Automatic speaker verification refers to process of verifying a speaker from the entire utterance without segregating the words. The process is text and language independent. The police get hold of some disputed voice. If they suspect some person, he/she is called to give sample utterance. The person may not be ready to utter same words as are there in the disputed voice. This system helps here. It has both the modes, text dependent and text independent. It also has a facility for speaker identification and for storing the suspected person’s voices in a database.

​

Features:

​

• ASV5000 is automatic speaker verification and identification system. No need to segregate the words.

• It has 2 modes of operation, text dependent and text independent mode. The dependent mode uses formats matching and text independent mode uses embedding score with parameters matching.

• It also has automatic speaker identification mode. This can match the control utterance with all suspect audios you may have to see if the culprit is one of the already identified criminals.

• Both the algorithms perform seamlessly with great accuracy. The system generates the detailed report automatically for both modes in word or in pdf format.

• It has a facility to find SNR of any audio, gender detection for any audio.

• There is a facility for all types of basic audio processing like record, play, export, split stereo channels, zoomin, zoomout, audio processing like normalize, show clipping with opening and closing of windows and changing their layout etc.

• All above features are supported for mono as well as stereo files of any formats and extensions.

• It has a facility to automatically divert all audios tested to a database folder after proper segregation as male and female. This database created can be directly used for speaker identification.

• It includes applications for noise cancellation, speaker diarization and manual speaker diarization.

• Facility to run on the device as a complete offline solution.