Electrical audio signal processing systems and devices – Feedback suppression
Reexamination Certificate
2001-09-04
2004-01-27
Mei, Xu (Department: 2644)
Electrical audio signal processing systems and devices
Feedback suppression
C381S056000
Reexamination Certificate
active
06683961
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention relates to a filter system for eliminating loudspeaker-generated acoustic signals from a microphone signal, said system comprising a stereo sampling unit for sampling the interference-bearing microphone signal and the interfering loudspeaker signal in time (t) and for generating samples x(t) and z(t) which are applied to a computer, wherein said computer ascertains the transfer function H(f,T) between the loudspeaker and the microphone and provides said transfer function to a filter unit (
12
) adapted to process x(t), z(t) and H(f,T) to generate the interference-free microphone signal y(t).
2. Description of the Related Art
In many applications in telecommunications and in machine voice processing, a problem arises in that the microphone used for voice input may pick up—in addition to the voice signal to be transmitted or to be processed—signals which are generated by one or several loudspeakers used within the same audio system. In a variety of applications, loudspeaker interference of this kind may cause a variety of problems. Best known of these is the problem which arises when a mobile telephone or cellphone user wishing to converse by telephone inside a motor vehicle, for example, uses a so-called hands-free unit which amplifies the called party's signal for output from a loudspeaker. The hands-free microphone will transmit this loudspeaker signal to the partner in addition to ambient noise and the driver's speech signal. This results in the called partner hearing an “echo” of his/her own voice, which echo will be all the more disturbing as the signal propagation time is long. In mobile telephone networks, where signal propagation may take up to 300 ms, this phenomenon constitutes a major problem. As decoders in mobile telephone networks cause non-linear distortion, subsequent echo elimination is all the more difficult in such a case. For this reason, it is absolutely necessary to eliminate the aforesaid loudspeaker interference where it arises, i.e. in the motor vehicle's hands-free device in the example given above.
Another example of the problems caused by loudspeaker interference relates to systems for the machine recognition of voice. The use of systems of this kind is increasing, so that the elimination of loudspeaker signals from the microphone signal controlling the voice recognition system has fundamental importance for these systems also. Voice recognition systems need an input signal as interference-free as possible so as to obtain satisfactory recognition rates. Again, conventional systems of the kind used in motor vehicles, for example, have a so-called “push-to-talk” button for muting the system loudspeaker before the system accepts voice input. This, then, is at variance with the desired hands-free operation, the main point of using a voice recognition system in the first place.
On the basis of the first-mentioned example above, the prior methods of solving the above-described problem are summed up under the term “echo cancelling”. All these prior methods are based on variants of an iterative process (LMS for “least mean square”) which seeks to minimize the echo signal by the successive adaptation of filter parameters.
The basic principle underlying these algorithms is known e.g. from the textbook by Peter Vary: Digitale Sprachsignalverarbeitung, chapter 13, Teubner Verlag, Stuttgart, 1998. It is described also in U.S. Pat. No. 5,475,731 and in EP 0 870 365; a variety of proposals for improvement are known by EP 0 988 744 and WO 00/16497. All these methods have in common that their results are optimum only if the microphone picks up the loudspeaker signal only. As soon as additional non-stationary acoustic disturbances are present, such as may occur in a motor vehicle in transit or in a so-called full-duplex situation, in which the partners at both ends of the link speak simultaneously, the performance of all these prior systems deteriorates considerably.
This also applies to a method which is known as an affine projection from U.S. Pat. No. 5,539,731. The improved convergence this method does in fact feature requires an increased computation effort, however.
As in all the aforesaid ones, a duplex situation or disturbing noise input to the microphone cause convergence problems in this method too. Also, the aforesaid methods have the problem of the convergence performance of the algorithms deteriorating rapidly as the filter length increases.
SUMMARY OF THE INVENTION
It is the object underlying the present invention to effectively and efficiently eliminate loudspeaker interference from the microphone signal used for voice input.
This object is attained by the measures set forth in claim
1
, in particular by a computation unit in which a Fourier transform unit determines the spectra X(f,T) and Z(f,T) of the microphone signal and the loudspeaker signal, respectively, and processes these to compute the transfer function H(f,T) in multiplying, smoothing and division layers as well as a convolution layer, with the signal propagation delay effective between the loudspeaker and the microphone—as determined in the delay unit—taken into account.
These measures act to effectively eliminate loudspeaker interference from the microphone signal used for voice input, obviating a push-to-talk key. If the loudspeaker signal is generated by a system controlled by voice input, such as a voice-controlled television receiver, the inventive system ensures that voice signals the television set itself generates will not be processed as voice control commands.
The range of applications in which the present invention is suitable is much broader, however. For this reason, reference will generally be had in the explanations that follow to the elimination of loudspeaker interference from a microphone signal.
The present invention presents a novel method in which a network-like computation unit is used to determine the transfer function of the loudspeaker-space system in a direct manner. The computation effort the method requires is quite modest, and the method is extremely resistant to noise disturbances and duplex situations as, unlike all other methods, it does not assume as a starting point the special case of other non-stationary signals—apart from the loudspeaker signal—being absent from the microphone.
These measures are convergent without problems for very long filters too, and the filter length is limited only by available memory. The system implements readily on any hardware platform; it is fully adaptive and does not have to be pre-configured for expected signal propagation delays. In contrast to other systems, it does not call for the prior detection of duplex situations. The system is stable as soon as a loudspeaker signal having sufficient intensity is available. Also, the system will automatically and rapidly track changes in the acoustic system conditions.
Additional advantageous measures are described in the dependent claims. The invention is illustrated in the attached drawings and is described in greater detail hereinbelow.
REFERENCES:
patent: 5694476 (1997-12-01), Klippel
patent: 5768398 (1998-06-01), Janse et al.
patent: 6111957 (2000-08-01), Thomasson
LandOfFree
Process and apparatus for eliminating loudspeaker... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Process and apparatus for eliminating loudspeaker..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Process and apparatus for eliminating loudspeaker... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3255161