Abstract: Embodiments are directed to a companding method and system for reducing coding noise in an audio codec. A compression process reduces an original dynamic range of an initial audio signal through a compression process that divides the initial audio signal into a plurality of segments using a defined window shape, calculates a wideband gain in the frequency domain using a non-energy based average of frequency domain samples of the initial audio signal, and applies individual gain values to amplify segments of relatively low intensity and attenuate segments of relatively high intensity. The compressed audio signal is then expanded back to the substantially the original dynamic range that applies inverse gain values to amplify segments of relatively high intensity and attenuating segments of relatively low intensity. A QMF filterbank is used to analyze the initial audio signal to obtain a frequency domain representation.

Abstract: An information processing device according to an embodiment includes a keyword extracting unit, a tag generating unit and a UI control unit. The keyword extracting unit extracts a keyword from time-series texts within a time range set by a user. The tag generating unit generates a tag corresponding to a time period from a first appearing time until a last appearing time of a same keyword appearing plural times within a duration set according to the time range. The UI control unit creates a UI screen including a first display area in which a time axis corresponding to the time range is displayed and a second display area in which the tag is displayed while causing the tag to correspond to the time period on the time axis, and resets, by selecting the tag, a time period of the selected tag in the time range to update the UI screen.

Abstract: A sound source localization unit determines a localized sound source direction that is a direction to a sound source on the basis of acoustic signals of a plurality of channels acquired from M (M is an integer equal to or greater than 3) sound pickup units being at different positions, and a sound source position estimation unit determines an intersection of straight lines to an estimated sound source direction, which is a direction from the sound pickup unit to an estimated sound source position of the sound source for each set of the two sound pickup units, classifies a distribution of intersections into a plurality of clusters, and updates the estimated sound source positions so that an estimation probability that is a probability of the estimated sound source positions being classified into clusters corresponding to the sound sources becomes high.

Abstract: Methods and apparatuses are comprising: a screen; an input device; at least one non-transitory memory storing instructions; and one or more processors in communication with the screen, the input device, and the at least one non-transitory memory, wherein the one or more processors execute the instructions to: display, utilizing the screen, a contactor window including: at least one contactor user interface element configured to have presented, in connection therewith, a plurality of contactor identifiers of a contactor communicant represented by a contactor email communications agent, at least one contactee user interface element configured to have presented, in connection therewith, a plurality of contactee identifiers of a plurality of contactee communicants each represented by a corresponding contactee email communications agent, a message user interface element configured to present a message addressed from one of the plurality of contactor identifiers of the contactor selected in connection with the at l

Abstract: Customer support, and other types of activities in which there is a dialog between two humans can generate large volumes of conversation records. Automated analysis of these records can provide information about high-level features of, for example, the workings of a customer service department. Analysis of these conversations between a customer and a customer-support agent may also allow identification of customer support activities that can be provided by virtual agents instead of actual human agents. The analysis may evaluate conversations in terms of complexity, duration, and sentiment of the participants. Additionally, the conversations may also be analyzed to identify the existence of selected concepts or keywords. Workflow characteristics, the extent to which the conversation represents a multi-step process intended to accomplish a task, may also be determined for the conversations.

Abstract: A system and method for associating an audio clip with an object is provided wherein the voice-based system, such as a voicemail system, is used to record the audio clips.

Abstract: Example embodiments disclosed herein relate to perception based multimedia processing. There is provided a method for processing multimedia data, the method includes automatically determining user perception on a segment of the multimedia data based on a plurality of clusters, the plurality of clusters obtained in association with predefined user perceptions and processing the segment of the multimedia data at least in part based on determined user perception on the segment. Corresponding system and computer program products are disclosed as well.

Abstract: Input context for a statistical dialog manager may be provided. Upon receiving a spoken query from a user, the query may be categorized according to at least one context clue. The spoken query may then be converted to text according to a statistical dialog manager associated with the category of the query and a response to the spoken query may be provided to the user.

Abstract: Methods and systems for managing a premises are described. A premises or devices at a premises may be associated with one or more premises zones. The one or more premises zones may be associated with corresponding content. If data is received from a device associated with a particular premises zone, then the content may be output. The content may be used to notify a user of an event, state change, or other indication associated with the particular premises zone.

