Alternative linear predictive analysis techniques with applications to speaker identification

1994 ◽  
Author(s):  
Ravi P. Ramachandran ◽  
M. S. Zilovic ◽  
Richard J. Mammone
Keyword(s):  
Speaker Identification ◽  
Predictive Analysis ◽  
Analysis Techniques ◽  
Linear Predictive Analysis

Automatic Recognition of Aircraft Noise with PLP Method

2012 ◽  
Vol 160 ◽  
pp. 145-149 ◽  
Author(s):  
Jian Li Ding ◽  
Yong Yang
Keyword(s):  
Feature Extraction ◽  
Filter Bank ◽  
Recognition Algorithm ◽  
Automatic Recognition ◽  
Aircraft Noise ◽  
Predictive Analysis ◽  
Modified Method ◽  
Extraction Algorithm ◽  
Auditory Feature ◽  
Linear Predictive Analysis

This paper proposes a modified auditory feature extraction algorithm based on perceptual linear predictive analysis which is more suitable for automatic recognition of aircraft noise. In this algorithm, a different distribution of filter-bank is introduced in order to fit the physical characteristic of aircraft noise and the result shows that the modified method indeed performs better. The effect of Gammatone filter in improving the robustness of recognition algorithm is also demonstrated in the experiment.

Voice Biometric for Learner Authentication

2019 ◽  
pp. 150-181
Author(s):  
Kamaraj A. ◽  
Selva Nidhyananthan S ◽  
Kalyana Sundaram C.
Keyword(s):  
Mixture Model ◽  
Gaussian Mixture ◽  
Predictive Analysis ◽  
Model Parameters ◽  
Training Phase ◽  
Prosodic Features ◽  
Speech Corpus ◽  
Log Likelihood ◽  
Testing Data ◽  
Linear Predictive Analysis

The objective of this chapter is to verify the identity of the claimed learner by extracting the prosodic features of the speech signal. TIMIT Acoustic-Phonetic Continuous Speech Corpus is used for learner verification using prosodic and articulation features such as energy, pitch, and formants. The prosodic feature includes pitch (F0), and articulation feature includes formants (F1-F7). From this database, for this project in the training phase, 200 learners were used and in the testing phase 160 learners were used. The pitch and formants were extracted using linear predictive analysis. The first seven formants were used for verification purpose. The feature set consists of eight features. The features are fed into the Guassian mixture model. In the Gaussian mixture model, parameters are estimated from the training and testing data using the iterative expectation-maximization. Log likelihood score is computed using these parameters, and then these scores are normalized to make decisions. The decision is made based on the threshold.

Sign in / Sign up

Export Citation Format

Share Document