FR2615057A1 - Method and device for echo path simulation for testing echo cancellers - Google Patents

Abstract

The echo path simulation device DS comprises two-wire/four-wire couplers 1, 2, cable transmittance simulation circuits 3, 3a and a remote supply circuit 8 rendering it equivalent to a cable joining two items of transmission equipment ET1, ET2. The items of transmission equipment each include an echo canceller to be tested. For each of the items of equipment, an echo generator 5, 5a simulates an echo path having a discrete impulse response of rectangular shape, the position of which with respect to an origin instant and the amplitude of which are determined by digital control words 53, 54; 53a, 54a. From the ability of the echo canceller to synthesise individually each of the terms of a plurality of discrete impulse responses approximating a given impulse response of an echo path is deduced its ability to synthesise the given impulse response of the echo path.

Description

Path simulation method and device
to test echo cancellers.

 The present invention relates generally to problems encountered in controlling the performance of an echo canceller, for example, at the end of its manufacture or during a maintenance operation.

 Currently, an echo canceller, also called echo canceller, is controlled by operating it with various types of transmission cables known from experience to be the most penalizing by the echo created. Besides the fact that such a control by its principle is not complete, it has several drawbacks.

 The transmission cables used must be of considerable length in order to generate an echo signal suitable for the co-check of the canceller. This results in the need to have at the place of control an appropriate infrastructure, such as a storage area for holds and handling means. In addition, the operations of handling, wiring and removing cables lead to a total time necessary for the execution of the control which is relatively high.

 Furthermore, the quality of control is not satisfactory.

In general, the measurements on. digital transmission systems require perfect martyrdom and knowledge of the signals present in the transmission medium and more particularly of the noise signals. A cable of considerable length is difficult to control from the point of view of parasitic external influences, for example of the electromagnetic type.

 The present invention aims to provide a method for testing an echo canceller and echo path simulation devices to overcome the drawbacks of the prior art mentioned above.

 To this end, a method according to the invention for evaluating the ability of an echo canceller to synthesize a given echo path impulse response, is characterized in that the echo canceller is tested for an approximation of said given impulse response constituted by a plurality of discrete echo path impulse responses all having a rectangular shape and deduced from a sampling at a determined frequency of the given impulse response, and in that of the aptitude of the echo canceller to synthesize each of said discrete impulse responses individually is evaluated its ability to synthesize the impulse response of given echo path.

An echo path simulation device for implementing the method according to the invention, interconnecting first and second transmission equipment transmitting the first and second digital signals to each other and each comprising an echo canceller to evaluate, is characterized in that it includes
first and second 2-wire / 4-wire adaptation means for adapting first and second unidirectional transmission paths of opposite directions included in said device to first and second bidirectional links serving said first and second equipment,
first means receiving the first digital signal transmitted in the first channel by the first equipment through said first adaptation means and having an impulse response of rectangular shape adjustable in amplitude and positions for shaping and injecting into the second channel first echo signal derived from the first signal transmitted in the first channel, and
second means receiving the second digital signal transmitted in the second channel by the second equipment through said second adaptation means and having an impulse response of rectangular shape adjustable in amplitude and in position, for shaping and injecting into the first channel a second echo signal derived from the second signal transmitted in the second channel.

 The method according to the invention has the advantage of requiring for its implementation an echo path simulation device having an impulse response of simple form. This simulation device is therefore of little complex structure and can be portable. In addition, the structure of the simulation device is symmetrical in order to be able to simultaneously test two echo cancellers under conditions close to real conditions of transmission between two pieces of equipment.

Other advantages and characteristics of the invention will appear more clearly on reading the following description of several preferred embodiments of a simulation device according to the invention with reference to the corresponding appended drawings in which
- Fig. 1 shows schematically a test configuration of the echo cancellers according to the prior art
- the Pig. 2 shows an actual waveform of an echo path impulse response in a telephone network;
- Fig. 3 is a functional block diagram of a device for simulating an echo path according to the invention
- Fig. 4 is a detailed block diagram of an echo generator included in a simulation device as shown in FIG, 3
- Fig. 5 shows waveforms of rhythm signals and impulse signals produced by the echo generator shown in FIG. 4; and
Fig. 6 shows additional means included in the echo generator, for regenerating a received bipolar signal.

