Abstract: A learning data acquisition device or the like, capable of acquiring learning data by superimposing noise data on clean voice data at an appropriate SN ratio, is provided. The learning data acquisition device includes a voice recognition influence degree calculation unit and a learning data acquisition unit. The voice recognition influence degree calculation unit calculates an influence degree on voice recognition accuracy caused by a change of a signal-to-noise ratio, based on a result of voice recognition on the kth noise superimposed voice data and a result of voice recognition on the k?1th noise superimposed voice data, where K is an integer of 2 or larger, k=2, 3, . . .

Abstract: The present document relates to audio coding systems which make use of a harmonic transposition method for high frequency reconstruction (HFR), and to digital effect processors, e.g. so-called exciters, where generation of harmonic distortion adds brightness to the processed signal. In particular, a system configured to generate a high frequency component of a signal from a low frequency component of the signal is described. The system may comprise an analysis filter bank (501) configured to provide a set of analysis subband signals from the low frequency component of the signal; wherein the set of analysis subband signals comprises at least two analysis subband signals; wherein the analysis filter bank (501) has a frequency resolution of ?f.

Abstract: System and methods are described for generating training data for a natural language search system, the training data including synthetic user queries and associated structured query language (SQL) statements. A method includes generating an abstract query representation from a probabilistic context-free grammar (PCFG), selecting a sample record from a database using the abstract query representation, generating a tagged user query from the abstract query representation and the sample record, and generating the one or more SQL statements from the tagged user query and one or more SQL templates.

Abstract: A quantization device includes: a trellis-structured vector quantizer which quantizes a first error vector between an N-dimensional (here, “N” is two or more) subvector and a first predictive vector; and an inter-frame predictor which generates a first predictive vector from the quantized N-dimensional subvector, wherein the inter-frame predictor uses a predictive coefficient comprising an N×N matrix and performs an inter-frame prediction using the quantized N-dimensional subvector of a previous stage.

Abstract: A processing system detects a period of non-voice activity and compares its duration to a cutoff period. The system adapts the cutoff period based on parsing previously-recognized speech to determine, according to a model, such as a machine-learned model, the probability that the speech recognized so far is a prefix to a longer complete utterance. The cutoff period is longer when a parse of previously recognized speech has a high probability of being a prefix of a longer utterance.

Abstract: A method for generation of comfort noise for at least two audio channels. The method comprises determining a spatial coherence between audio signals on the respective audio channels, wherein at least one spatial coherence value per frame and frequency band is determined to form a vector of spatial coherence values. A vector of predicted spatial coherence values is formed by a weighted combination of a first coherence prediction and a second coherence prediction that are combined using a weight factor ?. The method comprises signaling information about the weight factor ? to the receiving node, for enabling the generation of the comfort noise for the at least two audio channels at the receiving node.

Abstract: A laryngeal nerve exciting system includes a collar holding a bridge, or a neckband, pressing soft tissue nerve exciters against a patient's neck providing a source of vibrations to stimulate the laryngeal nerve through the larynx. At least one exciter, and preferably two exciters, provide vibrations at preferably 70 Hz to 110 Hz and sufficiently strong to penetrate to the laryngeal nerve. The exciters may be held by the collar circling the neck, or by the neck band partially circling the neck. The therapy system includes a Personal Digital Assistant (PDA) and software which wirelessly connects, monitors, and triggers the device. The system may be used to treat dysphagia, chronic cough, and spasmodic dysphonia.

Abstract: A method for decoding an encoded audio bitstream is disclosed. The method includes receiving the encoded audio bitstream and decoding the audio data to generate a decoded lowband audio signal. The method further includes extracting high frequency reconstruction metadata and filtering the decoded lowband audio signal with an analysis filterbank to generate a filtered lowband audio signal. The method also includes extracting a flag indicating whether either spectral translation or harmonic transposition is to be performed on the audio data and regenerating a highband portion of the audio signal using the filtered lowband audio signal and the high frequency reconstruction metadata in accordance with the flag. The high frequency regeneration is performed as a post-processing operation with a delay of 3010 samples per audio channel.

