公开(公告)号：JP2000250578A

公开(公告)日：2000-09-14

申请号：JP2000046073

申请日：2000-02-23

Applicant: IBM

Inventor： FRANK FAD ,  GUASTI PETER ,  AMADO NASUIFU ,  RONALD VAN BASUKURIKU ,  HARVEY ROEBUCK

IPC: G06F3/16 ,  G10L15/06 ,  G10L15/10 ,  G10L15/26

Abstract: PROBLEM TO BE SOLVED: To permit correct relation to a voice input device corresponding to user training data by generating a graphical user interface(GUI) display picture making it possible to select the user of at least one of a voice input device and audio environment. SOLUTION: A voice file of training data relating to at least one of a specific voice input device and specific audio environment as to a specific user is stored in enrollments. Then the GUI display picture which prompting and enabling the selection of the user of at least one of the voice input device and audio environment is generated. Namely, when the user has >=2 sound cards supporting speech recognition, the GUI picture for selecting one of the sound cards is presented to the user. This picture makes the user to select which sound card is used for input, i.e., recording and which sound card is used for output, i.e., reproduction.

公开(公告)号：JP2000250734A

公开(公告)日：2000-09-14

申请号：JP2000034466

申请日：2000-02-14

Applicant: IBM

Inventor： FRANK FAD ,  GUASTI PETER ,  AMADO NASHIFU ,  RONALD VAN BASUKAAKU ,  HARVEY ROEBUCK

IPC: G10L15/28 ,  G06F3/01 ,  G06F3/16 ,  G10L15/00 ,  G10L15/20 ,  G10L15/26 ,  G10L21/02 ,  G06F3/00

Abstract: PROBLEM TO BE SOLVED: To guide a procedure for changing a voice input device to a user by generating a GUI display picture while using information in a dynamic table including information extracted from the registry of dictation recording and dictation reproducing device descriptions. SOLUTION: The GUI picture for a user to select one of sound cards for supporting voice recognition is displayed to the user and the user selects respective sound cards to be used for recording and reproducing (S18). Afterwards, the type of the input device is judged by reference to user peculiar data such as user selection, command row parameter, registry item and registration identification information, for example, and a voice reproducing level is tested and adjusted into a comfortable setting value (S20 and S22). One or plural switch setting values capable of predicting prevention of voice feedback are selected, the voice feedback is prevented and the input device is selected and connected.