 A test configuration according to the prior art is shown in FIG. 1. This configuration uses two analog transmission equipment ET1 and ET2 connected by a capable C, for example of symmetrical pair types. The equipment ET1 and ET2 each include an echo canceller to be tested, AE1 and AE2 respectively.

 The first ET1 transmission equipment is shown in detail. It also includes the AE1 echo canceller, a Mol modulator, a 2-wire / 4-wire COI coupler, such as a differential transformer, and a DE1 demodulator.

The echo canceller AE1 essentially comprises, as is known, an adaptive digital transversal filter FT1 producing a convolution product between an input signal to be transmitted stl (t) to the equipment ET2, supplied in digital form, and a series of coefficients representing an impulse response of the synthesized echo path re'l (t), rel (t) being the real impulse response of the echo path. The AEI echo canceller calculates an estimate of the echo, known as synthesized echo, e1 (t) = st1 (t) * re'l (t) which is received by the ETI equipment in response to the signal transmission st1 (t), via the cable CB, to the remote equipment ET2. This estimate e1 (t) = st1 (t) * re'l (t) is subtracted from a signal sr1 (t) supplied by the coupler COl to the canceller AE1. The signal sr1 (t) is received by the coupler Col via the cable CB and includes a component corresponding to a signal st2 (t) transmitted by the distant equipment ET2, and an echo component. The AE1 echo canceller outputs the signal srl (t) - e1 (t) free of the echo component, due to the real response rel (t), and corresponding to the signal st2 (t) transmitted by l remote equipment ET2. Similarly, the second device ET2 performs analogous operations to recover a signal sr2 (t) - e2 (t) corresponding to the signal stl (t) transmitted by the first device
ET1.

 The modulators M01 and demodulator DE1 are of known type. The modulator M01 receives the digital input signal stl (t) to be transmitted, in the form of serial binary words at an equal rate, for example, at 160 kbit / s, and converts it into a bipolar line code at a frequency of rhythm F - 160 kHz for transmitting it via the coupler COl in the cable CB to the remote equipment ET2. The demodulator DE1 performs the opposite operation on the signal srl (t) - e1 (t).

 The error rate calculation from the transmitted signals, st1 (t) and st2 (t), and the received signals, sr1 (t) - e1 (t) and sr2 (t) - e2 (t) allows evaluate the effectiveness of the AE1 and AE2 echo cancellers.

 A waveform of an echo path impulse response re (t) such as that which can be obtained by making a connection of the equipment ET1 and ET2 via a telephone network, is shown in FIG. 2. This waveform can be measured in a test configuration as shown in Fig. 1, by applying a signal stl (t) = d (t) at the input of the coupler COl, where 6 (t) is a Dirac pulse. The signal st2 (t) being zero, the signal sr1 (t) measured at the output of the coupler COl then represents the impulse response of echo path re (t) = rel (t) = -srl (t}.

 Compared to an origin instant O, where the Dirac impulse is applied 6 (t), the response re (t) is delayed by a time tp necessary for the path of the echo path. In the case of the test configuration shown in Fig. 1, the time tp is essentially due to the propagation time introduced by the cable CB.

où T est un -temps prédéterminé définissant une période d'échantillonnage de la réponse re(t), où n est un entier compris entre un entier minimum K égal a la partie entière de tp/r et un entier.maximum N égal à la partie entière de (tp + ta)/r, où e (t) est une réponse impulsionnelle discrète de chemin d'écho ayant une forme rectangulaire et d'amplitude unité, telle que e(t) = 1 pour tCO,rC et e(t) - 0 pour tEe O,TE ,et AK à AN sont les amplitudes respectives des réponses impulsionnelles diacrètes composant la somme approximant la réponse re(t).
Un dispositif selon l'invention est basé sur un principe de décomposition en éléments simples. En effet, si l'annuleur d'écho est capable de synthétiser successivement chacune des réponses impulsionnelles discrètes composant la réponse impulsionnelle réelle de chemin d'écho pour laquelle on désire tester l'annuleur, alors l'annuleur est également capable de synthétiser la réponse impulsionnelle réelle re(t).As shown in Fig. 2, the actual waveform of an echo path impulse response is generally complex and therefore difficult to simulate. However, the impulse response re (t) can be approximated by a sum of discrete echo path impulse responses. The impulse response re (t) can be expressed by the following sum

