DESC:TECHNICAL FIELD
[1] The present disclosure generally relates to improve communication via an acoustic duplex communication device. In particular, the present disclosure relates to a means to attenuate acoustic echo in an acoustic duplex communication device.

BACKGROUND
[2] Background description includes information that may be useful in understanding the present invention. It is not an admission that any of the information provided herein is prior art or relevant to the presently claimed invention, or that any publication specifically or implicitly referenced is prior art.
[3] A duplex communication device is any device that allows for two-way communication, i.e., it allows a user to transmit and receive voice communication. Examples of such devices may include, without limitations, mobile phones, telephones, computer devices, laptops, tablets, smart media, etc. Such devices may generally be quite compact. The device may include a speaker and a microphone. The loudspeaker may generate first acoustic signals. The first acoustic signals may be communication to be received by the user. The microphone may be configured to receive, as input, voice signals. The microphone may receive a plurality of inputs, such as a target acoustic signal, which may be a communication to be transmitted from the user, ambient noise, and the first acoustic signals from the loudspeaker of the audio device. The communication to be transmitted by the user may be second acoustic signals. The first and second acoustic signals may be generated either concurrently or separately, relative to each other. As a result, the microphone may be receiving, along with the target communication signal, several other acoustic signals, which is undesirable. One of the undesirable signals that significantly impacts the communication is an echo signal, which is first acoustic signal captured by the microphone.
[4] There is, therefore, a requirement in the art for a means to attenuate acoustic echo in signals being transmitted by a duplex communication device.

OBJECTS OF INVENTION
[5] An object of the present invention is to provide a system and method for attenuating acoustic echo in a duplex communication device.
[6] Another object of the present invention is to provide a system to attenuate acoustic echo in both linear and non-linear domains.

SUMMARY
[7] The present disclosure generally relates to improve communication via an acoustic duplex communication device. In particular, the present disclosure relates to a means to attenuate acoustic echo in an acoustic duplex communication device.
[8] In a first aspect, the present disclosure provides a system for attenuating acoustic echo in an acoustic duplex communication device. The system includes an audio device provided on the acoustic duplex communication device, adapted to generate first acoustic signals. The system further includes a microphone provided on the acoustic duplex communication device, adapted to receive, as input, acoustic signals. The system further includes an acoustic source disposed in a vicinity of the acoustic duplex communication device, adapted to generate second acoustic signals, wherein the second acoustic signals are generated any one or both of separately and concurrently, relative to the generated first acoustic signals from the audio device. The system further includes a controller communicably coupled to the audio device and the microphone, the controller including a processor and a memory communicably coupled to the processor. The controller is configured to receive, from the microphone, input acoustic signals, wherein the input acoustic signals include at least the second acoustic signals, acoustic echo signals corresponding to the first acoustic signals, and ambient noise. The controller is further configured to model the acoustic echo signals from the received input acoustic signals using any one of a plurality of Cepstrum-based models to obtain a minimum phase component of the acoustic echo signals. The controller is further configured to determine a first minimum phase function corresponding to the modelled acoustic echo signals defined based on a Cepstrum decomposition of the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals. The controller is further configured to apply a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.
[9] In some embodiments, the controller is further configured to model the acoustic echo signals from the received input acoustic signals using any one of a plurality of models to obtain an all-pass component of the acoustic echo signals. The controller is further configured to apply an all-pass filter to attenuate the all-pass component of the acoustic signals from the input acoustic signals.
[10] In some embodiments, responsive to one model of the plurality of models further including a complex Cepstrum, the controller is further configured to determine a first minimum phase function corresponding to the modelled acoustic echo signals defined based on real and complex components of a Fourier transform the acoustic echo signals. The first minimum phase function is indicative of the minimum phase component of the acoustic echo signals. The controller is further configured to apply a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.
[11] In some embodiments, the controller is further configured to determine an impulse response of the first minimum phase function defined based on an exponential of a fast Fourier transform of a noise component of the input acoustic signals. The controller is further configured to apply an impulse response filter to improve a signal to noise ratio of the input acoustic signals.
[12] In some embodiments, responsive to one model of the plurality of models further including recursion, the controller is further configured to determine a first minimum phase function corresponding to the modelled acoustic echo signals defined based on an exponential of the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals. The controller is further configured to apply a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.
[13] In some embodiments, the controller is further configured to determine an impulse response of the first minimum phase function defined based on an exponential of the acoustic echo signals. The controller is further configured to apply an impulse response filter to improve a signal to noise ratio of the input acoustic signals.
[14] In some embodiments, responsive to one model of the plurality of models further including machine learning, the controller is further configured to determine the first minimum phase function corresponding to the modelled acoustic echo signals defined based on real and complex Cepstrum coefficients of the acoustic echo signals.
[15] In a second aspect, the present disclosure provides a method for attenuating acoustic echo in an acoustic duplex communication device. The method includes providing the acoustic duplex communication device provided with an audio device provided on the acoustic duplex communication device, adapted to generate first acoustic signals. The acoustic duplex communication device is further provided with a microphone provided on the acoustic duplex communication device, adapted to receive, as input, acoustic signals. The method further includes providing an acoustic source disposed in a vicinity of the acoustic duplex communication device, adapted to generate second acoustic signals, wherein the second acoustic signals are generated any one or both of separately and concurrently, relative to the generated first acoustic signals from the audio device. The method further includes receiving, by a controller, input acoustic signals, wherein the input acoustic signals include at least the second acoustic signals, acoustic echo signals corresponding to the first acoustic signals, and ambient noise. The method further includes modeling, by the controller, the acoustic echo signals from the received input acoustic signals using any one of a plurality of Cepstrum-based models to obtain a minimum phase component of the acoustic echo signals. The method further includes determining, by the controller, a first minimum phase function corresponding to the modelled acoustic echo signals defined based on a Cepstrum decomposition of the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals. The method further includes applying, by the controller, a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.
[16] In some embodiments, the method further includes modeling, by the controller, the acoustic echo signals from the received input acoustic signals using any one of a plurality of models to obtain an all-pass component of the acoustic echo signals. The method further includes applying, by the controller, an all-pass filter to attenuate the all-pass component of the acoustic signals from the input acoustic signals.
[17] In some embodiments, responsive to one model of the plurality of models further including a complex Cepstrum, the method further includes determining, by the controller, a first minimum phase function corresponding to the modelled acoustic echo signals defined based on real and complex components of a Fourier transform the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals. The method further includes applying, by the controller, an all-pass filter to attenuate the all-pass component of the acoustic signals from the input acoustic signals.
[18] In some embodiments, the method further includes determining, by the controller, an impulse response of the first minimum phase function defined based on an exponential of a fast Fourier transform of a noise component of the input acoustic signals. The method further includes applying, by the controller, an impulse response filter to improve a signal to noise ratio of the input acoustic signals.
[19] In some embodiments, responsive to one model of the plurality of models further including recursion, the method further includes determining, by the controller, a first minimum phase function corresponding to the modelled acoustic echo signals defined based on an exponential of the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals. The method further includes applying, by the controller, a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.
[20] In some embodiments, the method further includes determining, by the controller, an impulse response of the first minimum phase function defined based on an exponential of the acoustic echo signals. In some embodiments, the method further includes applying, by the controller, an impulse response filter to improve a signal to noise ratio of the input acoustic signals.
[21] In some embodiments, responsive to one model of the plurality of models further including machine learning, the method further includes determining, by the controller, the first minimum phase function corresponding to the modelled acoustic echo signals defined based on real and complex Cepstrum coefficients of the acoustic echo signals.
[22] In a third aspect, the present disclosure provides a non-transitory, computer readable storage device storing instructions, which when executed by a processor causes the processor to perform a process to attenuate acoustic echo in an acoustic duplex communication device. The process includes receiving, by a controller, input acoustic signals, wherein the input acoustic signals include at least the second acoustic signals, acoustic echo signals corresponding to the first acoustic signals, and ambient noise. The process further includes modeling, by the controller, the acoustic echo signals from the received input acoustic signals using any one of a plurality of Cepstrum-based models to obtain a minimum phase component of the acoustic echo signals. The process further includes determining, by the controller, a first minimum phase function corresponding to the modelled acoustic echo signals defined based on a Cepstrum decomposition of the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals. The process further includes applying, by the controller, a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.
[23] Various objects, features, aspects, and advantages of the inventive subject matter will become more apparent from the following detailed description of preferred embodiments, along with the accompanying drawing figures in which like numerals represent like components.

