Method and arrangement for compensating signal echoes during duplex data transmission with discrete multitone modulation

1. A method for compensating echoes during discrete multitone modulation DMT, a compensation being performed in a frequency domain, comprising:

during an initialization phase in which there is no signal from an opposite end, a first error of a frequency channel which is defined as a first reference channel is derived from a difference between a transmission value and reception value of the first reference channel, and

said first error is used to correct other DMT frequency channels by weighting said first error with coefficients which are specific to said other DMT frequency channels, prior to adding said first error to said other DMT frequency channels.

2. The method according to claim 1 wherein the error is added to the signals of the other DMT frequency channels.

3. The method according to claim 1 , wherein the signal from the opposite end is taken into account during a transmission phase.

4. The method according to claim 3 , wherein in order to determine the error of the first reference channel, a first difference is first formed between a reception signal of the first reference channel after a transformation into the frequency domain and the reception signal of this channel downstream of a decision element, and a second difference is subsequently formed from the first difference and a transmission signal of the first reference channel.

5. The method as claimed in claim 1 , wherein during the initialization phase, the first reference channel is corrected by means of a second error formed from the difference between a transmission signal and a reception signal of a second reference channel of the frequency channels.

6. The method according to claim 5 , wherein the second error is added to the signals of the first reference channel.

7. The method according to claim 6 , wherein the second error is weighted with a coefficient.

8. The method according to claim 5 , wherein the signal from the opposite end is taken into account in the second error during the transmission phase.

9. The method according to claim 8 , wherein in order to determine the error of the second reference channel, a first difference is first fanned between the reception signal of the reference channel after the transformation into the frequency domain and the reference signal of this channel downstream of the decision element, and a second difference is subsequently formed from the first difference and the transmission signal of the second reference channel.

10. The method according to claim 1 , wherein the coefficients are determined using an iterative algorithm for error minimization.

11. A device for compensating for echoes during discrete multitone modulation (DMT), said device comprising:

an echo-compensation system, which operates in a frequency domain and is connected to a plurality of frequency channels, having at least one of said plurality of frequency channels defined as a reference channel and others of the plurality of frequency channels defined as remaining DMT channels, said echo compensation system comprising:

an error determining unit, which is connected to a transmission path and a reception path of said reference channel for deriving a first error;

a weighting unit which is connected to an output of the error determining unit for weighting said first error;

an addition unit, which is connected between said weighting unit and the reception oath of said remaining DMT channels for adding said first error to said remaining DMT frequency channels;

a Fast Fourier Transformer unit, which is arranged upstream of a first and a second connection point; and

a decision element, which is arranged downstream of said first connection point and said second connection point, at which said determining unit is connected to said reception path and said addition unit is connected to said reception path, respectively.

12. A system for compensating for echoes during discrete multitone modulation DMT, a compensation being performed in a frequency domain, comprising:

means for deriving a first error of a frequency channel during an initialization phase in which there is no signal from an opposite end, wherein the first error of the frequency channel is defined as a first reference channel, and wherein the first error is derived from a difference between a transmission value and a reception value of the first reference channel, and

means for said first error to correct other DMT frequency channels by weighting said first error with coefficients which are specific to said other DMT frequency channels, prior to adding said first error to said other DMT frequency channels.