<< tfrsp Bilinear Time-Frequency Processing in the Cohen's Class tfrspwv >>

stftb >> stftb > Bilinear Time-Frequency Processing in the Cohen's Class > tfrspbk

tfrspbk

Smoothed Pseudo K-Bertrand time-frequency distribution.

Calling Sequence

[TFR,T,F] = tfrspbk(X)
[TFR,T,F] = tfrspbk(X, T)
[TFR,T,F] = tfrspbk(X, T, K)
[TFR,T,F] = tfrspbk(X, T, K, NH0)
[TFR,T,F] = tfrspbk(X, T, K, NH0,NG0)
[TFR,T,F] = tfrspbk(X, T, K, NH0,NG0, FMIN,FMAX)
[TFR,T,F] = tfrspbk(X, T, K, NH0,NG0, FMIN,FMAX, N)
[TFR,T,F] = tfrspbk(X, T, K, NH0,NG0, FMIN,FMAX, N, TRACE)
[TFR,T,F] = tfrspbk(...,'plot')

Parameters

X :

A Nx elements vector (auto-SPBK) or a Nx by 2 array signal (cross-SPBK).

T:

a real Nt vector : time instant(s) (default: 1:NX).

T must be a uniformly sampled vector whose elements are between 1 and Nx

K :

label of the K-Bertrand distribution. The distribution with parametrization function lambdak(u,K) = (K (exp(-u)-1)/(exp(-Ku)-1))^(1/(K-1)) is computed (default : 0).

K=-1 : Smoothed pseudo (active) Unterberger distribution

K=0 : Smoothed pseudo Bertrand distribution

K=0.5 : Smoothed pseudo D-Flandrin distribution

K=2 : Affine smoothed pseudo Wigner-Ville distribution

NH0 :

half length of the analyzing wavelet at coarsest scale. A Morlet wavelet is used. NH0 controles the frequency smoothing of the smoothed pseudo K-Bertrand distribution.(default : sqrt(Nx)).

NG0 :

half length of the time smoothing window. NG0 = 0 corresponds to the Pseudo K-Bertrand distribution. (default : 0).

FMIN:

a positive scalar in ]0 0.5], the normalized lower frequency bound in (Hz) of the analyzed signal. When unspecified, you have to enter it at the command line from the plot of the spectrum.

FMAX :

a positive scalar in ]0 0.5], the normalized upper frequency bound (in Hz) of the analyzed signal. When unspecified, you have to enter it at the command line from the plot of the spectrum.

N :

positive integer: the number of analyzed voices. When unspecified, you have to enter it at the command line from the plot of the spectrum.

TRACE :

A boolean (or a real scalar) if true (or nonzero),the progression of the algorithm is shown (default : %f).

'plot':

if one input parameter is 'plot', tfrqview is called and the time-frequency representation will be plotted.

TFR :

A complex N by Nt array: the time-frequency representation.

abscissa correspond to uniformly sampled time, and ordinates correspond to a geometrically sampled frequency). First row of TFR corresponds to the lowest frequency.

F :

A N vector of normalized frequencies (geometrically sampled from FMIN to FMAX).

Description

tfrspbk generates the auto or cross Smoothed Pseudo K-Bertrand distribution.

Examples

N = 64,
sig = altes(N,0.1,0.45); 
[TFR,T,F] = tfrspbk(sig,1:N,0,sqrt(N),0,0.1,0.35,8);
clf; gcf().color_map = jetcolormap(128);
Sgrayplot(T,F,TFR');

Authors


Report an issue
<< tfrsp Bilinear Time-Frequency Processing in the Cohen's Class tfrspwv >>