula, N is T2(S)=82+2αS+(7k)2 −・・・・
... αΦα is the minimum value of k for which the formula has an imaginary root. Furthermore, in formula 1Q... (1s~πU. However, in the derivation of formula 09, θ is -7k.

An approximation is made where k≧N. From the above (8), (
1◇.

(If formulas 12 and 90 are used, formula (6) becomes as follows.

・・・・・・・・・0Q ” In the second formula And then becomes. (The old formula is further expanded as follows.

・・・−・ (In the α formula, P□(s) and P2(s) have no real root for the subscriber line length, P,(s) = P2(s) = 1 − ・−−・
Since (c), the equation (19) becomes. That is, the transfer function H(S) of the echo path has a functional form approximately having one pole and one zero point on the negative real axis of the S plane. On the other hand, the transfer function H(S) of the echo canceller 8 shown in FIG. 3 can be expressed by equation (A). In other words, the pole -61 is on the real axis of the S thousand-sided negative
It has a zero point -do+dt. In other words, tap coefficient d.

. By adaptively controlling oCR dl, the echo path transfer function H
It is possible to match the zero point of (S). The pole position of the echo canceller 8 is fixed to -1 R, but if this fixed pole is fixed at the pole position for the maximum length of the subscriber line to which it is applied, then the subscriber line of less than The HS/N ratio does not deteriorate with respect to the length. Next, tap coefficient d. , dl will be described. First, let the evaluation function J be 1- . . . (g) J=Σe2(t). Here, - indicates the time average. At this time, the tap coefficient d using the maximum slope method. The update algorithm for +dl is derived as follows. The tap coefficient value after the νth (ν) (ν) update is d. , dl...
... (goods). Here, α is equal to the first update amount of the tap coefficient, so by substituting equation (C) into equation (C), the update algorithm of the tap coefficient becomes as follows. However, Δd−αe(t) x(t) rΔd,
=αe(t) P(t) (c) Ta, , f
A control system using the coefficient updating algorithm is an echo canceller shown in FIG.

(Effects of the Invention) As explained above, the present invention derives an approximate form of the transfer function of an echo path, and changes the circuit configuration for generating the transfer function form to RC1.
The present invention is realized with a simple configuration consisting of a circuit network, a multiplier, and an adder, and provides an echo canceller that has a large echo suppression effect, low power consumption, and is suitable for LSI implementation.

[Brief explanation of drawings]

FIG. 1 is a block diagram of a Dinotal subscriber line bidirectional transmission system using a conventional echo canceller, FIG. 2 is a conventional pseudo echo generating circuit, and FIG. 3 is a circuit diagram of an echo canceller according to the present invention. 1. 4-wire side input terminal, 2.. 4-wire side output terminal, 3.. Hybrid circuit A 4° pseudo power cord generation circuit, 5..
・Adder, 6...Subscriber line, 7, ~7-(N-1)・
... Delay 1 element, 8-0 to 8-N ... Multiplier, 9 ... Adder,
lo...Resistance, 11...Capacitor, 12.14-
1.14-2゜16 +, , 16-2... Multiplier, 13... Adder, 15... Cumulative adder, 17-1
．． 17.2 ・Terminal resistance of subscriber line. Written amendment (1-1 control) Display of the case 1982 Patent Application No. 154136 2 Name of the invention Echo Canceller 3 Relationship with the person making the amendment Patent applicant's office (105) Toranomon 1, Minato-ku, Tokyo Chome 7-12
No. 6 Contents of amendment (1) On page 4, line 11 of the specification, “17-2 is R(Ω
) is corrected to 1112 as “s(Ω) becomes”. (2) Formula (8) on page 1-7 of the same is amended as follows. (3) The entire text is on page 9, line 12 of the same book. (4) In the 8th line of page 12 of the same book, the phrase ``1st'' has been amended to read ``11th''. (5) In the second line from page 9, page 9 of the same book, “Δd1−α
. (t)P(t) Expression (26)'' is corrected to read, ``At Δd, αe(t)P(t) The echo canceller shown in FIG. 3 is expressed by Expression (26).'' [6, same book, page 13, 3rd item] also says 1st 4-ite month 4 ``Hayabusa'I Mitorikoku-de bru 2J.''
- E "$"1 @ system p eraser 1 romono 7, blue and correction T.

Claims (1)

[Claims] In an echo canceller that cancels the signals of the echo path and the pseudo-echo path, one of the two coefficients determined to minimize the echo cancellation error is multiplied by the input signal, and the other is the input signal via the RCI circuit network. two multipliers for multiplying by and an adder for adding the outputs of both multipliers;
An echo canceller comprising means for canceling an echo to a four-wire side output and an output of the adder.