構造体 Value

mfcc configuration parameter values [詳細]

#include <mfcc.h>


変数

long smp_period
 Sampling period in 100ns units
long smp_freq
 Sampling frequency
int framesize
 Window size in samples, similar to WINDOWSIZE in HTK (unit is different)
int frameshift
 Frame shift length in samples
float preEmph
 Pre-emphasis coefficient, corresponds to PREEMCOEF in HTK
int lifter
 Cepstral liftering coefficient, corresponds to CEPLIFTER in HTK
int fbank_num
 Number of filterbank channels, corresponds to NUMCHANS in HTK
int delWin
 Delta window size, corresponds to DELTAWINDOW in HTK
int accWin
 Acceleration window size, corresponds to ACCWINDOW in HTK
float silFloor
 Energy silence floor in dBs, corresponds to SILFLOOR in HTK
float escale
 Scaling coefficient of log energy, corresponds to ESCALE in HTK
int hipass
 High frequency cut-off in fbank analysis, -1 if disabled, corresponds to HIFREQ in HTK
int lopass
 Low frequency cut-off in fbank analysis, -1 if disabled, corresponds to LOFREQ in HTK
int enormal
 1 if normalise raw energy, 0 if disabled, corresponds to ENORMALISE in HTK
int raw_e
 1 if using raw energy, 0 if disabled, corresponds to RAWENERGY in HTK
int zmeanframe
 1 if apply zero mean frame like ZMEANSOURCE in HTK
int usepower
 1 if use power instead of magnitude in filterbank analysis
float vtln_alpha
 warping factor for VTLN, corresponds to WARPFREQ in HTK
float vtln_upper
 hi freq. cut off for VTLN, corresponds to WARPUCUTOFF in HTK
float vtln_lower
 low freq. cut off for VTLN, corresponds to WARPLCUTOFF in HTK
int delta
 1 if delta coef. needs to be computed
int acc
 1 if acceleration coef. needs to be computed
int energy
 1 if energy coef. needs to be computed
int c0
 1 if use 0'th cepstral parameter, 0 if disabled, corresponds to _0 qualifier in HTK
int absesup
 1 if absolute energy should be suppressed
int cmn
 1 if use Cepstrum Mean Normalization, 0 if disabled, corresponds to _Z qualifier in HTK
int cvn
 1 if use cepstral variance normalization, else 0 */
int mfcc_dim
 Number of MFCC dimensions
int baselen
 Number of base MFCC dimension with energies
int vecbuflen
 Vector length needed for computation
int veclen
 Resulting length of vector
int loaded
 1 if these parameters were loaded from HTK config file or binhmm header


説明

mfcc configuration parameter values

mfcc.h76 行で定義されています。


この構造体の説明は次のファイルから生成されました:
Juliusに対してThu Jul 23 12:17:29 2009に生成されました。  doxygen 1.5.1