Below is a reply by Jani Huhtanen in another old thread...
>Below is a diagram of the synthesis part of the wavelet transform (one
> + ---> z[n]
>Say that y0[n] and y1[n] are defined for n=1..8 (i.e. 8 samples). After
>UpSample x0[n] and x1[n] are defined for n=1..16 (i.e. 16 samples), where
> x0[2*n] = y0[n],
> x0[2*n+1] = 0,
> x1[2*n] = y1[n],
> x1[2*n+1] = 0.
>At this point the sample count has doubled from 8 to 16 for both detail
>approximation parts. Next these upsampled signals x0 and x1 are filtered
>G0 and G1 respectively. This does not increase the sample count.
>Tampere University of Technology, Pori
In regards to this topic, I am still trying to understand DWT.
After upsampling is done, Zeros are inserted between the coefficents(8
samples becomes 16 samples). After filtering/convolution is done, does the
zeros change to some other value(ie. IDWT change the zero to a non-zero