Title :
The effect of additive noise on speech amplitude spectra: a quantitative analysis
Author :
Zhu, Qifeng ; Alwan, Abeer
Author_Institution :
Henry Samueli Sch. of Eng. & Appl. Sci., California Univ., Los Angeles, CA, USA
Abstract :
This article analyzes the effect of additive noise on speech amplitude spectra, and introduces a method to estimate speech spectra from noisy observations. Estimated spectra are used to compute the Mel-frequency cepstral coefficients as a recognition front-end. Compared to linear spectral subtraction, this technique improves the performance of digit recognition in noise.
Keywords :
cepstral analysis; noise; speech processing; speech recognition; HMM-based system; Mel-frequency cepstral coefficients; additive noise; automatic speech recognition; digit recognition; linear spectral subtraction; noisy observations; speech amplitude spectra; speech recognition front-end; Acoustic noise; Additive noise; Automatic speech recognition; Cepstral analysis; Frequency; Hidden Markov models; Noise level; Random variables; Speech analysis; Speech enhancement;
Journal_Title :
Signal Processing Letters, IEEE
DOI :
10.1109/LSP.2002.801722
