The present invention relates to a method for automatically matching the levels of the signals exchanged between apparatuses such as telephones, videophones (system for transmitting voice and video via the telephone network), faxes or computers which are connected to a communication network. The invention also relates to an automatic matching device.
The invention relates in particular to a method for automatically matching the levels of the signals exchanged in a telephone network.
One solution for correcting the distortions introduced by the analogue transmission line consists in measuring a DC voltage Vdc across the terminals of a load 7 which is connected to the line through an inductor L1, given that the capacitors C1, C2 act as filters for low-frequency signals while the inductors L1, L2 filter the high-frequency signals. This voltage Vdc is then delivered to a calculation module 8 which, on the basis of the result of this measurement, determines a value for ZL. The calculation module also determines a gain G1, chosen so that the gain of IN1 at the point VL2 does not depend on ZL, a gain G2, chosen so that the gain of IN2 in the signal OUT2 does not depend on ZL either, and a gain G3 which is chosen so as to suppress the sent signal IN2 from the received signal OUT2 and acts as an echo canceller (G3 is not shown in
It can be determined that:
In this case, setting:
and
the following are obtained: OUT1=0.5*IN1 and OUT2=0.5*IN2
This solution is not suitable for compensating the signals exchanged by digital apparatuses, which need to be isolated from the subscriber loop and which do not therefore have access to the line impedance ZL via a direct voltage/current measurement.
Document U.S. Pat. No. 5,422,950 (Miller et al.) relates to the automatic compensation for attenuations in a telephone system. The estimate of the impedance of the line is made by measuring the voltage on the lines 28, 30 for a particular current using a simple resistance-measuring circuit 42. The subject matter of document U.S. Pat. No. 5,422,950 does not solve the problem mentioned above.
The object of the invention is to reduce the effect of the line impedance, and to do so even though the direct measurement described above is impossible.
This object is achieved by a method for automatically matching the levels of the signals exchanged between a first apparatus and a second apparatus which communicates with the said first apparatus via a transmission line, characterized in that it comprises the following steps:
With the method according to the invention, it is no longer necessary to measure a DC voltage in order to determine the gains needed for the compensation, since the solution employed is essentially digital, that is to say software-based, and can therefore be employed for compensating the level of the signals exchanged in applications using digital apparatuses which are isolated from the subscriber loop, such as videophones, faxes or computers. The method allows, dynamically, operation in full duplex mode which is independent of temperature variations so long as at least one signal transmission out from the apparatus has been made in order to ascertain the initial characteristics of the line.
It may be advantageous for the numerical estimate making it possible to evaluate the transfer function (K) to be made using a software calculation method.
According to one embodiment, this calculation method implements an identification algorithm.
Preferably, the identification algorithm is of the LMS (Least Mean Square), RLS (Recursive Least Square) or Kalman type.
The invention also relates to a device for automatically matching the levels of signals exchanged between a first apparatus (3) and a second apparatus communicating via a transmission line, characterized in that it has:
According to one embodiment, the numerical calculation block has a unit for identifying the transfer function interacting with a calculation module which is intended to supply a first amplification means with the first gain for matching the level of the signal transmitted by a user, and to supply a second amplification means with the second gain for matching the level of the signal received by the user.
Advantageously, the calculation block is a DSP (Digital Signal Processing) circuit implementing an identification algorithm.
Other characteristics and advantages of the invention will become apparent from the following description, made by way of nonlimiting example and with reference to the appended figures, in which
In order to avoid the attenuation due to the impedance ZL of the line 6 which the signals IN1 and OUT2 suffer, and in order to keep the transfer functions for the signal IN1, at the point VL2, and for the signal OUT2 independent of the line impedance, the method according to the invention has a step of digitizing the signal entering the said apparatus, a step of estimating the transfer function K as a function of the exchanged signals OUT2 and IN1, then a step of multiplying each signal by a suitable gain determined on the basis of the value of the transfer function K determined beforehand.
When the signal IN1 is transmitted, the signal OUT2 detected at the output of the subscriber loop is applied to an analogue/digital converter 44 which digitizes the said signal OUT2.
The estimate of the transfer function K is made numerically by an identification algorithm based, for example, on the method of least squares, the RLS (Recursive Least Square) algorithm or alternatively on the Kalman algorithm. The algorithm has the function of calculating the characteristic parameters of the transfer function K, which may in particular be a matrix (hi)1≦i≦n or a polynomial fraction in (ZL−i)1≦i≦n.
In the present embodiment, the calculation consists in firstly determining the ratio:
where
This being true in the present embodiment with an impedance ZL which is assumed to be constant. It is clear that the source impedance is equal to the input impedance of the line for a short line ZL=0 and the input impedance of the line is dependent on the characteristic impedance ZC and on the load impedance ZR; in the present case, ZR is equal to the source impedance R1. For the sake of simplicity, the condition ZL=ZC is set.
A step subsequent to this calculation consists in determining:
for the transmitter signal, a first gain
and for the received signal, a second gain
It can be seen that for these values of gains, the voltage OUT2 at the ends of the transmission line is equal to half the voltage VL2 (apart from echoes of IN1).
The device in
Preferably, the numerical calculation block 10 is a DSP (Digital Signal Processing) circuit employing one of the identification algorithms mentioned above. Another type of circuit may, of course, be used.
As can be seen in
During operation, the identification unit 12 supplies the calculation module 14 with an estimated value of the transfer function K, calculated on the basis of the values of the signals transmitted and of the signals received by the user. These signals are applied respectively to the first input 20 and to the second input 22 of the identification unit 12.
The calculation module 14 supplies the first amplification means 16 with the first gain G1 in order to compensate for the attenuation of the signals transmitted by the user, and supplies the second amplification means 18 with the second gain G2 in order to compensate for the attenuation of the signals received by the user.
The method and the device of the invention thus make it possible to perform automatic matching of the levels of the signals exchanged through a transmission line. Furthermore, this system is not sensitive to temperature drifts which can affect the voltage measurement advocated in the prior art, such as that across the terminals of the load 7 in
Knowledge of the transfer function K(ZL) can also be used to detect the presence of a parallel connection of the device of the invention in the transmission line. The said detection method includes a step of observing the sign of the gain of the identified transfer function K. When the sign is negative, then it is deduced that a second set is connected in parallel with the transmission line. This information can be used, for example, without implying any limitation, for security reasons in the case of using a modem and a telephone. If detection is made by the modem then the latter can hang up to free the line.
Number | Date | Country | Kind |
---|---|---|---|
97 01827 | Feb 1997 | FR | national |
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/FR98/00295 | 2/16/1998 | WO | 00 | 11/23/1999 |
Publishing Document | Publishing Date | Country | Kind |
---|---|---|---|
WO98/36549 | 8/20/1998 | WO | A |
Number | Name | Date | Kind |
---|---|---|---|
3781481 | Shaffer et al. | Dec 1973 | A |
4277655 | Surprenant | Jul 1981 | A |
4984265 | Connan et al. | Jan 1991 | A |
5287406 | Kakuishi | Feb 1994 | A |
5422950 | Miller et al. | Jun 1995 | A |
Number | Date | Country |
---|---|---|
2289188 | Nov 1995 | GB |
05-047099 | Feb 1993 | JP |
08-335976 | Dec 1996 | JP |