In accordance with the present invention, a method for providing generation of speech includes the steps of providing input to be acoustically produced, comparing the input to training data or application specific splice files to identify one of words and word sequences corresponding to the input for constructing a phone sequence, using a search algorithm to identify a segment sequence to construct output speech according to the phone sequence and concatenating segments and modifying characteristics of the segments to be substantially equal to requested characteristics. Application specific data is advantageously used to make pertinent information available to synthesize both the phone sequence and the output speech. Also, described is a system for performing operations in accordance with the disclosure.

Phrase splicing and variable substitution using a trainable speech synthesizer

A system for contextualizing disorganized content (2a) captured from any live session (1) using external devices 30-xd to first detect & record 30-1 session activities (1d) being conducted by session attendees (1c). Activities (1d) become normalized tracked object data 2-otd for differentiation 30-2 into normalized session marks 3-pm denoting thresholded activity (1d) changes. Normalized marks 3-pm are integrated 30-3 into normalized events 4-pe using a "mark creates, start or stops event" model. Events 4-pe may be synthesized 30-4 via waveform convolution forming new combined events 40se, or used as containers to summarize the occurrences of marks 3-pm or other events 4-pe, the results of which create new summary marks 3-sm. Calculation marks 3-tm may also be synthesized 30-4 for sampling various session data at various session times. During content expression 30- 5 events 4-pe and 4-se can be automatically named and foldered creating index (2i) and organized content (2b).

Session automated recording together with rules based indexing, analysis and expression of content

A method for generating a language component vocabulary VC for a speech recognition system having a language vocabulary V of a plurality of word forms is disclosed. The method includes: partitioning the language vocabulary V into subsets of word forms based on frequencies of occurrence of the respective word forms; and in at least one of the subsets, splitting word forms having frequencies less than a threshold to thereby generate word form components. Also disclosed is a method for use in speech recognition including: splitting an acoustic vocabulary comprising baseforms into baseform components and storing the baseform components; and, performing sound to spelling mapping on the baseform components so as to generate a baseform components to word parts table for use in subsequent decoding of speech. A method for decoding a speech utterance using language model components and acoustic components, includes the steps of: generating from the utterance a stack of baseform component paths; concatenating baseform components in a path to generate concatenated baseforms, when the concatenated baseform components correspond to a baseform found in an acoustic vocabulary; mapping the concatenated baseforms into words; computing language model (LM) scores associated with the words using a language model, and performing further decoding of the utterance based thereupon.

System and methods for acoustic and language modeling for automatic speech recognition with large vocabularies

The invention concerns a method and system for creating a named entity language model. The method may include recognizing input communications from a training corpus, parsing the training corpus, tagging the parsed training corpus, aligning the recognized training corpus with the tagged training corpus, and creating a named entity language model from the aligned corpus.

Method and system for creating a named entity language model

The invention concerns a method and system for detecting morphemes in a user's communication. The method may include recognizing a lattice of phone strings from the user's input communication, the lattice representing a distribution over the phone strings, and detecting morphemes in the user's input communication using the lattice. The morphemes may be acoustic and/or non-acoustic. The morphemes may represent any unit or sub-unit of communication including phones, diphones, phone-phrases, syllables, grammars, words, gestures, tablet strokes, body movements, mouse clicks, etc. The training speech may be verbal, non-verbal, a combination of verbal and non-verbal, or multimodal.

Method and system for automatically detecting morphemes in a task classification system using lattices

In a method of automatically recognizing data which comprises sequential data units represented as sequential tokens grouped into one or more items, known items are stored as respective finite state sequence models. Each state corresponds to a token and the models which have common prefix states are organized in a tree structure such that suffix states comprise branches from common prefix states and there are a plurality of tree structures each having a different prefix state. Each sequential data unit is compared with stored reference data units identified by reference tokens to generate scores indicating the similarity of the data units to reference data units. An accumulated score for the final state in the models is determined by steps of (a) sequentially calculating the accumulated score for a model to reach the final state comprising a leaf in the tree, (b) identifying the closest branch to the leaf corresponding to a next model for which an accumulated score for the final stage has not yet been calculated, and (c) accumulating the score from the identified closest branch for the next model to the final state. These steps are repeated for the branches of the trees. The item corresponding to a model having the highest accumulated score is recognized as the model best matching the data.

Recognition of sequential data using finite state sequence models organized in a tree structure

Audio source data is subjected to a pre-emphasis step (302) to perform gross decorrelation, followed by an adaptive linear prediction (306) to perform further decorrelation. A transform is performed on the residual of the linear prediction, to obtain transform coefficients representing the residual in the frequency domain. A number of tonal components are identified (310), subtracted from the transform coefficients and encoded by vector quantization. The transform coefficients are then grouped into sub-bands, and each sub-band encoded in the frequency domain by vector quantization. The sub-bands are of uniform width on an auditory scale, so that each vector may comprise a different number of transform coefficients.

Anthony John Robinson

Papers

Recognition of sequential data using finite state sequence models organized in a tree structure

Low bit rate audio coder and decoder operating in a transform domain using vector quantization

Low bit rate audio coder and decoder