Abstract: A method for decoding an encoded audio bitstream is disclosed. The method includes receiving the encoded audio bitstream and decoding the audio data to generate a decoded lowband audio signal. The method further includes extracting high frequency reconstruction metadata and filtering the decoded lowband audio signal with an analysis filterbank to generate a filtered lowband audio signal. The method also includes extracting a flag indicating whether either spectral translation or harmonic transposition is to be performed on the audio data and regenerating a highband portion of the audio signal using the filtered lowband audio signal and the high frequency reconstruction metadata in accordance with the flag. The high frequency regeneration is performed as a post-processing operation with a delay of 3010 samples per audio channel.

Abstract: A method for decoding an encoded audio bitstream is disclosed. The method includes receiving the encoded audio bitstream and decoding the audio data to generate a decoded lowband audio signal. The method further includes extracting high frequency reconstruction metadata and filtering the decoded lowband audio signal with an analysis filterbank to generate a filtered lowband audio signal. The method also includes extracting a flag indicating whether either spectral translation or harmonic transposition is to be performed on the audio data and regenerating a highband portion of the audio signal using the filtered lowband audio signal and the high frequency reconstruction metadata in accordance with the flag. The high frequency regeneration is performed as a post-processing operation with a delay of 3010 samples per audio channel.

Abstract: An encoder for encoding an audio signal has: an analyzer configured for deriving prediction coefficients and a residual signal from an unvoiced frame of the audio signal; a gain parameter calculator configured for calculating a first gain parameter information for defining a first excitation signal related to a deterministic codebook and for calculating a second gain parameter information for defining a second excitation signal related to a noise-like signal for the unvoiced frame; and a bitstream former configured for forming an output signal based on an information related to a voiced signal frame, the first gain parameter information and the second gain parameter information.

Abstract: A stereo signal encoding method includes performing spectrum broadening on a quantized line spectral frequency (LSF) parameter of a primary channel signal in a current frame in a stereo signal to obtain a spectrum-broadened LSF parameter of the primary channel signal, determining a prediction residual of an LSF parameter of a secondary channel signal in the current frame based on an original LSF parameter of the secondary channel signal and the spectrum-broadened LSF parameter of the primary channel signal, and performing a quantization on the prediction residual of the LSF parameter of the secondary channel signal.

Abstract: A radio frequency connection between a far field voice detection device and a further device is used to determine a first angular direction from the far field voice detection device to the further device. The determined first angular direction is then used to emphasize, during a noise processing of a plurality of sounds received via use of a plurality of microphones of the far field voice detection device, a first one of the plurality of sounds relative to a remainder of the plurality of sounds.

Abstract: A computer-implemented method for determining a relevance or rating score of a particular user in an electronic social network includes obtaining a digital representation of a local connection structure of the particular user in the network in a computer-readable non-volatile memory; determining a numerical value characteristic of the particular user's relevance; and outputting the numerical value. The numerical value is determined based on the connection structure of the particular user, and on degrees of all sets of users in the social network reachable from the particular user using at most k connections, where k is a fixed preset positive integer number, and based on a degree determined for each enumerated set, and where the degree of a set is determined as a number of connections connecting users inside the set to users outside.

Abstract: An encoding method includes determining an adaptive broadening factor based on a quantized line spectral frequency (LSF) vector of a first channel of a current frame of an audio signal and an LSF vector of a second channel of the current frame, and writing the quantized LSF vector and the adaptive broadening factor into a bitstream.

Abstract: A cognitive function evaluation device includes: an obtainment unit configured to obtain speech data indicating speech uttered by a subject; a calculation unit configured to extract a plurality of vowels from the speech data obtained by the obtainment unit, and calculate, for each of the plurality of vowels, a feature value based on a frequency and an amplitude of at least one formant obtained from a spectrum of the vowel; an evaluation unit configured to evaluate a cognitive function of the subject from the feature value calculated by the calculation unit; and an output unit configured to output an evaluation result of the evaluation unit.

Abstract: Systems, devices, and methods are described for reducing degradation of a voice recognition input. An always listening device may always be listening for voice commands via a microphone and may experience interference from unwanted audio such as from the output audio of television speakers. The always listening device may receive data associated with the output audio over a first communications channel. The always listening device may also receive, on a second communications channel, timing information associated with data. The always listening device may adjust admission of the audio received by the microphone to enable it to arrive at approximately the same time as the data received via the first communications channel. The unwanted output audio included in the audio received via the microphone may then be determined and may be removed so that a voice command in the audio received by the microphone may be processed.

