This is where navigation should be.

WPFBT2FILTERBANK - WPFBT equivalent non-iterated filterbank

Usage

[g,a] = wpfbt2filterbank(wt)

Input parameters

wt Wavelet filter tree definition

Output parameters

g Cell array containing filters
a Vector of sub-/upsampling factors

Description

wpfbt2filterbank(wt) calculates the impulse responses g and the subsampling factors a of non-iterated filterbank, which is equivalent to the wavelet packet filterbank tree described by wt. The returned parameters can be used directly in filterbank, ufilterbank or filterbank.

Please see help on wfbt for description of wt. The function additionally support the following flags:

'freq'`(default),'nat'`
The filters are ordered to produce subbands in the same order as wpfbt with the same flag.
'intsqrt'`(default),'intnoscale', `'intscale'
The filters in the filterbank tree are scaled to reflect the behavior of wpfbt and iwpfbt with the same flags.
'scaling_notset'`(default),'noscale','scale','sqrt'`
Support for scaling flags as described in uwpfbt. By default, the returned filterbank g and a is equivalent to wpfbt, passing any of the non-default flags results in a filterbank equivalent to uwpfbt i.e. scaled and with a(:)=1.

Examples:

The following two examples create a multirate identity filterbank using a tree of depth 3. In the first example, the filterbank is identical to the DWT tree:

[g,a] = wpfbt2filterbank({'db10',3,'dwt'});
filterbankfreqz(g,a,1024,'plot','linabs','posfreq');

In the second example, the filterbank is identical to the full wavelet tree:

[g,a] = wpfbt2filterbank({'db10',3,'full'});
filterbankfreqz(g,a,1024,'plot','linabs','posfreq');