Wavelet Fidelity Filter (video compression)

I have read a specification of Dirac (video compression format). A key element of this format is the use of the discrete wavelet transform. Section 15.6.3 of the specification defines individual wavelets by making use of a lifting scheme. I am interested in the Fidelity filter for improved down-conversion and anti-aliasing. Where did it come from? Is there a paper that describes it?

For convenience, the lifting filters are defined as:

1. [ −2, 10, −25, 81, 81, −25, 10, −2 ]
2. [ −8, 21, −46, 161, 161, −46, 21, −8 ]
• Hi @DaBler ! Can you be more specific about which aspect of this lifting FIR filter (that you provide its coefficients) you have a problem with? Specifically what are the exact requirements on the signal being processed through the vide encoding confinguration which uses these filters for achieving those requirements ? These are ordinary filters. Don't expect a miracle other than computational efficiency due to integer coefficients... – Fat32 May 12 '16 at 13:13
• @Fat32: The question is: Where these coefficients come from? Using Google, I can find no other mention of this filter. – DaBler May 13 '16 at 7:54
• If this is a mandatory filter, then its coefficients are either pre-computed by the standard or at least the Frequency Magnitude of the filter must be given so that the coefficients can be computed from there. If this is an optional filter then at least its frequency characteristics must be stated so that you can find the coefficients by a sutiable filter design procedure. – Fat32 May 13 '16 at 11:48