Abstract: Embodiments may be a standalone module or part of mobile devices, desktop computers, servers, stereo systems, or any other systems that might benefit from condensed audio presentations of item structures such as lists or tables. Embodiments may comprise logic such as hardware and/or code to adjust the temporal characteristics of items comprising words. The items maybe included in a structure such as a text listing or table, an audio listing or table, or a combination thereof, or may be individual words or phrases. For instance, embodiments may comprise a keyword extractor to extract keywords from the items and an abbreviations generator to generate abbreviations based upon the keywords. Further embodiments may comprise a text-to-speech generator to generate audible items based upon the abbreviations to render to a user while traversing the item structure.

Abstract: A handheld electronic device includes a reduced QWERTY keyboard and is enabled with disambiguation software that is operable to disambiguate text input. In addition to identifying and outputting representations of language objects that are stored in the memory and that correspond with a text input, the device provides a learning function which facilitates providing proposed corrected output by the device in certain circumstances of erroneous input.

Abstract: Disclosed is a method for training a spoken dialog service component from website data. Spoken dialog service components typically include an automatic speech recognition module, a language understanding module, a dialog management module, a language generation module and a text-to-speech module. The method includes selecting anchor texts within a website based on a term density, weighting those anchor texts based on a percent of salient words to total words, and incorporating the weighted anchor texts into a live spoken dialog interface, the weights determining a level of incorporation into the live spoken dialog interface.

Abstract: According to one embodiment, in a speech processing device, an extractor windows a part of the speech signal and extracts a partial waveform. A calculator performs frequency analysis of the partial waveform to calculate a frequency spectrum. An estimator generates an artificial waveform that is a waveform according to an interval between the pitch marks for each harmonic component having a frequency that is a predetermined multiple of a fundamental frequency of the speech signal and estimates harmonic spectral features representing characteristics of the frequency spectrum of the harmonic component from each of the artificial waveforms. A separator separates the partial waveform into a periodic component produced from periodic vocal-fold vibration as an acoustic source and an aperiodic component produced from aperiodic acoustic sources other than the vocal-fold vibration by using the respective harmonic spectral features and the frequency spectrum of the partial waveform.

Abstract: The voice-activated motorized remote control device relates to a motorized remote control, for consumer electronics, with wheels that will drive itself over to an end-user in response to the sound of his or her voice. The voice-activated motorized remote control device will also feature a beeping noise during normal operation and also a high pitched beeping noise that will alert the user if the device has become stuck against an object. The voice-activated motorized remote control device comprises a plurality of wheels; a remote control; a plurality of axles; a microphone; a speaker, a central processing unit with a voice recognition system, and a drive motor.

Abstract: A method, system and computer program for encoding speech according to a source-filter model. The method comprises deriving a spectral envelope signal representative of a modelled filter and a first remaining signal representative of a modelled source signal, and deriving a second remaining signal from the first remaining signal by, at intervals during the encoding: exploiting a correlation between approximately periodic portions in the first remaining signal to generate a predicted version of a later portion from a stored version of an earlier portion, and using the predicted-version of the later portion to remove an effect of said periodicity from the first remaining signal. The method further comprises, once every number of intervals, transforming the stored version of the earlier portion of the first remaining signal prior to generating the predicted version of the respective later portion.

Abstract: A connected text data system for efficiently and accurately translating connected text. The connected text data system includes inputting or receiving connected text, transmitting the connected text to a text iterator, scanning the connected text, identifying a plurality of words in the connected text comprising a coordinate logic to help parse connected text matches into separated text by invalidating words with overlapping coordinates, and translating the connected text to separated text by adding a space between each of the plurality of words.

Abstract: A speech recognition apparatus includes a speech recognition dictionary and a speech recognition unit. The speech recognition dictionary includes comparison data used to recognize a voice input. The speech recognition unit is adapted to calculate the score for each comparison data by comparing voice input data generated based on the voice input with each comparison data, recognize the voice input based on the score, and produce the recognition result of the voice input. The speech recognition apparatus further includes data indicating score weights associated with particular comparison data, used to weight the scores calculated for the particular comparison data. After the score is calculated for each comparison data, the score weights are added to the scores of the particular comparison data, and the voice input is recognized based on total scores including the added score weights.

Abstract: In audio processing for an audio or video conference, a terminal receives audio packets having transform coefficients for reconstructing an audio signal that has undergone transform coding. When receiving the packets, the terminal determines whether there are any missing packets and interpolates transform coefficients from the preceding and following good frames. To interpolate the missing coefficients, the terminal weights first coefficients from the preceding good frame with a first weighting, weights second coefficients from the following good frame with a second weighting, and sums these weighted coefficients together for insertion into the missing packets. The weightings can be based on the audio frequency and/or the number of missing packets involved. From this interpolation, the terminal produces an output audio signal by inverse transforming the coefficients.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for enhancing speech recognition accuracy. In one aspect, a method includes receiving an audio signal that corresponds to an utterance recorded by a mobile device, determining a geographic location associated with the mobile device, identifying a set of geotagged audio signals that correspond to environmental audio associated with the geographic location, weighting each geotagged audio signal of the set of geotagged audio signals based on metadata associated with the respective geotagged audio signal, and using the set of weighted geotagged audio signals to perform noise compensation on the audio signal that corresponds to the utterance.