where T is a predetermined time defining a sampling period of the response re (t), where n is an integer between a minimum integer K equal to the integer part of tp / r and a maximum integer N equal to the integer part of (tp + ta) / r, where e (t) is a discrete impulse response of echo path having a rectangular shape and of unit amplitude, such that e (t) = 1 for tCO, rC and e (t) - 0 for tEe O, TE, and AK to AN are the respective amplitudes of the diacretic impulse responses making up the sum approximating the response re (t).
A device according to the invention is based on a principle of decomposition into simple elements. Indeed, if the echo canceller is able to synthesize successively each of the discrete impulse responses composing the real impulse response of echo path for which one wishes to test the canceller, then the canceller is also capable of synthesizing the response real impulse re (t).

it follows from the above that an echo path simulation device according to the invention simulating an impulse echo path response of the type Ae (t - nn), where n and A are adjustable parameters, makes it possible to test the echo canceller for any real impulse response re (t). The test for canceling echo can, for example, be carried out using a computer memorizing the different values of the integer n as well as the different amplitudes i to
With reference to FIG. 3, a simulation device DS according to the invention replaces the cable CB and connects the equipment ETl and
ET2.

The DS device comprises two 2-wire / 4-wire analog couplers 1 and 2, connecting it respectively through bidirectional two-wire links Li and L2 to the 2-wire / 4-wire couplers, such as the coupler C01, in equipment ET1 and ET2 . The couplers 1 and 2 are intended to route the bipolar signals transmitted from the equipment ET1 to the equipment ET2 and vice versa in two unidirectional and symmetrical transmission channels. A first transmission channel 12-21, from the equipment ET1 to the equipment ET2, is constituted by a series combination of a cable transmittance simulation circuit 3, of an adder 4, and of two line transformers 6 and 7 and by an echo generator 5. A second transmission channel 22-11, from the equipment ET2 to the equipment ET1, is constituted by a series combination of a cable transmittance simulation circuit 3a, analogous to circuit 3, of an adder 4a, analogous to adder 4,
and two line transformers 6a and 7a, analogous to transformers 6 and 7, and by an echo generator 5a, analogous to generator 5.

 These different circuits included in the two channels are described in detail below with reference to the first channel.

 The coupler 1 connected to the coupler COl of the equipment ET1 through the link L1 is shown in detail. The coupler I essentially comprises an operational amplifier All operating as a subtractor. A direct input (+) of the amplifier Ali is connected to a first wire of the link Lla through a link capacitor C11 and to a unidirectional input il of the coupler 1 through an impedance Zil. The input 11 is connected to an output 49a of the summator 4a in the other channel. A reverse input (-) of the amplifier All is connected to the input 11 through an impedance Z12. An impedance Z13 is placed in feedback between the input (-) and a unidirectional output 12 of the amplifier Ail. A second wire of the link L1 is connected to a terminal, to a reference voltage of the device DS, through a link capacitor C12. The output 12 is connected to an input 51a of the echo generator 5a and to the transformer 6 to deliver a signal equivalent to the signal stl (t) transmitted by the equipment ET1.

The coupler 1 is balanced and ensures no spurious transmission from its input 11 to its output 12 when the equality
Z11 / Ze = Z12 / Z13 is verified, Ze being the input / output impedance of the coupler COl seen from the coupler 1 through the capacitors C11,
C12 and the L1 bond. The coupler C01 in the ETi equipment also respects an equilibrium equation of the same type such that the signal srl (t) delivered by the coupler COl to the echo canceller AE1 does not include any local echo component due has a direct transmission of the signal stl (t) to the echo canceller AE1 through the coupler C01.

 The function of the cable transmittance circuit 3 is to deform and attenuate the transmitted signal in the same way as a cable does. The circuit 3 thus simulates the transmittance of a cable and can be a filter comprising resistive, capacitive, and inductive elements, or more simply a cable such as the cable CB but of reduced length.

