Abstract: A non-transitory computer-readable recording medium having stored therein an update program that causes a computer to execute a procedure, the procedure includes calculating a selection rate of each of a plurality of quantization points included in a quantization table, based on quantization data obtained by quantizing features of a plurality of utterance data, and updating the quantization table by updating the plurality of quantization points based on the selection rate.

Abstract: A non-transitory computer-readable recording medium stores a noise determination program for causing a computer to execute a process including: comparing a sound pressure level for each frequency with a sound pressure level in a band having a frequency lower than a threshold value, in a spectrum of a voice signal; and determining whether a component corresponding to each frequency is a voice or noise, based on a similarity between the sound pressure level for each frequency and the sound pressure level in the band.

Abstract: A voice correction method implemented by a computer, the method includes: obtaining first voice information which is voice information recorded when noise is generated and on which noise suppression processing is performed and second voice information indicating voice information recorded in an environment in which no noise is generated, and generating emphasized information by emphasizing a component of a band corresponding to a band having a low signal noise ratio (SNR) of the first voice information, among bands of the second voice information; performing machine learning on a model based on the first voice information and the emphasized information; and generating corrected voice information by correcting third voice information on which noise suppression processing is performed, based on the machine-learned model.

Abstract: A non-transitory computer-readable recording medium having stored therein an update program that causes a computer to execute a procedure, the procedure includes calculating a selection rate of each of a plurality of quantization points included in a quantization table, based on quantization data obtained by quantizing features of a plurality of utterance data, and updating the quantization table by updating the plurality of quantization points based on the selection rate.

Abstract: An apparatus divides each of first and second sound-signals generated respectively by first and second sound-input devices, into frames having a predetermined time length, and converts each frame of the first and second sound-signals into first and second frequency-spectra, respectively, in a frequency domain. For each frame, the apparatus calculates, based on the first and second frequency-spectra, a probability that a sound of the frame is emitted only from a sound-source positioned in a second direction among a first direction prioritized with respect to sound reception and the second direction, and outputs a first directivity sound-signal including a sound coming from the first direction, while controlling, depending on the probability, output of the first directivity sound-signal and a second directivity sound-signal including a sound coming from the second direction, where each of the first and second directivity sound-signals is calculated based on the first and second frequency-spectra.

Abstract: A speech recognition method to be performed by a computer, the method including: detecting a first keyword uttered by a user from an audio signal representing voice of the user; detecting a term indicating a request of the user from sections that follow the first keyword in the audio signal; and determining a type of speech recognition processing applied to the following sections in accordance with the detected term indicating the request of the user.

Abstract: A sound processing method includes: executing a time frequency conversion process; executing a noise level evaluation process; executing a bandwidth controlling process; executing a sound source direction decision process; executing a gain setting process; executing a correction process; and executing a frequency time conversion process.

Abstract: A voice correction method implemented by a computer, the method includes: obtaining first voice information which is voice information recorded when noise is generated and on which noise suppression processing is performed and second voice information indicating voice information recorded In an environment in which no noise is generated, and generating emphasized information by emphasizing a component of a band corresponding to a band having a low signal noise ratio (SNR) of the first voice information, among bands of the second voice information; performing machine learning on a model based on the first voice information and the emphasized information; and generating corrected voice information by correcting third voice information on which noise suppression processing is performed, based on the machine-learned model.

Abstract: A noise suppression device including: a processor configured to: divide a voice signal into frames such that two successive frames overlap each other; compute a frequency spectrum for the voice signal; remove a noise component from an amplitude component of the frequency spectrum to compute a noise suppression amplitude component; compare amplitude components or noise suppression amplitude components for each pair of successive frames to compute a comparison value for each frequency; compute a gain according to the comparison value for each frequency; compute a corrected amplitude component for each frequency by multiplying the noise suppression amplitude component by the corresponding gain; and compute a corrected voice signal based on the corrected amplitude component for each frequency.

Abstract: A voice signal processing method includes: converting a first and a second voice signals to a first and a second frequency signals; setting a coefficient of existence representing degree of existence of a target sound and a coefficient of non-existence representing degree of existence of a non-target sound based on a phase difference for each of the predetermined frequencies between the first and the second frequency signals and a target sound existence region indicating an existence position of the target sound; and judging whether the first voice and/or the second voice include the target sound, based on the coefficient of existence, the coefficient of non-existence and a representative value corresponding to either one of the first and the second frequency signals.

