Functions > Signal Processing > Joint Time-Frequency Analysis > Time-Frequency Representations and Local Autocorrelation
Time-Frequency Representations and Local Autocorrelation
timecorr(v, t, l, s, [d])—Returns a matrix containing the samples of a time-dependent autocorrelation function for a signal v, type t, l number of lags, step size s, and an optional scalar or matrix argument d.
timefreq(v, t, f, s, [d])—Returns a matrix containing the bilinear time-frequency representation for a signal v, type t, f number of frequencies, step size s, and an optional scalar or matrix argument d.
 Both functions return a matrix with f rows and approximately length(v)/s columns, containing the time-dependent autocorrelation or the bilinear time-frequency representation matrix for signal v.
Arguments
v is a vector of real-valued data.
t is a string containing the name of the desired representation (optional for the custom type).
Value of t
Representation
"wigner"
Wigner
"cone"
Cone
"bornjordan"
Born-Jordan
"marghill"
Margenau-Hill
"choiwill"
Choi-Williams
"custom"
custom kernel in the time-lag plane
f is an integer number of lags (for timecorr) or number of frequencies (for timefreq) to be computed.
s is the step size, an integer giving the number of samples to skip between successive computations.
d (optional) is a positive dampening factor for the Choi-Williams distribution. Ignored for other representations. If omitted, d = 1.
 If t is set to "custom", the functions take on the following set of arguments: timecorr(v, ["custom"], s, r ) timefreq(v, ["custom"], s, r ) Where: • r is a matrix of real numbers giving the time-lag kernel.
Was this helpful?