Simple Multi-User Voice Recognition
This project consisted of implementing a simple multi-user speaker recognition system using Linear Prediction (LP). Three users each recorded an audio file containing thirty repetions of a spoken word to be used for training and testing of the system. The three audio samples were sampled at 48 kHz and stored as user1.mat, user2.mat, and user3.mat.
The first fifteen repitions of each recorded audio file served as the training data used to characterize each user's speech. The remaining fifteen repetitions were used for testing the systems ability to correctly identify a user.
The first fifteen repitions of each recorded audio file served as the training data used to characterize each user's speech. The remaining fifteen repetitions were used for testing the systems ability to correctly identify a user.
Stage 1: Word Isolation
...adding these sections shortly.
