The present invention relates generally to communication networks and in particular to an echo cancelling system and method for cancelling echos in a communication path.
Echo cancellers to cancel echos in a communication path are well known. A typical echo canceller generates a filter that models the transfer function of the echo signal path using a linear algorithm such as a least-mean-squared (LMS) algorithm. Network signals conveyed to the echo signal path are also conveyed to the echo canceller. The filter therein processes the network signals to generate an estimated copy of echo signals received from the echo signal path. The estimated echo signals and actual echo signals received from the echo signal path are subtracted. Differences between the estimated echo signals and the actual echo signals result in error signals, which are fed back to the echo canceller so that the echo canceller can converge to the correct transfer function.
When dealing with long delay networks, the echo signal path can change and can be different from call to call. As a result, an adaptive algorithm is required to allow the echo canceller to model adaptively the echo signal path. Unfortunately, the longer the time window or delay of the echo signals that are required to be cancelled, the longer the adaptive algorithm takes to adapt to the correct transfer function.
In addition, as the total time window increases, the number of coefficients in the filter generated by the adaptive algorithm increases, requiring more processing power. Placing conventional echo cancellers in long delay networks to cancel echos from traditional networks such as public switched telephone networks (PSTNs) or private branch exchanges (PBXs) increases the complexity and cost significantly.
It is therefore an object of the present invention to provide a novel echo cancelling system and method and a novel echo locator.
In accordance with one aspect of the present invention there is provided an echo cancelling system for cancelling echoes in a communication path comprising:
In accordance with another aspect of the present invention there is provided a method of cancelling echoes in a communication path comprising the steps:
In accordance with yet another aspect of the invention there is provided an echo locator to locate echoes in a communication path comprising:
The present invention provides advantages in that since the position of echos in the echo signal time window are determined, the echo canceller can be made active only in those areas in the time window. As a result, a reduced number of filter coefficients used to generate the filter modeling the echo signal path need to be active. This allows the echo canceller to adapt faster to changes in the echo signal path. In addition, by reducing the number of active filter coefficients, fewer processing cycles are required to generate the filter. Furthermore, the echo canceller is more robust in adverse conditions.
An embodiment of the present invention will now be described more fully with reference to the accompanying drawings in which:
Turning now to
As mentioned previously, the algorithm executed by the echo canceller 10 takes longer to converge to the correct transfer function as the time window of echo signals received from the communication path 12 increases. As a result, echo signals pass through the echo canceller 10.
To deal with the above, a network echo locator 30 in accordance with the present invention is provided and is used in conjunction with an echo canceller 32. As can be seen in
Turning now to
The peaks in the aliased transfer function of the echo locator 30 represent the areas of active echoes in the communication path 34. Only these areas need to be active in the echo canceller 32 in order to cancel the echoes. For example, if M=4 and the peaks in the aliased transfer function of the LMS processor 42 are at positions 3, 4 and 5 and 10, 11, and 12, then the filter coefficients or taps in the LMS algorithm of the echo canceller 32 which need to be made active by the echo locator 30 are:
3×M=12 up to 5×M=20; and
10×M=4 up to 12×M=48.
The remaining taps of the echo canceller are considered to be equal to zero and do not need to be used in the adaptation calculation or filtering of the echo canceller.
If desired, a weighting function can be applied to the taps in the echo canceller 32 based on the magnitude of the peaks in the aliased transfer function of the echo locator 30. In this case, higher peaks in the transfer function of the echo locator result in higher weightings applied to the corresponding taps in the echo canceller.
Turning now to
Although a preferred embodiment of the present invention has been described, those of skill in the art will appreciate that variations and modifications may be made without departing from the spirit and scope thereof as defined by the appended claims.
Number | Date | Country | Kind |
---|---|---|---|
9921555 | Sep 1999 | GB | national |
Number | Name | Date | Kind |
---|---|---|---|
4823382 | Martinez | Apr 1989 | A |
5295136 | Ashley et al. | Mar 1994 | A |
5343522 | Yatrou et al. | Aug 1994 | A |
5400394 | Raman et al. | Mar 1995 | A |
5649012 | Gupta et al. | Jul 1997 | A |
5724485 | Rainton | Mar 1998 | A |
5818945 | Makino et al. | Oct 1998 | A |
6185300 | Romesburg | Feb 2001 | B1 |
Number | Date | Country |
---|---|---|
0798877 | Oct 1997 | EP |
2123258 | Jan 1984 | GB |