Abstract: The disclosed solution includes a method for dynamically switching modalities based upon inferred conditions in a dialogue session involving a speech application. The method establishes a dialogue session between a user and the speech application. During the dialogue session, the user interacts using an original modality and a second modality. The speech application interacts using a speech modality only. A set of conditions indicative of interaction problems using the original modality can be inferred. Responsive to the inferring step, the original modality can be changed to the second modality. A modality transition to the second modality can be transparent the speech application and can occur without interrupting the dialogue session. The original modality and the second modality can be different modalities; one including a text exchange modality and another including a speech modality.

Abstract: A method for building and contracting a linguistic dictionary, the linguistic dictionary comprising a list of surface forms and a list of normalized forms, each normalized form being associated with a surface form, the method comprising the steps of: comparing each character of a surface form with each character of the surface form's normalized form; in response to the comparing step, determining an edit operation for each character compared; and generating a transform code from the set of the edit operations in order to transform the surface form to its normalized form.

Abstract: The present invention adopts the fundamental architecture of a statistical machine translation system which utilizes statistical models learned from the training data and does not require expert knowledge for rule-based machine translation systems. Out of the training parallel data, a certain amount of sentence pairs are selected for manual alignment. These sentences are aligned at the phrase level instead of at the word level. Depending on the size of the training data, the optimal amount for manual alignment may vary. The alignment is done using an alignment tool with a graphical user interface which is convenient and intuitive to the users. Manually aligned data are then utilized to improve the automatic word alignment component. Model combination methods are also introduced to improve the accuracy and the coverage of statistical models for the task of statistical machine translation.

Abstract: A computer program product for use with dictated medical patient information resides on a computer-readable medium and comprises computer-readable instructions for causing a computer to analyze the dictated information, identify likely confidential information in the dictated medical patient information, and treat the likely confidential information disparately from likely non-confidential information in the dictated medical patient information.

Abstract: A particular technique for classifying a name includes accessing a name; dividing the name into a series of first n-grams; forming multiple concatenated second n-grams by concatenating pairs of the first n-grams; for each of multiple groups, for each of the second n-grams, determining the term frequency-group frequency score; for each of the multiple groups, summing up the term frequency-group frequency scores for each second n-gram for that group; and determining a likelihood that the name belongs to one group of the multiple groups based on the summed scores, wherein a largest summed score indicates a greater likelihood that the name belongs to the one group.

Abstract: A device for suppressing ambient sounds from speech received by a microphone array is provided. One embodiment of the device comprises a microphone array, a processor, an analog-to-digital converter, and memory comprising instructions stored therein that are executable by the processor.

Abstract: In a voice processing device, a male voice index calculator calculates a male voice index indicating a similarity of the input sound relative to a male speaker sound model. A female voice index calculator calculates a female voice index indicating a similarity of the input sound relative to a female speaker sound model. A first discriminator discriminates the input sound between a non-human-voice sound and a human voice sound which may be either of the male voice sound or the female voice sound. A second discriminator discriminates the input sound between the male voice sound and the female voice sound based on the male voice index and the female voice index in case that the first discriminator discriminates the human voice sound.

Abstract: Systems, methods, and apparatuses including computer program products are provided for encoding and using a language model. In one implementation, a method is provided. The method includes generating a compact language model, including receiving a collection of n-grams, each n-gram having one or more associated parameter values, determining a fingerprint for each n-gram of the collection of n-grams, identifying locations in an array for each n-gram using a plurality of hash functions, and encoding the one or more parameter values associated with each n-gram in the identified array locations as a function of corresponding array values and the fingerprint for the n-gram.

Abstract: During operation an input signal to be coded is received and coded to produce a coded audio signal. The coded audio signal is then scaled with a plurality of gain values to produce a plurality of scaled coded audio signals, each having an associated gain value and a plurality of error values are determined existing between the input signal and each of the plurality of scaled coded audio signals. A gain value is then chosen that is associated with a scaled coded audio signal resulting in a low error value existing between the input signal and the scaled coded audio signal. Finally, the low error value is transmitted along with the gain value as part of an enhancement layer to the coded audio signal.

Abstract: The present invention discloses a method for performing a frame erasure concealment to a higher-band signal, including: calculating a periodic intensity of a higher-band signal with respect to a lower-band signal; judging whether the periodic intensity of the higher-band signal is higher than or equal to a preconfigured threshold; if the periodic intensity of the higher-band signal is higher than or equal to the preconfigured threshold, using a pitch period repetition method to perform the frame erasure concealment to the higher-band signal of a current lost frame; and if the periodic intensity of the higher-band signal is lower than the preconfigured threshold, using a previous frame data repetition method to perform the frame erasure concealment to the higher-band signal of the current lost frame. The present invention further discloses a device for performing a frame erasure concealment to a higher-band signal and a speech decoder. The problem that the quality of the voice signal is lowered is avoided.

