wnoise (Wavelet Toolbox)

Generate noisy wavelet test data.

X = wnoise(NUM,N)
[X,XN] = wnoise(NUM,N,SNRAT)
[X,XN] = wnoise(NUM,N,SNRAT,INIT)

Description

X = wnoise(NUM,N) returns values of test function number NUM, on a 2^N sample of [0,1].

[X,XN] = wnoise(NUM,N,SNRAT) returns a test vector X as above, rescaled such that std(x) = SNRAT. The returned vector XN contains the same test vector corrupted by additive Gaussian white noise N(0,1). XN has a signal-to-noise ratio of SNRAT.

[X,XN] = wnoise(NUM,N,SNRAT,INIT) returns previous vector X and XN, but the generator seed is set to INIT value.

The six functions are due to Donoho and Johnstone (See Reference):


`NUM` = 1	Blocks
`NUM` = 2	Bumps
`NUM` = 3	Heavy sine
`NUM` = 4	Doppler
`NUM` = 5	Quadchirp
`NUM` = 6	Mishmash

Examples

% Generate 2^10 samples of 'Heavy sine' (item 3). 
    x = wnoise(3,10); 

% Generate 2^10 samples of 'Doppler' (item 4) and of 
% noisy 'Doppler' with a signal-to-noise ratio of 7. 
    [x,noisyx] = wnoise(4,10,7);

% To introduce your own rand seed, a fourth 
% argument is allowed: 
    init = 2055415866; 
    [x,noisyx] = wnoise(4,10,7,init);

% Plot all the test functions. 
    ind = linspace(0,1,2^10); 
    for i = 1:6 
        x = wnoise(i,10); 
        subplot(6,1,i), plot(ind,x) 
    end