Abstract: A method, computer system, and a computer program product for adaptively selecting an acoustic feature extractor in an Artificial Intelligence system is provided. The present invention may include acquiring a frame of an acoustic signal. The present invention may include checking a status of a flag to be used to indicate a proper acoustic feature extractor to be selected. The present invention may include processing the frame of the acoustic signal by the selected acoustic feature extractor indicated by the checked status. The present invention may include determining, based on data generated in the processing of the frame of the acoustic signal, an actual status of the frame of the acoustic signal. The present invention may include updating the status of the flag according to the actual status.

Abstract: A language model is modified for a local speech recognition system using remote speech recognition sources. In one example, a speech utterance is received. The speech utterance is sent to at least one remote speech recognition system. Text results corresponding to the utterance are received from the remote speech recognition system. A local text result is generated using local vocabulary. The received text results and the generated text result are compared to determine words that are out of the local vocabulary and the local vocabulary is updated using the out of vocabulary words.

Abstract: A method for predicting a speech recognition quality of a phrase comprising at least one word includes: receiving, on a computer system including a processor and memory storing instructions, the phrase; computing, on the computer system, a set of features comprising one or more features corresponding to the phrase; providing the phrase to a prediction model on the computer system and receiving a predicted recognition quality value based on the set of features; and returning the predicted recognition quality value.

Abstract: Exemplary embodiments provide techniques for evaluating when words or phrases of a translation were generated with a low degree of confidence, and conveying this information when the translation is presented. For example, if a source language word is encountered in source material for translation, but the source language word was only encountered a few times (or not at all) in the training data used to train the translation system, then the resulting translation may be flagged as being of low confidence. Other situations, such as the generation of two equally-likely translations, or translation system model disagreement, may also indicate a questionable translation. When the translation is displayed, questionable words and phrases may be flagged, and possible alternative translations may be presented. If one of the alternatives is selected, this information may be used to update the translation system's models in order to improve translation quality in the future.

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: An approach to keyword spotting makes use of acoustic parameters that are trained on a keyword spotting task as well as on a second speech recognition task, for example, a large vocabulary continuous speech recognition task. The parameters may be optimized according to a weighted measure that weighs the keyword spotting task more highly than the other task, and that weighs utterances of a keyword more highly than utterances of other speech. In some applications, a keyword spotter configured with the acoustic parameters is used for trigger or wake word detection.

Abstract: The quality of encoded signals can be improved by reclassifying AUDIO signals carrying non-speech data as VOICE signals when periodicity parameters of the signal satisfy one or more criteria. In some embodiments, only low or medium bit rate signals are considered for re-classification. The periodicity parameters can include any characteristic or set of characteristics indicative of periodicity. For example, the periodicity parameter may include pitch differences between subframes in the audio signal, a normalized pitch correlation for one or more subframes, an average normalized pitch correlation for the audio signal, or combinations thereof. Audio signals which are re-classified as VOICED signals may be encoded in the time-domain, while audio signals that remain classified as AUDIO signals may be encoded in the frequency-domain.

Abstract: A system for adjusting lighting of a towing hitch region of a vehicle includes processor, user input device communicatively coupled to the processor, and light source comprising actuator communicatively coupled to the processor. The light source are configured to selectively illuminate the towing hitch region of the vehicle. The system further includes memory module communicatively coupled to the processor that stores logic that, when executed by the processor, cause the system to receive user instructions from the user input device; and adjust an angle of illumination of the light source with the actuator to illuminate the towing hitch region based on the user instructions.

Abstract: Systems and methods for providing video subtitling and text communications (e.g., real time text (RTT) and conventional text messaging) during video calls. The system can include video calling with voice recognition based subtitles. The system can also include a call log to provide a textual record of the audio portion of the video call. The system can utilize embedded or online (e.g., cloud-based) voice recognition systems to provide the subtitles and call log. The system can enable users to send RTT, standard text, or other messages to multiple users participating in a video call via a public text interface. The system can also enable users to send private RTT, standard text, or other messages to specified participants during video calls using parallel interfaces.