BRIEF DESCRIPTION OF DRAWINGS
[24] The accompanying drawings are included to provide a further understanding of the present disclosure and are incorporated in and constitute a part of this specification. The drawings illustrate exemplary embodiments of the present disclosure and, together with the description, serve to explain the principles of the present disclosure.
[25] FIG. 1 illustrates a schematic representation of an environment including an acoustic duplex communication device, according to an embodiment of the present disclosure;
[26] FIG. 2 illustrates a schematic representation of operation of the system to attenuate the acoustic echo, according to an embodiment of the present disclosure;
[27] FIG. 3 illustrates a schematic block diagram of a controller of the system to attenuate the acoustic echo, according to an embodiment of the present disclosure;
[28] FIG. 4 illustrates a schematic flow diagram for a method to attenuate the acoustic echo, according to an embodiment of the present disclosure;
[29] FIG. 5 illustrates an exemplary representation of a Cepstrum operation;
[30] FIG. 6 illustrates another representation of the Cepstrum operation;
[31] FIG. 7 illustrates an operation to model acoustic echo using complex Cepstrum, according to an embodiment of the present disclosure;
[32] FIG. 8 illustrates an operation to model acoustic echo using complex Cepstrum, according to an embodiment of the present disclosure;
[33] FIGs. 9 - 10 illustrate an operation to model acoustic echo using real Cepstrum, according to an embodiment of the present disclosure;
[34] FIG. 11 illustrates an operation to model acoustic echo using discrete Fourier Transform (DFT), according to an embodiment of the present disclosure;
[35] FIGs. 12A – 12D illustrate schematic representations of an operation to model acoustic echo using machine learning approach, according to an embodiment of the present disclosure; and
[36] FIG. 13 illustrates an exemplary schematic block diagram of a hardware platform for implementation of the system.