Abstract: An architecture and framework for speech/noise classification of an audio signal using multiple features with multiple input channels (e.g., microphones) are provided. The architecture may be implemented with noise suppression in a multi-channel environment where noise suppression is based on an estimation of the noise spectrum. The noise spectrum is estimated using a model that classifies each time/frame and frequency component of a signal as speech or noise by applying a speech/noise probability function. The speech/noise probability function estimates a speech/noise probability for each frequency and time bin. A speech/noise classification estimate is obtained by fusing (e.g., combining) data across different input channels using a layered network model.

Abstract: A system is described for providing biofeedback to hearing-impaired persons as to the degree of nasalization of vowel-like sounds in their speech, in order to monitor their own nasality and thus correct inappropriate nasalization. In a preferred embodiment, this feedback uses tactile vibration, with the vibration frequency or amplitude reflecting the nasalance of the speech.

Abstract: Techniques for generating and providing phrases are described herein. These techniques may include analyzing one or more sources to generate a first corpus of phrases, each of the phrases for use as an identifier and/or for association with a user. The first corpus of phrases may be generated based on characteristics associated with phrases. Once a first corpus of phrases has been generated, these phrases may be filtered to define a second corpus of phrases. Phrases of this second corpus may then be suggested to one or more users. In some instances, the phrases suggested to a particular user are personalized to the user based on information previously known about the user or based on information provided by the user.

Abstract: One or more embodiments include a speech recognition learning system for improved speech recognition. The learning system may include a speech optimizing system. The optimizing system may receive a first stimulus data package including spoken utterances having at least one phoneme, and contextual information. A number of result data packages may be retrieved which include stored spoken utterances and contextual information. A determination may be made as to whether the first stimulus data package requires improvement. A second stimulus data package may be generated based on the determination. A number of speech recognition implementation rules for implementing the second stimulus data package may be received. The rules may be associated with the contextual information. A determination may be made as to whether the second stimulus data package requires further improvement.

Abstract: A method for automatically detecting sentiments in an audio signal of an interaction held in a call center, including, receiving the audio signal from a logging and capturing unit. Performing audio analysis on the audio signal to obtain text spoken within the interaction. Segmenting the text into context units according to acoustic information acquired from the audio signal to identify units of speech bound by non-speech segments, wherein each context unit includes one or more words. Extracting a sentiment candidate context unit from the context units using a phonetic based search. Extracting linguistic features from the text of the sentiment candidate context unit and acoustic features from a segment of the audio signal associated with the sentiment candidate context unit. Determining in accordance with the linguistic features and acoustic features whether the sentiment candidate context unit is valid or erroneous, and determining sentiment polarity and intensity.

Abstract: A visual toolkit for prioritizing speech transcription is provided. The toolkit can include a logger (102) for capturing information from a speech recognition system, a processor (104) for determining an accuracy rating of the information, and a visual display (106) for categorizing the information and prioritizing a transcription of the information based on the accuracy rating. The prioritizing identifies spoken utterances having a transcription priority in view of the recognized result. The visual display can include a transcription category (156) having a modifiable textbox entry with a text entry initially corresponding to a text of the recognized result, and an accept button (157) for validating a transcription of the recognized result. The categories can be automatically ranked by the accuracy rating in an ordered priority for increasing an efficiency of transcription.

Abstract: A speech processing apparatus, including a segmenting unit to divide a fundamental frequency signal of a speech signal corresponding to an input text into pitch segments, based on an alignment between samples of at least one given linguistic level included in the input text and the speech signal. Character strings of the input text are divided into the samples based on each linguistic level. A parameterizing unit generates a parametric representation of the pitch segments using a predetermined invertible operator and generates a group of first parameters in correspondence with each linguistic level. A descriptor generating unit generates, for each linguistic level, a descriptor that includes a set of features describing each sample in the input text and a model learning unit classifies the first parameters of each linguistic level of all speech signals in a memory into clusters based on the descriptor corresponding to the linguistic level.

Abstract: Assisting apparatus that assists in a task for generating task-completed data from original data. The assisting apparatus includes: a conversion unit that converts the original data into converted data in order to keep a content of the original data secret from an operator in charge of the task; an interface unit that provides the converted data to the operator, and thereby allows the operator to perform the task by using the converted data and thus to generate a task-completed version of the converted data; and an inverse conversion unit that generates the task-completed data by performing inverse conversion of the task-completed version of the converted data.

Abstract: A method for generating a frequency warping function comprising preparing the training speech of a source and a target speaker; performing frame alignment on the training speech of the speakers; selecting aligned frames from the frame-aligned training speech of the speakers; extracting corresponding sets of formant parameters from the selected aligned frames; and generating a frequency warping function based on the corresponding sets of formant parameters. The step of selecting aligned frames preferably selects a pair of aligned frames in the middle of the same or similar frame-aligned phonemes with the same or similar contexts in the speech of the source speaker and target speaker. The step of generating a frequency warping function preferably uses the various pairs of corresponding formant parameters in the corresponding sets of formant parameters as key positions in a piecewise linear frequency warping function to generate the frequency warping function.

Abstract: A method of encoding speech, the method comprising: receiving a signal representative of speech to be encoded; at each of a plurality of intervals during the encoding, determining a pitch lag between portions of the signal having a degree of repetition; selecting for a set of said intervals a pitch lag vector from a pitch lag codebook of such vectors, each pitch lag vector comprising a set of offsets corresponding to the offset between the pitch lag determined for each said interval and an average pitch lag for said set of intervals, and transmitting an indication of the selected vector and said average over a transmission medium as part of the encoded signal representative of said speech.