 The cable transmittance circuit 3 is connected to the output 12 of the coupler 1 through the line transformer 6. A first winding of the transformer 6 is connected between the output 12 and the reference terminal. A second winding of the transformer 6 is connected between two inputs of the circuit 3. Between two outputs of the circuit 3 is connected a first winding of the transformer 7.

A second winding of the transformer 7 is connected between the reference terminal and a first input 41 of the adder 4.

 The summator 4 is a circuit essentially comprising an operational amplifier 42 operating as an adder. A direct input (+) of the amplifier 42 is connected to the input 41 of the summator 4 through a resistor 43 and to a second input 44 of the summator 4 through a resistor 45 and a capacitor 46 in series. A reverse input (-) of the amplifier 42 is connected to the reference terminal through a resistor 47. A feedback resistor 48 is placed between the reverse input (-) and an output 49 of the amplifier 42 delivering the sum of the signals present at the first and second inputs 41 and 44. The input 44 and the output 49 are respectively connected to an output 52 of the echo generator 5 and to a unidirectional input 2] of the 2-wire / 4-wire coupler 2.

 The 2-wire / 4-wire coupler 2, analogous to the coupler 1, delivers by a unidirectional output 22 a signal corresponding to the signal st2 (t) transmitted by the equipment ET2 via the bidirectional link L2. The output 22 is connected to an input 51 of the echo generator 5 and to a first winding of the transformer 6a. The first winding of the transformer 6a is connected between the output 22 and the reference terminal. A second winding of the transformer 6a is placed between two inputs of the cable transmittance circuit 3a.

 The cable transmittance circuit 3a, analogous to circuit 3, comprises two outputs between which is connected a first winding of the transformer 7a. A second winding of the transformer 7a is connected between the reference terminal and a first input 41a of the summator 4a. A second input 44a is connected to an output 52a of the echo generator Sa. As already said, the output 49a of the adder 4a is connected to the input 11 of the coupler 1.

 The device DS can ensure the passage of a remote supply current supplied by the first equipment ET1 to the second equipment ET2. To this end, a remote supply circuit 8 is provided in the DS device.

 The remote supply circuit 8 is connected to the links L1 and L2, in parallel to the transmission channels. The capacitors C11 and C12 of the coupler 1 and those of the coupler 2 prevent the passage of direct current through the transmission channels. Circuit 8 includes a choke coil 81 and a resistor 82 connected in series between the first wire of link L1 and a first wire of link L2. A choke coil 83 and a resistor 84 are connected in series between the second wire of the link L1 and a second wire of the link L2. In the equipment ET1, a supply TA injects a direct current of remote supply into secondary windings of a line transformer TL1 of the coupler C01.Resistors 82 and 84 limit this direct current and simulate losses in a cable, ie -to say the resistances of the wire of the cable simulated by circuits 3 and 3a. The stop coils 81 and 83 prevent the transmission signals from passing through the remote supply circuit 8.

 The echo generator 5a is described in detail in the following description. The echo generators 5 and 5a each receive, respectively by buses 53, 54, and 53a, 54a, two binary control words determining the parameters n and A of the discrete impulse response of the echo path to be simulated. The binary command words are delivered by a computer or a control panel and each correspond to a determined value of one of the parameters n and A.

 The signal srl (t) received in the equipment ET1 can be expressed as a function of the signal stl (t) and of a signal sl2 (t) present at input 41a of the adder 4a. The signal sl2 (t) comes from the equipment ET2 and corresponds to the signal st2 (t) transmitted by the latter.

la fonction notée TL étant la transformée de Laplace.We denote by SRl (p), STl (p) and SL2 (p) the transforms of
Laplace respectively of the signals srl (t), stl (t) and sl2 (t), by 111 (p) the transfer function in the direction ET1 to ET2 of the assembly constituted by the CO1 coupler, the L1 link and the coupler 1, by
H2 (p) the transfer function in the direction ET2 to ETI of the assembly constituted by the coupler CO1, the link L1 and the coupler 1, and by Hge (p) the transfer function of the echo generator 5a. As already said, the couplers C01 and 1 being balanced, that is to say each having a transfer function with practically no effect between unidirectional input and output, the transform of
Laplace SR1 (p) is given by the following relation
SR1 (p) = H1 (p) .H2 (p) .Hge (p) .ST1 (p) + H2 (p) .SL2 (p)
The term H2 (p) .SL2 (p) represents the component of the signal received in the equipment ET1 due to the signal st2 (t) transmitted by the equipment ET2. The term H1 (p) .H2 (p) .Hge (p) .ST1 (p) represents the echo signal introduced by the generator 5a. The product of the transfer functions Hl (p) .H2 (p) .Hge (p) must be equal to the Laplace transform of the simulated echo path discrete impulse response and can be expressed

the function noted TL being the Laplace transform.

