Adaptive nonlinear processor for echo cancellation

Details Adaptive nonlinear processor for echo cancellation Adaptive nonlinear processor for echo cancellation

1998-11-10

2002-07-23

Ton, Dang (Department: 2661)

Multiplex communications

Duplex

Transmit/receive interaction control

C379S406010

Reexamination Certificate

active

06424635

ABSTRACT:

BACKGROUND
1. Field
The present invention relates to the field of communications. In particular, this invention relates to a system and method for enhancing the quality of audio transferred over a network.
2. General Background
For many years, networks have been used to transfer voice data between a source and a destination. As shown in
FIG. 1
, a typical voice-based network
100
comprises a local (near-end) telephone
110
and a remote (far-end) telephone
120
coupled together over a bidirectional transmission path featuring a two-wire “local loop” line
130
in communication with a long-distance (four-wire) “trunk” line
131
of network
100
. The trunk line
131
includes a pair of separated two-wire lines to transmit information to and receive information from remote telephone
120
.
In particular, before reaching local telephone
110
, an incoming voice data is provided to a hybrid circuit
140
in a local terminal
150
. A “local terminal” is equipment maintained by a telephone company to route voice data to a telephone in a predetermined area. Hybrid circuit
140
is a transformer circuit that enables the voice data to be transmitted over different types of transmission lines. For example, hybrid circuit
140
may be designed to act as an interface between local loop line
130
(e.g., twisted pair) and trunk line
131
(e.g., fiber optics, microwave or any digital transport medium). Since hybrid circuit
140
is not perfectly terminated, a talker echo will occur.
“Talker echo” is generally defined as a particular amount of energy associated with incoming voice data that is reflected back to the caller from which the voice data originated. Typically, for voice-based networks, talker echo is experienced when a person speaking on the telephone can hear his or her own spoken words after a short delay. To mitigate talker echo, an echo canceler
160
is coupled to input and output transmission lines of trunk line
131
.
As further shown in
FIG. 1
, echo canceler
160
is a circuit that observes, but does not modify, an incoming voice data before transmission to local telephone
110
. Instead, echo canceler
160
performs an echo cancellation technique which includes the following: (1) estimating an amount of talker echo returning to the caller, (2) simulating the estimated echo, and (3) subtracting the estimate echo from outgoing voice data. Then, in most cases, a nonlinear processor (NLP)
161
of echo canceler
160
performs a center clipping transfer function to remove any echo residue as shown in FIG.
2
. The removal of the echo residue is accomplished by clipping a portion of the outgoing voice data if it falls below selected clipping amplitude threshold levels A1 205 and -A1 210. These levels are measured in decibels such as decibel power levels (dBm
0
) for example. The clipped portion of the outgoing voice data is substituted for comfort noise that emulates a background noise level in order to avoid perceivable, audible artifacts.
One problem with the above-described echo cancellation technique is that the conventional NLP analyzes numerous parameters to determine the clipping amplitude threshold. In many cases, these parameters vary substantially over a short time duration that, in some situations, causes clipping errors to occur. For example, presume that NLP uses the amplitude level of far-end incoming voice data to calculate (i) the clipping amplitude threshold and (ii) whether a double talk condition exists (e.g., the two callers are talking simultaneously). If both callers are talking at substantially different amplitude ranges (e.g., the voice level of the far-end caller is approximately 10 or more decibels “dBs” higher than the voice level of the near-end caller), the outgoing voice data may be accidentally clipped.
Hence, it would be desirable to provide a system and method that minimizes the complexity of the NLP and enhances the quality of all audio by reducing the likelihood of erroneous clipping of the audio. This is accomplished through dynamic adjustment of the clipping amplitude threshold.
SUMMARY OF THE INVENTION
In one embodiment, the present invention relates to an echo canceler comprising a nonlinear processor utilizing a dynamic clipping amplitude threshold to determine whether or not to clip outgoing digital signals. The dynamic clipping amplitude threshold is increased when the ratio between the residue echo and echo is decreased over a prescribed period of time. In one embodiment, the residue echo is taken as a difference between a true echo and an estimated echo initially emulated by the echo canceler. This mitigates the tendency of erroneous clipping of the digital signals. Likewise, when the ratio is increased over the prescribed period of time, the dynamic clipping amplitude threshold is decreased.


REFERENCES:
patent: 4064379 (1977-12-01), Horna
patent: 4679230 (1987-07-01), Lassaux et al.
patent: 4751730 (1988-06-01), Galand et al.
patent: 4912758 (1990-03-01), Arbel
patent: 4965822 (1990-10-01), Williams
patent: 4991166 (1991-02-01), Julstrom
patent: 4998241 (1991-03-01), Brox et al.
patent: 5274705 (1993-12-01), Younce et al.
patent: 5351291 (1994-09-01), Menez et al.
patent: 5535194 (1996-07-01), Ashley et al.
patent: 5668865 (1997-09-01), Duttweiler et al.
patent: 5687229 (1997-11-01), Sih
patent: 5809085 (1998-09-01), Goodson et al.
patent: 6061651 (2000-05-01), Nguyen

Affiliated with

Also associated with

Rating

Adaptive nonlinear processor for echo cancellation does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Adaptive nonlinear processor for echo cancellation, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Adaptive nonlinear processor for echo cancellation will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2891863