This paper describes a new artificial speech signal (ASVQ Artificial Speech by Vector Quantization technique) which reflects the average characteristics of the human voice. The ASVQ is intended for use as a test signal in the objective evaluation of speech coding system quality.To obtain the average characteristics, a very large speech data base is analyzed. The ASVQ generation method which reflects the extracted average characteristics of the human voice is formulated. This method applies vector quantizing analysis to the speech data base. The LPC speech synthesis circuit is used to reproduce the average characteristics. Finally, the new artificial speech signal is compared with a human voice and the estimation accuracy of the subjective quality of speech coding systems and nonlinear distortions is evaluated.
This paper proposes a method f o r reduci n g a d d i t i v e n o i s e from noisy speech s i g n a l . The e f f e c t i v e n e s s of t h i s comb f i l t e r i n g method i s demonstrated v i a experiments. The method i s a p p l i c a b l e t o n o i s e with varying power and non-flat spectrum. L i s t e n i n g experiments show t h a t t h e proposed method improves the s u b j e c t i v e SIN and t h e sound a r t i cu l a t i o n of extremely noisy speech. t i o n of t h i s method f o r t h e r e d u c t i o n of q u a n t i z a t i o n n o i s e and n o i s e caused by b i t e r r o r i n PCM speech i s i n v e s t i g a t e d . mental r e s u l t s i n d i c a t e t h a t : Applica-Experi-(1) For e r r o r -f r e e 2-to-5 b i t PCM t h e proposed method improves t h e s u b j e c t i v e quali t y of t h e speech equivalent t o one b i t ; (2) For b i t e r r o r r a t e 'ranging from 10-3 t o 10-1, t h e proposed method improves t h e q u a l i t y of 8-bit PCM speech e q u i v a l e n t t o t h e reduction of e r r o r rate down t o about 1 / 3 -1 / 5 . It i s very e f f e c t i v e f o r reducing t h e quantiz a t i o n noise i n 4-bit PCM speech. (3) The proposed method i s p a r t i c u l a r l y u s e f u l f o r improving t h e sound a r t i c u l a t i o n of extremely noisy speech. 1. I n t r o d u c t i o n Speech s i g n a l s i n communication and audio systems u s u a l l y are contaminated by noise. Noise causes degradations i n t h e q u a l i t y of speech s i g n a l s and reduces t h e performances of audio processing equipment. Therefore , r e s e a r c h and development of techniques f o r n o i s e r e d u c t i o n a r e v e r y important. Noise reduction methods proposed previously include Wiener f i l t e r i n g , a d d i t i v e noise c a n c e l i n g [13 , s p e c t r a l s u b t r a c t i o n [2-41 , method based on c o r r e l a t i o n [5] , and speech modeling [ 6 ] . These methods assume t h a t t h e power and spectrum of n o i s e are known v a l u e s . The comb f i l t e r i n g n o i s e red u c t i o n method proposed by t h i s paper does n o t have t o r e l y on such assumptions. Even though comb f i l t e r i n g method h a s been used i n t h e p a s t [ 7-91 , comprehensive q u a n t i t a t i v e e v a l u a t i o n p e r t i n e n t t o p i t c h e x t r a c t i o n of noisy speech and removal of harmonic noise has not been presented. The n o i s e reduction method proposed h e r e u t i l i z e s t h e p i t c h ext r a c t i o n [ l o ] and comb f i l t e r i n g techniques
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.