DETAILED DESCRIPTION
[37] The following is a detailed description of embodiments of the disclosure depicted in the accompanying drawings. The embodiments are in such details as to clearly communicate the disclosure. However, the amount of detail offered is not intended to limit the anticipated variations of embodiments; on the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the present disclosure as defined by the appended claims.
[38] In a first aspect, the present disclosure provides a system for attenuating acoustic echo in an acoustic duplex communication device. The system includes an audio device provided on the acoustic duplex communication device, adapted to generate first acoustic signals. The system further includes a microphone provided on the acoustic duplex communication device, adapted to receive, as input, acoustic signals. The system further includes an acoustic source disposed in a vicinity of the acoustic duplex communication device, adapted to generate second acoustic signals, wherein the second acoustic signals are generated any one or both of separately and concurrently, relative to the generated first acoustic signals from the audio device. The system further includes a controller communicably coupled to the audio device and the microphone, the controller including a processor and a memory communicably coupled to the processor. The controller is configured to receive, from the microphone, input acoustic signals, wherein the input acoustic signals include at least the second acoustic signals, acoustic echo signals corresponding to the first acoustic signals, and ambient noise. The controller is further configured to model the acoustic echo signals from the received input acoustic signals using any one of a plurality of Cepstrum-based models to obtain a minimum phase component of the acoustic echo signals. The controller is further configured to determine a first minimum phase function corresponding to the modelled acoustic echo signals defined based on a Cepstrum decomposition of the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals. The controller is further configured to apply a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.
[39] In some embodiments, the controller is further configured to model the acoustic echo signals from the received input acoustic signals using any one of a plurality of models to obtain an all-pass component of the acoustic echo signals. The controller is further configured to apply an all-pass filter to attenuate the all-pass component of the acoustic signals from the input acoustic signals.
[40] In some embodiments, responsive to one model of the plurality of models further including a complex Cepstrum, the controller is further configured to determine a first minimum phase function corresponding to the modelled acoustic echo signals defined based on real and complex components of a Fourier transform the acoustic echo signals. The first minimum phase function is indicative of the minimum phase component of the acoustic echo signals. The controller is further configured to apply a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.
[41] In some embodiments, the controller is further configured to determine an impulse response of the first minimum phase function defined based on an exponential of a fast Fourier transform of a noise component of the input acoustic signals. The controller is further configured to apply an impulse response filter to improve a signal to noise ratio of the input acoustic signals.
[42] In some embodiments, responsive to one model of the plurality of models further including recursion, the controller is further configured to determine a first minimum phase function corresponding to the modelled acoustic echo signals defined based on an exponential of the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals. The controller is further configured to apply a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.
[43] In some embodiments, the controller is further configured to determine an impulse response of the first minimum phase function defined based on an exponential of the acoustic echo signals. The controller is further configured to apply an impulse response filter to improve a signal to noise ratio of the input acoustic signals.
[44] In some embodiments, responsive to one model of the plurality of models further including machine learning, the controller is further configured to determine the first minimum phase function corresponding to the modelled acoustic echo signals defined based on real and complex Cepstrum coefficients of the acoustic echo signals.
[45] In a second aspect, the present disclosure provides a method for attenuating acoustic echo in an acoustic duplex communication device. The method includes providing the acoustic duplex communication device provided with an audio device provided on the acoustic duplex communication device, adapted to generate first acoustic signals. The acoustic duplex communication device is further provided with a microphone provided on the acoustic duplex communication device, adapted to receive, as input, acoustic signals. The method further includes providing an acoustic source disposed in a vicinity of the acoustic duplex communication device, adapted to generate second acoustic signals, wherein the second acoustic signals are generated any one or both of separately and concurrently, relative to the generated first acoustic signals from the audio device. The method further includes receiving, by a controller, input acoustic signals, wherein the input acoustic signals include at least the second acoustic signals, acoustic echo signals corresponding to the first acoustic signals, and ambient noise. The method further includes modeling, by the controller, the acoustic echo signals from the received input acoustic signals using any one of a plurality of Cepstrum-based models to obtain a minimum phase component of the acoustic echo signals. The method further includes determining, by the controller, a first minimum phase function corresponding to the modelled acoustic echo signals defined based on a Cepstrum decomposition of the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals. The method further includes applying, by the controller, a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.
[46] In some embodiments, the method further includes modeling, by the controller, the acoustic echo signals from the received input acoustic signals using any one of a plurality of models to obtain an all-pass component of the acoustic echo signals. The method further includes applying, by the controller, an all-pass filter to attenuate the all-pass component of the acoustic signals from the input acoustic signals.
[47] In some embodiments, responsive to one model of the plurality of models further including a complex Cepstrum, the method further includes determining, by the controller, a first minimum phase function corresponding to the modelled acoustic echo signals defined based on real and complex components of a Fourier transform the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals. The method further includes applying, by the controller, an all-pass filter to attenuate the all-pass component of the acoustic signals from the input acoustic signals.
[48] In some embodiments, the method further includes determining, by the controller, an impulse response of the first minimum phase function defined based on an exponential of a fast Fourier transform of a noise component of the input acoustic signals. The method further includes applying, by the controller, an impulse response filter to improve a signal to noise ratio of the input acoustic signals.
[49] In some embodiments, responsive to one model of the plurality of models further including recursion, the method further includes determining, by the controller, a first minimum phase function corresponding to the modelled acoustic echo signals defined based on an exponential of the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals. The method further includes applying, by the controller, a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.
[50] In some embodiments, the method further includes determining, by the controller, an impulse response of the first minimum phase function defined based on an exponential of the acoustic echo signals. In some embodiments, the method further includes applying, by the controller, an impulse response filter to improve a signal to noise ratio of the input acoustic signals.
[51] In some embodiments, responsive to one model of the plurality of models further including machine learning, the method further includes determining, by the controller, the first minimum phase function corresponding to the modelled acoustic echo signals defined based on real and complex Cepstrum coefficients of the acoustic echo signals.
[52] In a third aspect, the present disclosure provides a non-transitory, computer readable storage device storing instructions, which when executed by a processor causes the processor to perform a process to attenuate acoustic echo in an acoustic duplex communication device. The process includes receiving, by a controller, input acoustic signals, wherein the input acoustic signals include at least the second acoustic signals, acoustic echo signals corresponding to the first acoustic signals, and ambient noise. The process further includes modeling, by the controller, the acoustic echo signals from the received input acoustic signals using any one of a plurality of Cepstrum-based models to obtain a minimum phase component of the acoustic echo signals. The process further includes determining, by the controller, a first minimum phase function corresponding to the modelled acoustic echo signals defined based on a Cepstrum decomposition of the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals. The process further includes applying, by the controller, a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.
[53] FIG. 1 illustrates a schematic representation of an environment 100 including an acoustic duplex communication device 150, according to an embodiment of the present disclosure. A duplex communication device is any device that allows for two-way communication, i.e., it allows a user 190 to transmit and receive acoustic communication. Examples of such devices may include, without limitations, mobile phones, telephones, computer devices, laptops, tablets, smart media, etc. Such devices may generally be quite compact.
[54] The duplex communication device 150 may interchangeably be referred to as “the device 150”. The device 150 includes an audio device 152, such as a speaker, through which the device 150 may generate first acoustic signals. The first acoustic signals may be communication to be received by the user 190. The device 150 further includes a microphone 154 configured to receive, as input, acoustic signals. The microphone 154 may receive a plurality of inputs, such as a target acoustic signal, which may be a communication to be transmitted from the user 190, ambient noise, and the first acoustic signals from the audio device 152 of the device 150. The communication to be transmitted by the user 190 may be second acoustic signals. The first and second acoustic signals may be generated either concurrently or separately, relative to each other. As a result, the microphone 154 may be receiving, along with the target communication, several other acoustic signals, which may result in an acoustic echo that may also get transmitted, which is undesirable.
[55] The device 150 may be provisioned with a system 200 configured to attenuate acoustic echo. The system 200 may be configured on the device 150 or may be configured on an external device operatively coupled to the device 150.
[56] The second acoustic signal received by the microphone 154 at any instant n may be,

