CN1500260A

5G,4G,3G,2G

Title

METHOD FOR FORMATION OF SPEECH RECOGNITION PARAMETERS

Application Number:

CN2002807774

Publication Date:

26-05-2004

Current Assignee:

Family ID:

Application Date:

18-12-2002

Declaring Company:

Publication Country:

US

Priority Date:

30-01-2002

Title

METHOD FOR FORMATION OF SPEECH RECOGNITION PARAMETERS

Application Number:

CN2002807774

Family ID:

Publication Country:

US

Publication Date:

26-05-2004

Application Date:

18-12-2002

Priority Date:

30-01-2002

Current Assignee:

Declaring Company:

Abstract  Abstract

A voice sample characterization front end adapted for use in a distributed speech recognition environment. digitizing the voice sample (31) between the low frequency channel (32) and the high frequency channel (33) is divided. the two ways are used for determining spectral content which is applied to judging the voice identification parameter for identifying the case of characterizing speech samples (such as cepstrum coefficient). (32) the low frequency path has a thorough noise reduction capability. In one embodiment the high frequency path 33 with a noise reduction result to help eliminate noise and need not have the low frequency channel 32 used by the same level of resource capacity.

A voice sample characterization front-end suitable for use in a distributed speech recognition context. A digitized voice sample (31) is split between a low frequency path (32) and a high frequency path (33). Both paths are used to determine spectral content suitable for use when determining speech recognition parameters (such as cepstral coefficients) that characterize the speech sample for recognition purposes. The low frequency path (32) has a thorough noise reduction capability. In one embodiment the results of this noise reduction are used by the high frequency path (33) to aid in de-noising without requiring the same level of resource capacity as used by the low frequency path (32).

A voice sample characterization front-end suitable for use in a distributed speech recognition context. A digitized voice sample 31 is split between a low frequency path 32 and a high frequency path 33. Both paths are used to determine spectral content suitable for use when determining speech recognition parameters such as cepstral coefficients that characterize the speech sample for recognition purposes. The low frequency path 32 has a thorough noise reduction capability. In one embodiment the results of this noise reduction are used by the high frequency path 33 to aid in de-noising without requiring the same level of resource capacity as used by the low frequency path 32.

Note:

The information in blue was extracted from the third parties (Standard Setting Organisation, Espacenet)

The information in grey was provided by the patent holder

The information in purple was extracted from the FrandAvenue

Explicitly disclosed patent:openly and comprehensibly describes all details of the invention in the patent document.

Implicitly disclosed patent:does not explicitly state certain aspects of the invention, but still allows for these to be inferred from the information provided.

Basis patent:The core patent in a family, outlining the fundamental invention from which related patents or applications originate.

Family member:related patents or applications that share a common priority or original filing.