Abstract: A noise extraction device of the present invention includes: first and second microphone units (11 and 12) each picking up a sound; a directivity synthesis unit which performs a directivity synthesis on output signals respectively received from the first and second microphone units (11 and 12) and generates two directionally synthesized signals which have: different sensitivities to noise; the same directional pattern with respect to sound pressure; and the same effective acoustic center position; and an acoustic cancellation unit which cancels an acoustic component of one of the two directionally synthesized signals by subtracting the one of the two directionally synthesized signals from the other of the two directionally synthesized signal, so as to extract a noise component.

Abstract: A signal generating section generates a main signal and a noise reference signal. A determining section determines whether a level ratio is larger than a predetermined value. An adaptive filter section generates a signal indicative of a signal component of a target sound included in the noise reference signal generated by the signal generating section, and learns a filter coefficient only when the determining section determines that the level ratio is larger than the predetermined value. A subtracting section subtracts the signal generated by the adaptive filter section from the noise reference signal. A noise suppressing section suppresses a signal component of noise included in the main signal by using the main signal and the noise reference signal after subtraction by the subtracting section.

Abstract: A harmonic structure acoustic signal detection device not depending on the level fluctuation of the input signal including: an FFT unit which performs FFT on an input signal and calculates a power spectrum component for each frame; a harmonic structure extraction unit which leaves only a harmonic structure from the power spectrum component; a voiced feature evaluation unit which evaluates correlation between the frames of harmonic structures extracted by the harmonic structure extraction unit, thereby evaluates whether or not the segment is a vowel segment, and extracts the voiced segment; and a speech segment determination unit which determines a speech segment according to the continuity and durability of the output of the voiced feature evaluation unit.

Abstract: It is an object of the present invention to provide an echo suppressor which is simple in construction in comparison with the conventional echo suppressor, and which can suppress echoes corresponding to the spatial transfer paths, and reduce the amount of calculations necessary to suppress echoes. The echo suppressor is operative under the condition that transfer functions corresponding to spatial transfer paths between two or more loudspeakers (105, 106) and one or more microphones (107, 108) are estimated on the basis of symmetrical arrangement of the loudspeakers (105, 106) and the microphones (107, 108).

Abstract: A sound collection apparatus according to the present invention comprises: at least one target sound collection means for collecting a sound including a target sound generated from a target sound source, so as to output a collected-sound signal; a plurality of non-target sound collection means, provided at positions different from each other, each forming a dead zone of a sensitivity in a direction of the target sound source so as to collect a sound outside the dead zone and output a collected-sound signal; sensitivity suppression means for generating a sensitivity suppression signal for suppressing a sound collection sensitivity in an overlap region in which a plurality of the dead zones overlap each other, as compared to in a region surrounding the overlap region, by subjecting, to a predetermined signal processing, the collected-sound signal outputted by each of the plurality of non-target sound collection means; and extraction means for removing, from the collected-sound signal outputted by the at least o

Abstract: A wearable terminal is constantly worn by a user and continually picks up images and sounds from the surroundings. Even when using a directional microphone to sensitively pick up targeted audio, the wearable terminal can reduce noise occurring due to motion of the device, for example when the user is walking, and reduce the influence of a shift of the sound pickup direction. For this purpose, the wearable terminal includes a sensor for detecting motion, and performs microphone directivity control to use the directional microphone when the amount of motion is small, and to use an omnidirectional microphone that is not likely to be influenced by noise when the amount of motion is large.

Abstract: Howling, which occurs when amplifying a target sound collected by a first microphone through an amplification section and outputting the amplified sound as an intensified sound from a loudspeaker, is suppressed. A first power spectrum is produced according to a first acoustic signal output from the first microphone collecting a sound. A second power spectrum is produced according to a second acoustic signal of a sound including at least the intensified sound and not including the target sound. Then, the first acoustic signal is filtered based on the first power spectrum and the second power spectrum to output only an acoustic signal of the target sound to the amplification section.

