[M4IF Technotes] Re: AAC: PNS scale factors

[John Cox]

Dear Ralph
>> Ideally there would by synced 'random' number generators in the encoder
>> and decoder so the encoder knows exactly what to expect, in which case
>> there would be no need for the final renormalization.
>
>I guess you agree with me that this is a theoretical approach only. ;-)

Given the current protocol it has got to be theoretical only. However
the only extension required is a single bit to reset the 'random number'
generator to a known state.  After that you define a random number
generator (say random2 from the reference software) in the standard.  A
profile requirement to set the 'reset' bit once every (say) 8 frames
should make it all work.  This is a lot less complex to do than the
standard prediction tool.
Regards
John Cox