Abstract: An envelope sequence is provided that can improve approximation accuracy near peaks caused by the pitch period of an audio signal. A periodic-combined-envelope-sequence generation device according to the present invention takes, as an input audio signal, a time-domain audio digital signal in each frame, which is a predetermined time segment, and generates a periodic combined envelope sequence as an envelope sequence. The periodic-combined-envelope-sequence generation device according to the present invention comprises at least a spectral-envelope-sequence calculating part and a periodic-combined-envelope generating part. The spectral-envelope-sequence calculating part calculates a spectral envelope sequence of the input audio signal on the basis of time-domain linear prediction of the input audio signal.

Abstract: Provided is a sound collection loudspeaker apparatus that makes it possible to intuitively distinguish which talker is talking, and improve the comfort of conversations, when performing in-vehicle conversation and conversations with people outside of a vehicle. The sound collection loudspeaker apparatus is installed in a vehicle. Two or more sound collection and amplification positions are assumed to be present inside the vehicle.

Abstract: A coding method and a decoding method are provided which can use in combination a predictive coding and decoding method which is a coding and decoding method that can accurately express coefficients which are convertible into linear prediction coefficients with a small code amount and a coding and decoding method that can obtain correctly, by decoding, coefficients which are convertible into linear prediction coefficients of the present frame if a linear prediction coefficient code of the present frame is correctly input to a decoding device.

Abstract: An envelope sequence is provided that can improve approximation accuracy near peaks caused by the pitch period of an audio signal. A periodic-combined-envelope-sequence generation device according to the present invention takes, as an input audio signal, a time-domain audio digital signal in each frame, which is a predetermined time segment, and generates a periodic combined envelope sequence as an envelope sequence. The periodic-combined-envelope-sequence generation device according to the present invention comprises at least a spectral-envelope-sequence calculating part and a periodic-combined-envelope generating part. The spectral-envelope-sequence calculating part calculates a spectral envelope sequence of the input audio signal on the basis of time-domain linear prediction of the input audio signal.

Abstract: In encoding, a frequency-domain sample sequence derived from an acoustic signal is divided by a weighted envelope and is then divided by a gain, the result obtained is quantized, and each sample is variable-length encoded. The error between the sample before quantization and the sample after quantization is quantized with information saved in this variable-length encoding. This quantization is performed under a rule that specifies, according to the number of saved bits, samples whose errors are to be quantized. In decoding, variable-length codes in an input sequence of codes are decoded to obtain a frequency-domain sample sequence; an error signal is further decoded under a rule that depends on the number of bits of the variable-length codes; and from the obtained sample sequence, the original sample sequence is obtained according to supplementary information.

Abstract: A coding technology that efficiently codes an input sound signal irrespective of the characteristics thereof and can obtain a decoded sound signal that sounds less artificial to a listener. A coding method codes an input sound signal frame by frame of a predetermined time segment by a selected coding processing from a plurality of types of coding processing in the frequency domain, the coding method makes it possible for a selection unit to select coding processing which is different from the coding processing of the preceding frame as coding processing of the present frame if at least one of the magnitude of the energy of high frequency components of the input sound signal of the preceding frame and the magnitude of the energy of high frequency components of the input sound signal of the present frame is smaller than or equal to a predetermined threshold value.

Abstract: A coding technology that efficiently codes an input sound signal irrespective of the characteristics thereof and can obtain a decoded sound signal that sounds less artificial to a listener. A coding method codes an input sound signal frame by frame of a predetermined time segment by a selected coding processing from a plurality of types of coding processing in the frequency domain, the coding method makes it possible for a selection unit to select coding processing which is different from the coding processing of the preceding frame as coding processing of the present frame if at least one of the magnitude of the energy of high frequency components of the input sound signal of the preceding frame and the magnitude of the energy of high frequency components of the input sound signal of the present frame is smaller than or equal to a predetermined threshold value.

