stupid envelope follower idea

[Ian Fritz]

Martin --

Interesting ideas. And I agree that ear test are crucial here.  Which would
sound better (if you could do both perfectly): envelope following by
connecting peaks, or by tracking the instaneous power? Who knows!

  Ian


----- Original Message -----
From: "Martin Czech" <czech at Micronas.Com>

> I think you can arrange the different "measures" for function
> graphs in the following way:
>
> {INT |s(t)| dt} ^1
> {INT |s(t)|^2 dt} ^1/2
> {INT |s(t)|^3 dt} ^1/3
> .
> .
> .
> {INT |s(t)|^N dt} ^1/N
>
> for N-> +infinity this will be then peak. Right?
>
> So they are all differnt, but belong to the same family.
> Perhaps, if multiband works for RMS (N=2), it may sort of work
> also for just full wave rectify (N=1) or peak (N=inf) as
> well.
>
> Just a mathematical reasoning, gives some hope...
>
> As someone said: the ear will decide.