Prev: Need help in analysing equations-Active Power FIlter
Next: Optimal digital FSK demodulator
From: Rajkumar.R on 28 Jan 2010 08:13
Good morning to every one. . .

I am very new to this forum and not much expert in DSP field too. . .

I am trying to implement a IEEE paper named "A New Delayless Subband
Adaptive Filtering Algorithm for Active Noise Control Systems"
published in IEEE TRANSACTIONS ON AUDIO, SPEECH, AND LANGUAGE
PROCESSING july 2009.


I have not done any such projects before and so i am strugling where
to start and how to implement the given details.

This project is all about the following

1) A full-band filter that filters the input signal.
2) Decomposition of input and error signals into subbands.
3) Decimation in subbands.
4) Adaptive filters working in subbands.
5) A weight stacking method to combine all subbands weights into a
full-band filter.


These are the details of a filter specification given in the paper.
Could any one please help me to implement this in the matlab please?

x(n) is the input signal which is to be filtered into M subbands named

X'k(n)

where K ranges form 0 to M-1, using a analysis filter bank h(z) with a
decimation factor D.

Filter h(z) is given by

h(z)=[ H0(z) ]
| H1(z) |
| . |
| . |
| . |
[ Hm-1(z)]

Where Hk(z) is the transfer function of Kth subband.


Given the low pass prototype filter for filter banks is

H0(z)=1+(Z)^-1+(Z)^-2+.......+(Z)^-M+1


The resulting filter bank is the simplest FIR filter perfect
reconstruction filter bank which is made by


Hk(z)=H0(z.exp(-j*2*pi*k/M))

with a frequency response of

H0(exp(jw))={ M , when w=0;
{exp(-j*w(M-1)/2sin(wM/2)/sin(w/2)) , otherwise;


M is given about to be 16

D=M/4

For this proposed UDFT modulated filter banks, h(z) is defined by


h(z) = (1/M)F*[1 , (Z)^-1 , (Z)^-2 ,. . ., (Z)^-M+1] (transpose)

Where F is the DFT matix of order M. The central frequency of bandpass
filters Hk(Z) are located at

wk=2*pi*k/M for 0 ≤ k ≤ M-1.

This filter can be realized by tapped delay line of length M followed
by an inverse FFT block.
From: Rune Allnor on 28 Jan 2010 08:54
On 28 Jan, 14:13, "Rajkumar.R" <rraj...(a)gmail.com> wrote:
> Good morning to every one. . .
....
> Could any one please help me to implement this in the matlab please?

Helping with the programming would be consultancy work.
Several people here might be both able and willing to
help, but for a fee.

If you have questions about the algorithms and details
in the paper, by all means, we might help for free.
But do expect to have to pay for hands-on help.

Rune
 | 
Pages: 1
Prev: Need help in analysing equations-Active Power FIlter
Next: Optimal digital FSK demodulator