Telephonic communications – Echo cancellation or suppression – Using digital signal processing
Reexamination Certificate
1999-03-11
2004-03-16
Tsang, Fan (Department: 2644)
Telephonic communications
Echo cancellation or suppression
Using digital signal processing
C379S406020, C379S406070, C379S406010, C379S406090, C379S406050, C379S406100
Reexamination Certificate
active
06707912
ABSTRACT:
FIELD OF THE INVENTION
The present invention pertains to echo cancellers, and more particularly to an adaptive step-size thereof.
BACKGROUND OF THE INVENTION
Operation of bi-directional hands-free communication devices, such as two-way radios, speaker phones which are commonly referred to as hands-free telephones, teleconferencing devices and car-kits for cellular telephones, requires management of those signals emitted by an audio speaker that are coupled to a microphone. Such devices are utilized in systems having a communication channel, such as a cable, twisted-wire, optical fiber, a frequency bandwidth for signals transmitted by air, or the like, which connects a local, or “near-end,” device to a remote, or “far-end,” device. Signals output by the speaker of a device that are detected by the microphone of the device are commonly called echoes. Echoes which occur at one end of the communication link are generally a nuisance to users at the other end. In severe cases, echoes can result in a phenomena known as “howling” which is very unpleasant at both ends of a communication link.
Echo cancellers have been developed to cancel echo signals. Echo cancellers employ a filter to estimate the echo signal in a communication device. The echo canceller subtracts the echo estimate from the signal output by the microphone to produce an echo suppressed signal.
Although echo cancellers work well in some environments, the effective cancellation of echo signals in a hands-free vehicle environment is particularly challenging. Least means squares (LMS) error minimization recursive filters are often used for echo control. Although these filters are very stable, nonlinear system effects, as well as limitations of algorithmic and arithmetic precision, limit the effectiveness of echo cancellers incorporating such filters. Post processing stages are thus employed to suppress residual echoes. Post processing can include attenuation of the output signal using a variable gain control or a filter, or other known post processing techniques.
Unfortunately, post processing can result in significant degradation and attenuation of desired transmission signals that are present when both users are speaking simultaneously (double talk condition). For example, post processing attenuation can result in the echo canceled taking on half-duplex characteristics, such that only one user can speak at a time. Additionally, post processing typically introduces perceptible changes, or attenuation, of the background noise, which is present in noisy environments such as vehicle interiors. When this noise variation correlates with speech activity in the signal received at the far end, it is objectionable to far-end users.
A variety of methods have been developed to improve echo canceller performance. Such methods include storing initial tap values for the echo canceller. Storing such initial tap values may speed up echo canceller operation upon initialization. However, if the present initial conditions are dramatically different from the previous initial conditions, starting from the previous conditions will not be beneficial.
Another known method of improving echo canceller operation employs an adaptive step size, such that the error signal gain fed back to the adaptive filter is varied as a function of the magnitude of the error signal. If the error signal is large, the step size will be large. If the error signal is small, the step size is small. Although echo cancellers operating in this manner provide improved performance by adapting more quickly, the echo canceller adaptation must still occur without having initial state unique to the environment of the echo canceller.
Accordingly, there is a need or an improved echo canceller to provide more rapid convergence based upon the environment of the echo canceller.
REFERENCES:
patent: 4783818 (1988-11-01), Graupe et al.
patent: 4894820 (1990-01-01), Miyamoto et al.
patent: 5022082 (1991-06-01), Eriksson et al.
patent: 5475731 (1995-12-01), Rasmusson
patent: 5764753 (1998-06-01), McCaslin et al.
patent: RE35867 (1998-08-01), Sugiyama
patent: 5859907 (1999-01-01), Kawahara et al.
patent: 5933494 (1999-08-01), Yang et al.
patent: 5951626 (1999-09-01), Duttweiler
patent: 6061444 (2000-05-01), Kawahara et al.
patent: 6081593 (2000-06-01), Kim
patent: 6147979 (2000-11-01), Michel et al.
patent: 6421326 (2002-07-01), Lee et al.
patent: 0023056 (1981-01-01), None
patent: 0708536 (1996-04-01), None
patent: 0856834 (1998-08-01), None
patent: 070193529 (1995-07-01), None
Duttweiler, “Propotionate Normalized Least-Mean-Squares Adapataion in Echo Canclellers”, 2000, IEEE Trans. on Speech an Audio Processing, vol. 8, No. 5, Sep. 2000; pp. 508-518.*
Exponentially Weighted Stepsize NLMS Adaptive Filter Based on the Statistics of a Room Impulse Response, Shoji Makino, Member, IEEE, Yutaka Kaneda, Member, IEEE and Nobuo Koizumi, IEEE Translations on Speech and Audio Processing, vol. 1, No. 1, Jan. 1993.
Chen Ernest Pei-Ching
Piket James Brian
Stephens James Allen
Yip William Chunhung
Motorola Inc.
Singh Ramnandan
Tsang Fan
Vaas Randall S.
LandOfFree
Method and apparatus for setting a step size for an adaptive... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for setting a step size for an adaptive..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for setting a step size for an adaptive... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3231378