Abstract: In encoding, a frequency-domain sample sequence derived from an acoustic signal is divided by a weighted envelope and is then divided by a gain, the result obtained is quantized, and each sample is variable-length encoded. The error between the sample before quantization and the sample after quantization is quantized with information saved in this variable-length encoding. This quantization is performed under a rule that specifies, according to the number of saved bits, samples whose errors are to be quantized. In decoding, variable-length codes in an input sequence of codes are decoded to obtain a frequency-domain sample sequence; an error signal is further decoded under a rule that depends on the number of bits of the variable-length codes; and from the obtained sample sequence, the original sample sequence is obtained according to supplementary information.

Abstract: An autocorrelation calculation unit 21 calculates an autocorrelation Ro(i) from an input signal. A prediction coefficient calculation unit 23 performs linear prediction analysis by using a modified autocorrelation R?o(i) obtained by multiplying a coefficient wo(i) by the autocorrelation Ro(i). It is assumed here, for each order i of some orders i at least, that the coefficient wo(i) corresponding to the order i is in a monotonically increasing relationship with an increase in a value that is negatively correlated with a fundamental frequency of the input signal of the current frame or a past frame.

Abstract: The present disclosure relates to a method of machine learning regardless of the number of dimensions of the samples. The method provides model learning of a variational auto-encoder that uses AUC optimization criteria. The method includes learning parameters ?{circumflex over (?)} and ?{circumflex over (?)} of the a variational auto-encoder. The variational auto-encoder includes an encoder for constructing a latent variable from an observed variable and a decoder for reconstructing the observed variable. The method uses learning data set defined using based on normal data generated from sounds observed during normal operation and abnormal data generated from sounds observed during abnormal operation. The AUC value is based in part on a reconstruction probability. Incorporating aspects of the reconstruction error into the AUC value prevents the variational auto-encoder from divergence of the abnormality degree regarding the abnormal data.

Abstract: An autocorrelation calculation unit 21 calculates an autocorrelation RO(i) from an input signal. A prediction coefficient calculation unit 23 performs linear prediction analysis by using a modified autocorrelation R?O(i) obtained by multiplying a coefficient wO(i) by the autocorrelation RO(i). It is assumed here, for each order i of some orders i at least, that the coefficient wO(i) corresponding to the order i is in a monotonically increasing relationship with an increase in a value that is negatively correlated with a fundamental frequency of the input signal of the current frame or a past frame.

Abstract: A howling suppression apparatus includes: an integration processing part that obtains the maximum value among L values corresponding to n-th frames of L i-th signals, for i=1, 2, . . . , L, L being any integer equal to or greater than 2, the L i-th signals being frequency-domain signals obtained from sound signals collected by multiple microphones; and a howling suppression processing part that performs howling suppression processing on at least any of the L i-th signals using the maximum value.

Abstract: There is provided a model learning technique for learning a model which performs classification into three values by model learning using an AUC optimization criterion. A model learning unit is included which learns a parameter ?{circumflex over (?)} of a model by using a learning data set based on a criterion which uses a predetermined AUC value, the learning data set being defined using normal data generated from sound observed in a normal state and abnormal data generated from sound observed in an abnormal state, and the AUC value is defined from a difference between an abnormality degree of the normal data and an abnormality degree of the abnormal data using a two-stage step function T(x).

Abstract: An encoding/decoding technique is provided with which encoding can be performed using a small average number of bits even for a sequence of integer values including small non-zero values and having a distribution heavily biased toward small values. The present invention includes an integer encoding part 110 that, with respect to an inputted sequence of non-negative integer values x_n, n?{1, 2, . . . , N} (hereinafter referred to as an “integer sequence”), obtains a 1-bit code as a code corresponding to a run of integer values 0 of length L, a predetermined number which satisfies 2?L?2K?1, which is included in the integer sequence and obtains a (K×x_n)-bit or (K×x_n+1)-bit code as a code corresponding to a set of a run of integer values 0 of length from 0 to L?1 and one non-zero integer value x_n included in the integer sequence, where K is assumed to be an integer equal to or larger than 2.

