IPC 7 English version G10L !A-G10L 101:20

G10L-G10L10120

SECTION G – PHYSICS

G 10

MUSICAL INSTRUMENTS; ACOUSTICS

G 10 L

SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION (sound input/output for computers G06F 3/16; digital data processing methods or equipment specially adapted for handling natural language data G06F 17/20; teaching or communicating with the blind, deaf or mute G09B 21/00; telephonic communication H04M) [4]

Notes

(1)

In this subclass, it is desirable to add the indexing codes relating to speech classification or search methods. The indexing codes, which are chosen from groups G10L 15/08 to G10L 15/18, have the same numbers as the classification symbols, but a colon is used instead of the oblique stroke, and should be unlinked. [7]

(2)

In this subclass, it is desirable to add the indexing codes of group G10L 101:00. The indexing codes should be unlinked. [7]

(3)

Attention is drawn to Chapter IV of the Guide which sets forth the rules concerning the application and presentation of the different types of indexing code. [7]

(transferred to G10L 11/00 to G10L 11/06, G10L 13/04, G10L 15/04, G10L 17/00)

(transferred to G10L 21/00 to G10L 21/04)

(transferred to G10L 13/00, G10L 15/00, G10L 17/00)

(transferred to G10L 13/00 to G10L 13/08)

(transferred to G10L 15/00 to G10L 17/00)

(transferred to G10L 13/00, G10L 15/00, G10L 17/00, G10L 19/02)

(transferred to G10L 13/02, G10L 19/02)

(transferred to G10L 15/00 to G10L 17/00)

(transferred to G10L 21/06)

(transferred to G10L 15/02, G10L 15/08, G10L 15/16, G10L 19/02)

(transferred to G10L 15/00 to G10L 17/00)

(transferred to G10L 11/00 to G10L 21/06)

(transferred to G10L 13/00 to G10L 17/00, G10L 19/04 to G10L 19/14, G10L 21/00 to G10L 21/06)

(transferred to G10L 11/00 to G10L 21/06)

(transferred to G10L 15/24)

11/

Determination or detection of speech characteristics not restricted to a single one of groups G10L 15/00 to G10L 21/00 [7]

11/

Detection of presence or absence of speech signals (switching of direction of transmission by voice frequency in two-way loud-speaking telephone systems H04M 9/10) [7]

11/

Pitch determination of speech signals [7]

11/

Discriminating between voiced and unvoiced parts of speech signals (G10L 11/04 takes precedence) [7]

13/

Speech synthesis; Text to speech systems (electrophonic musical instruments G10H) [7]

13/

Methods for producing synthetic speech; Speech synthesisers [7]

13/

Details of speech synthesis systems, e.g. synthesiser structure, memory management [7]

13/

Elementary speech units used in speech synthesisers; Concatenation rules [7]

13/

Text analysis or generation of parameters for speech synthesis out of text, e.g. grapheme to phoneme translation, prosody generation, stress or intonation determination [7]

15/

Speech recognition (G10L 17/00 takes precedence) [7]

15/

Feature extraction for speech recognition; Selection of recognition unit [7]

15/

Segmentation or word limit detection [7]

15/

Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice (G10L 15/14 takes precedence) [7]

15/

Speech classification or search (pattern recognition G06K 9/00) [7]

15/

using distance or distortion measures between unknown speech and reference templates [7]

15/

using dynamic programming techniques, e.g. Dynamic Time Warping (DTW) [7]

15/

using statistical models, e.g. Hidden Markov Models (HMM) (G10L 15/18 takes precedence) [7]

15/

using artificial neural networks [7]

15/

using natural language modelling [7]

15/

Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech (G10L 21/02 takes precedence) [7]

15/

Procedures used during a speech recognition process, e.g. man-machine dialog [7]

15/

Speech recognition using non-acoustical features, e.g. position of the lips [7]

15/

Speech to text systems (G10L 15/08 takes precedence) [7]

15/

Constructional details of speech recognition systems [7]

17/

Speaker identification or verification [7]

Note

In group G10L 17/00, it is desirable to add the indexing codes relating to the feature extraction or training or procedure. The indexing codes, which are chosen from groups G10L 15/02, G10L 15/06 and G10L 15/22, have the same numbers as the classification symbols, but a colon is used instead of the oblique stroke, and should be unlinked. [7]

19/

Speech analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech [7]

19/

using spectral analysis, e.g. transform vocoders, subband vocoders [7]

19/

using predictive techniques [7]

19/

Determination or coding of the spectral characteristics, e.g. of the short term prediction coefficients [7]

19/

Determination or coding of the excitation function; Determination or coding of the long-term prediction parameters [7]

19/

Determination or coding of a multipulse excitation [7]

19/

Determination or coding of a code excitation, e.g. in code excited linear prediction (CELP) vocoders [7]

19/

Details not provided for in groups G10L 19/06 to G10L 19/12, e.g. gain coding, post filtering design, vocoder structure [7]

21/

Processing of the speech signal to produce another audible or non-audible signal, e.g. visual, tactile, in order to modify its quality or its intelligibility (G10L 19/00 takes precedence) [7]

21/

Speech enhancement, e.g. noise reduction, echo cancellation (echo suppression in hand-free telephones H04M 9/08; hearing aids H04R 25/00) [7]

21/

Time compression or expansion [7]

21/

Transformation of speech into a non-audible representation, e.g. speech visualisation, speech processing for tactile aids (G10L 15/26 takes precedence) [7]

Indexing scheme associated with groups G10L 11/00 to G10L 21/00, relating to speech signal processing or feature extraction. The indexing codes should be unlinked. [7]

Note

Attention is drawn to Chapter IV of the Guide which sets forth the rules concerning the application and presentation of the different types of indexing code. [7]

101:

Signal processing or feature extraction methods applied in speech analysis, synthesis or recognition [7]

101:

Spectral analysis [7]

101:

023

using filter banks [7]

101:

027

using Fourier or Walsh transformation [7]

101:

Cepstral analysis [7]

101:

Waveform coding [7]

101:

065

Energy determination [7]

101:

Zero crossing rate [7]

101:

Vector quantisation [7]

101:

Predictive techniques, e.g. linear prediction [7]

101:

Correlation techniques [7]

101:

Orthogonal transformations, e.g. wavelets (G10L 101:027 takes precedence) [7]

101:

Techniques based on formant identification [7]

101:

Techniques based on simulation of the vocal or the aural systems [7]