TeleFlow Step 0451
From TeleFlow
SR VoicePass Reg
Registers a biometric voice password.
This step allows a user to create his or her unique voice-print that can be used later to authenticate the user. This process involves taking 3 samples of the user's voice, speaking the exact same password (or phrase) each time. This process is analogous to creating a password or PIN for the first time. The results are compiled into a "pattern" and stored in a file that is later used to authenticate log-in attempts.
Note: Voice Password authentication is not available as a part of the standard TeleFlow Suite. You must purchase additional software to use this step.
Contents |
Properties
Voice pattern filename (or User ID): File name to store the voice-print pattern created by this step. Typically, the user's ID is also used as the file name to make it easier to match an ID to its associated voice-print pattern.
Do not play start recording tone (beep): If checked, this indicates the SR VoicePass Reg step should not play a tone/beep just prior to recording for each voice-print.
Play voice file / code for recording 1 of 3: The Script Code, or name of an audio file, to play to prompt the caller for the first utterance of his or her voice password.
Play voice file / code for recording 2 of 3: The Script Code, or name of an audio file, to play to prompt the caller for the second utterance of his or her voice password.
Play voice file / code for recording 3 of 3: The Script Code, or name of an audio file, to play to prompt the caller for the third utterance of his or her voice password.
Registration Factor (1-tight - 100-loose): The registration factor controls consistency checking between the password utterances as they are processed and compiled to form a voice print pattern. The lowest setting yields the strictest control; even small inconsistencies would be detected. The highest setting produces the most lenient control; some inconsistencies and noise are ignored.
The mid-range setting (50%) is recommended for most scenarios, providing good matching accuracy while filtering out some background noise and natural inconsistencies in a person's speech.
Maximum time recordable (ms): Defines the maximum recording length in milliseconds of each voice password utterance. The default is 4000 (4 seconds).
Action Steps
Time Out: Path followed if caller doesn't speak during the utterance recording.
Fail: Registration failed. Likely causes: The caller didn't speak consistently enough to successfully register, or one of the "Play voice file" entries was invalid.
Supported Telephony Hardware Platforms
AudioCodes (Ai-Logix)
NMS Communications
Related Steps