Abstract: A technology of accurately coding and decoding coefficients which are convertible into linear prediction coefficients even for a frame in which the spectrum variation is great while suppressing an increase in the code amount as a whole is provided. A coding device includes: a first coding unit that obtains a first code by coding coefficients which are convertible into linear prediction coefficients of more than one order; and a second coding unit that obtains a second code by coding at least quantization errors of the first coding unit if (A?1) an index Q commensurate with how high the peak-to-valley height of a spectral envelope is, the spectral envelope corresponding to the coefficients which are convertible into the linear prediction coefficients of more than one order, is larger than or equal to a predetermined threshold value Th1 and/or (B?1) an index Q? commensurate with how short the peak-to-valley height of the spectral envelope is, is smaller than or equal to a predetermined threshold value Th1?.

Abstract: An envelope sequence is provided that can improve approximation accuracy near peaks caused by the pitch period of an audio signal. A periodic-combined-envelope-sequence generation device according to the present invention takes, as an input audio signal, a time-domain audio digital signal in each frame, which is a predetermined time segment, and generates a periodic combined envelope sequence as an envelope sequence. The periodic-combined-envelope-sequence generation device according to the present invention comprises at least a spectral-envelope-sequence calculating part and a periodic-combined-envelope generating part. The spectral-envelope-sequence calculating part calculates a spectral envelope sequence of the input audio signal on the basis of time-domain linear prediction of the input audio signal.

Abstract: An envelope sequence is provided that can improve approximation accuracy near peaks caused by the pitch period of an audio signal. A periodic-combined-envelope-sequence generation device according to the present invention takes, as an input audio signal, a time-domain audio digital signal in each frame, which is a predetermined time segment, and generates a periodic combined envelope sequence as an envelope sequence. The periodic-combined-envelope-sequence generation device according to the present invention comprises at least a spectral-envelope-sequence calculating part and a periodic-combined-envelope generating part. The spectral-envelope-sequence calculating part calculates a spectral envelope sequence of the input audio signal on the basis of time-domain linear prediction of the input audio signal.

Abstract: Accuracy of unsupervised anomalous sound detection is improved using a small number of pieces of anomalous sound data. A threshold deciding part (13) calculates an anomaly score for each of a plurality of pieces of anomalous sound data, using a normal model learned with normal sound data and an anomaly model expressing the pieces of anomalous sound data, and decides a minimum value among the anomaly scores as a threshold. A weight updating part (14) updates, using a plurality of pieces of normal sound data, the pieces of anomalous sound data and the threshold, weights of the anomaly model so that all the pieces of anomalous sound data are judged as anomalous, and probability of the pieces of normal sound data being judged as anomalous is minimized.

Abstract: An encoder and a decoder are provided that are capable of reproducing a frequency-domain envelope sequence that provides high approximation accuracy around peaks caused by the pitch period of an audio signal by using a small amount of code. An encoder of the present invention comprises a periodic-combined-envelope generating part and a variable-length coding part. The periodic-combined-envelope generating part generates a periodic combined envelope sequence which is a frequency-domain sequence based on a spectral envelope sequence which is a frequency-domain sequence corresponding to a linear predictive coefficient code obtained from an input audio signal and on a frequency-domain period. The variable-length coding part encodes a frequency-domain sequence derived from the input audio signal. A decoder of the present invention comprises a periodic-combined-envelope generating part and a variable-length decoding part.

Abstract: A coding technology that efficiently codes an input sound signal irrespective of the characteristics thereof and can obtain a decoded sound signal that sounds less artificial to a listener. A coding method codes an input sound signal frame by frame of a predetermined time segment by a selected coding processing from a plurality of types of coding processing in the frequency domain, the coding method makes it possible for a selection unit to select coding processing which is different from the coding processing of the preceding frame as coding processing of the present frame if at least one of the magnitude of the energy of high frequency components of the input sound signal of the preceding frame and the magnitude of the energy of high frequency components of the input sound signal of the present frame is smaller than or equal to a predetermined threshold value.