Abstract: The reception side can easily acquire at least function enhancement information inside a codec. Image data of each picture of a base layer is encoded to generate a first encoded stream, and image data of each picture of an enhanced layer is encoded to generate a second encoded stream. The first encoded stream and the second encoded stream are transmitted. Function enhancement information including at least function enhancement information inside a codec is transmitted outside the encoded stream. For example, the function enhancement information further includes function enhancement information outside the codec.

Abstract: Methods, an encoder and a decoder are configured for transition between frames with different internal sampling rates. Linear predictive (LP) filter parameters are converted from a sampling rate S1 to a sampling rate S2. A power spectrum of a LP synthesis filter is computed, at the sampling rate S1, using the LP filter parameters. The power spectrum of the LP synthesis filter is modified to convert it from the sampling rate S1 to the sampling rate S2. The modified power spectrum of the LP synthesis filter is inverse transformed to determine autocorrelations of the LP synthesis filter at the sampling rate S2. The autocorrelations are used to compute the LP filter parameters at the sampling rate S2.

Abstract: Described herein is a technique to facilitate conjunctive filtering for an embedding-based information retrieval system. Each item in a content collection is initially analyzed to identify its property values for some set of properties. For each item, its property values are encoded in the form of a vector, and concatenated with an item embedding for the item to derive an enhanced item embedding. When a query is received, a query embedding is generated. Any filtering parameters received with or as part of the query are encoded in the form of a vector, and the vector is concatenated with the query embedding to derive an enhanced query embedding. The enhanced query embedding is then used in a “k” approximate nearest neighbor search to identify items relevant to the query and having property values that satisfy the filtering parameters.

Abstract: A wireless communication apparatus includes a radio frequency integrated circuit (RFIC) configured to receive an input signal to generate a digital sample signal from the input signal, a data compressor configured to compress the digital sample signal according to a compression manner based on a data probability distribution of the input signal varying based on a receivable signal amplitude range of the RFIC, a data decompressor configured to decompress the compressed digital sample signal on the basis of a decompression manner corresponding to the compression manner to generate a decompressed digital sample signal, a data transfer link configured to transfer the compressed digital sample signal to the data decompressor, and a processor configured to process the decompressed digital sample signal.

Abstract: Embodiments provide an audio processor for processing an audio signal to obtain a subband representation of the audio signal. The audio processor is configured to perform a cascaded lapped critically sampled transform on at least two partially overlapping blocks of samples of the audio signal, to obtain a set of subband samples on the basis of a first block of samples of the audio signal, and to obtain a corresponding set of subband samples on the basis of a second block of samples of the audio signal. Further, the audio processor is configured to perform a weighted combination of two corresponding sets of subband samples, one obtained on the basis of the first block of samples of the audio signal and one obtained on the basis on the second block of samples of the audio signal, to obtain an aliasing reduced subband representation of the audio signal.

Abstract: The present invention relates to audio coding systems which make use of a harmonic transposition method for high frequency reconstruction (HFR). A system and a method for generating a high frequency component of a signal from a low frequency component of the signal is described. The system comprises an analysis filter bank providing a plurality of analysis subband signals of the low frequency component of the signal. It also comprises a non-linear processing unit to generate a synthesis subband signal with a synthesis frequency by modifying the phase of a first and a second of the plurality of analysis subband signals and by combining the phase-modified analysis subband signals. Finally, it comprises a synthesis filter bank for generating the high frequency component of the signal from the synthesis subband signal.

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: The present document relates to audio coding systems which make use of a harmonic transposition method for high frequency reconstruction (HFR), and to digital effect processors, e.g. so-called exciters, where generation of harmonic distortion adds brightness to the processed signal. In particular, a system configured to generate a high frequency component of a signal from a low frequency component of the signal is described. The system may comprise an analysis filter bank (501) configured to provide a set of analysis subband signals from the low frequency component of the signal; wherein the set of analysis subband signals comprises at least two analysis subband signals; wherein the analysis filter bank (501) has a frequency resolution of ?f.

Abstract: A method for recovering a current frame of an audio stream includes detecting that a current packet is lost, the current packet including an audio signal; splitting one or more frames into respective high-band signals and respective low-band signals, the one or more frames precede the current frame in the audio stream; inferring a current low-band signal of the current frame using, as inputs to a machine-learning model, the respective low-band signals; combining the inferred current low-band signal with the respective high-band signals to obtain the current frame; and adding the current frame to a playout buffer.

