This preview shows pages 1–8. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Digital Speech Processing Digital Speech Processing Lecture 14A Lecture 14A Algorithms for Algorithms for Speech Processing Speech Processing peech Processing Algorithms peech Processing Algorithms Speech Processing Algorithms Speech Processing Algorithms Speech/Nonspeech detection Rulebased method using log energy and zero crossing rate i l h i t l i b k d i Single speech interval in background noise Voiced/Unvoiced/Background classification Bayesian approach using 5 speech parameters Needs to be trained (mainly to establish statistics for background signals) eeds o be a ed ( a y o es ab s s a s cs o bac g ou d s g a s) Pitch detection Estimation of pitch period (or pitch frequency) during regions of voiced speech plicitly needs classification of signal as voiced speech Implicitly needs classification of signal as voiced speech Algorithms in time domain, frequency domain, cepstral domain, or using LPCbased processing methods Formant estimation Estimation of the frequencies of the major resonances during voiced speech regions Implicitly needs classification of signal as voiced speech Need to handle birth and death processes as formants appear and disappear depending on spectral intensity Median Median moothing moothing Smoothing Smoothing nd nd and and Speech Speech Processing Processing hy Median Smoothing hy Median Smoothing Why Median Smoothing Why Median Smoothing Obvious pitch period discontinuities that need to be smoothed in a manner that preserves the character of the surrounding regions using a median (rather than a linear filter) smoother. unning Medians unning Medians Running Medians Running Medians 5 point median 5 point averaging on on inear Smoothing inear Smoothing Non Non Linear Smoothing Linear Smoothing linear smoothers (filters) are not always appropriate for smoothing arameter estimates because of smearing and blurring discontinuities parameter estimates because of smearing and blurring discontinuities pitch period smoothing would emphasize errors and distort the contour use combination of nonlinear smoother of running medians and g linear smoothing linear smoothing => separation of signals based on nonoverlapping frequency content on ear smoothing => separating signals based on their character nonlinear smoothing => separating signals based on their character (smooth or noiselike) ] ( [ ]) ( [ ]) smooth rough components = + + n S x n R x n 1 [ ] ( [ ]) ( [ ]) smooth rough components ( [ ]) medi an( [ ]) ( [ ]) ( [ ]) median of [ ] ... [ ] + + = = = + L x n S x n R x n y x n x n M x n M x n x n x n L 6 L roperties of Running Medians roperties of Running Medians Properties of Running Medians Properties of Running Medians unning medians of length Running medians of length L : 1. M L ( x [ n ]) = M L ( x [ n ]) edians will ot mear out discontinuities 2. Medians will not smear out discontinuities (jumps) in the signal if there are no iscontinuities within...
View
Full
Document
This note was uploaded on 12/29/2011 for the course ECE 259 taught by Professor Rabiner,l during the Fall '08 term at UCSB.
 Fall '08
 Rabiner,L
 Algorithms

Click to edit the document details