Wavelet Analysis - Torrence & Compo method

7 次查看（过去 30 天）

显示更早的评论

Harry 2014-8-15

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/150999-wavelet-analysis-torrence-compo-method

评论： Harry 2014-8-15

在 MATLAB Online 中打开

Hello

I have a (probably silly) question re the use of the Torrence and Compo CWT method.

I am plotting the WT of my data using the interactive tool available on the website, and this is what is looks like. The Y-axis is showing the period (in hertz) correctly, and the x axis is the correct time interval (fs = 0.001hz) for 1537 values.

Here is what I end up with when I plot this using the .m script:

I am plotting this as follows:

Y=data;
DT=900; PAD=1; DJ=-1; S0=-1; J1=-1; MOTHER=-1; PARAM=-1;
[WAVE,PERIOD,SCALE,COI,DJ] = wavelet(Y,DT,PAD,DJ,S0,J1,MOTHER,PARAM);
figure; imagesc(abs(WAVE)); colorbar

I am completely lost as to how to get the correct y axis. To me, the period values obtained from the wavelet function do not correspond to the actual period/s of the signal. When I try:

figure; imagesc(timeseries,PERIOD,abs(WAVE)); colorbar

The y-axis on the graph seems to equate to the cumulative time on the first example plot above. I have no idea what I am missing here, and think that I need to do something more to be able to get something looking like the first plot above, with the actual periods...

So I guess my question is, why does the period output obtained not look correct? Something to do with pseudo/instantaneous frequency difference? I am lost!

Thanks in advance!!