Abstract: A stereo signal encoding method includes determining a window length of an attenuation window based on an inter-channel time difference; determining a modified linear prediction analysis window based on the window length of the attenuation window, where values of at least some points from a point (L?sub_window_len) to a point (L?1) in the modified linear prediction analysis window are less than values of corresponding points from a point (L?sub_window_len) to a point (L?1) in an initial linear prediction analysis window, and the window length of the modified linear prediction analysis window is equal to a window length of the initial linear prediction analysis window; and performing linear prediction analysis on a to-be-processed sound channel signal based on the modified linear prediction analysis window.

Abstract: Data units of a dataset may be compressed by clustering the data units into clusters, selecting a reference unit for each unit cluster, and compressing data units of each unit cluster using the reference unit of the unit cluster as a dictionary. The computational efficiency of the clustering algorithm may be improved by not applying it to data units themselves, but rather to hash values of the data units, where the hash values have a much smaller size than the data units. The hash function may be a locality-sensitive hash (LSH) function. The reference unit of a cluster may be determined in any of a variety of ways, for example, by selecting a centroid or exemplar of the cluster. Clusters, including their references values, may be indexed in a cluster index (e.g., a Faiss index), which may be searched to assign future added or modified data units to clusters.

Abstract: Methods, apparatus, and processor-readable storage media for data integration demand management using artificial intelligence are provided herein. An example computer-implemented method includes obtaining at least one data integration demand, wherein the at least one data integration demand comprises textual information provided by at least one user; determining multiple parameters of the at least one data integration demand by applying one or more machine learning natural language processing techniques to at least a portion of the textual information provided by the at least one user; generating at least one delivery date prediction for the at least one data integration demand by applying one or more artificial intelligence techniques to the multiple determined parameters of the at least one data integration demand; and performing one or more automated actions based at least in part on the at least one generated delivery date prediction.

Abstract: State of the art systems that are used for time series data prediction have the disadvantage that perform only one step prediction, which has only limited application. Disadvantage of such systems is that extent of applications of such single step predictions are limited. The disclosure herein generally relates to time series data prediction, and, more particularly, to time series data prediction based on seasonal lags. The system processes collected input data and determines order of seasonality of the input data. The system further selects encoders based on the determined order of seasonality and generates input data for a decoder that forms encoder-decoder pair with each of the encoders. The system then generates time series data predictions based on seasonal lag information distributed without redundance between encoder and decoder inputs.

Abstract: Systems, methods, and computer-readable storage devices are disclosed for improved real-time audio processing. One method including: constructing a deep neural network model, including a plurality of at least one-bit neurons, configured to output a predicted label of audio data, the plurality of at least one-bit neurons arranged in a plurality of layers, including at least one hidden layer, and being connected by a plurality of connections, each connection having at least a one-bit weight, wherein one or both of the plurality of at least one-bit neurons and the plurality of connections have a reduced bit precision; receiving a training data set, the training data set including audio data; training the deep neural network model using the training data set; and outputting a trained deep neural network model configured to output a predicted label of real-time audio data.

Abstract: What is described is an apparatus for post-processing an audio signal, having: a time-spectrum-converter for converting the audio signal into a spectral representation having a sequence of spectral frames; a prediction analyzer for calculating prediction filter data for a prediction over frequency within a spectral frame; a shaping filter controlled by the prediction filter data for shaping the spectral frame to enhance a transient portion within the spectral frame; and a spectrum-time-converter for converting a sequence of spectral frames having a shaped spectral frame into a time domain.

Abstract: An apparatus for generating a bandwidth enhanced audio signal from an input audio signal having an input audio signal frequency range includes: a raw signal generator configured for generating a raw signal having an enhancement frequency range, wherein the enhancement frequency range is not included in the input audio signal frequency range; a neural network processor configured for generating a parametric representation for the enhancement frequency range using the input audio frequency range of the input audio signal and a trained neural network; and a raw signal processor for processing the raw signal using the parametric representation for the enhancement frequency range to obtain a processed raw signal having frequency components in the enhancement frequency range, wherein the processed raw signal or the processed raw signal and the input audio signal frequency range of the input audio signal represent the bandwidth enhanced audio signal.