Abstract: Method for identification of a interaction signature of a user, comprising the flowing steps of: acquisition of a direct data (10), acquisition of an indirect data (20), acquisition and mapping of the environment (30), to obtain an interaction data set characterized in that the method comprising further steps of establishing an interaction space comprising representation (50) of characteristics of an interaction build upon interaction data set; searching through an interaction space historic data for a pattern of characteristics of an interaction to identify a user, or storing interaction data in an interaction space for future identification.

Abstract: According to one embodiment, a request for seeding a predetermined number of small files with a predetermined locality in a storage system is received, each of the files to have a predetermined file size. In response to the request, a plurality of segments and fingerprints of the segments are generated. File trees representing the predetermined number of files respectively are generated based on the fingerprints of the segments, each of the files represented by the segments having the predetermined file size. A namespace representing one or more directories of the files is generated based on the file trees, where each of the directories of files satisfies the predetermined locality. The namespace and segments corresponding to the files of one or more directories are written to a storage device of the storage system.

Abstract: Technologies are generally described to develop and implement a searchable knowledge source to identify distributed user interface (DUI) elements. In some examples, a DUI identification system may receive a control record of an application and populate one or more searchable knowledge sources based on an application description retrieved. The application description may include keywords, input elements, and output elements, and the searchable knowledge sources may be generated from control records of a multitude of applications. The DUI identification system may execute a query on the searchable knowledge sources based on the received keywords, input elements, and output elements associated with a target workflow from a requesting client. A query result that includes one or more DUI elements may be provided to the requesting client. The DUI elements may connect the input elements to corresponding output elements and match the keywords associated with the target workflow.

Abstract: An encoder includes 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 includes a periodic-combined-envelope generating part and a variable-length decoding 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 and on a frequency-domain period.

Abstract: A spectrum coding method includes quantizing spectral data of a current band based on a first quantization scheme, generating a lower bit of the current band using the spectral data and the quantized spectral data, quantizing a sequence of lower bits including the lower bit of the current band based on a second quantization scheme, and generating a bitstream based on a upper bit excluding N bits, where N is 1 or greater, from the quantized spectral data and the quantized sequence of lower bits.

Abstract: An electronic notebook system is described that comprises a housing, a computing device, wireless interfaces, antennas, sensors, a touch display configured to receive input via a stylus and/or human digit input, the stylus comprising a pressure and/or an inclination sensor, a microphone, camera, the notebook system configured to provide a user condition interface, receive a user selection of a first user condition, provide an interface configured to receive user details, receive audible user details via the microphone, convert the audible user details received via the microphone to text, perform natural language processing to identify text keywords utilizing sentence segmentation, part-of-speech tagging, paraphrase recognition, and/or co-reference resolution, identify a condition based at least in part on the identified one or more keywords, dynamically generate an alert based at least in part on the identified condition, wirelessly transmit the generated alert to one or more destinations via at least a first

Abstract: A system and method of converting source speech to target speech using intermediate speech data is disclosed. The method comprises identifying intermediate speech data that match target voice training data based on acoustic features; performing dynamic time warping to match the second set of acoustic features of intermediate speech data and the first set of acoustic features of target voice training data; training a neural network to convert the intermediate speech data to target voice training data; receiving source speech data; converting the source speech data to an intermediate speech; converting the intermediate speech to a target speech sequence using the neural network; and converting the target speech sequence to target speech using the pitch from the target voice training data.

Abstract: An image forming apparatus includes an application on a framework separated into a core logic portion handling basic processing and a user interface frame portion handling rendering processing and operates; and a controller that executes the application and the framework. The core logic portion is implemented with an application programming interface defined by the framework. The application includes plural applications including a device application. The framework loads all core logic portions of the plural applications at activation of a system. A core logic portion of the device application monitors a state of a device relating to execution of another application and holds information on the device. A core logic portion of the other application acquires the information on the device from the device application and holds the information as display information.

Abstract: An autocorrelation calculating part calculates autocorrelation Ro(i) from an input signal. A predictive coefficient calculating part performs linear predictive analysis using modified autocorrelation R?o(i) obtained by multiplying the autocorrelation Ro(i) by a coefficient wo(i). Here, a case is comprised where, for at least part of each order i, the coefficient wo(i) corresponding to each order i monotonically decreases as a value having positive correlation with a pitch gain in an input signal of a current frame or a past frame increases.

