Indian Journal of Science and Technology
Year: 2015, Volume: 8, Issue: 29, Pages: 1-7
P. L. Chithra1 * and R. Aparna2
1 Department of Computer Science, University of Madras, Chennai - 600 005, Tamil Nadu, India; [email protected]
2 Department of Computer Science, MOP Vaishnav College, Chennai – 600034, Tamil Nadu, India; [email protected]
Background/Objectives: Automatic Speech Recognition (ASR) and Language Identification (LID) are the key areas of acoustic speech signal processing. Speech signals watermarking, steganography and cryptography are considered to be the emerging techniques to ensure information security in speech signal transmission. Efforts should be taken to retain every minute detail of the signal. In all above mentioned methods, it is necessary to preprocess the speech signal so as to get best results. To take into account, this paper presents the performance analysis of windowing techniques in automatic speech signal segmentation. Methods/Statistical Analysis: Speech signal is segmented into syllables as a first step. In the process of segmentation, a windowing technique is applied to enhance the syllable boundaries. Then window function is applied to the input signal before Discrete Fourier Transform has applied. There are many windowing techniques available. Proposed work is carried to analyze the performance of few windowing functions in order to retain every minute detail of the signal and to preprocess the speech signal effectively in the Automatic Speech Recognition (ASR) and Language Identification (LID). Findings: The results produced by the windowing and filtering techniques in segmentation process are plotted. The proposed method out performs well and the performance of different windowing and filtering techniques is analyzed. The number of peaks found is tabulated. Application/Improvements: Thus our experimental results shows the significance of segmenting speech signals effectively using windowing function with discrete filters than the adaptive filters. Further those segments can be used in the field of Automatic Speech Recognition (ASR), Language Identification (LID), Speech signals watermarking, steganography. The observed segments with windowing using discrete filters are highly useful for clustering and pattern matching techniques.
Keywords: Filtering, Speech Signal Processing, Segmentation, Windowing Function
Subscribe now for latest articles and news.