But any sinusoidal waveform will be wrapped around due to this DC offset,

which is an effect I don't want to have after a Chebychev waveshaping. A
highpass filter takes out the DC, but only after the transforming.

And if To = 1 and T2 = 2x*x-1 as Chebychev functions, than the sum To+T2
should eliminate the DC offset, but this doesn't happen with a GEN13 function
even if h0 = 0 .

Any comments?

Sinan Boekesoy

>omni wrote:
>> Is there anyway to eliminate the DC offset, which I get everytime after a 
>> Chebychev waveshaping of any waveform? A simple example is below.
>
>A high-pass filter.
>
>  e
>
>

omni wrote:
> But any sinusoidal waveform will be wrapped around due to this DC offset,

Are you thinking of fixed-point calculation?  In Csound, wrapping only
happens if you hit the rails at output.

> And if To = 1 and T2 = 2x*x-1 as Chebychev functions, than the sum To+T2
> should eliminate the DC offset, 

That's equivalent to running the signal through T2 and then trying to
fix up the DC offset by adding 1.  But the offset after T2 is -1 only
in the small-signal case.  (Note that the "fixed" function never goes
below zero...)

-- 
     Eli Brandt  |  eli+@cs.cmu.edu  |  http://www.cs.cmu.edu/~eli/