Re: [Tftb-help] RE: Tftb-help Digest, Vol 25, Issue 1

[Eric Chassande-Mottin]

dear Hamid,

> I have sometimes faced to this error. I think the number of input points 
> not to exceed about 3000 points.(According to my experience in TFR 
> analysis) Of course the number of points is related to the number of 
> window points, too.

this is not correct. the limitation is that the result should fit in the 
memory of your computer. there is no hard threshold on the number of input 
samples what so ever. the memory usage is essentially determined by the 
size of the output TF matrix. in your experiences, you were probably 
computing a TF matrix approaching the size of the largest matrix your 
system can handle.

for the short-term Fourier analysis, the window size is disconnected from 
the size of the input signal vector (within a range of acceptable values). 
it can be adjusted in order to evidence the signal properties of interest.

cheers,
eric