 The echo generator 5a is shown in FIG. 4. The generator Sa comprises a pulse regeneration and rhythm recovery circuit 55, two analog programmable delay registers 56 and 57, an echo signal shaping circuit by level adjustment 58, and a filter pre-emphasis 59.

 The rhythm regeneration and recovery circuit 55 receives by the input 51a a bipolar signal equivalent to the signal st1 (t) and having positive and negative voltage marks. The circuit 55 is conventional and comprises at input two threshold comparators 551 and 552 for producing two binary signals H + and B corresponding respectively to the positive and negative marks, at output two decision flip-flops 553 and 554 for regenerating the binary signals as a function of a recovered clock signal and a phase lock loop (PLL) to recover the rhythm of the incoming bipolar signal.

 The bipolar signal is supplied through a connecting capacitor C55 to a direct input (+) of the first comparator 551 and to an inverse input (-) of the second comparator 552. The comparators 551 and 552 constitute voltage threshold detectors. Comparator 551 has an inverse input (-) fixed at a positive continuous threshold voltage + V less than the nominal voltage amplitude of the positive marks of the bipolar signal and preferably equal to half of this nominal amplitude. Comparator 552 has a direct input (+) fixed at a negative continuous threshold voltage -V lower in absolute value than the nominal amplitude: of voltage of the negative marks of the bipolar signal. The input (+) of comparator 551 and the input (-) of comparator 552 are connected together and connected to the reference terminal through a resistor R55. The capacitor C55 and the resistor R55 constitute a limiting high-pass RC filter possible low frequency oscillations. An output of comparator 551 is connected to a data input D of the D type flip-flop 553 and delivers the binary signal B + corresponding to the positive marks detected. An output of comparator 552 is connected to a data input D of the other D type flip-flop 554 and delivers a binary signal H corresponding to the negative marks detected.

The binary signals B + and H are also supplied to first and second inputs, respectively, of a phase comparator 555 included in the circuit 55. A third input of the phase comparator 555 receives an impulse signal from a frequency divider 556 of rhythm H recovered at the frequency
F = 160 kHz. The phase comparator 555 supplies a control voltage as output to a voltage controlled oscillator (VCO) 557.

The oscillator delivers a clock pulse signal Hc at a frequency set to that of the oscillator equal to Fo = PF, where P is a predetermined integer, for example equal to 4. The signal Hc is applied to an input of the divider of frequency 556 which divides by
P = 4 the frequency Fo to produce the rhythm signal R to compare with the signals B + and B in the comparator 555.

The phase comparator 555, the voltage controlled oscillator 557, and the frequency divider 556 constitute the aforementioned phase locked loop. Two signals are recovered, the rhythm signal H at frequency F = 160 kKz intended for the synchronization of signals B + and B and the clock signal Bc of frequency Fo = PF = 640 kHz intended for sampling the signals B + and B
The rhythm signal H is applied to clock inputs CK of flip-flops 553 and 554. The signal Hc is slightly delayed by passing through a non-inverting logic gate 558, in order to be put in phase with the signal H produced by the divider 556 and be applied to reset inputs R of flip-flops 553 and 554.

Outputs Q of flip-flops 553 and 554 respectively deliver pulse trains Bc + and Bc synchronous with the signal H and calibrated and sampled by the signal Hc.