where,
– first signal from the audio device;
– acoustic echo path response;
– input signal to be preserved and transmitted;
– ambient noise; and
– acoustic echo component.
[57] The acoustic echo component is to be attenuated and the system 200 is configured to attenuate the acoustic echo component by applying a suitable filter. Considering ambient noise to be negligible, the input signal may be given as,

where, is the acoustic echo estimate.
[58] Generally, the first signal generated by the audio device may be in a linear or non-linear domain. The system 200 is configured to address estimation of acoustic echo as a deconvolution problem in the non-linear domain.
[59] Assuming that s[n] is absent,

[60] The system 200 is configured to utilize the deconvolution property of Cepstrum. The acoustic echo path h[n] may be factorized into a minimum phase filter in cascade with a causal all pass filter.

where,
– minimum phase sequence;
- is an allpass sequence;
[61] Here, h[n] is a real sequence, i.e., the poles and zeroes of H(z) are a complex conjugate pair, so that, sum of the poles are real. Further, is real, causal, and stable, i.e., whose poles and zeroes are inside the unit circle. Furthermore, is a stable sequence with poles inside the unit circle and zeroes outside the unit circle.
[62] FIG. 2 illustrates a schematic representation of operation of the system 200 to attenuate the acoustic echo, according to an embodiment of the present disclosure. The system 200 is configured to model the acoustic echo in two stages. Firstly, the magnitude or minimum phase response of is modelled by an improved Cepstrum approach. The minimum phase response includes information, such that subsequent frequency bins are spaced based on the phase of the input signal, even though individual frequency bins may not retain all of the phase information.
[63] Secondly, the all-pass response of is modelled by a known methodologies continuous adaptive algorithm based on frequency bins selective adaptation.
[64] The system 200 may include a controller 300 configured to implement protocols and techniques to attenuate the acoustic echo.
[65] FIG. 3 illustrates a schematic block diagram of the controller 300 of the system 200 to attenuate the acoustic echo, according to an embodiment of the present disclosure. The controller 300 is communicably coupled to the system 200 and includes a processor 302 communicably coupled to a memory 304. The memory 304 stores instructions executable by the processor 302. The controller 300 further includes a processing engine 310 configured to execute steps to attenuate the acoustic echo. The processing engine 310 includes an input acoustic signals engine 312, a modelling engine 314, a minimum phase function engine 316, a minimum phase filter engine 318, and other engine(s) 320. The other engine(s) 320 are configured to perform functions ancillary or supplemental to the processing engine 310.
[66] The input acoustic signals engine 312 is configured to receive, from the microphone, input acoustic signals. The input acoustic signals include at least the second acoustic signals, acoustic echo signals corresponding to the first acoustic signals, and ambient noise.
[67] The modelling engine 314 is configured to model the acoustic echo signals from the received input acoustic signals using any one of a plurality of Cepstrum-based models to obtain a minimum phase component of the acoustic echo signals.
[68] The minimum phase function engine 316 is configured to determine a first minimum phase function corresponding to the modelled acoustic echo signals defined based on an inverse Fourier transform of a logarithm of a Fourier transform of the acoustic echo signals. The first minimum phase function is indicative of the minimum phase component of the acoustic echo signals.
[69] The minimum phase filter engine 318 is configured to apply a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.
[70] FIG. 4 illustrates a schematic flow diagram of a method 400 to attenuate the acoustic echo, according to an embodiment of the present disclosure. Referring now to FIGs. 1 to 4, at step 402, the method 400 includes providing the acoustic duplex communication device 150 provided with the audio device 152 provided on the acoustic duplex communication device 150, adapted to generate first acoustic signals. At step 404, the method 400 further includes providing an acoustic source disposed in a vicinity of the acoustic duplex communication device 150, adapted to generate second acoustic signals, wherein the second acoustic signals are generated any one or both of separately and concurrently, relative to the generated first acoustic signals from the audio device 152. At step 406, the method 400 further includes receiving, by the controller 300, input acoustic signals, wherein the input acoustic signals include at least the second acoustic signals, acoustic echo signals corresponding to the first acoustic signals, and ambient noise. At step 408, the method 400 further includes modeling, by the controller 300, the acoustic echo signals from the received input acoustic signals using any one of a plurality of Cepstrum-based models to obtain a minimum phase component of the acoustic echo signals. At step 410, the method 400 further includes determining, by the controller 300, a first minimum phase function corresponding to the modelled acoustic echo signals defined based on an inverse Fourier transform of a logarithm of a Fourier transform of the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals. At step 412, the method 400 further includes applying, by the controller 300, a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.
[71] In some embodiments, the method 400 further includes modeling, by the controller 300, the acoustic echo signals from the received input acoustic signals using any one of a plurality of models to obtain an all-pass component of the acoustic echo signals. The method further includes applying, by the controller, an all-pass filter to attenuate the all-pass component of the acoustic signals from the input acoustic signals.
[72] In some embodiments, responsive to one model of the plurality of models further including a complex Cepstrum, the method 400 further includes determining, by the controller 300, a first minimum phase function corresponding to the modelled acoustic echo signals defined based on real and complex components of a Fourier transform the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals. The method 400 further includes applying, by the controller 300, an all-pass filter to attenuate the all-pass component of the acoustic signals from the input acoustic signals.
[73] In some embodiments, the method 400 further includes determining, by the controller 300, an impulse response of the first minimum phase function defined based on an exponential of a fast Fourier transform of a noise component of the input acoustic signals. The method 400 further includes applying, by the controller 300, an impulse response filter to improve a signal to noise ratio of the input acoustic signals.
[74] In some embodiments, responsive to one model of the plurality of models further including recursion, the method 400 further includes determining, by the controller 300, a first minimum phase function corresponding to the modelled acoustic echo signals defined based on an exponential of the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals. The method 400 further includes applying, by the controller 300, a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.
[75] In some embodiments, the method 400 further includes determining, by the controller 300, an impulse response of the first minimum phase function defined based on an exponential of the acoustic echo signals. In some embodiments, the method 400 further includes applying, by the controller 300, an impulse response filter to improve a signal to noise ratio of the input acoustic signals.
In some embodiments, responsive to one model of the plurality of models further including machine learning, the method 400 further includes determining, by the controller 300, the first minimum phase function corresponding to the modelled acoustic echo signals defined based on real and complex components of a Fourier transform of the acoustic echo signals.
[76] FIG. 5 illustrates an exemplary representation of a Cepstrum operation. The Cepstrum is defined as follows,

