1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
|
fbmfwt(2) Scilab Function fbmfwt(2)
NAME
fbmfwt - Discrete wavelet based synthesis of a fBm
Author: Paulo Goncalves
Generates a 1/f Gaussian process from a discrete wavelet transform
Usage
[x] = fbmfwt(N,H,[noctave,Q,randseed]) ;
Input parameters
o N : Positive integer Sample size of the fBm
o H : Real in [0,1] Holder exponent
o noctave : integer Maximum resolution level (should not exceeed
log2(N))
o Q : real vector. Analyzing QMF (e.g. Q = MakeQMF('daubechies',4)
)
o randseed : real scalar Random seed generator
Output parameters
o x : real vector [1,N] Time samples of the 1/f Gaussian process
Algorithm details
Generates a 1/f Gaussian process by inverting a discrete wavelet transform.
Step 1: generates y a [1,N] i.i.d. standard Gaussian noise
Step 2: computes the discrete wavelet coefficients y
Step 3: weigth the wavelet coefficients y with the corresponding scale
power law
Step 4: invert the weighted discrete wavelet transform
See also:
fbmlevinson, synth2, FWT, MakeQMF
Example:
Q = MakeQMF('daubechies',4) ;
[x] = fbmfwt(1024,0.5,10,Q) ;
[wt,scale,f] = contwt(x,2^(-8),2^(-1),64,8) ;
[H] = cwttrack(wt,scale,0,1,1,8,1,1) ;
|
