![]() The authors also thank him for imparting the signal processing knowledge necessary for this endeavor and for his guidance and clarifications. Ananthakrishna Chintanpalli, Department of Electrical and Electronics Engineering, Birla Institute of Technology and Science, Pilani for providing them an opportunity to undertake this project work. In the LPC spectrum of the signal, the peaks correspond to the formant frequencies. In the pole-zero plot of A(z), the zeros lying close to or on the unit circle correspond to the formant frequencies of the speech signal. On carrying out LPC analysis for a speech signal, a prediction polynomial A(z) in z^(-1) is obtained. To identify these formants, linear predictive coding proves extremely useful. The formant frequencies correspond to local maxima in the spectrum. On analyzing the frequency spectrum of a speech signal, various peaks can be observed. These resonance frequencies are called the formant frequencies of the sound. The resonance frequencies resulting from a particular configuration of the articulators are instrumental in forming the sound corresponding to a given phoneme. The sounds created in the vocal tract are shaped in the frequency domain by the frequency response of the vocal tract. That is, the changes in vocal tract configuration occur relatively slowly compared to the detailed time variation of the speech signal. The general character of the speech signal varies at the phoneme rate, which is on the order of 10 phonemes per second, while the detailed time variations of the speech waveform are at a much higher rate. Taru Kapoor, Kshitij Khandelwal, Anubhav Sachan Introduction Further, an attempt has been made to understand which vowels (both single and concurrent) are more susceptible to noise. An analysis has been carried out to investigate the effects of the different SNR levels on the formants. The formant frequencies were then again estimated for these corrupted signals having different SNR levels. After this, the noise was added to the vowels for 3 different SNR values. Following this, using various speech signals, Speech Spectrum Shaped Noise was generated. With the help of Linear Predictive Coding (LPC), the formant frequencies were first found for these vowels. The dataset used for this purpose included values of single and concurrent vowels at two fundamental frequencies, 100 Hz and 126 Hz respectively. The project focuses on understanding the effects of noise on the formant representations of both single and concurrent vowels. Use of Linear Predictive Coding for Formant Analysis of Concurrent Vowels
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |