| Zhipeng.Gao 2005-11-29, 7:07 pm |
| This problem is about how to set correct frequency bins when using
'tfrsp.m' and 'tfrview.m' (functions in time-frequency tool
box--TFTB).
I tried to analysis a 3-component signal using Spectrogram(tfrsp.m).
clear
clc
%----------produce the signal----------%
fm=4e14; % the maximum frequency
fs=2.5*fm; Ts=1/fs; % sampling frequency
tm=0.3e-12; % the maximum time
N = round(tm/Ts); % sampling points
t=linspace(0,tm,N); % time duration
%----------------------------------------------%
b1=exp(-((t-0.5e-13)/0.1e-13).^2);
b2=exp(-((t-1.5e-13)/0.1e-13).^2);
b3=exp(-((t-2.5e-13)/0.1e-13).^2);
f1=exp(-i*2*pi*1e14*t); % modulation frequency 1
f2=exp(-i*2*pi*4e14*t); % modulation frequency 2
f3=exp(-i*2*pi*2.5e14*t); % modulation frequency 3
%%%%%%%%%%%%%%%%%
s=b1.*f1+b2.*f2+b3.*f3;
s=s'; % transpose
%-------------------------------------%
wl=N/4; % the length of analysis window
h=tftb_window(wl,'hamming'); % analysis window
tt=1:length(s); % time instants
M=length(s); % number of frequency bins
tfr = tfrsp(s,tt,M,h);
NF2 = M/2;
% NF2 is the number of frequency bins displayed
% question here question here question here
param=[1, 0, 0, 18, NF2, 4, fs, 1, 0, 0.5];
figure
tfrview(tfr,s,tt,'tfrsp',param);
The problem is why I had to set 'NF2=M/2' is I want to get a correct
display of the tfr(time-frequency representation).
Compared with the problem, when I use 'tfrwv.m', I can get the
correct dislay of tfr by using the following codes:
(....the same as the former case)
[tfr,t,f]=tfrwv(s,tt,M);
NF2=M;
param=[1, 0, 0, 18, NF2, 4, fs, 1, 0, 0.5];
figure
tfrview(tfr,s,tt,'tfrwv',param);
Please help me~~~~~~~~~~~
no9_magician@yahoo.com.cn
|