Abstract: Methods and systems are provided for visualizing spatial audio using determined properties for time segments of the spatial audio. Such properties include the position sound is coming from, intensity of the sound, focus of the sound, and color of the sound at a time segment of the spatial audio. These properties can be determined by analyzing the time segment of the spatial audio. Upon determining these properties, the properties are used in rendering a visualization of the sound with attributes based on the properties of the sound(s) at the time segment of the spatial audio.

Abstract: An autocorrelation calculating part calculates autocorrelation Ro(i) from an input signal. A predictive coefficient calculating part performs linear predictive analysis using modified autocorrelation. R?o(i) obtained by multiplying the autocorrelation Ro(i) by a coefficient wo(i). Here, a case is comprised where, for at least part of each order i, the coefficient wo(i) corresponding to each order i monotonically decreases as a value having positive correlation with a pitch gain in an input signal of a current frame or a past frame increases.

Abstract: Methods, systems, and computer readable media for reconfiguring a session binding repository (SBR) database for a data communications network. In one example, a method includes, before reconfiguring the SBR database, selecting a first plurality of SBR servers for storage of a first plurality of SBR records using a first assignment algorithm. While reconfiguring the SBR database, the method includes selecting a second plurality of SBR servers for storage of a second plurality of SBR records using a second assignment algorithm and searching for a first plurality of stored records in the SBR database based on both the first and second assignment algorithms.

Abstract: An autocorrelation calculating part calculates autocorrelation Ro(i) from an input signal. A predictive coefficient calculating part performs linear predictive analysis using modified autocorrelation R?o(i) obtained by multiplying the autocorrelation Ro(i) by a coefficient wo(i). Here, it is assumed that a case where, for at least part of each order i, the coefficient wo(i) corresponding to each order i monotonically increases as a value having negative correlation with a fundamental frequency of an input signal in a current frame or a past frame increases and a case where the coefficient wo(i) monotonically decreases as a value having positive correlation with a pitch gain in a current frame or a past frame increases, are included.

Abstract: A server, method, and system for providing information to an electronic device are provided. An audio-based connection is established with the electronic device. A first identifier of the electronic device is determined in association with the audio-based connection. The electronic device is detected as being able to support a visual-based connection based on the first identifier. A second identifier of the electronic device is determined based on the first identifier. Visual information is provided to the electronic device via the visual-based connection after detecting that the electronic device is able to support the visual-based connection and determining the second identifier. The visual information is provided to the electronic device based on the second identifier.

Abstract: A digital broadcasting system including a transmitting system and a receiving system, and a method of processing data are disclosed. A method of processing data of a transmitting system includes sequentially grouping N number of columns (Kc) configured of A number of enhanced data bytes having information included therein, thereby creating a frame having a size of N (rows)*Kc (columns), wherein N and A are integers, encoding the created frame, and multiplexing and transmitting enhanced data included in the encoded frame and main data.

Abstract: A method and electronic device for adaptive processing of sound data is provided. An electronic device includes a speaker, a communication module configured to communicate with an external electronic device, and a processor connected to the communication module, wherein the processor is configured to receive data from the external electronic device using the communication module, when the data corresponds to speech, decode the data using a first decoding scheme and change the quality of the decoded data using a first signal processing scheme, when the data corresponds to music, decode the data using a second decoding scheme and change the quality of the decoded data using a second signal processing scheme, and output, through the speaker, an audio signal corresponding to the data changed using the first signal processing scheme or the second signal processing scheme.

Abstract: An apparatus for decoding to obtain a reconstructed audio signal envelope includes a signal envelope reconstructor for generating the reconstructed audio signal envelope depending on one or more splitting points and an output interface for outputting the reconstructed audio signal envelope. The signal envelope reconstructor is configured to generate the reconstructed audio signal envelope such that the one or more splitting points divide the reconstructed audio signal envelope into two or more audio signal envelope portions, and to generate the reconstructed audio signal envelope such that, for each of the two or more signal envelope portions, an absolute value of its signal envelope portion value is greater than half of an absolute value of the signal envelope portion value of each of the other signal envelope portions.

