Mike Berry wrote:
> 
> There has been a lot of discussion about maximizers, so I thought I would add
> my thoughts.


> 
> a1 (... an), arest              filtbank        kbands, kbase, kspacing
> 
> where:
>         kbands = the number of bands
>         kbase = the top frequency of the first band
>         kspacing = the width in octaves of each band
>         a1...an = the output of each band
>         arest = the remaining signal, delayed to be in sync with the bands
> 
>         Then each band could be manipulated separately and then recombined at will.
> With this, it would be easy to experiment with your own maximizer orchestra.
> This is not a trivial opcode, since designed the necessary filters is rather
> tricky.  But there must be some research somewhere on how to do this.

Let me chime in for that. 
optionally run the filtbank at krate and obtain rms of each band as output?


-- 
________________________________________________________
Jean Piche
Universite de Montreal
http://mistral.ere.umontreal.ca/~pichej