Abstract: Systems, devices, media, and methods are presented for converting sounds in an audio stream. The systems and methods receive an audio conversion request initiating conversion of one or more sound characteristics of an audio stream from a first state to a second state. The systems and methods access an audio conversion model associated with an audio signature for the second state. The audio stream is converted based on the audio conversion model and an audio construct is compiled from the converted audio stream and a base audio segment. The compiled audio construct is presented at a client device.

Abstract: Embodiments are disclosed for transliteration based on a machine translation model training pipeline. A method according to some embodiments includes steps of: training a probabilistic model for transliteration from a first script system to a second script system using a machine translation model training pipeline; segmenting, using the probabilistic model, an input string in the first script system into phonemes that correspond to characters in the second script system; converting the phonemes in the first script system into the characters in the second script system, the characters forming a word or a word prefix in the second script system; and outputting the word or the word prefix in the second script system.

Abstract: A method and apparatus for performing coding and decoding for high-frequency bandwidth extension. The decoding apparatus may include: a mode checking unit to check mode information of each of frames included in a bitstream; a first core decoding unit to perform code excited linear prediction (CELP) decoding on a CELP coded frame, when a core coding mode of a low-frequency signal indicates a CELP coding mode; a first extension decoding unit to generate a decoded signal of a high-frequency band by using at least one of a result of the performing the CELP decoding and an excitation signal of the low-frequency signal; a second core decoding unit to perform audio decoding on an audio coded frame, when the core coding mode indicates an audio coding mode; and a second extension decoding unit to generate a decoded signal of the high-frequency band by performing frequency-domain (FD) extension decoding.

Abstract: The disclosed system continuously refines a model used by an Automatic Speech Recognition (ASR) system to enable fast and accurate transcriptions of detected speech activity. The ASR system analyzes speech activity to generate text transcriptions and associated metrics (such as minimum Bayes risk and/or perplexity) that correspond to the quality of or confidence in each generated transcription. The system employs a filtering process to select certain text transcriptions based in part on one or more associated quality metrics. In addition, the system corrects for known systemic errors within the ASR system and provides a mechanism for manual review and correction of transcriptions. The system selects a subset of transcriptions based on factors including confidence score, and uses the selected subset of transcriptions to re-train the ASR model. By continuously retraining the ASR model, the system is able to provide ever faster and more accurate text transcriptions of detected speech activity.

Abstract: A system and method for tokenizing sentences in transcriptions of electronic communication audio files includes one or more electronic data storage systems coupled in electronic communication with an audio receiver. A collection of audio files is stored in the electronic data storage system(s). The audio receiver is configured to receive one or more audio files from the electronic data storage systems. The system also includes a speech-to-text processing engine configured to transcribe the one or more audio files received by the audio receiver to produce input transcriptions. In addition, the system includes a call tokenization engine that stores a list of candidate expressions and expression pairs that are indicative of turn taking and change of context in phone conversations.

Abstract: A method performed in an audio decoder for decoding M encoded audio channels representing N audio channels is disclosed. The method includes receiving a bitstream containing the M encoded audio channels and a set of spatial parameters, decoding the M encoded audio channels, and extracting the set of spatial parameters from the bitstream. The method also includes analyzing the M audio channels to detect a location of a transient, decorrelating the M audio channels, and deriving N audio channels from the M audio channels and the set of spatial parameters. A first decorrelation technique is applied to a first subset of each audio channel and a second decorrelation technique is applied to a second subset of each audio channel. The first decorrelation technique represents a first mode of operation of a decorrelator, and the second decorrelation technique represents a second mode of operation of the decorrelator.

Abstract: A device may receive a set of audio data files corresponding to a set of calls, wherein the set of audio data files includes digital representations of one or more segments of respective calls of the set of calls, and wherein the set of calls includes audio data relating to a particular industry. The device may receive a set of transcripts corresponding to the set of audio data files. The device may determine a plurality of text-audio pairs within the set of calls, wherein a text-audio pair, of the plurality of text-audio pairs, comprises: a digital representation of a segment a call of the set of calls, and a corresponding excerpt of text from the set of transcripts. The device may train, using a machine learning process, an industry-specific text-to-speech model, tailored for the particular industry, based on the plurality of text-audio pairs.

Abstract: The present disclosure discloses an audio data processing performed by a computing device. The computing device obtains song information of a song, the song information comprising an accompaniment file, a lyric file, and a music score file that correspond to the song and then determines a predefined portion of the song and music score information corresponding to the predefined portion according to the song information. After receiving audio data that is input by a user, the computing device determines time information of each word in the audio data and then processes the audio data according to the time information of each word in the audio data and the music score information of the predefined portion of the song. Finally, the computing device obtains mixed audio data by mixing the processed audio data and the accompaniment file.

Abstract: A decoder generates decoded signals based on quantized signals. The decoder includes an inverse quantizer and a predictor circuit. The quantized signals are generated in an encoder by low-pass filtering an input signal and encoding the filtered signal using adaptive differential pulse code modulation. The predictor circuit has filter coefficients based on a frequency response of the low-pass filter used to filter the input signal.