Waveforms of the signals B +> 8, H, Hc, and pulses Bc + and Bc are shown in Fig. 5. A Bc + pulse,
Bc, comes to a logical state "1" on rising edges of the recovered rhythm signal H, when a logical state "1" of the signal B +,
B-, is applied to the entry D of the respective rocker; more preclslement, this reading occurs substantially in the middle of the logical state "I" corresponding to the maximum opening of the eye in the signal, bipolar in order to benefit from the best noise immunity possible. A pulse Bc +, Bc, comes to a logic state "0" on the first rising edge of the signal Hc following the rising edge of the signal H having commanded its setting to the state "1". The pulses Bc + and Bc in logic state "i" are thus calibrated for a duration equal to a period TI 1 / Fo = 1 / 4.F = 1.56 ps of the signal
Hc.

 The delay registers 56 and 57 are intended to introduce an adjustable delay nT. respectively on the pulses Bc +, Bc. The registers 56 and 57 are programmable type shift registers. The register 56, 57, receives at a data input 561, 571, the pulse Bc +, Bc of the rhythm regeneration and recovery circuit 55,, and at a clock input 562, 572, the signal Hc available in output of oscillator 557. A delayed pulse RBc +, RBc, is delivered by an output 564, 574, of register 56, 57. A binary delay adjustment word MR is supplied to programming inputs 563 and 573 of registers 56 and 57 respectively. The word MR is conveyed by the bus 53a and includes, for example, b 8 6 bits. The word MR at b = 6 bits represents the number n in binary code and controls the adjustment of a delay equal to nT.

The adjustment resolution is equal to T = 1.56 ps and is acceptable for bipolar signals at the rate of d = 160 kbit / s. Waveforms of the delayed pulses RBc +, RBc are shown at the bottom of FIG. 5 for a delay equal to 3T.

 The level shaping and adjusting circuit 58 comprises a digital analog converter 581, a potentiometer 582 and two switching elements 583 and 584.

 The converter 581 is connected to the bus 54a and receives a binary amplitude adjustment word MA having, for example, f = 8 bits.

The converter 581 provides respectively by first and second analog outputs a positive voltage + Va and a negative voltage -Va having an amplitude corresponding to the digital value of the binary word MA. The voltages + seen, and -Va, are applied to a first terminal of the potentiometer 582 through the switching elements 583 and 584 respectively. A second terminal of potentiometer 582 is connected to the reference terminal. A third terminal of potentiometer 582 takes a fraction of the voltage + Va, -Va, between the first and second terminals. The switching elements 581 and 582, shown in FIG. 3 in the form of contacts, are, for example, switching transistors of the FET or MOS type, having terminals or control gates connected to the outputs 564 and 574 of the shift registers respectively.

 The closing of the switching elements 583 and 584 is controlled by the pulses RBc + and Roc, respectively. A pulse RBc +, Roc, commands the closing of 11 switching elements 583, 584, to apply a voltage level + Va, -Va between. the first and second terminals of the potentiometer 582. A bipolar echo signal having positive and negative voltage marks, of level + Va and -Va respectively is thus formed between the first and second terminals of the potentiometer 582.

correspondant à une réponse impulsionnelle discrète de forme rectangulaire. La fonction à réaliser par le générateur d'écho 5a est définie par l'égalité

le terme 1/((H1(p).H2(p)) compensant éventuellement effet des coupleurs
COl et 1 sur les signaux transmis. Cet effet est généralement négligeable, de telle sorte que ce terme peut être omis. Toutefois, si tel n'est pas le cas, il peut être introduit par le filtre de préaccentuation 59.Between the input 51a of the echo generator 5a and an output of the echo generator 5a considered at the potentiometer 582 is carried out a transfer function of the type

corresponding to a discrete impulse response of rectangular shape. The function to be performed by the echo generator 5a is defined by the equality

the term 1 / ((H1 (p) .H2 (p)) possibly compensating for the effect of the couplers
COl and 1 on the transmitted signals. This effect is usually negligible, so this term can be omitted. However, if this is not the case, it can be introduced by the pre-emphasis filter 59.

 The main function of the pre-emphasis filter 59 is to compensate for the effect of a high-pass filter located at the input to an echo canceller, and pre-accentuates the echo signal reproduced by the third terminal of the potentiometer 582, in a loop low frequency. The filter 59 can be calculated to also introduce the term l / ((H1 (p) .H2 (p)) in the transfer function Hge (p) of the echo generator.

