Granted Patent B2

US 7,734,472 · App. 10/951,594 · Granted Jun 8, 2010

Assignee: Alcatel

Abstract

Claims (47)

1. A speech recognition system comprising:

an audio input unit arranged in a terminal;

a speech recognizer;

an adjustable speech recognition enhancer, separate from and arranged between the audio input unit and the speech recognizer, comprising:

a parametrizable pre-filtering unit,

a parametrizable dynamic voice level control unit,

a parametrizable noise reduction unit,

a parametrizable voice level control unit, and

a parameter setting unit;

wherein the parameters of the parametrizable pre-filtering unit, the parametrizable dynamic voice level control unit, the parametrizable noise reduction unit and the parametrizable voice level control unit are adjusted by the parameter setting unit to the characteristics of the specific audio input unit and/or the characteristics of the specific speech recognizer of the speech recognition system for adapting the audio input unit to the speech recognizer,

wherein parameters set by the parameter setting unit are stored in a data storage, and

wherein the adjustable speech recognition enhancer is adapted to process speech wave forms only in the time domain.

2. The speech recognition system of claim 1 ,

wherein the speech recognition system is a distributed speech recognition system,

wherein the speech recognizer comprises:

a central speech recognition server and at least one remote distributed speech recognition front-end performing the process of feature extraction,

wherein the distributed speech recognition front-end is located in a respective terminal,

wherein the terminal comprises a respective speech recognition enhancer and a respective audio input unit,

wherein each speech recognition enhancer is arranged in-between the audio input unit and the distributed speech recognition front-end of the respective terminal, and

wherein the parameters of the parametrizable pre-filtering unit, the parametrizable dynamic voice level control unit, the parametrizable noise reduction unit and the parametrizable voice level control unit of each speech recognition enhancer are adjusted to the characteristics of the respective audio input unit and/or the characteristics of the respective distributed speech recognition front-end of the respective terminal for adapting the audio input unit of this terminal to the distributed speech recognition front-end of this terminal.

3. The speech recognition system of claim 1 ,

wherein the speech recognition system is a stand-alone speech recognition system, wherein the speech recognizer is embedded in the terminal.

4. The speech recognition system of claim 1 ,

wherein the audio input unit comprises a microphone and an analogue to digital converter, and

wherein the speech recognizer is adapted to perform a transformation of a speech signal from the time domain to the frequency domain and to perform further processing of the transformed speech signal in the frequency domain.

5. The speech recognition system of claim 1 ,

wherein the pre-filtering unit is adapted to perform high-pass filtering of the speech signal provided by the audio input unit, and

wherein the parameters of the pre-filtering unit are adjusted to the characteristics of the audio input unit, in particular to the characteristics of a microphone arranged in the terminal.

6. The speech recognition system of claim 1 ,

wherein the dynamic voice level control unit is adapted to provide a dynamic voice level compression of the output signal of the pre-filtering unit, the dynamic voice level compression depending on parameters specifying a compression factor and a nominal voice level.

7. The speech recognition system of claim 1 ,

wherein the noise reduction unit comprises a voice activity detector and an amplifier controlled by the voice activity detector, and

wherein the voice activity detector is adapted to reduce the amplification factor of the amplifier when detecting a speech pause.

8. The speech recognition system of claim 1 ,

wherein the voice level control unit contains an amplifier for adapting the voice level of the output signal of the noise reduction unit to a preset voice level adapted to the characteristics of the speech recognizer.

9. A speech recognition enhancer for arrangement separate from and between an audio input unit and a speech recognizer of a speech recognition system comprising

a parametrizable pre-filtering unit,

a parametrizable dynamic voice level control unit,

a parametrizable dynamic voice level control unit,

a parametrizable noise reduction unit, and

a parametrizable voice level control unit,

wherein the parameters of the parametrizable pre-filtering unit, the parametrizable dynamic voice level control unit, the parametrizable noise reduction unit and the parametrizable voice level control unit are adjustable by a parameter setting unit to the characteristics of the specific audio input unit and/or the characteristics of the specific speech recognizer for adapting the audio input unit to the speech recognizer,

wherein parameters set by the parameter setting unit are stored in a data storage,

wherein the speech recognition enhancer is adapted to process speech wave forms only in the time domain.

10. The speech recognition system of claim 2 , wherein the central speech recognition server and the at least one remote distributed speech recognition front-end communicate via a communication network.

11. The speech recognition system of claim 10 , wherein the communication network is one of a Global System for Mobile Communications (GSM) network, a Universal Telecommunications System (UMTS) network or a DCMA 2000 network.

12. The speech recognition system of claim 10 , wherein the communication network is one of Public Switched Telecommunication Network (PSTN) or Integrated Service Digital Network (ISDN).

Assignments (9)

Priority Claims (1)

Continuity (1)