where F is the Fourier Transform and is the Inverse FT.

[77] FIG. 6 illustrates an operation to model acoustic echo using recursion, according to an embodiment of the present disclosure. The impulse response and Cepstrum can be related as




Similar to and can be written as power series





Taking inverse transform






[78] The impulse response and Ceptrum can be related as

[79] Assuming that is analytic, differentiating equation,



[80] Inverse transform of is given as,
where, c denotes closed contour.

for

where, is the complex cepstrum of all pass response
is the complex cepstrum of minimum phase response
is the complex cepstrum of
[81] Using the recursion, allpass response can be estimated as below

[82] FIG. 7 illustrates an operation to model acoustic echo using complex Cepstrum, according to an embodiment of the present disclosure. For a sequence , the Cepstrum may be given as,

[83] The spectra contains harmonics at evenly spaced intervals, whose magnitude decreases quite quickly as frequency bin index increases. Log operation on compresses its dynamic range and reduces the amplitude differences in the harmonics. Complex logarithm in place of real log in provides complex cepstrum. However, they are real sequence for real input sequence . The complex cepstrum involves the use of the complex log and the cepstrum involves only the log magnitude. Log function converts product into sums. Hence, Cepstrum is used to retrieve the impulse response which is a product or ratio or combination of product and ratio of different spectrums.



where
[84] The value of N may preferably be high (more than 512) in order to minimalize aliasing.
[85] For minimum phase signals (no poles or zeros outside unit circle) the complex cepstrum can be completely represented by the real part of the Fourier transforms. In other words, the complex cepstrum of minimum phase signals may be represented by the log of the magnitude of the FT alone. Since the real part of the FT is the FT of the even part of the sequence, thus, the complex cepstrum (for minimum phase signals) may be computed by computing the cepstrum and using the equations above.
[86] The impulse response to be modelled may be given as,




[87] Hence,

where and are the complex cepstrum of and respectively.
[88] Even though estimated from complex Cepstral coefficients is independent of error signal, the noise component in input acoustic signal to the microphone during noisy speech is not available explicitly for modeling the numerator Cepstrum. Hence, should be frozen during noisy speech regions. Thus, this approach can only improve the accuracy of modelling noise alone regions response but may not provide adequate echo modelling for preserving speech intelligibility.
[89] FIGs. 8 - 10 illustrate an operation to model acoustic echo using real Cepstrum, according to an embodiment of the present disclosure. Using the property of the Fourier Transform for the real sequence , the magnitude of the Fourier transform is even function of , and even part of transforms to . In other words, .
[90] Further, the real part of l is the cepstrum generating function . Therefore, the Cepstrum is corresponding to even part of the
[91] Hence, the complex cepstrum corresponding to may be given as,



where . is a constrained operator and is a frequency invariant function. Thus, if is causal, then it can be recovered form by a frequency invariant linear filtering .
[92] However, is based on only the Fourier transform magnitude it is not invertible, i.e., original cannot be re-obtained from . The complex cepstrum is somewhat more difficult to compute, but it is invertible. In other words, the properties of Cepstrum can be derived from the properties of the complex cepstrum and vice versa.
[93] Above equation indicates how the complex cepstrum can be computed from the cepstrum and consequently log magnitude alone if is real and causal. Therefore is the complex cepstrum corresponding to minimum phase frequency response and that preserves .




