The fIrst two experiments are based upon the 100- words of dataset for comparison with the previous research work, while the other two experiments are based on the 250 words dataset. We have used individual speakers in training and testing as well as mixture of speakers in training
and testing as explained later.