EP0366584A2
Expired / Lapsed
EP
⏱️ 0.0 years remaining
Full-duplex digital speakerphone
Assignee: Siemens Business Communication Systems Inc
•
Filing: 1989-09-12
•
Granted: —
•
Expires: 2009-09-12
Abstract
Abstract A full-duplex digital speakerphone 10 includes a transmit signal path having an output coupled to a telephone trunk and a receive signal path having an input coupled to the telephone trunk and an output coupled to a loudspeaker means. The speakerphone further includes a room echo cancellation adaptive filter 56 and a trunk echo cancellation adaptive filter 66. Serially coupled within the transmit signal path is a selective suppression block 50 for suppressing a component of a Mu-Law or an A-Law quantization error signal. A second selective suppression block 52 is serially coupled within the receive signal path. Suppression of non-linearities due to Mu-Law or A-Law signal conversion is also accommodated by providing a non-linear signal processing block 40 at an input to an adaptive filter and an optional non-linear signal processing block at an output of the adaptive filter. Each of the blocks emulates and compensates for signal converter non-linearity. The speakerphone facilitates adaptive filter coefficient initialization by beginning a call in a half-duplex mode and switching to full-duplex when filter coefficients are adapted. The speakerphone also has a variable adaptation step size which is a function of a short-term estimate of signal power within the associated transmit or receive signal paths.
Inventors
- 👤 Ygal Arbel
Bibliographic Details
Application Number
—
Jurisdiction
EP
Data Source
google_patents
Last Synchronized
2026-11-04 21:27
CPC Classifications
H; H04; H04M; H04M9/00; H04M9/08; H04M9/082
Claims Analysis
Total Claims: 12
Claim 1
Independent Claim
299 words
A speakerphone including a transmit signal path having an output coupled to a communications link, the speakerphone further including a receive signal path having an input coupled to the communications link and an output coupled to a loudspeaker means, the speakerphone comprising: means for generating a transmit path signal, the transmit path signal including a first echo signal component; means for converting the transmit path signal to a digital transmit signal expressed by a plurality of bits, the digital transmit signal including a first digital quantization error signal; first adaptive filter means having an input coupled to the receive signal path and an output coupled to the transmit signal path for cancelling most of the first echo signal component; means for coupling the receive signal path to the communications link and generating a receive path signal including a second echo signal component; means for converting the receive path signal to a digital receive signal expressed by a plurality of bits, the digital receive signal including a second digital quantization error signal; second adaptive filter means having an input coupled to the transmit signal path and an output coupled to the receive signal path for cancelling most of the second echo signal component; the speakerphone further comprising first means for selectively suppressing the first digital quantization error signal, said first selective suppression means being serially coupled within the transmit signal path and having an input coupled to the transmit signal path at a point after the output of said first adaptive filter means; and second means for selectively suppressing the second digital quantization error signal, said second selective suppression means being serially coupled within the receive signal path and having an input coupled to the receive signal path at a point after the output of said second adaptive filter means.
Claim 2
Dependent Claim
299 words
A speakerphone including a transmit signal path having an output coupled to a communications link, the speakerphone further including a receive signal path having an input coupled to the communications link and an output coupled to a loudspeaker means, the speakerphone comprising: means for generating a transmit path signal, the transmit path signal including a first echo signal component; means for converting the transmit path signal to a digital transmit signal expressed by a plurality of bits, the digital transmit signal including a first digital quantization error signal; first adaptive filter means having an input coupled to the receive signal path and an output coupled to the transmit signal path for cancelling most of the first echo signal component; means for coupling the receive signal path to the communications link and generating a receive path signal including a second echo signal component; means for converting the receive path signal to a digital receive signal expressed by a plurality of bits, the digital receive signal including a second digital quantization error signal; second adaptive filter means having an input coupled to the transmit signal path and an output coupled to the receive signal path for cancelling most of the second echo signal component; the speakerphone further comprising first means for selectively suppressing the first digital quantization error signal, said first selective suppression means being serially coupled within the transmit signal path and having an input coupled to the transmit signal path at a point after the output of said first adaptive filter means; and second means for selectively suppressing the second digital quantization error signal, said second selective suppression means being serially coupled within the receive signal path and having an input coupled to the receive signal path at a point after the output of said second adaptive filter means.
Claim 3
Dependent Claim
67 words
A speakerphone as set forth in Claim 1 wherein said transmit signal path converting means and said receive signal path converting means are each either a Mu-Law or a A-Law converting means and wherein the first and the second selective suppression means each comprise threshold means responsive to either a Mu-Law or a A-Law quantization step size employed by the associated Mu-Law or a A-Law converting means.
Claim 4
Dependent Claim
74 words
A speakerphone as set forth in Claim 2 and further comprising transmit signal path output Mu-Law or A-Law conversion means for converting an output of the first selective suppression means to an analog signal, the speakerphone further comprising first Mu-Law or A-Law Effect compensating means coupled between the receive signal path and the input to the first adaptive filter means for compensating for a nonlinear effect of the output Mu-Law or A-Law conversion means.
Claim 5
Dependent Claim
49 words
A speakerphone as set forth in Claim 3 and further comprising second Mu-Law or A-Law compensating means coupled between the output of the first adaptive filter means and the transmit signal path for compensating for a nonlinear Mu-Law or A-Law effect of the transmit signal path input converting means.
