Blog posts
— In this paper, we describe the use of prosodic features for speaker identification in real environment (noisy). Features based on pitch and energy of the speech are combined with standard acoustic parameters to make the acoustic models more robust to the variability in the speech signal under real...