Abstract: A method and system are provided for automated ontology building. The method includes creating contextual tokens from text, parsing the text into at least one parse tree, and calculating a dependency graph across the contextual tokens using the at least one parse tree. The method further includes generating concept instance candidates and parent-child relationships based on pattern matching and transformation of the at least one parse tree. The method also includes grouping concept instance candidates into concept candidates. The method additionally includes arranging the concept candidates into a tree having tree nodes and creating predicate-based relationships between the tree nodes based on patterns and predicates identified in the text. The method further includes scoring and sorting the tree nodes. The method also includes performing an analysis of the tree nodes and rebalancing the tree based on the analysis to provide an ontology based on the text.

Abstract: A method and system are provided for automated ontology building. The method includes creating contextual tokens from text, parsing the text into at least one parse tree, and calculating a dependency graph across the contextual tokens using the at least one parse tree. The method further includes generating concept instance candidates and parent-child relationships based on pattern matching and transformation of the at least one parse tree. The method also includes grouping concept instance candidates into concept candidates. The method additionally includes arranging the concept candidates into a tree having tree nodes and creating predicate-based relationships between the tree nodes based on patterns and predicates identified in the text. The method further includes scoring and sorting the tree nodes. The method also includes performing an analysis of the tree nodes and rebalancing the tree based on the analysis to provide an ontology based on the text.

Abstract: The present invention provides a method and apparatus for the production and labeling of objects in a manner suitable for the prevention and detection of counterfeiting. Thus, the system incorporates a variety of features that make unauthorized reproduction difficult. In addition, the present invention provides a system and method for providing a dynamically reconfigurable watermark, and the use of the watermark to encode a stochastically variable property of the carrier medium for self-authentication purposes.

Abstract: A system for biometrically securing business transactions uses speech recognition and voiceprint authentication to biometrically secure a transaction from a variety of client devices in a variety of media. A voiceprint authentication server receives a request from a third party requestor to authenticate a previously enrolled end user of a client device. A signature collection applet presents the user a randomly generated signature string, prompting the user to speak the string, and recording the user's as he speaks. After transmittal to the authentication server, the signature string is recognized using voice recognition software, and compared with a stored voiceprint, using voiceprint authentication software. An authentication result is reported to both user and requestor. Voiceprints are stored in a repository along with the associated user data. Enrollment is by way of a separate enrollment applet, wherein the end user provides user information and records a voiceprint, which is subsequently stored.

Abstract: An objective of the present invention is to correct a temporal envelope shape of a decoded signal with a small information volume and to reduce perceptible distortions.

Abstract: An electronic device and a method are provided. The electronic device includes an audio input module configured to receive a speech of a user as a voice input, an audio output module configured to output content corresponding to the voice input, and a processor configured to determine an output scheme of the content based on at least one of a speech rate of the speech, a volume of the speech, and a keyword included in the speech, which is obtained from an analysis of the voice input.

Abstract: Audio and video home theater ceiling projector system that optionally replaces any type of home theater video projector with the advantage of ten audio channels; the system simplifies in one main unit and two peripheral devices a cinema audio and video projection, it also allows the user to install, use, and distribute a complicated system of this type in a very simple way and to change the loudspeaker driver layout just by rotating the main unit installed at the electrical outlet box of the lamp, the invention is equipped with a bridged wall switch which allows controlling the main unit lights without interrupting its operation and the system receives the audio and video signal via wireless to decode all audio channels to emit sounds and effects in different directions to immerse the listener in a three-dimensional sound during a video playing.

Abstract: The present document relates an audio encoding and decoding system (referred to as an audio codec system). In particular, the present document relates to a transform-based audio codec system which is particularly well suited for voice encoding/decoding. A transform-based speech encoder (100, 170) configured to encode a speech signal into a bitstream is described. The encoder (100, 170) comprises a framing unit (101) configured to receive a set (132, 332) of blocks; wherein the set (132, 332) of blocks comprises a plurality of sequential blocks (131) of transform coefficients; wherein the plurality of blocks (131) is indicative of samples of the speech signal; wherein a block (131) of transform coefficients comprises a plurality of transform coefficients for a corresponding plurality of frequency bins (301).

