|
Purpose
|
Instantaneous frequency density.
|
Synopsis
tifd = friedman(tfr,hat)
tifd = friedman(tfr,hat,t)
tifd = friedman(tfr,hat,t,method)
tifd = friedman(tfr,hat,t,method,trace)
|
Description
friedman computes the time-instantaneous frequency density
(defined by Friedman [1]) of a reassigned time-frequency
representation.
| Name |
Description |
Default value |
| tfr |
time-frequency representation, (N,M) matrix |
|
| hat |
complex matrix of the reassignment vectors |
|
| t |
time instant(s) |
(1:M) |
| method |
chosen representation |
'tfrrsp' |
| trace |
if nonzero, the progression of the algorithm is shown |
0 |
| tifd |
time instantaneous-frequency density. When called without
output arguments, friedman runs tfrqview |
|
Warning : tifd is not an energy distribution, but an estimated
probability distribution.
|
Example
Here is an example of such an estimated probability distribution operated
on the reassigned pseudo-Wigner-Ville distribution of a linear frequency
modulation :
sig=fmlin(128,0.1,0.4);
[tfr,rtfr,hat]=tfrrpwv(sig);
friedman(tfr,hat,1:128,'tfrrpwv',1);
The result is almost perfectly concentrated on a line in the time-frequency
plane.
|
See Also
Reference
|
[1] D.H. Friedman, "Instantaneous Frequency vs Time : An
Interpretation of the Phase Structure of Speech", Proc. IEEE
ICASSP, pp. 29.10.1-4, Tampa, 1985.
|
Eric Chassande-Mottin
2005-10-26
|
