Method and apparatus for the voice control of a device appertaining to consumer electronics

• US 20020035477A1
• Filed: 09/14/2001
• Published: 03/21/2002
• Est. Priority Date: 09/19/2000
• Status: Active Grant

• Current Assignee: InterDigital Patent Holdings, Inc. (InterDigital, Inc.)
• Original Assignee: Thomson Licensing (Vantiva SA)
• Inventors: Platte, Hans-Joachim, Schroder, Ernst F.
• Granted Patent: US 7,136,817 B2
• Time in Patent Office: Days
• US Class Current: 704/275
• CPC Class Codes: G06F 3/16 Sound input; Sound output speech processing G10L
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 17/00 Speaker identification or verification techniques

Abstract

In the voice control of a device appertaining to consumer electronics, speech inputs of a user are converted into digitized voice signals. First features, which are characteristic of the individual sounds of the speech, and thus permit recognition of the spoken sounds, are extracted from the digitized voice signals. Furthermore, second features, which permit a characterization of the voice of the respective user and are used for distinguishing between the speech inputs of different users, are extracted from the digitized voice signals. This enables the device being operated to distinguish between speech inputs of the present user, which represent operator-control commands, and utterances by other persons, and to respond only to operator-control commands of the present user. Furthermore, the quality of an analysis of the viewing habits of different users can be improved by a distinction between the users.

Claims

1. Method for the voice control of a device appertaining to consumer electronics, comprising:

converting speech inputs of a user into digitized voice signals;

extracting first features, which are characteristic of individual sounds of the speech, and thus permit recognition of the spoken sounds, from the digitized voice signals;

extracting second features, which permit a characterization of the voice of the respective user and are used for distinguishing between the speech inputs of different users, from the digitized voice signals; and

accepting after a voice command from a first user further voice commands only from this first user, by testing the further speech inputs for characteristic voice features and only accepting them if they can be assigned to the same speaker on the basis of these features.

2. Method according to claim 1, wherein a voice command for switching on the device is accepted from any first user and, after that, temporarily only voice command inputs from this first user are accepted.

3. Method according to claim 2, wherein a voice command for switching off the device is accepted only from the first user and in that, after switching off the device, voice commands are accepted again from any user.

4. Method according to claim 2, wherein a voice command for switching off the device is accepted from any user.

5. Method according to claim 1, wherein an operator-control command which, after its input by the first user, allows voice commands from a second user to be accepted is provided.

6. Method according to claim 1, wherein an identification of the various users takes place and an analysis of the viewing habits of the identified users is performed in order to create user profiles of the various users from this analysis.

7. Method according to claim 6, wherein the user profile is used in a buffer storage of television programmes in order to permit separate buffer storage of preferred programmes for different users.

8. Method according to claim 6, wherein the user profile is used to make proposals for programmes to be viewed to the users identified.

9. Apparatus for the voice control of a device appertaining to consumer electronics, comprising:

means for converting speech inputs of a user into digitized voice signals;

means for extracting first features, which are characteristic of individual sounds of the speech, and thus permit recognition of the spoken sounds, from the digitized voice signals;

means for extracting second features, which permit a characterization of the voice of the respective user and are used for distinguishing between the speech inputs of different users, from the digitized voice signals; and

means for accepting after a voice command from a first user further voice commands only from this first user, by testing the further speech inputs for characteristic voice features and only accepting them if they can be assigned to the same speaker on the basis of these features.