Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronic Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

white and 1/f noise, randn MATLAB

Status
Not open for further replies.

usualsuspect

Newbie level 3
Joined
Jul 21, 2004
Messages
4
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,281
Activity points
38
randn matlab

In MATLAB

a = randn(1,1000) results in a vector of 1000 elements with a Gaussian distribution, mean=0 and variance=1.

b = cumsum(a) what is "b"? It seems to be meaningless.

c = sqrt(cumsum(a.^2)) If "a" is white noise, then is "c" 1/f noise? How is the mean and variance of "c" related to that of "a"?

Thanks.
 

Aya2002

Advanced Member level 4
Joined
Dec 12, 2006
Messages
1,142
Helped
184
Reputation
376
Reaction score
118
Trophy points
1,343
Location
Iraq
Activity points
8,019
noise matlab

you are generating (a) in your first step,

now, according to your application you will use it. any way, cumsum(a) is a vector containing the cumulative sum of the elements of a, this is according to the help of matlab.

in your second step, you said that (b) is meaningless. My friend, this is depends on your application's requirements, may be your applications asking you to do a cumulative some to your noise vector!!!

at the end, can you please be more clear so that may be some one else can help you more and more?

regards
 

usualsuspect

Newbie level 3
Joined
Jul 21, 2004
Messages
4
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,281
Activity points
38
white noise in matlab

By using "cumsum", i.e. integrating (or accumulating) I was trying to create something similar to an ideal 1/f noise. However, this is probably not the right way to do it.

As an example, assume vector a = randn(1,1000) represent the period of a clock observed over 1000 cycles (period-to-period jitter), now if I synthesize a clock waveform with these random period; take the FFT of the clock waveform, remove the hamonics of the clock, I'm left with white noise, from which I can calculate the noise power and I should get the variance of my original vector.

Since each element of vector "a" is independent (memoryless), the noise is white, that's why I started thinking about accumulating, then I'm confused by what am I getting when I "integrate" a vector with Gaussian distribution.

Thanks
 

Status
Not open for further replies.
Toggle Sidebar

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Top