Abstract: This encoding device is able to encode an S signal efficiently in MS prediction encoding. An M signal encoding unit generates first encoding information by encoding a sum signal indicating a sum of a left channel signal and a right channel signal that constitute a stereo signal. An energy difference calculation unit calculates a prediction parameter for predicting a difference signal indicating a difference between the left channel signal and the right channel signal by using a parameter regarding an energy difference between the left channel signal and the right channel signal. An entropy encoding unit generates second encoding information by encoding the prediction parameter.

Abstract: The present disclosure relates to a method of decoding audio scene content from a bitstream by a decoder that includes an audio renderer with one or more rendering tools.

Abstract: An encoding method includes determining a target adaptive broadening factor based on a quantized line spectral frequency (LSF) parameter of a primary channel signal in a current frame and an LSF parameter of a secondary channel signal in the current frame, and writing the quantized LSF parameter of the primary channel signal in the current frame and the target adaptive broadening factor into a bitstream.

Abstract: A computer may train a single-class machine learning using normal speech recordings. The machine learning model or any other model may estimate the normal range of parameters of a physical speech production model based on the normal speech recordings. For example, the computer may use a source-filter model of speech production, where voiced speech is represented by a pulse train and unvoiced speech by a random noise and a combination of the pulse train and the random noise is passed through an auto-regressive filter that emulates the human vocal tract. The computer leverages the fact that intentional modification of human voice introduces errors to source-filter model or any other physical model of speech production. The computer may identify anomalies in the physical model to generate a voice modification score for an audio signal. The voice modification score may indicate a degree of abnormality of human voice in the audio signal.

Abstract: A method for generation of comfort noise for at least two audio channels. The method comprises determining a spatial coherence between audio signals on the respective audio channels, wherein at least one spatial coherence value per frame and frequency band is determined to form a vector of spatial coherence values. A vector of predicted spatial coherence values is formed by a weighted combination of a first coherence prediction and a second coherence prediction that are combined using a weight factor a. The method comprises signaling information about the weight factor a to the receiving node, for enabling the generation of the comfort noise for the at least two audio channels at the receiving node.

Abstract: A stereo signal encoding method includes performing spectrum broadening on a quantized line spectral frequency (LSF) parameter of a primary channel signal in a current frame in a stereo signal to obtain a spectrum-broadened LSF parameter of the primary channel signal, determining a prediction residual of an LSF parameter of a secondary channel signal in the current frame based on an original LSF parameter of the secondary channel signal and the spectrum-broadened LSF parameter of the primary channel signal, and performing a quantization on the prediction residual of the LSF parameter of the secondary channel signal.

Abstract: An apparatus for encoding a multi-channel signal having at least two channels, has: a downmixer for calculating a downmix signal from the multi-channel signal; a parameter calculator for calculating a side gain from a first channel of the at least two channels and a second channel of the at least two channels and for calculating a residual gain from the first channel and the second channel; and an output interface for generating an output signal, the output signal having information on the downmix signal, and on the side gain and the residual gain.

Abstract: The present invention relates to a method for encoding a candidate vector for searching for a neighbor that is nearest to a query in a candidate dataset, the method comprising a normalization step of normalizing an input vector to obtain a direction vector and vector energy; a quantization step of quantizing the direction vector to obtain a code word and a residual vector; a step of repeating the normalization step and the quantization step, as many times as a predetermined number of encoding times, by using the residual vector as an input vector; and a step of encoding the candidate vector by using one or more code words and energy of one or more vectors resulting from the repetition. According to the present invention, a dataset having a very wide range of energy values can be effectively approximated and higher precision thereof can be obtained.

Abstract: A laryngeal nerve exciting system includes a collar holding a bridge, or a neckband, pressing soft tissue nerve exciters against a patient's neck providing a source of vibrations to stimulate the laryngeal nerve through the larynx. At least one exciter, and preferably two exciters, provide vibrations at preferably 70 Hz to 110 Hz and sufficiently strong to penetrate to the laryngeal nerve. The exciters may be held by the collar circling the neck, or by the neck band partially circling the neck. The therapy system includes a Personal Digital Assistant (PDA) and software which wirelessly connects, monitors, and triggers the device. The system may be used to treat dysphagia, chronic cough, and spasmodic dysphonia.

