[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Tftb-help] problem with TFTB

From: Alessandra Iacobucci
Subject: Re: [Tftb-help] problem with TFTB
Date: Tue, 21 Mar 2006 18:11:36 +0100

Thank you for your fast answer.
the nonstationarity of the series (which is exactly what I want to study, that is, the evolution of its frequency content) is another problem.

What I am trying to say is that the software gives an energy spectral density (well-defined or not) with peaks only at low frequencies while the contour plot of the spectrogram distribution (well defined or not) shows component only around the Nyquist frequency, so it seems to me that there is an incoherence.

If the use of a non analytic signal disturbs you, you may check this example (p. 56 of the tutorial)
>> sig2=fmconst(128,.15)+fmconst(128,.35);
>> tfrscalo(sig2,1:128,6,0.05,0.45,128,1);
the right results should be the one in Fig 3.20, p.56, but I got the following:

Attachment: untitled.pdf
Description: Adobe PDF document

So in my opinion there might be a problem.


On Mar 21, 2006, at 7:36 PM, George A. Bertos wrote:

In the same topic, is the spectral density calculated legitimate ? - since it is using the fft() - which is not applicable for non-stationary signals. I think one should use integrate in time and frequency domain based on the


-----Original Message-----
From: address@hidden
[mailto:address@hidden On Behalf Of
Alessandra Iacobucci
Sent: Tuesday, March 21, 2006 7:27 AM
To: address@hidden
Subject: [Tftb-help] problem with TFTB


I used previous versions of this toolbox and I am starting to use its
last version.

I obtained the picture I am sending you by the comand:


I think that either the contour or the energy spectral density plot
is upside down.
I tried also the demos and the same thing happens with command like
the plot that is produced differs from that shown in the tutorial.

Is this a bug?

Many thanks for you help,


reply via email to

[Prev in Thread] Current Thread [Next in Thread]