[94] Therefore, the minimum phase echo path response is estimated by filtering the cepstrum corresponding to cepstrum generating function using frequency invariant linear lifter.
[95] Further, the all pass component may be modelled as,

[96] In some embodiments, may be modelled using Power Ceptrum. L be the cepstrum generating function corresponding to power spectrum of ).








where
[97] The minimum phase echo path response is estimated by filtering the cepstrum estimated from the power spectrum using another frequency invariant linear lifter .
[98] FIG. 11 illustrates an operation to model acoustic echo using discrete Fourier Transform (DFT), according to an embodiment of the present disclosure. The practical implementation generally uses Discrete Fourier Transform DFT of finite length frames of signal. Since Causality or one sidedness of a real sequence has some strong constraints on Fourier transform of it, the Fourier transform of finite sequence to be more constrained and can be represented better by Discrete Fourier Transform (DFT).
[99] Since DFT involves sums rather integrals, the problem of improper integral disappears. However, DFT is in reality a representation of a periodic sequence and for finite sequence, hence the last N/2-1 points k[n] are zero. Because k[n] is zero in the second half of each period, k[-n] is zero in the first half of each period and consequently, except for n = 0 and n = N/2 there is no overlap between the non-zero portion of k[n] and k[-n]. Therefor for causal periodic sequences, the complex cepstrum obtained from from its even part is given as

where,

Thus,

]

where, ] is another frequency invariant liftering function.
[100] is the point discrete form of cepstrum generating function from power spectrum of impulse response. It can be written as a power series as below,

[101] The minimum phase response of the impulse response can be retrieved by replacing anti-causal exponents with the corresponding causal component, i.e., anti-causal Cepstrum exponentials to be flipped about time zero so that it adds to causal part. Therefore,


[102] From the above equations, the cepstrum function corresponding to min phase part that maintain the magnitude response of the impulse response is given by,