Abstract: A howling detection device detects a dominance ratio, which indicates a risk of howling to be occurred when a mixed signal obtained by a sound mixing section for mixing a plurality of sound signals respectively collected by a plurality of microphones is outputted by a speaker, for each of the sound signals.

Abstract: A multichannel echo canceller according to the present invention includes an echo cancellation section which receives loudspeaker input signals sp1, sp2 including a second acoustic signal to be respectively inputted to a plurality of loudspeakers 10, 20 provided in a first location and detection signals m1, m2 detected by a plurality of microphones 11, 21 provided in the first location; separates a first acoustic signal and a second acoustic signal, which are included in the detection signals m1, m2, by performing signal processing based on an independent component analysis; cancels as an echo the second acoustic signal, which is included in the detection signals m1, m2, by outputting only the separated first acoustic signal to a plurality of loudspeakers 30, 40 provided in a second location.

Abstract: The evaluating apparatus of the present invention includes: a digital filter for filtering a signal in according with one or more tap coefficients of the digital filter. The evaluating apparatus further includes: a detecting section for detecting an index to be used for evaluating quality of the signal based on the filtered signal; and a controlling section for controlling the one ox more tap coefficients of the digital filter to be within a pre-determined range such that a value of the detected index includes an optimal value of the index.

Abstract: There is provided a harmonic structure acoustic signal detection device not depending on the level fluctuation of the input signal, having an excellent real time property and noise resistance. The device includes: an FFT unit (200) which performs FFT on an input signal and calculates a power spectrum component for each frame; a harmonic structure extraction unit (201) which leaves only a harmonic structure from the power spectrum component; a voiced feature evaluation unit (210) which evaluates correlation between the frames of harmonic structures extracted by the harmonic structure extraction unit (201), thereby evaluates whether or not the segment is a vowel segment, and extracts the voiced segment; and a speech segment determination unit (205) which determines a speech segment according to the continuity and durability of the output of the voiced feature evaluation unit (210).

Abstract: A pickup section transduces sounds to audio signals. A zoom control section outputs a zoom position signal corresponding to a zoom position. A directivity control section alters the directivity characteristics under a telescopic operation so as to mainly pick up sounds coming from a frontal direction with an enhancement which is in accordance with the zoom position signal, thereby outputting an R channel audio signal and an L channel audio signal. In accordance with the zoom position signal, a noise suppression section applies a greater degree of suppression to the background noise contained in the respective channel audio signals under the telescopic operation than under a wide-angle operation. As a result, a target sound from a remote location can be picked up with a sufficient enhancement in accordance with the zoom position under the telescopic operation.

Abstract: A signal generating section generates a main signal and a noise reference signal. A determining section determines whether a level ratio is larger than a predetermined value. An adaptive filter section generates a signal indicative of a signal component of a target sound included in the noise reference signal generated by the signal generating section, and learns a filter coefficient only when the determining section determines that the level ratio is larger than the predetermined value. A subtracting section subtracts the signal generated by the adaptive filter section from the noise reference signal. A noise suppressing section suppresses a signal component of noise included in the main signal by using the main signal and the noise reference signal after subtraction by the subtracting section.

Abstract: A low-cost speech recognition apparatus for AV equipment capable of speech recognition with high accuracy while 2-channel sound is being produced from loudspeakers is achieved. A monaural conversion part converts 2-channel signals to be inputted to the loudspeakers into a monaural signal. A single echo canceller is provided with an output from a microphone and an output from the monaural conversion part (monaural signal). The echo canceller estimates an echo of multichannel sound based on the monaural signal, and then eliminates the echo sound from the microphone output. Thus, with only a single echo canceller, speech recognition can be carried out while 2-channel sound is being produced from the loudspeakers. Moreover, unlike the case where two echo cancellers are provided, the present invention can prevent the occurrence of mutual interference between the echo cancellers that leads to deterioration in speech recognition performance.