Abstract: A laryngeal nerve exciting system includes a collar holding a bridge, or a neckband, pressing soft tissue nerve exciters against a patient's neck providing a source of vibrations to stimulate the laryngeal nerve through the larynx. At least one exciter, and preferably two exciters, provide vibrations at preferably 70 Hz to 110 Hz and sufficiently strong to penetrate to the laryngeal nerve. The exciters may be held by the collar circling the neck, or by the neck band partially circling the neck. The therapy system includes a Personal Digital Assistant (PDA) and software which wirelessly connects, monitors, and triggers the device. The system may be used to treat dysphagia, chronic cough, and spasmodic dysphonia.

Abstract: An apparatus for determining an estimated pitch lag is provided. The apparatus includes an input interface for receiving a plurality of original pitch lag values, and a pitch lag estimator for estimating the estimated pitch lag. The pitch lag estimator is configured to estimate the estimated pitch lag depending on a plurality of original pitch lag values and depending on a plurality of information values, wherein for each original pitch lag value of the plurality of original pitch lag values, an information value of the plurality of information values is assigned to the original pitch lag value.

Abstract: A method and system for controlling air-interface resource utilization in a wireless communication system in which an access node provides service on a carrier defining air-interface resources. An example method includes monitoring a rate of voice muting on the carrier, detecting when the monitored rate of voice muting on the carrier is at least predefined threshold high, and responsive to detecting that the monitored rate of voice muting on the carrier is at least predefined threshold high, starting to apply a maximum limit on how many user equipment devices (UEs) the access node will concurrently serve with a resource-intensive service on the carrier.

Abstract: There is provided an error concealment unit, method, and computer program, for providing an error concealment audio information for concealing a loss of an audio frame in an encoded audio information. In one embodiment, the error concealment unit provides an error concealment audio information for a lost audio frame on the basis of a properly decoded audio frame preceding the lost audio frame. The error concealment unit derives a damping factor on the basis of characteristics of a decoded representation of the properly decoded audio frame preceding the lost audio frame. The error concealment unit performs a fade out using the damping factor.

Abstract: In apparatus, methods, and programs for selecting pitch lag, an encoder obtains a first and a second estimates of a pitch lag for a current frame. A selected value is chosen by selection between the first and the second estimates, based on a first and a second correlation measurements. The second estimate is conditioned by the pitch lag selected at the previous frame. The selection is based on a comparison between: a downscaled version of a first correlation measurement associated to the current frame and obtained at a lag corresponding to the first estimate; and a second correlation measurement associated to the current frame and obtained at a lag corresponding to the second estimate.

Abstract: An audio coding method, comprising: obtaining an ith audio frame in n consecutive audio frames and obtaining an ith piece of coded data and an ith piece of redundant data based on the ith audio frame, wherein the ith piece of coded data is obtained by coding the ith audio frame, and the ith piece of redundant data is obtained by coding and buffering the ith audio frame, wherein n is a positive integer, and 1?i?n; and packing the ith piece of coded data and at most m pieces of redundant data before the ith piece of redundant data into an ith audio data packet, wherein m is a preset positive integer. An audio decoding method and a computer device are further provided.

Abstract: A processing system detects a period of non-voice activity and compares its duration to a cutoff period. The system adapts the cutoff period based on parsing previously-recognized speech to determine, according to a model, such as a machine-learned model, the probability that the speech recognized so far is a prefix to a longer complete utterance. The cutoff period is longer when a parse of previously recognized speech has a high probability of being a prefix of a longer utterance.

Abstract: A linear prediction residual energy tilt-based audio signal classification method and apparatus, where the method includes: determining, according to voice activity of a current audio frame, whether to obtain a linear prediction residual energy tilt of a current audio frame of the current audio frame and store a frequency spectrum fluctuation of the current frame in a frequency spectrum fluctuation memory, where the linear prediction residual energy tilt denotes an extent to which an audio signal's linear prediction residual energy changes as a linear prediction order inscreases; updating, according to whether the audio frame is percussive music or activity of a historical audio frame, frequency spectrum fluctuations stored in the frequency spectrum fluctuation memory; and classifying the current audio frame as a speech frame or a music frame according to statistics of some or all of effective data of the frequency spectrum fluctuations stored in the frequency spectrum fluctuation memory.