EP 2372707 B1 20130313 - Adaptive spectral transformation for acoustic speech signals
Title (en)
Adaptive spectral transformation for acoustic speech signals
Title (de)
Adaptive Spektraltransformation für akustische Sprachsignale
Title (fr)
Transformation spectrale adaptative pour signaux vocaux acoustiques
Publication
Application
Priority
EP 10156530 A 20100315
Abstract (en)
[origin: EP2372707A1] The present invention relates to a method for adaptive transformation of the frequency spectrum of a windowed speech signal. According to the present invention, frequency compression is achieved by applying frequency compression functions, which are dependent upon characteristics of the current frame. Formant boosting takes place using formant-dependent functions to increase the contrast between the formants and the non-formant frequencies in the frequency spectrum of the current frame. Generally, frequencies below a defined threshold are compressed linearly, corresponding to no compression when the slope is equal to 1, while the compression rate of higher frequencies varies over time and frequency depending on the current frame features. Thus features of the input signal located above the frequency threshold are moved into the low-pass range and are audible in the transmitted output signal. The feature based processing is designed to enhance the understandability of speech in the transmitted bandwidth and to increase recognition rates in an automatic speech recognition module.
IPC 8 full level
G10L 21/02 (2013.01)
CPC (source: EP)
G10L 21/0364 (2013.01)
Designated contracting state (EPC)
AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO SE SI SK SM TR
DOCDB simple family (publication)
DOCDB simple family (application)
EP 10156530 A 20100315