Abstract: An apparatus divides each of first and second sound-signals generated respectively by first and second sound-input devices, into frames having a predetermined time length, and converts each frame of the first and second sound-signals into first and second frequency-spectra, respectively, in a frequency domain. For each frame, the apparatus calculates, based on the first and second frequency-spectra, a probability that a sound of the frame is emitted only from a sound-source positioned in a second direction among a first direction prioritized with respect to sound reception and the second direction, and outputs a first directivity sound-signal including a sound coming from the first direction, while controlling, depending on the probability, output of the first directivity sound-signal and a second directivity sound-signal including a sound coming from the second direction, where each of the first and second directivity sound-signals is calculated based on the first and second frequency-spectra.

Abstract: A sound processing method includes: executing a time frequency conversion process; executing a noise level evaluation process; executing a bandwidth controlling process; executing a sound source direction decision process; executing a gain setting process; executing a correction process; and executing a frequency time conversion process.

Abstract: A microphone array device includes a first sound reception unit configured to obtain a first sound signal that is input from a first microphone, a second sound reception unit configured to obtain a second sound signal that is input from a second microphone, a noise state evaluation unit configured to compare the first sound signal and the second sound signal and to obtain an evaluation parameter to evaluate an influence of a non-target sound included in the second sound signal on a target sound included in the first sound signal according to a result of the comparison, a subtraction adjustment unit configured to set a suppression amount for the second sound signal based on the evaluation parameter and to generate a third sound signal; and a subtraction unit configured to generate a signal to be output based on the third sound signal and the first sound signal.

Abstract: A speech recognition method to be performed by a computer, the method including: detecting a first keyword uttered by a user from an audio signal representing voice of the user; detecting a term indicating a request of the user from sections that follow the first keyword in the audio signal; and determining a type of speech recognition processing applied to the following sections in accordance with the detected term indicating the request of the user.

Abstract: A noise suppression device includes: an adaptive filter unit that suppresses, using an adaptive filter, a noise component contained in a voice signal generated from a voice captured by a voice input unit to generate a corrected voice signal; a noise generation detection unit that detects timing of generation of the noise component in the voice signal; and a period suppression unit that suppresses the corrected voice signal during a predetermined period of time after the timing of the generation of the noise component.

Abstract: A voice signal processing method includes: converting a first and a second voice signals to a first and a second frequency signals; setting a coefficient of existence representing degree of existence of a target sound and a coefficient of non-existence representing degree of existence of a non-target sound based on a phase difference for each of the predetermined frequencies between the first and the second frequency signals and a target sound existence region indicating an existence position of the target sound; and judging whether the first voice and/or the second voice include the target sound, based on the coefficient of existence, the coefficient of non-existence and a representative value corresponding to either one of the first and the second frequency signals.

Abstract: A noise suppression device including: a processor configured to: divide a voice signal into frames such that two successive frames overlap each other; compute a frequency spectrum for the voice signal; remove a noise component from an amplitude component of the frequency spectrum to compute a noise suppression amplitude component; compare amplitude components or noise suppression amplitude components for each pair of successive frames to compute a comparison value for each frequency; compute a gain according to the comparison value for each frequency; compute a corrected amplitude component for each frequency by multiplying the noise suppression amplitude component by the corresponding gain; and compute a corrected voice signal based on the corrected amplitude component for each frequency.

Abstract: An audio signal processing device that includes: a processor configured to execute a procedure, the procedure comprising: detecting a speech segment of an audio signal; suppressing noise in the audio signal; and adjusting an amount of suppression of noise such that the amount of suppression during a specific period, which starts from a position based on a terminal end of the detected speech segment and is a period shorter than a period spanning from the terminal end of the detected speech segment to a starting end of a next speech segment, becomes greater than in other segments, and a memory configured to store audio signals before and after noise suppression and the amount of suppression before and after adjustment.

Abstract: A noise suppression device includes: an adaptive filter unit that suppresses, using an adaptive filter, a noise component contained in a voice signal generated from a voice captured by a voice input unit to generate a corrected voice signal; a noise generation detection unit that detects timing of generation of the noise component in the voice signal; and a period suppression unit that suppresses the corrected voice signal during a predetermined period of time after the timing of the generation of the noise component.

Abstract: A speech enhancement device which includes: a speech production section detection unit configured to detect a speech production section in which a speaker produces speech, from an input signal generated by a speech input unit; a timer unit configured to measure an elapsed time from a starting point of the speech production section; a gain determination unit configured to determine a gain, which represents a level of enhancement of the input signal, according to the elapsed time; and an enhancement unit configured to enhance the input signal or a spectrum signal of the input signal in the speech production section according to the gain, whereby the input signal is enhanced only at necessary portions thereof.