[103] The cepstrum generating function corresponding to power spectrum of the sequence will be obtained by multiplying the power spectrum of the sequence by two.
[104] FIGs. 12A – 12D illustrate schematic representations of an operation to model acoustic echo using machine learning (ML) approach, according to an embodiment of the present disclosure. Specifically, the minimum phase part of the echo is modeled using the ML approach. This model uses real cepstrum of the microphone and far-end signal and predicts the real cepstrum of acoustic echo. The proposed method uses a temporal convolution network (TCN) using dilated 1-D convolutional layers that are stacked together to create a large temporal receptive field with fewer parameters. The instant echo path modeling is achieved in the cepstral domain using minimum phase and all pass decomposition. The minimum phase impulse response is estimated using the difference of the corresponding cepstral coefficients between the reference and actual signal. Conventionally, true echo during the double talk is not available explicitly. Hence, it is significantly complex using a typical signal processing approach. However, ML methods can easily model the same.
[105] The proposed model uses the architecture of the temporal convolutional neural network to predict the real cepstrum coefficient corresponding to the minimum phase part of the echo path. Initially, the reference and microphone signals (y(n) and x(n), hereafter called input signal), sampled at 16 kHz are divided into 25ms window with a frame shift (hope length) of 10ms. Then, the windowed input is converted into real cepstrum coefficients. The dimension of concatenated cepstral coefficients from reference and microphone signal is 514. The output label is the cepstrum coefficients corresponding to Hmin as depicted in FIG. 12A and of dimension 257.
[106] The complete network is depicted in FIG. 12C. It has an encoder to encode the input cepstral coefficients, a separator to suppress near-end signal, and a decoder to retrieve the difference cepstral coefficients corresponding to Hmin. The encoder block is a linear layer used to reduce the dimension from 257 to B. B value is fixed as 128. TCN consists of dilated 1-D convolutional layers and the proposed architecture uses a stack of seven TCN blocks with two rounds to maintain a 1 sec large temporal receptive field with fewer parameters. The output of each TCN stack is recombined with the input using a skip connection to avoid losing low-level details. Finally, the decoder block which is a linear layer uses a linear layer to convert the dimensionality from B (128) to 257 is the dimensionality of the estimated echo output.
[107] The detailed TCN block is shown in FIG. 12D. Each TCN block comprises a 1×1 convolutional layer to increase the dimensionality from B to H, a dilated depth-wise convolutional (D-conv) layer with kernel size P and varying dilation factors, and another 1×1 convolutional layer to reduce the dimensionality from H back to B. The dilation factor of the D-conv layer exponentially increases the dilation factors to form such a large receptive field that a TCN. A PReLU activation layer and a batch normalization layer are inserted both before and after each D-Conv layer.
[108] The model has 1.95 million parameters. The estimated echo of the microphone signal to reconstruct the time domain signals. We use a mean squared error (MSE) loss between the predicted and estimated echo signals. The Adam optimizer with a learning rate of 0.0005 and batch size of 128 are used to train the model.
[109] FIG. 13 illustrates an exemplary schematic block diagram of a hardware platform for implementation of the system 200. As shown in FIG. 13, a computer system 1300 can include an external storage device 1310, a bus 1320, a main memory 1330, a read only memory 1340, a mass storage device 1250, communication port 1360, and a processor 1370. A person skilled in the art will appreciate that the computer system may include more than one processor and communication ports. Examples of processor 1370 include, but are not limited to, an Intel® Itanium® or Itanium 2 processor(s), or AMD® Opteron® or Athlon MP® processor(s), Motorola® lines of processors, FortiSOC™ system on chip processors or other future processors. Processor 1370 may include various modules associated with embodiments of the present invention. Communication port 1360 can be any of an RS-232 port for use with a modem-based dialup connection, a 10/100 Ethernet port, a Gigabit or 10 Gigabit port using copper or fiber, a serial port, a parallel port, or other existing or future ports. Communication port 1360 may be chosen depending on a network, such a Local Area Network (LAN), Wide Area Network (WAN), or any network to which computer system connects. Memory 1330 can be Random Access Memory (RAM), or any other dynamic storage device commonly known in the art. Read-only memory 1340 can be any static storage device(s) e.g., but not limited to, a Programmable Read Only Memory (PROM) chips for storing static information e.g., start-up or BIOS instructions for processor 1370. Mass storage 1250 may be any current or future mass storage solution, which can be used to store information and/or instructions. Exemplary mass storage solutions include, but are not limited to, Parallel Advanced Technology Attachment (PATA) or Serial Advanced Technology Attachment (SATA) hard disk drives or solid-state drives (internal or external, e.g., having Universal Serial Bus (USB) and/or Firewire interfaces), e.g. those available from Seagate (e.g., the Seagate Barracuda 7102 family) or Hitachi (e.g., the Hitachi Deskstar 7K1000), one or more optical discs, Redundant Array of Independent Disks (RAID) storage, e.g. an array of disks (e.g., SATA arrays), available from various vendors including Dot Hill Systems Corp., LaCie, Nexsan Technologies, Inc. and Enhance Technology, Inc.
[110] Bus 1320 communicatively couples processor(s) 1370 with the other memory, storage, and communication blocks. Bus 1320 can be, e.g., a Peripheral Component Interconnect (PCI) / PCI Extended (PCI-X) bus, Small Computer System Interface (SCSI), USB or the like, for connecting expansion cards, drives and other subsystems as well as other buses, such a front side bus (FSB), which connects processor 1370 to software system.
[111] Optionally, operator and administrative interfaces, e.g., a display, keyboard, and a cursor control device, may also be coupled to bus 1320 to support direct operator interaction with a computer system. Other operator and administrative interfaces can be provided through network connections connected through communication port 1360. The external storage device 1310 can be any kind of external hard-drives, floppy drives, IOMEGA® Zip Drives, Compact Disc - Read Only Memory (CD-ROM), Compact Disc-Re-Writable (CD-RW), Digital Video Disk-Read Only Memory (DVD-ROM). Components described above are meant only to exemplify various possibilities. In no way should the aforementioned exemplary computer system limit the scope of the present disclosure.
[112] It should be apparent to those skilled in the art that many more modifications besides those already described are possible without departing from the inventive concepts herein. The inventive subject matter, therefore, is not to be restricted except in the spirit of the appended claims. Moreover, in interpreting both the specification and the claims, all terms should be interpreted in the broadest possible manner consistent with the context. In particular, the terms “comprise” and “comprising” should be interpreted as referring to elements, components, or steps in a non-exclusive manner, indicating that the referenced elements, components, or steps may be present, or utilized, or combined with other elements, components, or steps that are not expressly referenced. Where the specification claims refer to at least one of something selected from the group consisting of A, B, C ….and N, the text should be interpreted as requiring only one element from the group, not A plus N, or B plus N, etc. The foregoing description of the specific embodiments will so fully reveal the general nature of the embodiments herein that others can, by applying current knowledge, readily modify and/or adapt for various applications such specific embodiments without departing from the generic concept, and, therefore, such adaptations and modifications should and are intended to be comprehended within the meaning and range of equivalents of the disclosed embodiments. It is to be understood that the phraseology or terminology employed herein is for the purpose of description and not of limitation. Therefore, while the embodiments herein have been described in terms of preferred embodiments, those skilled in the art will recognize that the embodiments herein can be practiced with modification within the spirit and scope of the appended claims.
[113] While the foregoing describes various embodiments of the invention, other and further embodiments of the invention may be devised without departing from the basic scope thereof. The scope of the invention is determined by the claims that follow. The invention is not limited to the described embodiments, versions, or examples, which are included to enable a person having ordinary skill in the art to make and use the invention when combined with information and knowledge available to the person having ordinary skill in the art.

ADVANTAGES OF INVENTION
[114] The present invention provides a system and method for attenuating acoustic echo in a duplex communication device.
[115] The present invention provides a system to attenuate acoustic echo in linear and non-linear domains.

,CLAIMS:1. A system for attenuating acoustic echo in an acoustic duplex communication device, the system comprising:
an audio device provided on the acoustic duplex communication device, adapted to generate first acoustic signals;
a microphone provided on the acoustic duplex communication device, adapted to receive, as input, acoustic signals;
an acoustic source disposed in a vicinity of the acoustic duplex communication device, adapted to generate second acoustic signals, wherein the second acoustic signals are generated any one or both of separately and concurrently, relative to the generated first acoustic signals from the audio device; and
a controller communicably coupled to the audio device and the microphone, the controller comprising a processor and a memory communicably coupled to the processor, the controller configured to:
receive, from the microphone, input acoustic signals, wherein the input acoustic signals comprise at least the second acoustic signals, acoustic echo signals corresponding to the first acoustic signals, and ambient noise; and
model the acoustic echo signals from the received input acoustic signals using any one of a plurality of Cepstrum-based models to obtain a minimum phase component of the acoustic echo signals,
wherein, the controller is further configured to:
determine a first minimum phase function corresponding to the modelled acoustic echo signals defined based on a Cepstrum decomposition of the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals,
wherein the controller is further configured to apply a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.
2. The system as claimed in claim 1, wherein the controller is further configured to:
model the acoustic echo signals from the received input acoustic signals using any one of a plurality of models to obtain an all-pass component of the acoustic echo signals,
wherein the controller is further configured to apply an all-pass filter to attenuate the all-pass component of the acoustic signals from the input acoustic signals.
3. The system as claimed in claim 1, wherein responsive to one model of the plurality of models further comprising a complex Cepstrum, the controller is further configured to:
determine a first minimum phase function corresponding to the modelled acoustic echo signals defined based on real and complex components of a Fourier transform the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals,
wherein the controller is further configured to apply a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.
4. The system as claimed in claim 3, wherein the controller is further configured to:
determine an impulse response of the first minimum phase function defined based on an exponential of a fast Fourier transform of a noise component of the input acoustic signals,
wherein the controller is further configured to apply an impulse response filter to improve a signal to noise ratio of the input acoustic signals.

