manishmadugula / speech_processing-phonological-processes-

Intern at NITK

Speech_processing-Phonological-processes-

Intern at NITK All the assimilation's directory is attached to init2.m just open it and comment out the sound you would like to analyse.

By using following commands you can find minimum path,area etc feature for MFCCs,Pitch,Formants,GCI,Jitter,Shimmer,Zero frequency signal,lpresidual,zcr

mfcc_dtwcomparision.m formants_dtwcomparision.m pitch_dtwcomparision.m shimmer_dtwcomparision.m jitter_dtwcomparision.m zcr_dtwcomparision.m zfSig_dtwcomparision.m gci_dtwcomparision.m lpresidual_dtwcomparision.m

(pitch uses praat which is called from command line using praatcon.exe so make sure you have praat in your environmental variable)

normalization tried: mean_normalization.m duration-normalization.m cdf_normalization.m segment_normalization.m

DTW features are all extracted in dtwFeatureExtraction.m

dynamic_area.m is used to calculate area between diagonal and dtw path and return the region for which area value is maximum

DATA.ods has data and results related to identification of region of assimilation. mfccs and formants.ods have data related to similarity of words.