Telia Ab

Bertil Lyberg

Mats Wiren

A device for prosody generation at visual synthesis. A number of half-syllables are stored together with registered movement patterns in a face. When synthesizing speech, a number of half-syllables are put together into words and sentences. The words and sentences are given a stress and pattern of intonation corresponding to the intended language. In the face, a number of points and their movement patterns are further registered. In connection with the generation of words and sentences, the movement patterns of the different points are amplified depending on a given stress and sentence intonation. The given movement patterns are after that applied to a model, which is applied to a real face at which a life-like animation is obtained, at for instance a translation of a person's speech in a first language to a second language.

Device and method for prosody generation at visual synthesis

A method of speech synthesis for reproduction of facial movements of a person with synchronized speech synthesis. The speech is put together of polyphones that are taken from a database. A databank is established containing polyphones with the polyphones belonging to facial movement patterns of the face of a first person. Polyphones from a second person are also registered and stored in the database. The duration of sound segments in corresponding polyphones in the databank and the database are compared and the facial movements in the databank are modified in accordance with the deviation. The modified facial movements are stored in the databank and related to the polyphone in question. These registered polyphones are then utilized to put together words and sentences at the same time as corresponding facial movements build up a face model from the modified facial movements in the databank.

Method of speech synthesis

A device and method in which polyphones of speech of a first language is received and stored as well as a movement pattern in a person's face and/or body is registered. The registration of the movement pattern is made by measuring movement at a number of measuring points in the face/body of the speaker, where the measurements are made at the same time that the polyphones are registered. In connection with translation of a person's speech from one language into another, the polyphones and corresponding movement patterns in the face are linked up to a movement model in the face. A picture image of a face of the real person is after that pasted over the model, at which one to the language corresponding movement pattern is obtained. The invention consequently gives the impression that the person really speaks the language in question.

Device and method for dubbing an audio-visual presentation which

A method and device for determining quality of speech. The speech to be evaluated is listened to by a person who reproduces the speech. The end of vowel sounds in the produced and reproduced speech respectively are determined. The difference between the ends of the vowel sounds is registered. From the obtained time differences an average value is determined. The average value indicates the quality of the produced speech. The invention can be used for evaluation of different speech sources.

Method and device for rating of speech quality by calculating time

A method and device for speech-to-text conversion. From a given speech, the fundamental tone is extracted. A model of the speech is further created from the speech. In the model, a duration reproduction in words and sentences is obtained. The duration reproduction is compared with a segment duration in the speech. Information which decides which type of accent exists is obtained from the comparison from which a text with sentence accent information is produced. Further, in creating the model lexical information, which may indicate accents in words, can be utilized. Syntax analysis of the model can further by utilized to model a fundamental tone in a sentence. Such operations allow the re-recognition utilizing the accent and stress after a decision as to the type of accent which exists.

Syllable duration and pitch variation to determine accents and stresses

The present invention relates to a method and arrangement for speech to text conversion. A string of phonemes is identified from a given input speech. The different phonemes are identified and are joined together to form words and phrases/sentences. The words are checked lexically, any words which are not found in the language concerned being excluded. The phrases/sentences are checked syntactically, any word combinations which do not occur in the language concerned being excluded. A model of the speech is obtained by the process. The intonation patterns of the model and of the input speech are determined, and compared, the words and phrases/sentences of the model, whose intonation patterns do not correspond with those of the input speech, are excluded from the model. A representation of the words, and/or word combinations, which best corresponds with the input speech is then provided, preferably in the providing in the form of a print out of the related text.

Method and arrangement for speech to text conversion

Method and device for speech recognition

The present invention relates to a method and arrangement for determining stresses in a spoken sequence. From a sequence recognized in the spoken speech, a model of the speech is created. By comparing the spoken sequence with the modelled speech, a difference between them is obtained. The difference is utilized partly for correcting the modelled speech and partly for determining stresses in the spoken sequence. After having determined the stresses in the speech, it is possible to determine the meaning of individual words and sentences in an unambiguous manner. This is then utilized in different contexts, for example when translating a first language to a second language whilst retaining meaning and intonation. The invention can also be used in verbal man-to-machine communication.

Method and arrangement in automatic extraction of prosodic information

The present invention relates to a process for evaluating speech quality in speech synthesizers with the aid of a speech recognition system. The recognition system is programmed using a number of persons. It then receives synthetic or natural speech from speech synthesizers or persons respectively, displaying differing speech quality. The speech recognition system determines a level of recognition for each received speech quality. In order to evaluate the speech quality in a speech synthesizer, speech is received from the synthesizer by the speech recognition system, is allocated a level of recognition and is ranked in comparison to the levels of recognition for previously received speech.

Process for evaluating speech quality in speech synthesis

An arrangement for improved speech comprehension in artificial translation of one language to a second language. The arrangement comprises an analysis unit which carries out an analysis of duration and fundamental tone of the speech in the first language. A prosody-interpreting unit determines, on the basis of the analysis and language-characteristic information, prosody-dependent information in the first speech which is used by a prosody-generating unit for the second language for controlling the speech synthesis. A speech synthesis element thus produces stresses in the speech translated in the second language which, from a language point of view, correspond to stresses in the first language.

Growing community of inventors

Vagnharad, Sweden

Bertil Lyberg