 The filter 59 comprises two operational amplifiers 591 and 592. The amplifier 591 is wired as a voltage follower. A direct input (+) of the amplifier 591 is connected to the third terminal of the potentiometer 582 and receives a fraction of the shaped bipolar echo signal. An inverse input (-) and an output of amplifier 591 are directly connected. An inverse input (-) of the amplifier 592 is connected to the output of the amplifier 59] through a resistor 593. A direct input (+) of the amplifier 592 is directly connected to the reference terminal. A feedback circuit consists of a resistor 594 and a series combination of a capacitor 595 and a resistor 596 which are connected in parallel between the inverse input (-) and an output of the amplifier 592 constituting the output 52a of the echo generator 5a.

 A DS simulation device as just described with reference to Figs. 3 to 5 is entirely satisfactory. However, if this proves necessary, it is possible to make an improvement in order to compensate for a possible local echo transmitted by the couplers 1 and 2 from input 11 to output 12 and from input 21 to output 22, respectively. A bipolar signal equivalent to the signal st1 (t), st2 (t), can indeed be reconstructed in the echo generator 5a, 5, from the signals B + and B. This is no longer the bipolar signal delivered by the output 12, 22, of the coupler 1, 2, which is transmitted to the equipment ET2, ET1, but the bipolar signal reconstructed from the signals B + and B. According to this variant, the line transformer 6, 6a, is disconnected from the coupler 1, 2, and connected to an additional output of the echo generator 5a, 5, as shown in FIG. 6.

With reference to FIG. 6, the echo generator 5a further comprises two flip-flops of the D type 553 'and 554', two switching elements 583 'and 584' and a follower amplifier 591 '. The flip-flops 553 'and 554' have inputs D connected to the outputs of comparators 551 and 552 to receive binary signals B + and clock inputs CK to receive the recovered clock signal
H, and outputs Q connected to control terminals of the switching elements 583 'and 584', respectively. The switching elements 583 ′ and 584 ′ have inputs brought respectively to voltages dependent on the minimum voltages + Ve and -Ve of the positive and negative marks of the bipolar signal to be regenerated and retransmitted, and of outputs connected to an input of the amplifier. 591 'and to the reference terminal through a potentiometer 582'. An output of the amplifier 591 'is connected to the first winding of the transformer 6 through a resistor 593'.

 Thus, it appears that the decision flip-flops 553 'and 554', the switching elements 583 'and 584' and other additional components 582 ', 591' and 593 'respectively play similar roles by means of regeneration and shaping. constituted by the elements 553 and 554, 583 and 584, 582, 591 had 593, but to regenerate a bipolar signal having a peak-to-peak amplitude constructed equal to 2Ve> 2V and binary elements of width equal to 1 / F. All of these additional elements with the comparators 551 and 552 and the locking and phase loop 555-556-557 constitute a repeater-regenerator.

Claims (10)