Claim 6
Independent Claim
234 words
A method of operating a speakerphone, the speakerphone comprising a first echo cancellation adaptive filter having an input coupled to a receive signal path and an output coupled to a transmit signal path, the speakerphone further comprising a second echo cancellation adaptive filter having an input coupled to the transmit signal path and an output coupled to the receive signal path, the method comprising the steps of: converting a transmit signal path input signal to a digital representation of the transmit signal path input signal, the digital representation including the first echo noise component and also a quantization error component; converting a receive signal path input signal to a digital representation of the receive signal path input signal, the digital representation including the second echo noise component and also a quantization error component; cancelling at least a portion of the first echo noise component and the second echo noise component with the adaptive filters, the step of cancelling including the steps of, for each of the adaptive filters, generating an adaptive filter output error signal E(n) having a magnitude which is a function of the input signal, a reference input signal and a magnitude of adaptive filter coefficients; and adapting the magnitude of the adaptive filter coefficients as a function of an adaptation step size Beta, wherein Beta is a function of a constant BetaO which is divided by a signal power estimate Ps(n).
Claim 7
Independent Claim
234 words
A method of operating a speakerphone, the speakerphone comprising a first echo cancellation adaptive filter having an input coupled to a receive signal path and an output coupled to a transmit signal path, the speakerphone further comprising a second echo cancellation adaptive filter having an input coupled to the transmit signal path and an output coupled to the receive signal path, the method comprising the steps of: converting a transmit signal path input signal to a digital representation of the transmit signal path input signal, the digital representation including the first echo noise component and also a quantization error component; converting a receive signal path input signal to a digital representation of the receive signal path input signal, the digital representation including the second echo noise component and also a quantization error component; cancelling at least a portion of the first echo noise component and the second echo noise component with the adaptive filters, the step of cancelling including the steps of, for each of the adaptive filters, generating an adaptive filter output error signal E(n) having a magnitude which is a function of the input signal, a reference input signal and a magnitude of adaptive filter coefficients; and adapting the magnitude of the adaptive filter coefficients as a function of an adaptation step size Beta, wherein Beta is a function of a constant BetaO which is divided by a signal power estimate Ps(n).
Claim 8
Dependent Claim
88 words
A method as set forth in Claim 5 wherein a magnitude of Ps(n) is given by the function Ps(n) = As(n) * As(n) wherein As(n) is determined in accordance with the relationship if As(n-1) > abs(S(n)) then As(n) = As(n-1)'(1-Tf) + Tf * abs(S(n)) else As(n) = As(n-1)'(1-Tr) + Tr * abs(S(n)) where As(n) is an amplitude estimate of the adaptive filter input signal, abs(S(n)) is an absolute value of the filter input, Tf is a fall-time of the filter, and Tr is a rise-time of the filter.
Claim 9
Dependent Claim
46 words
A method as set forth in Claim 5 and further comprising a step of determining a state of the speakerphone to be a CALLER state, a USER state, an IDLE state or an OTHER state in accordance with the state of a plurality of binary variables.
Claim 10
Dependent Claim
258 words
A method as set forth in Claim 7 wherein the plurality of binary variables are b0, b1, b2 and b3, wherein when b2 and b3 are both equal to zero the speakerphone is determined to be in the IDLE state, when b0 and b2 are both equal to one, the speakerphone is determined to be in the CALLER state and when b1 and b3 are both equal to one the speakerphone is determined to be in the USER state, and wherein the state of each of the binary variables is determined in accordance with the following relationships: b3 = 1 if K1*Sx - Nx - K2 > 0, otherwise b3 = 0, b2 = 1 if K1*Sr - Nr - K2 > 0, otherwise b2 = 0, b1 = 1 if SPx * VOLx > K3 * SPr , otherwise b1 = 0 , and b0 = 1 if SPr * VOLr > K3 * SPx , otherwise b0 = 0 where K1 and K2 are constants that define a threshold above which a signal is considered to be speech, K3 defines a ratio of transmit and receive speech levels to distinguish between the USER and the CALLER states, VOLx and VOLr are transmit and receive volume levels, respectively, and wherein Sx = Transmit Channel total signal amplitude estimate, Sr = Receive Channel total signal amplitude estimate, Nx = Transmit Channel Noise amplitude estimate, Nr = Receive Channel Noise amplitude estimate, SPx = Transmit Channel Speech signal amplitude estimate, and SPr = Receive Channel Speech signal amplitude estimate.
Claim 11
Dependent Claim
39 words
A method as set forth in Claim 8 wherein for all other combinations of b0, b1, b2 and b3 except b1=1, b2=1, b3=0 and b0=1, b2=0, b3=1 and b0=b1=b2=b3=1 the speakerphone is determined to be in the OTHER state.
Claim 12
Dependent Claim
24 words
A method as set forth in Claim 8 wherein VOLx and VOLr each varies within a range of approximately + 12dB to approximately -24dB.
References Cited (Backward Citations)
No citation records indexed.
USPTO Annuity & Maintenance Windows
Maintenance schedule is calculated for granted US utility patents.