5. The system as claimed in claim 1, wherein responsive to one model of the plurality of models further comprising recursion, the controller is further configured to:
determine a first minimum phase function corresponding to the modelled acoustic echo signals defined based on an exponential of the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals,
wherein the controller is further configured to apply a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.
6. The system as claimed in claim 5, wherein the controller is further configured to:
determine an impulse response of the first minimum phase function defined based on an exponential of the acoustic echo signals,
wherein the controller is further configured to apply an impulse response filter to improve a signal to noise ratio of the input acoustic signals.
7. The system as claimed in claim 1, wherein responsive to one model of the plurality of models further comprising machine learning, the controller is further configured to:
determine the first minimum phase function corresponding to the modelled acoustic echo signals defined based on real and complex Cepstrum coefficients of the acoustic echo signals.
8. A method for attenuating acoustic echo in an acoustic duplex communication device, the method comprising:
providing the acoustic duplex communication device provided with:
an audio device provided on the acoustic duplex communication device, adapted to generate first acoustic signals; and
a microphone provided on the acoustic duplex communication device, adapted to receive, as input, acoustic signals;
providing an acoustic source disposed in a vicinity of the acoustic duplex communication device, adapted to generate second acoustic signals, wherein the second acoustic signals are generated any one or both of separately and concurrently, relative to the generated first acoustic signals from the audio device;
receiving, by a controller, input acoustic signals, wherein the input acoustic signals comprise at least the second acoustic signals, acoustic echo signals corresponding to the first acoustic signals, and ambient noise; and
modelling, by the controller, the acoustic echo signals from the received input acoustic signals using any one of a plurality of Cepstrum-based models to obtain a minimum phase component of the acoustic echo signals,
wherein the method further comprises:
determining, by the controller, a first minimum phase function corresponding to the modelled acoustic echo signals defined based on a Cepstrum decomposition of the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals; and
applying, by the controller, a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.
9. The method as claimed in claim 8, further comprising:
modelling, by the controller, the acoustic echo signals from the received input acoustic signals using any one of a plurality of models to obtain an all-pass component of the acoustic echo signals; and
applying, by the controller, an all-pass filter to attenuate the all-pass component of the acoustic signals from the input acoustic signals.
10. The method as claimed in claim 8, wherein responsive to one model of the plurality of models further comprising a complex Cepstrum, the method further comprises:
determining, by the controller, a first minimum phase function corresponding to the modelled acoustic echo signals defined based on real and complex components of a Fourier transform the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals; and
applying, by the controller, an all-pass filter to attenuate the all-pass component of the acoustic signals from the input acoustic signals.
11. The method as claimed in claim 10, further comprising:
determining, by the controller, an impulse response of the first minimum phase function defined based on an exponential of a fast Fourier transform of a noise component of the input acoustic signals; and
applying, by the controller, an impulse response filter to improve a signal to noise ratio of the input acoustic signals.
12. The method as claimed in claim 8, wherein responsive to one model of the plurality of models further comprising recursion, the method further comprises:
determining, by the controller, a first minimum phase function corresponding to the modelled acoustic echo signals defined based on an exponential of the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals; and
applying, by the controller, a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.
13. The method as claimed in claim 12, wherein further comprising:
determining, by the controller, an impulse response of the first minimum phase function defined based on an exponential of the acoustic echo signals; and
applying, by the controller, an impulse response filter to improve a signal to noise ratio of the input acoustic signals.
14. The method as claimed in claim 8, wherein responsive to one model of the plurality of models further comprising machine learning, the method further comprises:
determining, by the controller, the first minimum phase function corresponding to the modelled acoustic echo signals defined based on real and complex Cepstrum coefficients of the acoustic echo signals.
15. A non-transitory, computer readable storage device storing instructions, which when executed by a processor causes the processor to perform a process to attenuate acoustic echo in an acoustic duplex communication device, the process comprising:
receiving, by a controller, input acoustic signals, wherein the input acoustic signals comprise at least the second acoustic signals, acoustic echo signals corresponding to the first acoustic signals, and ambient noise; and
modelling, by the controller, the acoustic echo signals from the received input acoustic signals using any one of a plurality of Cepstrum-based models to obtain a minimum phase component of the acoustic echo signals,
wherein the process further comprises:
determining, by the controller, a first minimum phase function corresponding to the modelled acoustic echo signals defined based on a Cepstrum decomposition of the acoustic echo signals, wherein the first minimum phase function is indicative of the minimum phase component of the acoustic echo signals; and
applying, by the controller, a minimum phase filter to attenuate the minimum phase component of the acoustic signals from the input acoustic signals.