1mlib_SignalMelCepstral_F32(m3eMdLiIaBL)ib Library Funmcltiibo_nSsignalMelCepstral_F32(3MLIB)
2
3
4

NAME

6       mlib_SignalMelCepstral_F32 - perform cepstral analysis in mel frequency
7       scale
8

SYNOPSIS

10       cc [ flag... ] file... -lmlib [ library... ]
11       #include <mlib.h>
12
13       mlib_status mlib_SignalMelCepstral_F32(mlib_f32 *cepst,
14            const mlib_f32 *signal, void *state);
15
16

DESCRIPTION

18       The mlib_SignalMelCepstral_F32() function performs cepstral analysis in
19       mel frequency scale.
20
21
22       The  first  two  steps of mel scale cepstral analysis is the same as in
23       general cepstral anaysis. After the logarithm of the spectrum magnitude
24       is  obtained,  it  is  converted  into  mel  frequency scale before the
25       inverse Fourier transform.
26
27                    +-----------+        +-----------+
28                    |  Linear   |        |  Inverse  |
29         ... ------>|    to     |------->|  Fourier  |----->
30              X'(k) | Mel Scale | X''(m) | Transform | c(n)
31                    +-----------+        +-----------+
32
33
34
35       where X'(k) is defined in linear frequency scale and X''(m) is  defined
36       in mel frequency scale.
37
38
39       The mel frequency scale is defined as following.
40
41         freq_mel = melmul * LOG10(1 + freq_linear / meldiv)
42
43
44
45       where  freq_mel  is the frequency in mel scale, freq_linear is the fre‐
46       quency in linear scale, melmul is the multiplying factor, muldiv is the
47       dividing factor.
48
49
50       Optionally,  a  bank of band pass filters in linear frequency scale can
51       be used below the bank of band pass filters in mel frequency scale,  as
52       shown below in linear frequency scale.
53
54         0   f1  f2  f3    fp  fp+1 fp+2  fp+3  fp+q
55         |---|---|---| ... |---|----|-----| ... | ... -> freq
56
57
58
59       where  fp  =  melbgn, fp+q = melend, p = nlinear, q = nmel; the filters
60       number 1 to p are defined in linear frequency scale  which  have  equal
61       bandwidth  in linear frequency scale; the filters number p+1 to p+q are
62       defined in mel frequency scale which have equal bandwidth in  mel  fre‐
63       quency scale and increasing bandwidth in linear frequency scale.
64
65
66       See  Digital  Signal  Processing  by  Alan  V.  Oppenheim and Ronald W.
67       Schafer, Prentice Hall, 1974.
68
69
70       See Fundamentals of Speech Recognition by Lawrence Rabiner  and  Biing-
71       Hwang Juang, Prentice Hall, 1993.
72

PARAMETERS

74       The function takes the following arguments:
75
76       cepst     The cepstral coefficients.
77
78
79       signal    The input signal vector.
80
81
82       state     Pointer to the internal state structure.
83
84

RETURN VALUES

86       The  function  returns MLIB_SUCCESS if successful. Otherwise it returns
87       MLIB_FAILURE.
88

ATTRIBUTES

90       See attributes(5) for descriptions of the following attributes:
91
92
93
94
95       ┌─────────────────────────────┬─────────────────────────────┐
96       │      ATTRIBUTE TYPE         │      ATTRIBUTE VALUE        │
97       ├─────────────────────────────┼─────────────────────────────┤
98       │Interface Stability          │Committed                    │
99       ├─────────────────────────────┼─────────────────────────────┤
100       │MT-Level                     │MT-Safe                      │
101       └─────────────────────────────┴─────────────────────────────┘
102

SEE ALSO

104       mlib_SignalMelCepstralInit_F32(3MLIB),          mlib_SignalMelCepstral‐
105       Free_F32(3MLIB), attributes(5)
106
107
108
109SunOS 5.11                        2 Mar 2007 mlib_SignalMelCepstral_F32(3MLIB)
Impressum