Abstract: Generic and specific C-to-E binaural cue coding (BCC) schemes are described, including those in which one or more of the input channels are transmitted as unmodified channels that are not downmixed at the BCC encoder and not upmixed at the BCC decoder. The specific BCC schemes described include 5-to-2, 6-to-5, 7-to-5, 6.1-to-5.1, 7.1-to-5.1, and 6.2-to-5.1, where “0.1” indicates a single low-frequency effects (LFE) channel and “0.2” indicates two LFE channels.

Abstract: A communications system and method, comprising means for receiving a speech input from a user; converting the received speech input to a text representation thereof; communicating the text representation remotely from the user; and at least one of reproducing the speech input and displaying the text representation remotely from the user; and converting the text representation into speech remotely from the user.

Abstract: A home evergreen index and frontier evergreen indexes are maintained. The indexes cover topically-limited subject areas, which include digital information. Each index defines a hierarchy of topics. Each index further matches a topic model to the topic hierarchy's topics. Each topic model includes a pattern evaluable against the digital information, which identifies such digital information matching the topic model's topic. At least one frontier evergreen index that includes topics that are at least partially distinct from the topics included in the home evergreen index is identified. Vetted assessments for articles identified by the at least one frontier evergreen index are obtained. The articles corresponding to the vetted assessments that are favorable are selected. The patterns that include the topic models of the home evergreen index are matched against the selected articles of the digital information. The selected articles of the digital information that were matched are provided on a display.

Abstract: Disclosed herein is a computer implemented method and system for word sense disambiguation in a natural language sentence. The natural language sentence is parsed for identifying possible parts of speech for each term and identifying possible phrase structures. Terms comprising one or more linguistic roles are identified. The possible sense combinations for the terms with linguistic roles are identified. Emergent categories are applied to identify possible valid senses for each of the terms with identified linguistic roles. Linguistic role pairs are identified from among the terms identified with linguistic roles. The correspondence functions with the correspondence function types matching the identified linguistic role pairs are identified from an emergent categories database. The pair-wise senses for each term are compared with the identified linguistic roles to identify the possible sense combinations.

Abstract: A unified filter bank for performing signal conversions may include an interface that receives signal conversion commands in relation to multiple types of compressed audio bitstreams. The unified filter bank may also include a reconfigurable transform component that performs a transform as part of signal conversion for the multiple types of compressed audio bitstreams. The unified filter bank may also include complementary modules that perform complementary processing as part of the signal conversion for the multiple types of compressed audio bitstreams. The unified filter bank may also include an interface command controller that controls the configuration of the reconfigurable transform component and the complementary modules.

Abstract: A speech processing apparatus, medium, and method recognizing speech and responding to the speech. The speech processing apparatus may includes an entity extracting unit which extracts entity information and an upper entity corresponding to the entity information from input speech, a focus determination unit which determines a focus using the extracted entity information requiring a response, a mapping unit which maps lower entity corresponding to the focus with the extracted entity information, and a recognition unit which recognizes a result of arranging the extracted entity information according to semantic association among the lower entities as the input speech. Thus, the speech processing apparatus can accurately recognize grammatically correct speech as well as grammatically incorrect speech and then respond to the speech.

Abstract: In some embodiments, a reading device is provided with a shortcut read mode in which a user can instruct the reading device of the type of document (e.g., invoice, package label, newspaper, etc.) that is to be read so that the device can more efficiently find and read back to the user desired (target) information from the document.

Abstract: The present research can decrease the amount of computation and enhance speech quality by using a global pulse replacement method in a fixed codebook search. The fixed codebook search method in a speech encoder based upon global pulse replacement, includes the steps of: (a) computing absolute values of the pulse-position likelihood-estimator vectors; (b) temporarily obtaining a codebook vector; (c) computing a mathematical equation by replacing a pulse; (d) determining whether a value computed based upon the mathematical equation is increased after pulse replacement; (e) obtaining a new codebook vector by replacing the pulse; and (f) maintaining a previous codebook vector.

Abstract: A voice communication apparatus includes a communication portion that receives a plurality of frames including at least a first frame having first voice data and a second frame having second voice data subsequent to the first frame, the first voice data and the second voice data being encoded by a predetermined encoding system, a decoding portion that decodes the first voice data and the second voice data received by the communication portion, a buffer that retains the first voice data and the second voice data decoded by the decoding portion, a calculation portion that calculates an amplitude envelope based on the first voice data decoded by the decoding portion, and a controlling portion that judges whether or not the second voice data decoded by the decoding portion exceeds the amplitude envelope and corrects the second voice data that exceeds the amplitude envelope.

Abstract: A text-to-speech (TTS) system implemented in an automotive vehicle is dynamically tuned to increase intelligibility over a wide variety of vehicle operating states and environmental conditions by tuning characteristics of the synthesized voice in response to measured operating states. To decrease distractions to an operator of the vehicle, an embodiment of the invention prevents updates to the synthesized voice character from taking effect while a message phrase is being played. Instead, voice characteristics are updated only during natural phrase breaks. In another embodiment of the invention, a damping filter is applied to calculated changes in voice characteristics to prevent excessively rapid changes from being applied, reducing the likelihood of distracting the vehicle operator. In another embodiment of the invention, both phrase-break detectors and damping filters are employed.