Abstract: An encoder generates quantized signal words based on a difference signal. The encoder includes an adaptive quantizer. A step size applied by the adaptive quantizer is generated in a feedback loop and based on a loading factor and quantized signal words generated by the adaptive quantizer. The encoder includes coding circuitry which generates code words based on quantized signal words generated by the adaptive quantizer. The coding circuitry generates an escape code in response to a quantized signal word not being associated with a corresponding coding code word.

Abstract: According to one embodiment, an interactive electronic device control system includes an interactive electronic device and an authentication manager. The interactive electronic device sends input voice data to a voice recognition service server and receives a response generated based at least in part on a result of recognizing the voice data by the voice recognition service server and the authentication manager authenticates a mobile terminal connected to the interactive electronic device via a network as an authenticated user.

Abstract: Disclosed example apparatus select a first frequency from a set of frequencies based on a first symbol in a code, and select a first block size based on the first symbol and the code, a combination of the first block size and the first frequency to represent the first symbol. Disclosed example apparatus also synthesize a code frequency according to the first block size and the first frequency. Disclosed example apparatus further embed the code frequency in a first block of input audio samples of the audio having the first block size to form a block of encoded audio samples encoded with the first symbol, the code frequency and the first block of input audio samples to overlap in time.

Abstract: Various embodiments may include systems and non-transitory computer-readable media for assigning three-dimensional spatial data to sounds and audio files. In one embodiment, a method can include receiving at least one audio signal, receiving sonic spatial data via a foot pedal, associating the sonic spatial data with the at least one audio signal, associating the at least one audio signal and sonic spatial data with a time code generated by a video camera, storing the sonic spatial data, the at least one audio signal, and time code in an encoded sound file, and directing playback of the encoded sound file.

Abstract: Systems and methods of diarization using linguistic labeling include receiving a set of diarized textual transcripts. A least one heuristic is automatedly applied to the diarized textual transcripts to select transcripts likely to be associated with an identified group of speakers. The selected transcripts are analyzed to create at least one linguistic model. The linguistic model is applied to transcripted audio data to label a portion of the transcripted audio data as having been spoken by the identified group of speakers. Still further embodiments of diarization using linguistic labeling may serve to label agent speech and customer speech in a recorded and transcripted customer service interaction.

Abstract: The intelligent automated assistant system engages with the user in an integrated, conversational manner using natural language dialog, and invokes external services when appropriate to obtain information or perform various actions. The system can be implemented using any of a number of different platforms, such as the web, email, smartphone, and the like, or any combination thereof. In one embodiment, the system is based on sets of interrelated domains and tasks, and employs additional functionally powered by external services with which the system can interact.

Abstract: An encoder includes a low-pass filter to filter input audio signals. The low-pass filter has fixed filter coefficients. The encoder generates quantized signals based on a difference signal. The encoder includes an adaptive quantizer and a decoder to generate feedback signals. The decoder has an inverse quantizer and a predictor. The predictor has fixed control parameters which are based on a frequency response of the low-pass filter. The predictor may include a finite impulse response filter having fixed filter coefficients. The decoder may include an adaptive noise shaping filter coupled between the low-pass filter and the encoder. The adaptive noise shaping filter flattens signals within a frequency spectrum corresponding to a frequency spectrum of the low-pass filter.

Abstract: Systems and methods of diarization using linguistic labeling include receiving a set of diarized textual transcripts. A least one heuristic is automatedly applied to the diarized textual transcripts to select transcripts likely to be associated with an identified group of speakers. The selected transcripts are analyzed to create at least one linguistic model. The linguistic model is applied to transcripted audio data to label a portion of the transcripted audio data as having been spoken by the identified group of speakers. Still further embodiments of diarization using linguistic labeling may serve to label agent speech and customer speech in a recorded and transcripted customer service interaction.

Abstract: Systems and method of diarization of audio files use an acoustic voiceprint model. A plurality of audio files are analyzed to arrive at an acoustic voiceprint model associated to an identified speaker. Metadata associate with an audio file is used to select an acoustic voiceprint model. The selected acoustic voiceprint model is applied in a diarization to identify audio data of the identified speaker.

Abstract: Systems and methods of automated ontology development include a corpus of communication data. The corpus of communication data includes communication data from a plurality of interactions and is processed. A plurality of terms are extracted from the corpus. Each term of the plurality is a plurality of words that identify a single concept within the corpus. An ontology is automatedly generated from the extracted terms.

Abstract: A soft decision audio decoding system for preserving audio continuity in a digital wireless audio receiver is provided that deduces the likelihood of errors in a received digital signal, based on generated hard bits and soft bits. The soft bits may be utilized by a soft audio decoder to determine whether the digital signal should be decoded or muted. The soft bits may be generated based on the detected point and a detected noise power, or by using a soft-output Viterbi algorithm. The value of the soft bits may indicate confidence in the strength of the hard bit generation. The soft decision audio decoding system may infer errors and decode perceptually acceptable audio without requiring error detection, as in conventional systems, as well as have low latency and improved granularity.