Abstract: A built-in microphone device for use in an apparatus having a mechanism section generating internal noise inside a housing of the apparatus includes a main microphone for picking up an external sound; a noise reference microphone for picking up the internal noise; an adaptive filter member for generating a control audio signal based on an output signal from the noise reference microphone using a filter coefficient; a signal subtraction section for subtracting the control audio signal generated by the adaptive filter member from an output signal from the main microphone to generate a subtraction result; and a filter coefficient update control section for receiving an operation signal generated at the time of an operation of the mechanism section, and in response to the operation signal, updating the filter coefficient of the adaptive filter member based on the subtraction result generated by the signal subtraction section and an output signal from the noise reference microphone.

Abstract: A built-in microphone device for use in an apparatus having a mechanism section generating internal noise inside a housing of the apparatus includes a main microphone for picking up an external sound; a noise reference microphone for picking up the internal noise; an adaptive filter member for generating a control audio signal based on an output signal from the noise reference microphone using a filter coefficient; a signal subtraction section for subtracting the control audio signal generated by the adaptive filter member from an output signal from the main microphone to generate a subtraction result; and a filter coefficient update control section for receiving an operation signal generated at the time of an operation of the mechanism section, and in response to the operation signal, updating the filter coefficient of the adaptive filter member based on the subtraction result generated by the signal subtraction section and an output signal from the noise reference microphone.

Abstract: Speech issued by a speaker is collected by a microphone 1, and applied to a signal delay unit 3 and a sound level estimator 4 through an A/D converter 2. The sound level estimator 4 calculates a sound level estimation value based on the applied digital sound signal. The signal delay unit 3 applies the digital sound signal delayed by a predetermined sound level rising time period to a sound level adjuster 5. The sound level adjuster 5 adjusts the sound level of the digital sound signal based on the sound level estimation value, and applies the adjusted sound level output to the speech recognition unit 6. The speech recognition unit 6 performs speech recognition in response to the applied adjusted sound level output.

Abstract: A pickup section 11 transduces sounds to audio signals. A zoom control section 12 outputs a zoom position signal corresponding to a zoom position. A directivity control section 13 alters the directivity characteristics under telescopic operation so as to mainly pickup sounds coming from a frontal direction with an enhancement which is in accordance with the zoom position signal, thereby outputting an R channel audio signal and an L channel audio signal. In accordance with the zoom position signal, a noise suppression section 14 applies a greater degree of suppression to the background noise contained in the respective channel audio signals under telescopic operation than under wide-angle operation. As a result, a target sound from a remote location can be picked up with a sufficient enhancement in accordance with the zoom position under telescopic operation.

Abstract: A stereo ultradirectional microphone apparatus for detecting a sound to produce stereo sound signals, comprises: first and second ultradirectional microphones arranged side by side with a given distance in parallel for converting a sound into first and second sound signals respectively, first and second delays for delaying an output of the first and second microphones by a delay time .tau. respectively, and first and second subtractors for subtracting for subtraction between the first sound signal and an output of the second delay and subtracting for subtraction between the second sound signal and an output of the first delay. The delay time .tau. corresponds to a difference between the timings of a sound from a sound source in a direction making a clockwise angel .theta. from the front where a dead angel should be made. The subtraction provides the dead angle. Similarly, a dead angle on the side is also made to obtain a stereo characteristic.

Abstract: In an echo canceller, a first adaptive filter (1) estimates an impulse response of an echo path between reception and transmission side by an NLMS algorithm to synthesize a pseud echo, while a second adaptive filter (3) determines whether there is an echo path change or a double talk. A voice detector (8) detects a short time power of the received input signal (x) to determine whether far-end speech is present or absent. The voice detector controls the second adaptive filter (3) to execute adaptation of the second adaptive filter for renewal of an impulse response of an echo path when far-end speech is present and otherwise to suspend the adaptation of the second adaptive filter when far-end speech is absent, while a double talk detector (9) controls the first adaptive filter to execute or suspend adaptation of the first adaptive filter according to the determination results of the voice detector.