Speaker Recognition System

 
This product ALL is prepared for recognizing the voices of speaking persons within the speech passages of digitally recorded sound files. The main functions of the system include
o    detecting those speech segments within a sound file in which a given person is speaking, and
o    identifying those persons who are speaking in a given speech segment.
As a prerequisite to recognizing the voice of a given person, the voice model of that person must be available within the system. The voice model of a person can be trained from a short voice-sample by using one of the auxiliary services of the system.
The Speaker Recognition System of ALL may serve for speech segment retrieval, monitoring and statistics making purposes – whenever being deployed to a regularly updated sound archive.
The Speaker Recognition System can be used as a standalone product, or it can be embodied into the Speech Retrieval System of ALL. As a standalone product the software focuses only on the detection/identification of named persons’ voices within the target archive. Within the Speech Retrieval System, the speaker recognition possibility enriches the usual services of speech retrieval by allowing the end users to seek for the relevant speech segments of a sound archive also by names of persons.
The speaker recognition product of ALL works with high voice recognizing accuracy.