Abstract: A method, computer program product, and system are disclosed for facilitating access by a first person to a secure region within an environment having a plurality of items, wherein the secure region is at least partly defined by an access control device. The method acquires, using at least one visual sensor disposed within the environment, first image information including the first person and the access control device. The method identifies the first person using image analysis performed on the first image information. Further, the method identifies, using image analysis a first behavior of the first person relative to the access control device. Upon determining the first behavior corresponds to a predefined visual access behavior, and the security level of the first person satisfies a predetermined threshold security level associated with the access control device, deactivating a security element to permit the first person to physically access the secure region.

Abstract: Computationally efficient accelerated D2-clustering algorithms are disclosed for clustering discrete distributions under the Wasserstein distance with improved scalability. Three first-order methods include subgradient descent method with re-parametrization, alternating direction method of multipliers (ADMM), and a modified version of Bregman ADMM. The effects of the hyper-parameters on robustness, convergence, and speed of optimization are thoroughly examined. A parallel algorithm for the modified Bregman ADMM method is tested in a multi-core environment with adequate scaling efficiency subject to hundreds of CPUs, demonstrating the effectiveness of AD2-clustering.

Abstract: Systems and methods are provided for providing voice authentication of a candidate speaker. Training data sets are accessed, where each training data set comprises data associated with a training speech sample of a speaker and a plurality of speaker metrics, where the plurality of speaker metrics include a native language of the speaker. The training data sets are used to train a neural network, where the data associated with each training speech sample is a training input to the neural network, and each of the plurality of speaker metrics is a training output to the neural network. Data associated with a speech sample is provided to the neural network to generate a vector that contains values for the plurality of speaker metrics, and the values contained in the vector are compared to values contained in a reference vector associated with a known person to determine whether the candidate speaker is the known person.

Abstract: A computer operable method is described for transforming phonemes, graphemes, and other language structures into interactive elements. The method may comprise, receiving a word, wherein the word consists of a group of phonemes; forming a group of graphemes, wherein the group of graphemes is constructed using information relating to the group of phonemes; and forming a group of manipulatives, wherein the group of manipulatives is constructed using information relating to the group of phonemes or the group of graphemes.

Abstract: Methods and arrangements in a codec for supporting bandwidth extension, BWE, of a harmonic audio signal. The method in the decoder part of the codec comprises receiving a plurality of gain values associated with a frequency band b and a number of adjacent frequency bands of band b. The method further comprises determining whether a reconstructed corresponding frequency band b? comprises a spectral peak. When the band b? comprises a spectral peak, a gain value associated with the band b? is set to a first value based on the received plurality of gain values; and otherwise the gain value is set to a second value based on the received plurality of gain values. The suggested technology enables bringing gain values into agreement with peak positions in a bandwidth extended frequency region.

Abstract: A computer numerical processing method for encoding and decoding audio information for use in conjunction with human hearing is described. The method comprises approximating an eigenfunction equation representing a model of human hearing, calculating the approximation to each of a plurality of eigenfunctions from at least one aspect of the eigenfunction equation, and storing the approximation to each of a plurality of eigenfunctions for use in encoding and decoding. The approximation to each of a plurality of eigenfunctions represents a perception-oriented basis functions for mathematically representing audio information in a Hilbert-space representation of an audio signal space. The model of human hearing can include a bandpass operation with a bandwidth having the frequency range of human hearing and a time-limiting operation approximating the time duration correlation window of human hearing.

Abstract: Methods and apparatus are provided for coding and decoding a digital audio signal. Decoding includes: decoding according to an inverse transform decoding of a previous frame of samples of the digital signal, which is received and coded according to a transform coding; and decoding according to a predictive decoding of a current frame of samples of the digital signal, which is received and coded according to a predictive coding. The predictive decoding of the current frame is a transition predictive decoding which does not use any adaptive dictionary arising from the previous frame. At least one state of the predictive decoding is reinitialized to a predetermined default value, and an add-overlap step combines a signal segment synthesized by predictive decoding of the current frame and a signal segment synthesized by inverse transform decoding, corresponding to a stored segment of the decoding of the previous frame.