Help with question on Wavelet Decomposition

Hi all,

I am newbie in wavelets..
Just read the MATLAB wavelet tutorial as I found it the simplest amongst others to understand..

I want to ask that during the DECOMPOSITION phase in the DWT, when a signal is convolved with HPF and LPF to yield details and approximation..

Q: How are the co-efficients chosen for the HPF and LPF so that they can be convolved with a given signal , say s. Assume, say I use daubechies 2 wavelet.

Q. If we go deeper down the decomposition level, say till 3 or 4, whatever.. do the values of the filter co-effiecients remain the same for HPF and LPF throughout different levels, or do they change ? If they change, how are these values determined then ?

Can someone please help me with these concepts..
Help would be appreciated

Regards

You can read this paper

**broken link removed**

first part of it is enough for intuitive understanding basics of wavelet transform.

>>Q: How are the co-efficients chosen for the HPF and LPF so that they can be >>convolved with a given signal , say s. Assume, say I use daubechies 2 wavelet.
You can use any wavelet. But for every task there is optimal wavelets. Unfortunately, there is not too much papers about optimality of wavelets in different cases.

>>Q. If we go deeper down the decomposition level, say till 3 or 4, whatever.. do >>the values of the filter co-effiecients remain the same for HPF and LPF >>throughout different levels, or do they change ? If they change, how are these >>values determined then ?
Coefficients remain the same from stage to stage.

PS.:Some similarity you can see in fractals