| 摘要 |
An improved audio browser is disclosed. In an exemplary embodiment, a plurality of hypertext links (hereafter called "hyperlink") available from, for example, a World Wide Web document, are used as a vocabulary of a speech recognizer for an audio browser. These hyperlink words are read to the user in the ordinary course of the audio browser's "speaking voice" - such hyperlink words being identified to the user by, for example, a change in voice characteristics for the "speaking voice". When a user wishes to select a hyperlink word, the user merely repeats the hyperlink word itself, rather than speaking a command or using a DTMF tone. The speech recognizer, which has as its vocabulary some or all of the hyperlink words of the document, recognizes the spoken hyperlink word and causes the jump to the linked address associated with the recognized hyperlink word.
|