REVEBICATIONS  1 - Method for evaluating the aptitude of an echo canceller (AE1, AE2) to synthesize a given echo path impulse response (re (t)), characterized in that the echo canceller (AE1, AE2) is tested for an approximation of said given impulse response (re (t) ) constituted by a plurality of discrete echo path impulse responses all having a rectangular shape and deduced from sampling at a determined frequency (1 / T) from the given impulse response (re (t)), and in that the ability of the echo canceller (AE1, AE2) to individually synthesize each of said discrete impulse responses is evaluated its ability to synthesize the. impulse response of echo path given (re (t)).  2 - Device for implementing the method according to claim 1, interconnecting first and second transmission equipment (ET1, ET2) transmitting one to the other of the first and second digital signals (st1, st2) and comprising each an echo canceller (AE1, AE2) to be evaluated, characterized in that it comprises  first and second 2-wire / 4-wire adaptation means (1,2) for adapting first and second unidirectional transmission channels of opposite directions (12-21, 22-11) included in said device (DS) A of the first and second bidirectional links (L1, L2) serving said first and second equipment (ET1, ET2),  first means (5a, 4a) receiving the first digital signal (st1) transmitted in the first channel (12-21) by the first equipment (ET1) through said first adaptation means (1) and having an impulse response of rectangular shape adjustable in amplitude (AK at) and position (KT at NT) to shape and inject into. the second channel (22-11) a first echo signal derived from the first signal transmitted in the first channel, and  second means (5, 4) receiving the second digital signal (eut2) transmitted in the second channel (22-11) by the second equipment (ET2) through said second adaptation means (2) and having an impulse response of rectangular shape adjustable in amplitude (Pk to A, T) and in position (KT to NT), to shape and inject into the first channel (12-21) a second echo signal derived from the second signal transmitted in the second way.  3 - Device according to claim 2, characterized in that the first and second 2-wire / 4-wire adaptation means (1, 2) each comprise an operational amplifier (All) and connection capacitors (C11, C12) for filter any DC component of the first and second digital signals (stl, st2) and retransmit these first and second signals respectively in the first and second channels (12-21, 22-11) respectively,  4 - Device according to claim 2 or 3, characterized in that the first and second means for shaping and for injecting (5, 4; 5a, 4a) each comprise a summator (4, 4a) to. operational amplifier (42) for adding the first and second derived echo signals to the second and first transmitted signals (stl, st2) respectively.  means (58) controlled by the second means for selecting and receiving the calibrated and delayed pulses (RBc +, Roc) for reproducing the first, respectively second echo signal so that it is composed of the width calibrated and delayed pulses and having an amplitude (Va) proportional to the selected digital amplitude value (A).  second means (54a; 54) for selecting a digital amplitude value (A), and  means (56, 57) programmable as a function of the selected delay value (n) and receiving the calibrated pulses (hic +, Bc) and the clock signal (lit) to delay said calibrated pulses (hic +, Bc) by a duration (nT) proportional to the selected digital delay value (n), and to a period (T) of the clock signal (hic),  means (53a; 53) for selecting a digital delay value (n),  means (55) connected to the first, respectively second transmission path (12-21; 22-11) for regenerating and calibrating in width of the pulses (B +, B) of the first, respectively second digital signal and for recovering a signal d clock (hic) at a frequency (Fo t PF) multiple of the frequency (F) of rhythm of the digital signals,  5 - Device according to any one of claims 2 to 4, characterized in that the first and second means for shaping and injecting (5, 4; 5a, 4a) each comprise  6 - Device according to claim 5, characterized in that the first and second digital signals (st1, st2) are bipolar.  7 - Device according to claim 6, characterized in that the means for regenerating, calibrating, and recovering (55) comprise two threshold comparators (551, 552) for converting the bipolar signal into two corresponding binary signals (B +, B) respectively to positive and negative marks in the bipolar signal, a phase locked loop (555, 557, 556) connected to the comparator for recovering signals (Hc, B) at said multiple frequency and rhythm frequency, and two flip-flops (553 , 554) controlled by the locking loop to calibrate in width the pulses of the binary signals (B +, B),  in that the means for delaying (56, 57 comprise two shift registers programmable by the first means for selecting (53a; 53) for delaying respectively the calibrated pulses (hic +, Bc) of the two binary signals, and  in that the means for reproducing (58) comprise a digital-analog converter (581) controlled by the second means for selecting (54a; 54) to supply a positive voltage level (+ Va) and a negative voltage level (- Va) function of the digital value selected amplitude (A) and two switching means (583, 584) controlled respectively by the shift registers to deliver said positive (+ Va) and negative (-Va) voltage levels respectively for the pulses calibrated and delayed binary signals to produce the first and second echo signals respectively  8 - Device according to any one of claims 5 to 7, characterized in that first and second means for shaping and injecting (5, 4; 5a, 4a) each comprise filtering means (59) placed in output of the reproducing means (58) for pre-emphasizing the echo signal in a determined low frequency band  9 - Device according to any one of claims 2 to 8, characterized in that it comprises in each of the first and second transmission channels (12-21, 22-11), means (3, 3a) for simulating the transmittance of a cable.  10 - Device according to any one of claims 2 to 9, characterized in that it comprises a remote supply circuit (8) in parallel with the first and second transmission channels (12-21, 22-11) to allow the passage of a direct remote supply current from one of the transmission equipment (ET1) to the other (ET2).  1I - Device according to any one of claims 2 to 10, characterized in that it comprises in each of the first and second transmission channels (12-21, 22-11), means (551, 552, 553 ' , 554 ', 583', 584te 591 ') to regenerate the first, respectively the second transmitted signal.