Abstract: An audio device and a method thereof are provided. The method is adopted by an audio device to detect a voice, wherein the audio device is coupled to a host device. The method includes an acoustic conversion circuit converting an acoustic wave into an analog audio signal; an analog-to-digital converter (ADC) converting the analog audio signal into digital audio data; a first-level voice detection circuit detecting voice activity in the analog audio signal; a second-level voice detection circuit detecting a beginning syllable of a key phrase in the digital audio data when the voice activity is detected in the digital audio data; and a third-level voice detection circuit detecting the key phrase from the digital audio data only when the beginning syllable of the key phrase is detected in the digital audio data.

Abstract: An audio device and a method thereof are provided. The method is adopted by an audio device to detect a voice, wherein the audio device is coupled to a host device. The method includes an acoustic conversion circuit converting an acoustic wave into an analog audio signal; an analog-to-digital converter (ADC) converting the analog audio signal into digital audio data; a first-level voice detection circuit detecting voice activity in the analog audio signal; a second-level voice detection circuit detecting a beginning syllable of a key phrase in the digital audio data when the voice activity is detected in the digital audio data; and a third-level voice detection circuit detecting the key phrase from the digital audio data only when the beginning syllable of the key phrase is detected in the digital audio data.

Abstract: An embodiment of an acoustic echo cancellation system is disclosed. The system comprises an echo cancellation unit, a second filter and a subtraction unit. The echo cancellation unit comprises a first attenuator, a first filter and a first subtractor. The first attenuator has a first down-scaling factor for attenuating a first signal. The first filter generates a first echo signal estimate based on the attenuated first signal. The first subtractor generates a third signal by subtracting the first echo signal estimate from a second signal. The second filter generates a second echo signal estimate based on the first signal. The subtraction unit subtracts the second echo signal estimate from the third signal.

Abstract: A telephone system comprising an analog telephone and a full-duplex speakerphone adapter is disclosed. The analog telephone comprises an ordinary analog telephone subscriber circuit for transmitting and receiving analog signals and a handset for users. The full-duplex speakerphone adapter is coupled between the analog telephone and a central office, uses a subscriber loop interface circuit through the first telephone line to couple to the ordinary analog telephone subscriber circuit of the analog telephone and uses a telephone hybrid interface circuit through the second telephone line to couple to a wall jack to communicate with the central office.

Abstract: A voice communication device with an integrated framework structure for echo cancellation and noise reduction is disclosed. A microphone receives a local input signal while a speaker is outputting a local output signal. The local input signal and output signal are all decomposed into a plurality of subband signals by filter banks for conducting individual processing of echo cancellation and noise reduction per subband. The subband echo canceller is followed by a DFT unit to split the cancellation result into a plurality of narrow frequency bins whereby the noise reduction is performed. The noise reduction results are recombined by an IDFT unit for residual echo removal in a subband non-linear processor. The final output is obtained from a synthesis filter bank that synthesizes the subband signals after echo cancellation and noise reduction into a full-band signal.

Abstract: A telephone system comprising an analog telephone and a full-duplex speakerphone adapter is disclosed. The analog telephone comprises an ordinary analog telephone subscriber circuit for transmitting and receiving analog signals and a handset for users. The full-duplex speakerphone adapter is coupled between the analog telephone and a central office, uses a subscriber loop interface circuit through the first telephone line to couple to the ordinary analog telephone subscriber circuit of the analog telephone and uses a telephone hybrid interface circuit through the second telephone line to couple to a wall jack to communicate with the central office.

Abstract: A telephony system equipped with an echo cancellation module is disclosed. A hybrid interface circuit outputs an outbound signal to a wall jack while receiving an inbound signal from the wall jack. The inbound signal may contain a line echo of the outbound signal caused by the impedance mismatches in the hybrid interface circuit. A line echo canceller containing an adaptive filter is used to cancel the line echo in the inbound signal based on the learned line echo path characteristics captured in the training period. During a brief period after the telephony system is activated, the line echo canceller enters into a calibration mode wherein its adaptive filter is trained to learn the line echo path characteristics. During the calibration mode, the line echo canceller generates a calibration signal as the outbound signal and receives the line echo from the hybrid interface circuit to perform learning of the line echo path characteristics.

Abstract: A voice communication device with an integrated framework structure for echo cancellation and noise reduction is disclosed. A microphone receives a local input signal while a speaker is outputting a local output signal. The local input signal and output signal are all decomposed into a plurality of subband signals by filter banks for conducting individual processing of echo cancellation and noise reduction per subband. The subband echo canceller is followed by a DFT unit to split the cancellation result into a plurality of narrow frequency bins whereby the noise reduction is performed. The noise reduction results are recombined by an IDFT unit for residual echo removal in a subband non-linear processor. The final output is obtained from a synthesis filter bank that synthesizes the subband signals after echo cancellation and noise reduction into a full-band signal.

Abstract: An embodiment of an acoustic echo cancellation system is disclosed. The system comprises an echo cancellation unit, a second filter and a subtraction unit. The echo cancellation unit comprises a first attenuator, a first filter and a first subtractor. The first attenuator has a first down-scaling factor for attenuating a first signal. The first filter generates a first echo signal estimate based on the attenuated first signal. The first subtractor generates a third signal by subtracting the first echo signal estimate from a second signal. The second filter generates a second echo signal estimate based on the first signal. The subtraction unit subtracts the second echo signal estimate from the third signal.

Abstract: A method and apparatus under software control for pattern recognition utilizes a neural network implementation to recognize two dimensional input images which are sufficiently similar to a database of previously stored two dimensional images. Images are first image processed and subjected to a Fourier transform which yields a power spectrum. An in-class to out-of-class study is performed on a typical collection of images in order to determine the most discriminatory regions of the Fourier transform. A feature vector consisting of the highest order (most discriminatory) magnitude information from the power spectrum of the Fourier transform of the image is formed. Feature vectors are input to a neural network having preferably two hidden layers, input dimensionality of the number of elements in the feature vector and output dimensionality of the number of data elements stored in the database. Unique identifier numbers are preferably stored along with the feature vector.

Abstract: A pattern recognition method and apparatus utilizes a neural network to recognize input images which are sufficiently similar to a database of previously stored images. Images are first processed and subjected to a Fourier transform which yields a power spectrum. An in-class to out-of-class study is performed on a typical collection of images in order to determine the most discriminatory regions of the Fourier transform. A feature vector consisting of the (most discriminatory) information from the power spectrum of the Fourier transform of the image is formed. Feature vectors are input to a neural network having preferably two hidden layers, input dimensionality of the number of elements in the feature vector and output dimensionality of the number of data elements stored in the database. Unique identifier numbers are preferably stored along with the feature vector. Application of a query feature vector to the neural network results in an output vector.

Abstract: A method and apparatus under software control for pattern recognition utilizes a neural network implementation to recognize two dimensional input images which are sufficiently similar to a database of previously stored two dimensional images. Images are first image processed and subjected to a Fourier transform which yields a power spectrum. An in-class to out-of-class study is performed on a typical collection of images in order to determine the most discriminatory regions of the Fourier transform. A feature vector consisting of the highest order (most discriminatory) magnitude information from the power spectrum of the Fourier transform of the image is formed. Feature vectors are input to a neural network having preferably two hidden layers, input dimensionality of the number of elements in the feature vector and output dimensionality of the number of data elements stored in the database. Unique identifier numbers are preferably stored along with the feature vector.